ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1582 - (show annotations)
Wed Jul 22 09:33:04 2015 UTC (7 days, 6 hours ago) by ph10
File size: 281825 byte(s)
Fix (?R- not being diagnosed correctly.
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.
102 Version 8.37 28-April-2015
103 --------------------------
105 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
106 for those parentheses to be closed with whatever has been captured so far.
107 However, it was failing to mark any other groups between the hightest
108 capture so far and the currrent group as "unset". Thus, the ovector for
109 those groups contained whatever was previously there. An example is the
110 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
112 2. If an assertion condition was quantified with a minimum of zero (an odd
113 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
115 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
116 unrecognized modifier, a crash could occur.
118 4. An attempt to do global matching in pcretest with a zero-length ovector
119 caused a crash.
121 5. Fixed a memory leak during matching that could occur for a subpattern
122 subroutine call (recursive or otherwise) if the number of captured groups
123 that had to be saved was greater than ten.
125 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
126 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
127 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
129 7. A UTF pattern containing a "not" match of a non-ASCII character and a
130 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
132 8. When a pattern is compiled, it remembers the highest back reference so that
133 when matching, if the ovector is too small, extra memory can be obtained to
134 use instead. A conditional subpattern whose condition is a check on a
135 capture having happened, such as, for example in the pattern
136 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
137 setting the highest backreference number. This mattered only if pcre_exec()
138 was called with an ovector that was too small to hold the capture, and there
139 was no other kind of back reference (a situation which is probably quite
140 rare). The effect of the bug was that the condition was always treated as
141 FALSE when the capture could not be consulted, leading to a incorrect
142 behaviour by pcre_exec(). This bug has been fixed.
144 9. A reference to a duplicated named group (either a back reference or a test
145 for being set in a conditional) that occurred in a part of the pattern where
146 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
147 to be incorrectly calculated, leading to overwriting.
149 10. A mutually recursive set of back references such as (\2)(\1) caused a
150 segfault at study time (while trying to find the minimum matching length).
151 The infinite loop is now broken (with the minimum length unset, that is,
152 zero).
154 11. If an assertion that was used as a condition was quantified with a minimum
155 of zero, matching went wrong. In particular, if the whole group had
156 unlimited repetition and could match an empty string, a segfault was
157 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
158 assertions to be quantified, but not if they are being used as conditions,
159 so the above pattern is faulted by Perl. PCRE has now been changed so that
160 it also rejects such patterns.
162 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
163 failed to allow the zero-repeat case if pcre2_exec() was called with an
164 ovector too small to capture the group.
166 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
167 Red Hat Product Security:
169 (a) A crash if /K and /F were both set with the option to save the compiled
170 pattern.
172 (b) Another crash if the option to print captured substrings in a callout
173 was combined with setting a null ovector, for example \O\C+ as a subject
174 string.
176 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
177 forward reference repeated a large (but limited) number of times within a
178 repeated outer group that has a zero minimum quantifier, caused incorrect
179 code to be compiled, leading to the error "internal error:
180 previously-checked referenced subpattern not found" when an incorrect
181 memory address was read. This bug was reported as "heap overflow",
182 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
183 CVE-2015-2325.
185 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
186 call within a group that also contained a recursive back reference caused
187 incorrect code to be compiled. This bug was reported as "heap overflow",
188 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
189 number CVE-2015-2326.
191 24. Computing the size of the JIT read-only data in advance has been a source
192 of various issues, and new ones are still appear unfortunately. To fix
193 existing and future issues, size computation is eliminated from the code,
194 and replaced by on-demand memory allocation.
196 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
197 adjacent to the end of the range, and the range contained characters with
198 more than one other case, caused incorrect behaviour when compiled in UTF
199 mode. In that example, the range a-j was left out of the class.
201 26. Fix JIT compilation of conditional blocks, which assertion
202 is converted to (*FAIL). E.g: /(?(?!))/.
204 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
205 discovered by the LLVM fuzzer.
207 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
208 when this assertion was used as a condition, for example (?(?!)a|b). In
209 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
210 error about an unsupported item.
212 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
213 possessification code could take exponential time to complete. A recursion
214 depth limit of 1000 has been imposed to limit the resources used by this
215 optimization.
217 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
218 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
219 because \S ensures they are all in the class. The code for doing this was
220 interacting badly with the code for computing the amount of space needed to
221 compile the pattern, leading to a buffer overflow. This bug was discovered
222 by the LLVM fuzzer.
224 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
225 other kinds of group caused stack overflow at compile time. This bug was
226 discovered by the LLVM fuzzer.
228 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
229 between a subroutine call and its quantifier was incorrectly compiled,
230 leading to buffer overflow or other errors. This bug was discovered by the
231 LLVM fuzzer.
233 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
234 assertion after (?(. The code was failing to check the character after
235 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
236 was discovered by the LLVM fuzzer.
238 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
239 a fixed maximum following a group that contains a subroutine reference was
240 incorrectly compiled and could trigger buffer overflow. This bug was
241 discovered by the LLVM fuzzer.
243 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
244 caused a stack overflow instead of the diagnosis of a non-fixed length
245 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
247 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
248 (e.g. /(?<=\Ka)/) could make pcregrep loop.
250 37. There was a similar problem to 36 in pcretest for global matches.
252 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
253 and a subsequent item in the pattern caused a non-match, backtracking over
254 the repeated \X did not stop, but carried on past the start of the subject,
255 causing reference to random memory and/or a segfault. There were also some
256 other cases where backtracking after \C could crash. This set of bugs was
257 discovered by the LLVM fuzzer.
259 39. The function for finding the minimum length of a matching string could take
260 a very long time if mutual recursion was present many times in a pattern,
261 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
262 method has been implemented. This infelicity was discovered by the LLVM
263 fuzzer.
265 40. Static linking against the PCRE library using the pkg-config module was
266 failing on missing pthread symbols.
269 Version 8.36 26-September-2014
270 ------------------------------
272 1. Got rid of some compiler warnings in the C++ modules that were shown up by
273 -Wmissing-field-initializers and -Wunused-parameter.
275 2. The tests for quantifiers being too big (greater than 65535) were being
276 applied after reading the number, and stupidly assuming that integer
277 overflow would give a negative number. The tests are now applied as the
278 numbers are read.
280 3. Tidy code in pcre_exec.c where two branches that used to be different are
281 now the same.
283 4. The JIT compiler did not generate match limit checks for certain
284 bracketed expressions with quantifiers. This may lead to exponential
285 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
286 issue should be resolved now.
288 5. Fixed an issue, which occures when nested alternatives are optimized
289 with table jumps.
291 6. Inserted two casts and changed some ints to size_t in the light of some
292 reported 64-bit compiler warnings (Bugzilla 1477).
294 7. Fixed a bug concerned with zero-minimum possessive groups that could match
295 an empty string, which sometimes were behaving incorrectly in the
296 interpreter (though correctly in the JIT matcher). This pcretest input is
297 an example:
299 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
302 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
303 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
304 for an empty string was breaking the inner loop and carrying on at a lower
305 level, when possessive repeated groups should always return to a higher
306 level as they have no backtrack points in them. The empty string test now
307 occurs at the outer level.
309 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
310 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
312 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
313 doesn't).
315 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
316 to hold a pattern to be incorrectly computed (too small) when there were
317 named back references to duplicated names. This could cause "internal
318 error: code overflow" or "double free or corruption" or other memory
319 handling errors.
321 11. When named subpatterns had the same prefixes, back references could be
322 confused. For example, in this pattern:
324 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
326 the reference to 'Name' was incorrectly treated as a reference to a
327 duplicate name.
329 12. A pattern such as /^s?c/mi8 where the optional character has more than
330 one "other case" was incorrectly compiled such that it would only try to
331 match starting at "c".
333 13. When a pattern starting with \s was studied, VT was not included in the
334 list of possible starting characters; this should have been part of the
335 8.34/18 patch.
337 14. If a character class started [\Qx]... where x is any character, the class
338 was incorrectly terminated at the ].
340 15. If a pattern that started with a caseless match for a character with more
341 than one "other case" was studied, PCRE did not set up the starting code
342 unit bit map for the list of possible characters. Now it does. This is an
343 optimization improvement, not a bug fix.
345 16. The Unicode data tables have been updated to Unicode 7.0.0.
347 17. Fixed a number of memory leaks in pcregrep.
349 18. Avoid a compiler warning (from some compilers) for a function call with
350 a cast that removes "const" from an lvalue by using an intermediate
351 variable (to which the compiler does not object).
353 19. Incorrect code was compiled if a group that contained an internal recursive
354 back reference was optional (had quantifier with a minimum of zero). This
355 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
356 caused segmentation faults because of stack overflows at compile time.
358 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
359 group that is quantified with an indefinite repeat, caused a compile-time
360 loop which used up all the system stack and provoked a segmentation fault.
361 This was not the same bug as 19 above.
363 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
364 Patch by Mike Frysinger.
367 Version 8.35 04-April-2014
368 --------------------------
370 1. A new flag is set, when property checks are present in an XCLASS.
371 When this flag is not set, PCRE can perform certain optimizations
372 such as studying these XCLASS-es.
374 2. The auto-possessification of character sets were improved: a normal
375 and an extended character set can be compared now. Furthermore
376 the JIT compiler optimizes more character set checks.
378 3. Got rid of some compiler warnings for potentially uninitialized variables
379 that show up only when compiled with -O2.
381 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
382 of a match later then the end of the match. The pcretest program was not
383 handling the case sensibly - it was outputting from the start to the next
384 binary zero. It now reports this situation in a message, and outputs the
385 text from the end to the start.
387 5. Fast forward search is improved in JIT. Instead of the first three
388 characters, any three characters with fixed position can be searched.
389 Search order: first, last, middle.
391 6. Improve character range checks in JIT. Characters are read by an inprecise
392 function now, which returns with an unknown value if the character code is
393 above a certain threshold (e.g: 256). The only limitation is that the value
394 must be bigger than the threshold as well. This function is useful when
395 the characters above the threshold are handled in the same way.
397 7. The macros whose names start with RAWUCHAR are placeholders for a future
398 mode in which only the bottom 21 bits of 32-bit data items are used. To
399 make this more memorable for those maintaining the code, the names have
400 been changed to start with UCHAR21, and an extensive comment has been added
401 to their definition.
403 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
404 to the export list in Makefile.am (they were accidentally omitted from the
405 8.34 tarball).
407 9. The informational output from pcretest used the phrase "starting byte set"
408 which is inappropriate for the 16-bit and 32-bit libraries. As the output
409 for "first char" and "need char" really means "non-UTF-char", I've changed
410 "byte" to "char", and slightly reworded the output. The documentation about
411 these values has also been (I hope) clarified.
413 10. Another JIT related optimization: use table jumps for selecting the correct
414 backtracking path, when more than four alternatives are present inside a
415 bracket.
417 11. Empty match is not possible, when the minimum length is greater than zero,
418 and there is no \K in the pattern. JIT should avoid empty match checks in
419 such cases.
421 12. In a caseless character class with UCP support, when a character with more
422 than one alternative case was not the first character of a range, not all
423 the alternative cases were added to the class. For example, s and \x{17f}
424 are both alternative cases for S: the class [RST] was handled correctly,
425 but [R-T] was not.
427 13. The configure.ac file always checked for pthread support when JIT was
428 enabled. This is not used in Windows, so I have put this test inside a
429 check for the presence of windows.h (which was already tested for).
431 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
432 The algorithm provides a way to skip certain starting offsets, and usually
433 faster than linear prefix searches.
435 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
436 as for 'fr_FR' and 'french'. For some reason, however, it then used the
437 Windows-specific input and output files, which have 'french' screwed in.
438 So this could never have worked. One of the problems with locales is that
439 they aren't always the same. I have now updated RunTest so that it checks
440 the output of the locale test (test 3) against three different output
441 files, and it allows the test to pass if any one of them matches. With luck
442 this should make the test pass on some versions of Solaris where it was
443 failing. Because of the uncertainty, the script did not used to stop if
444 test 3 failed; it now does. If further versions of a French locale ever
445 come to light, they can now easily be added.
447 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
448 there was a message during ./configure, but it did not stop. This now
449 provokes an error. The invalid example in README has been corrected.
450 If a value less than the minimum is given, the minimum value has always
451 been used, but now a warning is given.
453 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
454 was a bug in the test system, which is now fixed. Also, the list of various
455 configurations that are tested for each release did not have one with both
456 16/32 bits and --enable-bar-anycrlf. It now does.
458 18. pcretest was missing "-C bsr" for displaying the \R default setting.
460 19. Little endian PowerPC systems are supported now by the JIT compiler.
462 20. The fast forward newline mechanism could enter to an infinite loop on
463 certain invalid UTF-8 input. Although we don't support these cases
464 this issue can be fixed by a performance optimization.
466 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
467 not take account if existing stack usage. There is now a new global
468 variable called pcre_stack_guard that can be set to point to an external
469 function to check stack availability. It is called at the start of
470 processing every parenthesized group.
472 22. A typo in the code meant that in ungreedy mode the max/min qualifier
473 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
474 not match "ab".
476 23. When UTF was disabled, the JIT program reported some incorrect compile
477 errors. These messages are silenced now.
479 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
480 compiler.
482 25. Change all the temporary files used in RunGrepTest to be different to those
483 used by RunTest so that the tests can be run simultaneously, for example by
484 "make -j check".
487 Version 8.34 15-December-2013
488 -----------------------------
490 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
491 executable memory. Patch inspired by Carsten Klein.
493 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
494 this macro is never tested and has no effect, because the work to support
495 coverage involves only compiling and linking options and special targets in
496 the Makefile. The comment in config.h implied that defining the macro would
497 enable coverage support, which is totally false. There was also support for
498 setting this macro in the CMake files (my fault, I just copied it from
499 configure). SUPPORT_GCOV has now been removed.
501 3. Make a small performance improvement in strlen16() and strlen32() in
502 pcretest.
504 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
505 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
506 these cases). There was also one in pcretest.c.
508 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
510 6. In UTF mode, the code for checking whether a group could match an empty
511 string (which is used for indefinitely repeated groups to allow for
512 breaking an infinite loop) was broken when the group contained a repeated
513 negated single-character class with a character that occupied more than one
514 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
515 UTF-8 mode). The effect was undefined: the group might or might not be
516 deemed as matching an empty string, or the program might have crashed.
518 7. The code for checking whether a group could match an empty string was not
519 recognizing that \h, \H, \v, \V, and \R must match a character.
521 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
522 an empty string. If it can, pcretest shows this in its information output.
524 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
525 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
526 matched by pcre_exec() without using JIT:
528 (a) If the rest of the pattern did not match after a maximal run of
529 grapheme clusters, the code for backing up to try with fewer of them
530 did not always back up over a full grapheme when characters that do not
531 have the modifier quality were involved, e.g. Hangul syllables.
533 (b) If the match point in a subject started with modifier character, and
534 there was no match, the code could incorrectly back up beyond the match
535 point, and potentially beyond the first character in the subject,
536 leading to a segfault or an incorrect match result.
538 10. A conditional group with an assertion condition could lead to PCRE
539 recording an incorrect first data item for a match if no other first data
540 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
541 first data item, and therefore matched "ca" after "c" instead of at the
542 start.
544 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
545 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
547 12. The source of pcregrep now includes z/OS-specific code so that it can be
548 compiled for z/OS as part of the special z/OS distribution.
550 13. Added the -T and -TM options to pcretest.
552 14. The code in pcre_compile.c for creating the table of named capturing groups
553 has been refactored. Instead of creating the table dynamically during the
554 actual compiling pass, the information is remembered during the pre-compile
555 pass (on the stack unless there are more than 20 named groups, in which
556 case malloc() is used) and the whole table is created before the actual
557 compile happens. This has simplified the code (it is now nearly 150 lines
558 shorter) and prepared the way for better handling of references to groups
559 with duplicate names.
561 15. A back reference to a named subpattern when there is more than one of the
562 same name now checks them in the order in which they appear in the pattern.
563 The first one that is set is used for the reference. Previously only the
564 first one was inspected. This change makes PCRE more compatible with Perl.
566 16. Unicode character properties were updated from Unicode 6.3.0.
568 17. The compile-time code for auto-possessification has been refactored, based
569 on a patch by Zoltan Herczeg. It now happens after instead of during
570 compilation. The code is cleaner, and more cases are handled. The option
571 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
572 options in pcretest are provided to set it. It can also be set by
573 (*NO_AUTO_POSSESS) at the start of a pattern.
575 18. The character VT has been added to the default ("C" locale) set of
576 characters that match \s and are generally treated as white space,
577 following this same change in Perl 5.18. There is now no difference between
578 "Perl space" and "POSIX space". Whether VT is treated as white space in
579 other locales depends on the locale.
581 19. The code for checking named groups as conditions, either for being set or
582 for being recursed, has been refactored (this is related to 14 and 15
583 above). Processing unduplicated named groups should now be as fast at
584 numerical groups, and processing duplicated groups should be faster than
585 before.
587 20. Two patches to the CMake build system, by Alexander Barkov:
589 (1) Replace the "source" command by "." in CMakeLists.txt because
590 "source" is a bash-ism.
592 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
593 without these the CMake build does not work on Solaris.
595 21. Perl has changed its handling of \8 and \9. If there is no previously
596 encountered capturing group of those numbers, they are treated as the
597 literal characters 8 and 9 instead of a binary zero followed by the
598 literals. PCRE now does the same.
600 22. Following Perl, added \o{} to specify codepoints in octal, making it
601 possible to specify values greater than 0777 and also making them
602 unambiguous.
604 23. Perl now gives an error for missing closing braces after \x{... instead of
605 treating the string as literal. PCRE now does the same.
607 24. RunTest used to grumble if an inappropriate test was selected explicitly,
608 but just skip it when running all tests. This make it awkward to run ranges
609 of tests when one of them was inappropriate. Now it just skips any
610 inappropriate tests, as it always did when running all tests.
612 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
613 character types such as \d or \w, too many callouts were inserted, and the
614 data that they returned was rubbish.
616 26. In UCP mode, \s was not matching two of the characters that Perl matches,
617 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
618 were matched by \h. The code has now been refactored so that the lists of
619 the horizontal and vertical whitespace characters used for \h and \v (which
620 are defined only in one place) are now also used for \s.
622 27. Add JIT support for the 64 bit TileGX architecture.
623 Patch by Jiong Wang (Tilera Corporation).
625 28. Possessive quantifiers for classes (both explicit and automatically
626 generated) now use special opcodes instead of wrapping in ONCE brackets.
628 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
629 (because it's meaningless), this was not happening when PCRE_CASELESS was
630 set. Not wrong, but inefficient.
632 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
633 properties for \w, \d, etc) is present in a test regex. Otherwise if the
634 test contains no characters greater than 255, Perl doesn't realise it
635 should be using Unicode semantics.
637 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
638 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
639 does in Unicode mode.
641 32. Added the "forbid" facility to pcretest so that putting tests into the
642 wrong test files can sometimes be quickly detected.
644 33. There is now a limit (default 250) on the depth of nesting of parentheses.
645 This limit is imposed to control the amount of system stack used at compile
646 time. It can be changed at build time by --with-parens-nest-limit=xxx or
647 the equivalent in CMake.
649 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
650 errors. Perl warns for these when in warning mode, but PCRE has no facility
651 for giving warnings.
653 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
654 However, this was not working for (?!) because it is optimized to (*FAIL),
655 for which PCRE does not allow quantifiers. The optimization is now disabled
656 when a quantifier follows (?!). I can't see any use for this, but it makes
657 things uniform.
659 36. Perl no longer allows group names to start with digits, so I have made this
660 change also in PCRE. It simplifies the code a bit.
662 37. In extended mode, Perl ignores spaces before a + that indicates a
663 possessive quantifier. PCRE allowed a space before the quantifier, but not
664 before the possessive +. It now does.
666 38. The use of \K (reset reported match start) within a repeated possessive
667 group such as (a\Kb)*+ was not working.
669 40. Document that the same character tables must be used at compile time and
670 run time, and that the facility to pass tables to pcre_exec() and
671 pcre_dfa_exec() is for use only with saved/restored patterns.
673 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
674 features for Builds with MSVC:
676 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
677 builds). The .rc files can be used to set FileDescription and many other
678 attributes.
680 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
681 This allows higher-level build scripts which want .pdb files to avoid
682 hard-coding the exact files needed."
684 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
685 mean "start of word" and "end of word", respectively, as a transition aid.
687 43. A minimizing repeat of a class containing codepoints greater than 255 in
688 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
689 compiled to use the heap for recursion.
691 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
692 is configured.
695 Version 8.33 28-May-2013
696 ------------------------
698 1. Added 'U' to some constants that are compared to unsigned integers, to
699 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
700 variables that are added to signed variables, to ensure the result is
701 signed and can be negated.
703 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
704 CMake config files.
706 3. Revise the creation of config.h.generic so that all boolean macros are
707 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
708 overriding via -D on the command line possible.
710 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
711 to unsigned int is reported to make a quite noticeable speed difference in
712 a specific Windows environment. Testing on Linux did also appear to show
713 some benefit (and it is clearly not harmful). Also fixed the definition of
714 Xop which should be unsigned.
716 5. Related to (4), changing the definition of the intermediate variable cc
717 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
718 improvements.
720 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
721 unnecessary spaces.
723 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
724 and later.
726 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
728 9. Optimizing fast_forward_start_bits in JIT.
730 10. Adding support for callouts in JIT, and fixing some issues revealed
731 during this work. Namely:
733 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
735 (b) Minimum length was not checked before the matching is started.
737 11. The value of capture_last that is passed to callouts was incorrect in some
738 cases when there was a capture on one path that was subsequently abandoned
739 after a backtrack. Also, the capture_last value is now reset after a
740 recursion, since all captures are also reset in this case.
742 12. The interpreter no longer returns the "too many substrings" error in the
743 case when an overflowing capture is in a branch that is subsequently
744 abandoned after a backtrack.
746 13. In the pathological case when an offset vector of size 2 is used, pcretest
747 now prints out the matched string after a yield of 0 or 1.
749 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
750 Only supported by the JIT compiler at the moment.
752 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
754 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
755 offset of the starting point of the matching process, provided the offsets
756 vector is large enough.
758 17. The \A escape now records a lookbehind value of 1, though its execution
759 does not actually inspect the previous character. This is to ensure that,
760 in partial multi-segment matching, at least one character from the old
761 segment is retained when a new segment is processed. Otherwise, if there
762 are no lookbehinds in the pattern, \A might match incorrectly at the start
763 of a new segment.
765 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
767 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
768 gives some modest performance improvement in 8-bit mode.
770 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
771 be expressed in certain programming languages using Universal Character
772 Names.
774 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
775 which points out that "non characters" are not "characters that may not
776 appear in Unicode strings" but rather "characters that are reserved for
777 internal use and have only local meaning".
779 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
780 there was a conditional group that depended on an assertion, if the
781 assertion was false, the callout that immediately followed the alternation
782 in the condition was skipped when pcre_exec() was used for matching.
784 23. Allow an explicit callout to be inserted before an assertion that is the
785 condition for a conditional group, for compatibility with automatic
786 callouts, which always insert a callout at this point.
788 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
789 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
791 25. (*PRUNE) is now supported by the JIT compiler.
793 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
795 27. Fix the case where there are two or more SKIPs with arguments that may be
796 ignored.
798 28. (*SKIP) is now supported by the JIT compiler.
800 29. (*THEN) is now supported by the JIT compiler.
802 30. Update RunTest with additional test selector options.
804 31. The way PCRE handles backtracking verbs has been changed in two ways.
806 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
807 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
808 backtracking. In some cases this makes it more Perl-compatible, but Perl's
809 rather obscure rules do not always do the same thing.
811 (2) Previously, backtracking verbs were confined within assertions. This is
812 no longer the case for positive assertions, except for (*ACCEPT). Again,
813 this sometimes improves Perl compatibility, and sometimes does not.
815 32. A number of tests that were in test 2 because Perl did things differently
816 have been moved to test 1, because either Perl or PCRE has changed, and
817 these tests are now compatible.
819 32. Backtracking control verbs are now handled in the same way in JIT and
820 interpreter.
822 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
823 contained a forward subroutine reference caused a compile error.
825 34. Auto-detect and optimize limited repetitions in JIT.
827 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
828 blocking (*UTF) etc.
830 36. In the interpreter, maximizing pattern repetitions for characters and
831 character types now use tail recursion, which reduces stack usage.
833 37. The value of the max lookbehind was not correctly preserved if a compiled
834 and saved regex was reloaded on a host of different endianness.
836 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
837 of the compiled pattern block, expand the flags field from 16 to 32 bits
838 because it was almost full.
840 39. Try madvise first before posix_madvise.
842 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
843 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
844 original change locked out all matches of empty strings. This has been
845 changed so that one match of an empty string per line is recognized.
846 Subsequent searches on the same line (for colouring or for --only-matching,
847 for example) do not recognize empty strings.
849 41. Applied a user patch to fix a number of spelling mistakes in comments.
851 42. Data lines longer than 65536 caused pcretest to crash.
853 43. Clarified the data type for length and startoffset arguments for pcre_exec
854 and pcre_dfa_exec in the function-specific man pages, where they were
855 explicitly stated to be in bytes, never having been updated. I also added
856 some clarification to the pcreapi man page.
858 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
859 a segmentation fault.
862 Version 8.32 30-November-2012
863 -----------------------------
865 1. Improved JIT compiler optimizations for first character search and single
866 character iterators.
868 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
869 Patch by Daniel Richard G.
871 3. Single character iterator optimizations in the JIT compiler.
873 4. Improved JIT compiler optimizations for character ranges.
875 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
876 Reported by Giuseppe D'Angelo.
878 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
879 of a line, was being set incorrectly in cases where .* appeared inside
880 atomic brackets at the start of a pattern, or where there was a subsequent
881 *PRUNE or *SKIP.
883 7. Improved instruction cache flush for POWER/PowerPC.
884 Patch by Daniel Richard G.
886 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
887 grep:
889 (a) There is now no limit to the number of patterns to be matched.
891 (b) An error is given if a pattern is too long.
893 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
894 are now supported.
896 (d) --exclude-from and --include-from (multiple use) have been added.
898 (e) Exclusions and inclusions now apply to all files and directories, not
899 just to those obtained from scanning a directory recursively.
901 (f) Multiple uses of -f and --file-list are now supported.
903 (g) In a Windows environment, the default for -d has been changed from
904 "read" (the GNU grep default) to "skip", because otherwise the presence
905 of a directory in the file list provokes an error.
907 (h) The documentation has been revised and clarified in places.
909 9. Improve the matching speed of capturing brackets.
911 10. Changed the meaning of \X so that it now matches a Unicode extended
912 grapheme cluster.
914 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
915 out POSIX threads when JIT support is configured.
917 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
919 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
920 structure could go wrong in environments where size_t is not the same size
921 as int.
923 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
924 set.
926 15. The EBCDIC support had decayed; later updates to the code had included
927 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
928 general tidy up of EBCDIC-related issues, and the documentation was also
929 not quite right. There is now a test that can be run on ASCII systems to
930 check some of the EBCDIC-related things (but is it not a full test).
932 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
933 in a small tidy to the code.
935 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
937 18. If the --only-matching (-o) option in pcregrep is specified multiple
938 times, each one causes appropriate output. For example, -o1 -o2 outputs the
939 substrings matched by the 1st and 2nd capturing parentheses. A separating
940 string can be specified by --om-separator (default empty).
942 19. Improving the first n character searches.
944 20. Turn case lists for horizontal and vertical white space into macros so that
945 they are defined only once.
947 21. This set of changes together give more compatible Unicode case-folding
948 behaviour for characters that have more than one other case when UCP
949 support is available.
951 (a) The Unicode property table now has offsets into a new table of sets of
952 three or more characters that are case-equivalent. The MultiStage2.py
953 script that generates these tables (the pcre_ucd.c file) now scans
954 CaseFolding.txt instead of UnicodeData.txt for character case
955 information.
957 (b) The code for adding characters or ranges of characters to a character
958 class has been abstracted into a generalized function that also handles
959 case-independence. In UTF-mode with UCP support, this uses the new data
960 to handle characters with more than one other case.
962 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
963 whose other case is greater than 256 are now correctly matched
964 caselessly. Previously, the high codepoint matched the low one, but not
965 vice versa.
967 (d) The processing of \h, \H, \v, and \ in character classes now makes use
968 of the new class addition function, using character lists defined as
969 macros alongside the case definitions of 20 above.
971 (e) Caseless back references now work with characters that have more than
972 one other case.
974 (f) General caseless matching of characters with more than one other case
975 is supported.
977 22. Unicode character properties were updated from Unicode 6.2.0
979 23. Improved CMake support under Windows. Patch by Daniel Richard G.
981 24. Add support for 32-bit character strings, and UTF-32
983 25. Major JIT compiler update (code refactoring and bugfixing).
984 Experimental Sparc 32 support is added.
986 26. Applied a modified version of Daniel Richard G's patch to create
987 pcre.h.generic and config.h.generic by "make" instead of in the
988 PrepareRelease script.
990 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
991 pcre_compile.c when checking for a zero character.
993 28. Introducing a native interface for JIT. Through this interface, the compiled
994 machine code can be directly executed. The purpose of this interface is to
995 provide fast pattern matching, so several sanity checks are not performed.
996 However, feature tests are still performed. The new interface provides
997 1.4x speedup compared to the old one.
999 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1000 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1001 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1003 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1004 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1005 the "old" RFC 2279). Instead, it ended up passing a negative length to
1006 pcre_exec().
1008 31. Add support for GCC's visibility feature to hide internal functions.
1010 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1011 "unknown -C option" after outputting 0 or 1.
1013 33. There is now support for generating a code coverage report for the test
1014 suite in environments where gcc is the compiler and lcov is installed. This
1015 is mainly for the benefit of the developers.
1017 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1018 unaddressable using valgrind annotations, allowing valgrind to detect
1019 invalid memory accesses. This is mainly for the benefit of the developers.
1021 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1023 26. Give configure error if --enable-cpp but no C++ compiler found.
1026 Version 8.31 06-July-2012
1027 -------------------------
1029 1. Fixing a wrong JIT test case and some compiler warnings.
1031 2. Removed a bashism from the RunTest script.
1033 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1034 to unsigned type, result still unsigned" that was given by an MS compiler
1035 on encountering the code "-sizeof(xxx)".
1037 4. Partial matching support is added to the JIT compiler.
1039 5. Fixed several bugs concerned with partial matching of items that consist
1040 of more than one character:
1042 (a) /^(..)\1/ did not partially match "aba" because checking references was
1043 done on an "all or nothing" basis. This also applied to repeated
1044 references.
1046 (b) \R did not give a hard partial match if \r was found at the end of the
1047 subject.
1049 (c) \X did not give a hard partial match after matching one or more
1050 characters at the end of the subject.
1052 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1053 a partial match for the string "\r".
1055 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1056 a partial match for a CR character at the end of the subject string.
1058 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1059 running pcretest, the text "(JIT)" added to the output whenever JIT is
1060 actually used to run the match.
1062 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1063 or /S+[+] with a digit between 1 and 7.
1065 8. OP_NOT now supports any UTF character not just single-byte ones.
1067 9. (*MARK) control verb is now supported by the JIT compiler.
1069 10. The command "./RunTest list" lists the available tests without actually
1070 running any of them. (Because I keep forgetting what they all are.)
1074 12. Applied a (slightly modified) user-supplied patch that improves performance
1075 when the heap is used for recursion (compiled with --disable-stack-for-
1076 recursion). Instead of malloc and free for each heap frame each time a
1077 logical recursion happens, frames are retained on a chain and re-used where
1078 possible. This sometimes gives as much as 30% improvement.
1080 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1081 call.
1083 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1085 15. It is now possible to link pcretest with libedit as an alternative to
1086 libreadline.
1088 16. (*COMMIT) control verb is now supported by the JIT compiler.
1090 17. The Unicode data tables have been updated to Unicode 6.1.0.
1092 18. Added --file-list option to pcregrep.
1094 19. Added binary file support to pcregrep, including the -a, --binary-files,
1095 -I, and --text options.
1097 20. The madvise function is renamed for posix_madvise for QNX compatibility
1098 reasons. Fixed by Giuseppe D'Angelo.
1100 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1102 22. Changed the option for creating symbolic links for 16-bit man pages from
1103 -s to -sf so that re-installing does not cause issues.
1105 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1107 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1108 matches in certain environments (the workspace was not being correctly
1109 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1110 some of the workspace data at the beginning of a restart.
1112 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1113 was not doing so when it should - probably a typo introduced by SVN 528
1114 (change 8.10/14).
1116 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1117 \w+ when the character tables indicated that \x{c4} was a word character.
1118 There were several related cases, all because the tests for doing a table
1119 lookup were testing for characters less than 127 instead of 255.
1121 27. If a pattern contains capturing parentheses that are not used in a match,
1122 their slots in the ovector are set to -1. For those that are higher than
1123 any matched groups, this happens at the end of processing. In the case when
1124 there were back references that the ovector was too small to contain
1125 (causing temporary malloc'd memory to be used during matching), and the
1126 highest capturing number was not used, memory off the end of the ovector
1127 was incorrectly being set to -1. (It was using the size of the temporary
1128 memory instead of the true size.)
1130 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1131 ovector size, it uses memory at the end of the block that it has got.
1133 29. Check for an overlong MARK name and give an error at compile time. The
1134 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1136 30. JIT compiler update.
1138 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1139 Rill for the patch.
1141 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1143 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1145 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1146 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1147 was enabled.
1149 35. Improve JIT code generation for greedy plus quantifier.
1151 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1152 1 to "aa" instead of to an empty string. The bug affected repeated groups
1153 that could potentially match an empty string.
1155 37. Optimizing single character iterators in JIT.
1157 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1158 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1159 codepoints that are too big for the mode are faulted, and in a UTF mode,
1160 disallowed codepoints are also faulted.
1162 39. If PCRE was compiled with UTF support, in three places in the DFA
1163 matcher there was code that should only have been obeyed in UTF mode, but
1164 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1165 processing when bytes with values greater than 127 were present. In 16-bit
1166 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1167 both cases the values are those that cannot be the first data item in a UTF
1168 character. The three items that might have provoked this were recursions,
1169 possessively repeated groups, and atomic groups.
1171 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1172 and pcregrep, because some OS require shared objects to be explicitly
1173 passed to ld, causing the link step to fail if they are not.
1175 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1176 mode, patterns that started with \h* or \R* might be incorrectly matched.
1179 Version 8.30 04-February-2012
1180 -----------------------------
1182 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1183 name is defined in ctype.h.
1185 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1186 only in quite long subpatterns.
1188 3. Removed the function pcre_info(), which has been obsolete and deprecated
1189 since it was replaced by pcre_fullinfo() in February 2000.
1191 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1192 match a (*MARK), and the match failed at the start of the subject, a
1193 reference to memory before the start of the subject could occur. This bug
1194 was introduced by fix 17 of release 8.21.
1196 5. A reference to an unset group with zero minimum repetition was giving
1197 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1198 /(another)?(\1?)test/ matched against "hello world test". This bug was
1199 introduced in release 8.13.
1201 6. Add support for 16-bit character strings (a large amount of work involving
1202 many changes and refactorings).
1204 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1205 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1206 from a file.
1208 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1209 rounding is not applied in this particular case).
1211 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1212 if they appear, or are escaped, in patterns.
1214 10. Get rid of a number of -Wunused-but-set-variable warnings.
1216 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1217 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1218 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1219 also returns the mark "x". This bug applied to capturing parentheses,
1220 non-capturing parentheses, and atomic parentheses. It also applied to some
1221 assertions.
1223 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1224 information out of configure.ac instead of relying on pcre.h.generic, which
1225 is not stored in the repository.
1227 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1228 -lreadline.
1230 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1232 15. Applied Graycode's patch to put the top-level frame on the stack rather
1233 than the heap when not using the stack for recursion. This gives a
1234 performance improvement in many cases when recursion is not deep.
1236 16. Experimental code added to "pcretest -C" to output the stack frame size.
1239 Version 8.21 12-Dec-2011
1240 ------------------------
1242 1. Updating the JIT compiler.
1244 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1245 are added as well.
1247 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1248 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1249 calling _pcre_jit_exec. Some extra comments are added.
1251 4. (*MARK) settings inside atomic groups that do not contain any capturing
1252 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1253 was introduced by change 18 for 8.20.
1255 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1256 ECMA-262 standard.
1258 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1259 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1260 This bug was probably introduced by change 9 of 8.13.
1262 7. While fixing 6 above, I noticed that a number of other items were being
1263 incorrectly rejected as "not fixed length". This arose partly because newer
1264 opcodes had not been added to the fixed-length checking code. I have (a)
1265 corrected the bug and added tests for these items, and (b) arranged for an
1266 error to occur if an unknown opcode is encountered while checking for fixed
1267 length instead of just assuming "not fixed length". The items that were
1268 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1269 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1270 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1272 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1273 being incorrectly compiled and would have given unpredicatble results.
1275 9. A possessively repeated subpattern with minimum repeat count greater than
1276 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1277 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1278 the first (A) could occur when it should not.
1280 10. Add a cast and remove a redundant test from the code.
1282 11. JIT should use pcre_malloc/pcre_free for allocation.
1284 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1285 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1286 is anything other than /usr, -L is still shown).
1288 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1290 14. Perl does not support \N without a following name in a [] class; PCRE now
1291 also gives an error.
1293 15. If a forward reference was repeated with an upper limit of around 2000,
1294 it caused the error "internal error: overran compiling workspace". The
1295 maximum number of forward references (including repeats) was limited by the
1296 internal workspace, and dependent on the LINK_SIZE. The code has been
1297 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1298 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1299 of around 200,000 forward references. While doing this, I also speeded up
1300 the filling in of repeated forward references.
1302 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1303 incorrectly expecting the subject to contain another "a" after the start.
1305 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1306 in the match, the SKIP should be ignored. This was not happening; instead
1307 the SKIP was being treated as NOMATCH. For patterns such as
1308 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1309 tested.
1311 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1312 now much more compatible with Perl, in particular in cases where the result
1313 is a non-match for a non-anchored pattern. For example, if
1314 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1315 "m", where previously it did not return a name. A side effect of this
1316 change is that for partial matches, the last encountered mark name is
1317 returned, as for non matches. A number of tests that were previously not
1318 Perl-compatible have been moved into the Perl-compatible test files. The
1319 refactoring has had the pleasing side effect of removing one argument from
1320 the match() function, thus reducing its stack requirements.
1322 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1323 subsequent uses of /S (without +) incorrectly behaved like /S+.
1325 21. Retrieve executable code size support for the JIT compiler and fixing
1326 some warnings.
1328 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1329 not work when the shorter character appeared right at the end of the
1330 subject string.
1332 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1333 systems.
1335 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1336 output it when the /M option is used in pcretest.
1338 25. The CheckMan script was not being included in the distribution. Also, added
1339 an explicit "perl" to run Perl scripts from the PrepareRelease script
1340 because this is reportedly needed in Windows.
1342 26. If study data was being save in a file and studying had not found a set of
1343 "starts with" bytes for the pattern, the data written to the file (though
1344 never used) was taken from uninitialized memory and so caused valgrind to
1345 complain.
1347 27. Updated RunTest.bat as provided by Sheri Pierce.
1349 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1351 29. Computation of memory usage for the table of capturing group names was
1352 giving an unnecessarily large value.
1355 Version 8.20 21-Oct-2011
1356 ------------------------
1358 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1359 a POSIX class. After further experiments with Perl, which convinced me that
1360 Perl has bugs and confusions, a closing square bracket is no longer allowed
1361 in a POSIX name. This bug also affected patterns with classes that started
1362 with full stops.
1364 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1365 captured substring, but while checking the failing first alternative,
1366 substring 1 is temporarily captured. If the output vector supplied to
1367 pcre_exec() was not big enough for this capture, the yield of the function
1368 was still zero ("insufficient space for captured substrings"). This cannot
1369 be totally fixed without adding another stack variable, which seems a lot
1370 of expense for a edge case. However, I have improved the situation in cases
1371 such as /(a)(b)x|abc/ matched against "abc", where the return code
1372 indicates that fewer than the maximum number of slots in the ovector have
1373 been set.
1375 3. Related to (2) above: when there are more back references in a pattern than
1376 slots in the output vector, pcre_exec() uses temporary memory during
1377 matching, and copies in the captures as far as possible afterwards. It was
1378 using the entire output vector, but this conflicts with the specification
1379 that only 2/3 is used for passing back captured substrings. Now it uses
1380 only the first 2/3, for compatibility. This is, of course, another edge
1381 case.
1383 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1384 main code base, and can be used by building with --enable-jit. When this is
1385 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1386 runtime --no-jit option is given.
1388 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1389 ovector, the return from the function was zero, implying that there were
1390 other matches that did not fit. The correct "exactly full" value is now
1391 returned.
1393 6. If a subpattern that was called recursively or as a subroutine contained
1394 (*PRUNE) or any other control that caused it to give a non-standard return,
1395 invalid errors such as "Error -26 (nested recursion at the same subject
1396 position)" or even infinite loops could occur.
1398 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1399 computing the minimum length on reaching *ACCEPT, and so ended up with the
1400 wrong value of 1 rather than 0. Further investigation indicates that
1401 computing a minimum subject length in the presence of *ACCEPT is difficult
1402 (think back references, subroutine calls), and so I have changed the code
1403 so that no minimum is registered for a pattern that contains *ACCEPT.
1405 8. If (*THEN) was present in the first (true) branch of a conditional group,
1406 it was not handled as intended. [But see 16 below.]
1408 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1409 Sheri Pierce.
1411 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1412 the first byte in a match must be "a".
1414 11. Change 17 for 8.13 increased the recursion depth for patterns like
1415 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1416 pattern contains any instances of (*THEN). If it does not, the old
1417 optimizations are restored. It would be nice to do this on a per-group
1418 basis, but at the moment that is not feasible.
1420 12. In some environments, the output of pcretest -C is CRLF terminated. This
1421 broke RunTest's code that checks for the link size. A single white space
1422 character after the value is now allowed for.
1424 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1425 For "fr", it uses the Windows-specific input and output files.
1427 14. If (*THEN) appeared in a group that was called recursively or as a
1428 subroutine, it did not work as intended. [But see next item.]
1430 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1431 pattern fragments (but not containing any | characters). If A and B are
1432 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1433 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1434 D. In other words, Perl considers parentheses that do not contain any |
1435 characters to be part of a surrounding alternative, whereas PCRE was
1436 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1437 differently. PCRE now behaves in the same way as Perl, except in the case
1438 of subroutine/recursion calls such as (?1) which have in any case always
1439 been different (but PCRE had them first :-).
1441 16. Related to 15 above: Perl does not treat the | in a conditional group as
1442 creating alternatives. Such a group is treated in the same way as an
1443 ordinary group without any | characters when processing (*THEN). PCRE has
1444 been changed to match Perl's behaviour.
1446 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1447 RunGrepTest script failed.
1449 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1450 inevitable for groups that contain captures, but it can lead to a lot of
1451 stack use in large patterns. The old behaviour has been restored for atomic
1452 groups that do not contain any capturing parentheses.
1454 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1455 suppress the check for a minimum subject length at run time. (If it was
1456 given to pcre_exec() or pcre_dfa_exec() it did work.)
1458 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1459 fail to work when decoding hex characters in data strings in EBCDIC
1460 environments.
1462 21. It appears that in at least one Mac OS environment, the isxdigit() function
1463 is implemented as a macro that evaluates to its argument more than once,
1464 contravening the C 90 Standard (I haven't checked a later standard). There
1465 was an instance in pcretest which caused it to go wrong when processing
1466 \x{...} escapes in subject strings. The has been rewritten to avoid using
1467 things like p++ in the argument of isxdigit().
1470 Version 8.13 16-Aug-2011
1471 ------------------------
1473 1. The Unicode data tables have been updated to Unicode 6.0.0.
1475 2. Two minor typos in pcre_internal.h have been fixed.
1477 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1478 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1479 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1481 4. There were a number of related bugs in the code for matching backrefences
1482 caselessly in UTF-8 mode when codes for the characters concerned were
1483 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1484 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1485 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1486 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1487 2-byte code at the end of the subject (it thought there wasn't enough data
1488 left).
1490 5. Comprehensive information about what went wrong is now returned by
1491 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1492 as the output vector has at least 2 elements. The offset of the start of
1493 the failing character and a reason code are placed in the vector.
1495 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1496 now returned is for the first byte of the failing character, instead of the
1497 last byte inspected. This is an incompatible change, but I hope it is small
1498 enough not to be a problem. It makes the returned offset consistent with
1499 pcre_exec() and pcre_dfa_exec().
1501 7. pcretest now gives a text phrase as well as the error number when
1502 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1503 failure, the offset and reason code are output.
1505 8. When \R was used with a maximizing quantifier it failed to skip backwards
1506 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1507 back over a single character (\n). This seems wrong (because it treated the
1508 two characters as a single entity when going forwards), conflicts with the
1509 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1510 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1511 has been changed.
1513 9. Some internal refactoring has changed the processing so that the handling
1514 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1515 time (the PCRE_DOTALL option was changed this way some time ago: version
1516 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1517 which was always a bit of a fudge. It also means that there is one less
1518 argument for the match() function, which reduces its stack requirements
1519 slightly. This change also fixes an incompatibility with Perl: the pattern
1520 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1522 10. More internal refactoring has drastically reduced the number of recursive
1523 calls to match() for possessively repeated groups such as (abc)++ when
1524 using pcre_exec().
1526 11. While implementing 10, a number of bugs in the handling of groups were
1527 discovered and fixed:
1529 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1530 (a|)*(?1) gave a compile-time internal error.
1531 ((a|)+)+ did not notice that the outer group could match an empty string.
1532 (^a|^)+ was not marked as anchored.
1533 (.*a|.*)+ was not marked as matching at start or after a newline.
1535 12. Yet more internal refactoring has removed another argument from the match()
1536 function. Special calls to this function are now indicated by setting a
1537 value in a variable in the "match data" data block.
1539 13. Be more explicit in pcre_study() instead of relying on "default" for
1540 opcodes that mean there is no starting character; this means that when new
1541 ones are added and accidentally left out of pcre_study(), testing should
1542 pick them up.
1544 14. The -s option of pcretest has been documented for ages as being an old
1545 synonym of -m (show memory usage). I have changed it to mean "force study
1546 for every regex", that is, assume /S for every regex. This is similar to -i
1547 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1548 using it. It makes it easier to run collections of tests with and without
1549 study enabled, and thereby test pcre_study() more easily. All the standard
1550 tests are now run with and without -s (but some patterns can be marked as
1551 "never study" - see 20 below).
1553 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1554 restoration of the capturing data to the outer values was not happening
1555 correctly.
1557 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1558 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1559 pattern had matched an empty string, and so incorrectly returned a no
1560 match.
1562 17. There was optimizing code for the last branch of non-capturing parentheses,
1563 and also for the obeyed branch of a conditional subexpression, which used
1564 tail recursion to cut down on stack usage. Unfortunately, now that there is
1565 the possibility of (*THEN) occurring in these branches, tail recursion is
1566 no longer possible because the return has to be checked for (*THEN). These
1567 two optimizations have therefore been removed. [But see 8.20/11 above.]
1569 18. If a pattern containing \R was studied, it was assumed that \R always
1570 matched two bytes, thus causing the minimum subject length to be
1571 incorrectly computed because \R can also match just one byte.
1573 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1574 was incorrectly computed.
1576 20. If /S is present twice on a test pattern in pcretest input, it now
1577 *disables* studying, thereby overriding the use of -s on the command line
1578 (see 14 above). This is necessary for one or two tests to keep the output
1579 identical in both cases.
1581 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1582 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1584 22. When an atomic group that contained a capturing parenthesis was
1585 successfully matched, but the branch in which it appeared failed, the
1586 capturing was not being forgotten if a higher numbered group was later
1587 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1588 group 1 to "a", when in fact it should be unset. This applied to multi-
1589 branched capturing and non-capturing groups, repeated or not, and also to
1590 positive assertions (capturing in negative assertions does not happen
1591 in PCRE) and also to nested atomic groups.
1593 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1594 subject after a captured substring, to make it easier to tell which of a
1595 number of identical substrings has been captured.
1597 24. The way atomic groups are processed by pcre_exec() has been changed so that
1598 if they are repeated, backtracking one repetition now resets captured
1599 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1600 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1601 "aaa". Previously, it would have been "a". As part of this code
1602 refactoring, the way recursive calls are handled has also been changed.
1604 25. If an assertion condition captured any substrings, they were not passed
1605 back unless some other capturing happened later. For example, if
1606 (?(?=(a))a) was matched against "a", no capturing was returned.
1608 26. When studying a pattern that contained subroutine calls or assertions,
1609 the code for finding the minimum length of a possible match was handling
1610 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1611 group 1 called group 2 while simultaneously a separate group 2 called group
1612 1). A stack overflow occurred in this case. I have fixed this by limiting
1613 the recursion depth to 10.
1615 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1616 Fortmann. This supports explicit test numbers on the command line, and has
1617 argument validation and error reporting.
1619 28. An instance of \X with an unlimited repeat could fail if at any point the
1620 first character it looked at was a mark character.
1622 29. Some minor code refactoring concerning Unicode properties and scripts
1623 should reduce the stack requirement of match() slightly.
1625 30. Added the '=' option to pcretest to check the setting of unused capturing
1626 slots at the end of the pattern, which are documented as being -1, but are
1627 not included in the return count.
1629 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1630 compiled something random. Now it gives a compile-time error (as does
1631 Perl).
1633 32. A *MARK encountered during the processing of a positive assertion is now
1634 recorded and passed back (compatible with Perl).
1636 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1637 had alternative anchored branches, the search for a second match in a line
1638 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1639 matched the line "0102" twice. The same bug affected patterns that started
1640 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1641 twice.
1643 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1644 does, and because of capturing effects, quantifying parenthesized
1645 assertions may at times be useful. Quantifiers are now allowed for
1646 parenthesized assertions.
1648 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1650 36. \g was being checked for fancy things in a character class, when it should
1651 just be a literal "g".
1653 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1654 appearance of a nested POSIX class supersedes an apparent external class.
1655 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1656 unescaped square brackets may also appear as part of class names. For
1657 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1658 more like Perl. (But see 8.20/1 above.)
1660 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1661 was because it thought it was \N{name}, which is not supported).
1663 39. Add minix to OS list not supporting the -S option in pcretest.
1665 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1666 cannot analyze patterns in sufficient detail to catch mutual recursions
1667 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1668 subgroup is called recursively as a subpattern for a second time at the
1669 same position in the subject string. In previous releases this might have
1670 been caught by the recursion limit, or it might have run out of stack.
1672 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1673 happen only once. PCRE was, however incorrectly giving a compile time error
1674 "recursive call could loop indefinitely" because it cannot analyze the
1675 pattern in sufficient detail. The compile time test no longer happens when
1676 PCRE is compiling a conditional subpattern, but actual runaway loops are
1677 now caught at runtime (see 40 above).
1679 42. It seems that Perl allows any characters other than a closing parenthesis
1680 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1681 has been changed to be the same.
1683 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1684 as not to get warnings when autogen.sh is called. Also changed
1685 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1687 44. To help people who use pcregrep to scan files containing exceedingly long
1688 lines, the following changes have been made:
1690 (a) The default value of the buffer size parameter has been increased from
1691 8K to 20K. (The actual buffer used is three times this size.)
1693 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1694 PCRE is built.
1696 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1697 to be set at run time.
1699 (d) Numerical values in pcregrep options can be followed by K or M, for
1700 example --buffer-size=50K.
1702 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1703 given and the return code is set to 2.
1705 45. Add a pointer to the latest mark to the callout data block.
1707 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1708 partial match of an empty string instead of no match. This was specific to
1709 the use of ".".
1711 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1712 complete match instead of a partial match. This bug was dependent on both
1713 the PCRE_UTF8 and PCRE_DOTALL options being set.
1715 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1716 starting byte set, because \b was not being ignored.
1719 Version 8.12 15-Jan-2011
1720 ------------------------
1722 1. Fixed some typos in the markup of the man pages, and wrote a script that
1723 checks for such things as part of the documentation building process.
1725 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1726 --match-limit and --recursion-limit options (added for 8.11). In
1727 particular, this made one of the standard tests fail. (The integer value
1728 went into the wrong half of a long int.)
1730 3. If the --colour option was given to pcregrep with -v (invert match), it
1731 did strange things, either producing crazy output, or crashing. It should,
1732 of course, ignore a request for colour when reporting lines that do not
1733 match.
1735 4. Another pcregrep bug caused similar problems if --colour was specified with
1736 -M (multiline) and the pattern match finished with a line ending.
1738 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1739 matched in multiline mode, the following line was shown as part of the
1740 match. This seems wrong, so I have changed it.
1742 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1743 the check for further matches in the same line (so they could be coloured)
1744 to overrun the end of the current line. If another match was found, it was
1745 incorrectly shown (and then shown again when found in the next line).
1747 7. If pcregrep was compiled under Windows, there was a reference to the
1748 function pcregrep_exit() before it was defined. I am assuming this was
1749 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1750 reported by a user. I've moved the definition above the reference.
1753 Version 8.11 10-Dec-2010
1754 ------------------------
1756 1. (*THEN) was not working properly if there were untried alternatives prior
1757 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1758 backtracked to try for "b" instead of moving to the next alternative branch
1759 at the same level (in this case, to look for "c"). The Perl documentation
1760 is clear that when (*THEN) is backtracked onto, it goes to the "next
1761 alternative in the innermost enclosing group".
1763 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1764 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1765 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1766 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1767 (*THEN).
1769 3. If \s appeared in a character class, it removed the VT character from
1770 the class, even if it had been included by some previous item, for example
1771 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1772 of \s, but is part of the POSIX "space" class.)
1774 4. A partial match never returns an empty string (because you can always
1775 match an empty string at the end of the subject); however the checking for
1776 an empty string was starting at the "start of match" point. This has been
1777 changed to the "earliest inspected character" point, because the returned
1778 data for a partial match starts at this character. This means that, for
1779 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1780 (previously it gave "no match").
1782 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1783 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1784 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1785 has an implication that the given string is incomplete (because a partial
1786 match is preferred over a full match). For this reason, these items now
1787 give a partial match in this situation. [Aside: previously, the one case
1788 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1789 match rather than a full match, which was wrong by the old rules, but is
1790 now correct.]
1792 6. There was a bug in the handling of #-introduced comments, recognized when
1793 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1794 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1795 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1796 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1797 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1798 places in pcre_compile().
1800 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1801 comments when looking ahead for named forward references to subpatterns,
1802 the only newline sequence it recognized was NL. It now handles newlines
1803 according to the set newline convention.
1805 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1806 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1807 cater for a lack of strerror(). These oversights have been fixed.
1809 9. Added --match-limit and --recursion-limit to pcregrep.
1811 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1813 11. When the -o option was used, pcregrep was setting a return code of 1, even
1814 when matches were found, and --line-buffered was not being honoured.
1816 12. Added an optional parentheses number to the -o and --only-matching options
1817 of pcregrep.
1819 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1820 can match an empty string. The code to do it in pcretest and pcredemo
1821 needed fixing:
1823 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1824 only one byte after an empty string match just before CRLF (this case
1825 just got forgotten; "any" and "anycrlf" were OK).
1827 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1828 mode when an empty string match preceded an ASCII character followed by
1829 a non-ASCII character. (The code for advancing by one character rather
1830 than one byte was nonsense.)
1832 (c) The pcredemo.c sample program did not have any code at all to handle
1833 the cases when CRLF is a valid newline sequence.
1835 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1836 as a starting offset was within the subject string. There is now a new
1837 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1838 negative or greater than the length of the string. In order to test this,
1839 pcretest is extended to allow the setting of negative starting offsets.
1841 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1842 starting offset points to the beginning of a UTF-8 character was
1843 unnecessarily clumsy. I tidied it up.
1845 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1846 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1848 17. Nobody had reported that the --include_dir option, which was added in
1849 release 7.7 should have been called --include-dir (hyphen, not underscore)
1850 for compatibility with GNU grep. I have changed it to --include-dir, but
1851 left --include_dir as an undocumented synonym, and the same for
1852 --exclude-dir, though that is not available in GNU grep, at least as of
1853 release 2.5.4.
1855 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1856 characters from a string of bytes) have been redefined so as not to use
1857 loops, in order to improve performance in some environments. At the same
1858 time, I abstracted some of the common code into auxiliary macros to save
1859 repetition (this should not affect the compiled code).
1861 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1862 compile-time error is now given if \c is not followed by an ASCII
1863 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1864 different, and any byte value is allowed.)
1866 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1867 START_OPTIMIZE option, which is now allowed at compile time - but just
1868 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1869 to pcregrep and other applications that have no direct access to PCRE
1870 options. The new /Y option in pcretest sets this option when calling
1871 pcre_compile().
1873 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1874 back references. Groups containing recursive back references were forced to
1875 be atomic by that change, but in the case of named groups, the amount of
1876 memory required was incorrectly computed, leading to "Failed: internal
1877 error: code overflow". This has been fixed.
1879 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1880 pcretest.c, to avoid build problems in some Borland environments.
1883 Version 8.10 25-Jun-2010
1884 ------------------------
1886 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1887 THEN.
1889 2. (*ACCEPT) was not working when inside an atomic group.
1891 3. Inside a character class, \B is treated as a literal by default, but
1892 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1893 causes the error). The code is unchanged, but I tidied the documentation.
1895 4. Inside a character class, PCRE always treated \R and \X as literals,
1896 whereas Perl faults them if its -w option is set. I have changed PCRE so
1897 that it faults them when PCRE_EXTRA is set.
1899 5. Added support for \N, which always matches any character other than
1900 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1902 6. When compiling pcregrep with newer versions of gcc which may have
1903 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1904 declared with attribute warn_unused_result" were given. Just casting the
1905 result to (void) does not stop the warnings; a more elaborate fudge is
1906 needed. I've used a macro to implement this.
1908 7. Minor change to pcretest.c to avoid a compiler warning.
1910 8. Added four artifical Unicode properties to help with an option to make
1911 \s etc use properties (see next item). The new properties are: Xan
1912 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1914 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1915 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1916 this option. Modified pcretest to add /W to test this facility. Added
1917 REG_UCP to make it available via the POSIX interface.
1919 10. Added --line-buffered to pcregrep.
1921 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1922 studied, and the match started with a letter with a code point greater than
1923 127 whose first byte was different to the first byte of the other case of
1924 the letter, the other case of this starting letter was not recognized
1925 (#976).
1927 12. If a pattern that was studied started with a repeated Unicode property
1928 test, for example, \p{Nd}+, there was the theoretical possibility of
1929 setting up an incorrect bitmap of starting bytes, but fortunately it could
1930 not have actually happened in practice until change 8 above was made (it
1931 added property types that matched character-matching opcodes).
1933 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1934 possible starting bytes for non-anchored patterns.
1936 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1937 \R, and also a number of cases that involve Unicode properties, both
1938 explicit and implicit when PCRE_UCP is set.
1940 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1941 input, it could crash or give wrong results if characters with values
1942 greater than 0xc0 were present in the subject string. (Detail: it assumed
1943 UTF-8 input when processing these items.)
1945 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1946 size_t is 64-bit (#991).
1948 17. Added a check for running out of memory when PCRE is compiled with
1949 --disable-stack-for-recursion (#990).
1951 18. If the last data line in a file for pcretest does not have a newline on
1952 the end, a newline was missing in the output.
1954 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1955 less than 128) in its various bitmaps. However, there is a facility for
1956 generating tables according to the current locale when PCRE is compiled. It
1957 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1958 characters, are recognized by isspace() and therefore were getting set in
1959 these tables, and indeed these tables seem to approximate to ISO 8859. This
1960 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1961 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1962 which of course cannot start UTF-8 characters. I have changed the code so
1963 that only real ASCII characters (less than 128) and the correct starting
1964 bytes for UTF-8 encodings are set for characters greater than 127 when in
1965 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1966 altogether.)
1968 20. Added the /T option to pcretest so as to be able to run tests with non-
1969 standard character tables, thus making it possible to include the tests
1970 used for 19 above in the standard set of tests.
1972 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1973 reference to a subpattern the other side of a comment that contains an
1974 opening parenthesis caused either an internal compiling error, or a
1975 reference to the wrong subpattern.
1978 Version 8.02 19-Mar-2010
1979 ------------------------
1981 1. The Unicode data tables have been updated to Unicode 5.2.0.
1983 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1984 configured.
1986 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1987 original author of that file, following a query about its status.
1989 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1990 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1992 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1993 quantifier applied to a forward-referencing subroutine call, could compile
1994 incorrect code or give the error "internal error: previously-checked
1995 referenced subpattern not found".
1997 6. Both MS Visual Studio and Symbian OS have problems with initializing
1998 variables to point to external functions. For these systems, therefore,
1999 pcre_malloc etc. are now initialized to local functions that call the
2000 relevant global functions.
2002 7. There were two entries missing in the vectors called coptable and poptable
2003 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2004 I've fixed the data, and added a kludgy way of testing at compile time that
2005 the lengths are correct (equal to the number of opcodes).
2007 8. Following on from 7, I added a similar kludge to check the length of the
2008 eint vector in pcreposix.c.
2010 9. Error texts for pcre_compile() are held as one long string to avoid too
2011 much relocation at load time. To find a text, the string is searched,
2012 counting zeros. There was no check for running off the end of the string,
2013 which could happen if a new error number was added without updating the
2014 string.
2016 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2018 11. \K was not working if it appeared in an atomic group or in a group that
2019 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2020 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2021 the assertion is positive, but not if it is negative.
2023 12. Change 11 fortuitously reduced the size of the stack frame used in the
2024 "match()" function of pcre_exec.c by one pointer. Forthcoming
2025 implementation of support for (*MARK) will need an extra pointer on the
2026 stack; I have reserved it now, so that the stack frame size does not
2027 decrease.
2029 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2030 item in branch that calls a recursion is a subroutine call - as in the
2031 second branch in the above example - was incorrectly given the compile-
2032 time error "recursive call could loop indefinitely" because pcre_compile()
2033 was not correctly checking the subroutine for matching a non-empty string.
2035 14. The checks for overrunning compiling workspace could trigger after an
2036 overrun had occurred. This is a "should never occur" error, but it can be
2037 triggered by pathological patterns such as hundreds of nested parentheses.
2038 The checks now trigger 100 bytes before the end of the workspace.
2040 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2043 Version 8.01 19-Jan-2010
2044 ------------------------
2046 1. If a pattern contained a conditional subpattern with only one branch (in
2047 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2048 computed the wrong minimum data length (which is of course zero for such
2049 subpatterns). This could cause incorrect "no match" results.
2051 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2052 the pattern is reset in the first branch, pcre_compile() failed with
2053 "internal error: code overflow at offset...". This happened only when
2054 the reset was to the original external option setting. (An optimization
2055 abstracts leading options settings into an external setting, which was the
2056 cause of this.)
2058 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2059 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2060 assertion pattern did not match (meaning that the assertion was true), it
2061 was incorrectly treated as false if the SKIP had been reached during the
2062 matching. This also applied to assertions used as conditions.
2064 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2065 assertion subpattern, including such a pattern used as a condition,
2066 unpredictable results occurred, instead of the error return
2069 5. The C++ GlobalReplace function was not working like Perl for the special
2070 situation when an empty string is matched. It now does the fancy magic
2071 stuff that is necessary.
2073 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2074 removed. (These were left over from very, very early versions of PCRE.)
2076 7. Some cosmetic changes to the code to make life easier when compiling it
2077 as part of something else:
2079 (a) Change DEBUG to PCRE_DEBUG.
2081 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2082 called "current" as "current_branch", to prevent a collision with the
2083 Linux macro when compiled as a kernel module.
2085 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2086 prevent a collision with the Linux macro when compiled as a kernel
2087 module.
2089 8. In pcre_compile() there are some checks for integer overflows that used to
2090 cast potentially large values to (double). This has been changed to that
2091 when building, a check for int64_t is made, and if it is found, it is used
2092 instead, thus avoiding the use of floating point arithmetic. (There is no
2093 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2094 double.
2096 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2097 2005 (difference between two addresses compared to an unsigned value).
2099 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2100 custom one, because of the following reported problem in Windows:
2102 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2103 under Win32.
2104 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2105 therefore missing the function definition.
2106 - The compiler thus generates a "C" signature for the test function.
2107 - The linker fails to find the "C" function.
2108 - PCRE fails to configure if asked to do so against libbz2.
2110 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2111 messages were output:
2113 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2114 rerunning libtoolize, to keep the correct libtool macros in-tree.
2115 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2117 I have done both of these things.
2119 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2120 most of the time, it *can* run out if it is given a pattern that contains a
2121 runaway infinite recursion. I updated the discussion in the pcrestack man
2122 page.
2124 13. Now that we have gone to the x.xx style of version numbers, the minor
2125 version may start with zero. Using 08 or 09 is a bad idea because users
2126 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2127 interpreted as invalid octal numbers. I've updated the previous comment in
2128 configure.ac, and also added a check that gives an error if 08 or 09 are
2129 used.
2131 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2132 causing partial matching to fail when the end of the subject matched \W
2133 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2135 15. There were some discrepancies between the declarations in pcre_internal.h
2136 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2137 their definitions. The declarations used "const uschar *" and the
2138 definitions used USPTR. Even though USPTR is normally defined as "const
2139 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2140 reported that: "This difference in casting confuses some C++ compilers, for
2141 example, SunCC recognizes above declarations as different functions and
2142 generates broken code for hbpcre." I have changed the declarations to use
2143 USPTR.
2145 16. GNU libtool is named differently on some systems. The autogen.sh script now
2146 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2147 (FreeBSD).
2149 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2150 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2151 comment: "Figure out how to create a longlong from a string: strtoll and
2152 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2153 instance, but it only takes 2 args instead of 3!"
2155 18. A subtle bug concerned with back references has been fixed by a change of
2156 specification, with a corresponding code fix. A pattern such as
2157 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2158 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2159 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2160 same bug. Such groups have to be quantified to be useful, or contained
2161 inside another quantified group. (If there's no repetition, the reference
2162 can never match.) The problem arises because, having left the group and
2163 moved on to the rest of the pattern, a later failure that backtracks into
2164 the group uses the captured value from the final iteration of the group
2165 rather than the correct earlier one. I have fixed this in PCRE by forcing
2166 any group that contains a reference to itself to be an atomic group; that
2167 is, there cannot be any backtracking into it once it has completed. This is
2168 similar to recursive and subroutine calls.
2171 Version 8.00 19-Oct-09
2172 ----------------------
2174 1. The table for translating pcre_compile() error codes into POSIX error codes
2175 was out-of-date, and there was no check on the pcre_compile() error code
2176 being within the table. This could lead to an OK return being given in
2177 error.
2179 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2180 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2181 in a Windows environment.
2183 3. The pcregrep --count option prints the count for each file even when it is
2184 zero, as does GNU grep. However, pcregrep was also printing all files when
2185 --files-with-matches was added. Now, when both options are given, it prints
2186 counts only for those files that have at least one match. (GNU grep just
2187 prints the file name in this circumstance, but including the count seems
2188 more useful - otherwise, why use --count?) Also ensured that the
2189 combination -clh just lists non-zero counts, with no names.
2191 4. The long form of the pcregrep -F option was incorrectly implemented as
2192 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2193 but it seems right to fix it, and I didn't think it was worth preserving
2194 the old behaviour.
2196 5. The command line items --regex=pattern and --regexp=pattern were not
2197 recognized by pcregrep, which required --regex pattern or --regexp pattern
2198 (with a space rather than an '='). The man page documented the '=' forms,
2199 which are compatible with GNU grep; these now work.
2201 6. No libpcreposix.pc file was created for pkg-config; there was just
2202 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2204 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2205 when UCP support is not needed, by modifying the Python script that
2206 generates it from Unicode data files. This should not matter if the module
2207 is correctly used as a library, but I received one complaint about 50K of
2208 unwanted data. My guess is that the person linked everything into his
2209 program rather than using a library. Anyway, it does no harm.
2211 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2212 was a minimum greater than 1 for a wide character in a possessive
2213 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2214 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2215 character. Chaos in the form of incorrect output or a compiling loop could
2216 result.
2218 9. The restrictions on what a pattern can contain when partial matching is
2219 requested for pcre_exec() have been removed. All patterns can now be
2220 partially matched by this function. In addition, if there are at least two
2221 slots in the offset vector, the offset of the earliest inspected character
2222 for the match and the offset of the end of the subject are set in them when
2223 PCRE_ERROR_PARTIAL is returned.
2225 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2226 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2227 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2228 and may be more useful for multi-segment matching.
2230 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2231 used to be given if ever the end of the subject was reached; now it is
2232 given only if matching could not proceed because another character was
2233 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2234 string "Z", which now yields "no match" instead of "partial match". In the
2235 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2236 final character ended with (*FAIL).
2238 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2239 if the pattern had a "must contain" character that was already found in the
2240 earlier partial match, unless partial matching was again requested. For
2241 example, with the pattern /dog.(body)?/, the "must contain" character is
2242 "g". If the first part-match was for the string "dog", restarting with
2243 "sbody" failed. This bug has been fixed.
2245 13. The string returned by pcre_dfa_exec() after a partial match has been
2246 changed so that it starts at the first inspected character rather than the
2247 first character of the match. This makes a difference only if the pattern
2248 starts with a lookbehind assertion or \b or \B (\K is not supported by
2249 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2250 matching functions compatible, and I think it's the right thing to do.
2252 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2253 so that the demonstration program is easily available in environments where
2254 PCRE has not been installed from source.
2256 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2257 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2258 library.
2260 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2261 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2262 is not the first non-POSIX option to be added. Clearly some people find
2263 these options useful.
2265 17. If a caller to the POSIX matching function regexec() passes a non-zero
2266 value for nmatch with a NULL value for pmatch, the value of
2267 nmatch is forced to zero.
2269 18. RunGrepTest did not have a test for the availability of the -u option of
2270 the diff command, as RunTest does. It now checks in the same way as
2271 RunTest, and also checks for the -b option.
2273 19. If an odd number of negated classes containing just a single character
2274 interposed, within parentheses, between a forward reference to a named
2275 subpattern and the definition of the subpattern, compilation crashed with
2276 an internal error, complaining that it could not find the referenced
2277 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2278 [The bug was that it was starting one character too far in when skipping
2279 over the character class, thus treating the ] as data rather than
2280 terminating the class. This meant it could skip too much.]
2282 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2283 /g option in pcretest when the pattern contains \K, which makes it possible
2284 to have an empty string match not at the start, even when the pattern is
2285 anchored. Updated pcretest and pcredemo to use this option.
2287 21. If the maximum number of capturing subpatterns in a recursion was greater
2288 than the maximum at the outer level, the higher number was returned, but
2289 with unset values at the outer level. The correct (outer level) value is
2290 now given.
2292 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2293 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2294 make it do so. The string so far is captured, making this feature
2295 compatible with Perl.
2297 23. The tests have been re-organized, adding tests 11 and 12, to make it
2298 possible to check the Perl 5.10 features against Perl 5.10.
2300 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2301 pattern matches a fixed length string. PCRE did not allow this; now it
2302 does. Neither allows recursion.
2304 25. I finally figured out how to implement a request to provide the minimum
2305 length of subject string that was needed in order to match a given pattern.
2306 (It was back references and recursion that I had previously got hung up
2307 on.) This code has now been added to pcre_study(); it finds a lower bound
2308 to the length of subject needed. It is not necessarily the greatest lower
2309 bound, but using it to avoid searching strings that are too short does give
2310 some useful speed-ups. The value is available to calling programs via
2311 pcre_fullinfo().
2313 26. While implementing 25, I discovered to my embarrassment that pcretest had
2314 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2315 study optimizations had never been tested with that matching function.
2316 Oops. What is worse, even when it was passed study data, there was a bug in
2317 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2318 were also very few tests of studied patterns with pcre_dfa_exec().
2320 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2321 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2322 on the other side of the coin, they are no longer allowed to have different
2323 names, because these cannot be distinguished in PCRE, and this has caused
2324 confusion. (This is a difference from Perl.)
2326 28. When duplicate subpattern names are present (necessarily with different
2327 numbers, as required by 27 above), and a test is made by name in a
2328 conditional pattern, either for a subpattern having been matched, or for
2329 recursion in such a pattern, all the associated numbered subpatterns are
2330 tested, and the overall condition is true if the condition is true for any
2331 one of them. This is the way Perl works, and is also more like the way
2332 testing by number works.
2335 Version 7.9 11-Apr-09
2336 ---------------------
2338 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2339 (pcretest), all targets were linked against these libraries. This included
2340 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2341 libraries. This caused unwanted dependencies to be created. This problem
2342 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2343 pcretest is linked with readline.
2345 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2346 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2347 moved outside it again, because FALSE and TRUE are already defined in AIX,
2348 but BOOL is not.
2350 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2351 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2353 4. The pcregrep documentation said spaces were inserted as well as colons (or
2354 hyphens) following file names and line numbers when outputting matching
2355 lines. This is not true; no spaces are inserted. I have also clarified the
2356 wording for the --colour (or --color) option.
2358 5. In pcregrep, when --colour was used with -o, the list of matching strings
2359 was not coloured; this is different to GNU grep, so I have changed it to be
2360 the same.
2362 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2363 each matching line was coloured. Now it goes on to look for further matches
2364 of any of the test patterns, which is the same behaviour as GNU grep.
2366 7. A pattern that could match an empty string could cause pcregrep to loop; it
2367 doesn't make sense to accept an empty string match in pcregrep, so I have
2368 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2369 seems to be how GNU grep behaves. [But see later change 40 for release
2370 8.33.]
2372 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2373 start or after a newline", because the conditional assertion was not being
2374 correctly handled. The rule now is that both the assertion and what follows
2375 in the first alternative must satisfy the test.
2377 9. If auto-callout was enabled in a pattern with a conditional group whose
2378 condition was an assertion, PCRE could crash during matching, both with
2379 pcre_exec() and pcre_dfa_exec().
2381 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2382 used for matching.
2384 11. Unicode property support in character classes was not working for
2385 characters (bytes) greater than 127 when not in UTF-8 mode.
2387 12. Added the -M command line option to pcretest.
2389 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2391 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2393 16. Added comments and documentation about mis-use of no_arg in the C++
2394 wrapper.
2396 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2397 from Martin Jerabek that uses macro names for all relevant character and
2398 string constants.
2400 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2401 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2402 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2403 these, but not everybody uses configure.
2405 19. A conditional group that had only one branch was not being correctly
2406 recognized as an item that could match an empty string. This meant that an
2407 enclosing group might also not be so recognized, causing infinite looping
2408 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2409 with the subject "ab", where knowledge that the repeated group can match
2410 nothing is needed in order to break the loop.
2412 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2413 exec(), but without supplying a callout function, matching went wrong.
2415 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2416 leak if the size of the offset vector was greater than 30. When the vector
2417 is smaller, the saved offsets during recursion go onto a local stack
2418 vector, but for larger vectors malloc() is used. It was failing to free
2419 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2420 error, in fact).
2422 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2423 heapframe that is used only when UTF-8 support is enabled. This caused no
2424 problem, but was untidy.
2426 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2427 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2428 included within another project.
2430 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2431 slightly modified by me:
2433 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2434 not building pcregrep.
2436 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2437 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2439 25. Forward references, both numeric and by name, in patterns that made use of
2440 duplicate group numbers, could behave incorrectly or give incorrect errors,
2441 because when scanning forward to find the reference group, PCRE was not
2442 taking into account the duplicate group numbers. A pattern such as
2443 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2445 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2446 the feature of a custom pointer more persuasive (as requested by a user).
2448 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2449 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2450 pre-defined.
2452 28. Added support for (*UTF8) at the start of a pattern.
2454 29. Arrange for flags added by the "release type" setting in CMake to be shown
2455 in the configuration summary.
2458 Version 7.8 05-Sep-08
2459 ---------------------
2461 1. Replaced UCP searching code with optimized version as implemented for Ad
2462 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2463 stage table and inline lookup instead of a function, giving speed ups of 2
2464 to 5 times on some simple patterns that I tested. Permission was given to
2465 distribute the MultiStage2.py script that generates the tables (it's not in
2466 the tarball, but is in the Subversion repository).
2468 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2469 scripts.
2471 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2472 a group with a zero qualifier. The result of the study could be incorrect,
2473 or the function might crash, depending on the pattern.
2475 4. Caseless matching was not working for non-ASCII characters in back
2476 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2477 It now works when Unicode Property Support is available.
2479 5. In pcretest, an escape such as \x{de} in the data was always generating
2480 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2481 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2482 truncation.
2484 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2486 7. Added two (int) casts to pcregrep when printing the difference of two
2487 pointers, in case they are 64-bit values.
2489 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2490 test 2 if it fails.
2492 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2493 and a #define of that name to empty if it is not externally set. This is to
2494 allow users of MSVC to set it if necessary.
2496 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2497 the convenience functions in the pcre_get.c source file.
2499 11. An option change at the start of a pattern that had top-level alternatives
2500 could cause overwriting and/or a crash. This command provoked a crash in
2501 some environments:
2503 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2505 This potential security problem was recorded as CVE-2008-2371.
2507 12. For a pattern where the match had to start at the beginning or immediately
2508 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2509 pcre_dfa_exec() could read past the end of the passed subject if there was
2510 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2511 pcretest so that it places the subject at the end of its malloc-ed buffer.
2513 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2514 exec() might read past the end of the data buffer in UTF-8 mode.
2516 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2517 the data contained the byte 0x85 as part of a UTF-8 character within its
2518 first line. This applied both to normal and DFA matching.
2520 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2521 /^[^d]*?$/8 failed to match "abc".
2523 16. Added a missing copyright notice to pcrecpp_internal.h.
2525 17. Make it more clear in the documentation that values returned from
2526 pcre_exec() in ovector are byte offsets, not character counts.
2528 18. Tidied a few places to stop certain compilers from issuing warnings.
2530 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2531 supplied by Stefan Weber. I made a further small update for 7.8 because
2532 there is a change of source arrangements: the pcre_searchfuncs.c module is
2533 replaced by pcre_ucd.c.
2536 Version 7.7 07-May-08
2537 ---------------------
2539 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2540 a string to a long long, pretend we don't even have a long long." This is
2541 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2543 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2544 pre-7.6 versions, which defined a global no_arg variable instead of putting
2545 it in the RE class. (See also #8 below.)
2547 3. Remove a line of dead code, identified by coverity and reported by Nuno
2548 Lopes.
2550 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2552 (1) The include/exclude patterns were being applied to the whole pathnames
2553 of files, instead of just to the final components.
2555 (2) If there was more than one level of directory, the subdirectories were
2556 skipped unless they satisfied the include/exclude conditions. This is
2557 inconsistent with GNU grep (and could even be seen as contrary to the
2558 pcregrep specification - which I improved to make it absolutely clear).
2559 The action now is always to scan all levels of directory, and just
2560 apply the include/exclude patterns to regular files.
2562 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2563 --exclude_dir in the tests to avoid scanning .svn directories.
2565 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2566 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2567 doesn't support NULs in patterns.
2569 7. Added some missing "const"s to declarations of static tables in
2570 pcre_compile.c and pcre_dfa_exec.c.
2572 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2573 caused by fix #2 above. (Subsequently also a second patch to fix the
2574 first patch. And a third patch - this was a messy problem.)
2576 9. Applied Craig's patch to remove the use of push_back().
2578 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2579 matching function regexec().
2581 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2582 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2583 references. PCRE supports relative numbers with this syntax (I don't think
2584 Oniguruma does).
2586 12. Previously, a group with a zero repeat such as (...){0} was completely
2587 omitted from the compiled regex. However, this means that if the group
2588 was called as a subroutine from elsewhere in the pattern, things went wrong
2589 (an internal error was given). Such groups are now left in the compiled
2590 pattern, with a new opcode that causes them to be skipped at execution
2591 time.
2593 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2594 to the way PCRE behaves:
2596 (a) A lone ] character is dis-allowed (Perl treats it as data).
2598 (b) A back reference to an unmatched subpattern matches an empty string
2599 (Perl fails the current match path).
2601 (c) A data ] in a character class must be notated as \] because if the
2602 first data character in a class is ], it defines an empty class. (In
2603 Perl it is not possible to have an empty class.) The empty class []
2604 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2605 The negative empty class [^] matches any one character, independently
2606 of the DOTALL setting.
2608 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2609 non-existent subpattern following a character class starting with ']' and
2610 containing () gave an internal compiling error instead of "reference to
2611 non-existent subpattern". Fortunately, when the pattern did exist, the
2612 compiled code was correct. (When scanning forwards to check for the
2613 existence of the subpattern, it was treating the data ']' as terminating
2614 the class, so got the count wrong. When actually compiling, the reference
2615 was subsequently set up correctly.)
2617 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2618 it was being rejected as not supported by pcre_dfa_exec(), even though
2619 other assertions are supported. I have made pcre_dfa_exec() support
2620 (*FAIL).
2622 16. The implementation of 13c above involved the invention of a new opcode,
2623 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2624 cannot be changed at match time, I realized I could make a small
2625 improvement to matching performance by compiling OP_ALLANY instead of
2626 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2627 on the OP_ANY path.
2629 17. Compiling pcretest on Windows with readline support failed without the
2630 following two fixes: (1) Make the unistd.h include conditional on
2631 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2633 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2634 ncurses library to be included for pcretest when ReadLine support is
2635 requested, but also to allow for it to be overridden. This patch came from
2636 Daniel Bergström.
2638 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2639 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2640 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2641 spotting this.
2644 Version 7.6 28-Jan-08
2645 ---------------------
2647 1. A character class containing a very large number of characters with
2648 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2649 overflow.
2651 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2652 HAVE_LONG_LONG is not defined.
2654 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2655 bring them up to date and include new features. This patch includes:
2657 - Fixed PH's badly added libz and libbz2 support.
2658 - Fixed a problem with static linking.
2659 - Added pcredemo. [But later removed - see 7 below.]
2660 - Fixed dftables problem and added an option.
2661 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2663 - Added readline support for pcretest.
2664 - Added an listing of the option settings after cmake has run.
2666 4. A user submitted a patch to Makefile that makes it easy to create
2667 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2668 Makefile.am that cause it to include this special target, without
2669 affecting anything else. Note that the same mingw target plus all
2670 the other distribution libraries and programs are now supported
2671 when configuring with CMake (see 6 below) instead of with
2672 Configure/Make.
2674 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2675 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2676 exported in the Windows port". It has not yet been confirmed that the patch
2677 solves the problem, but it does no harm.
2679 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2680 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2681 with CMake, and also correct the comment about stack recursion.
2683 7. Remove the automatic building of pcredemo from the ./configure system and
2684 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2685 of a program that users should build themselves after PCRE is installed, so
2686 building it automatically is not really right. What is more, it gave
2687 trouble in some build environments.
2689 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2692 Version 7.5 10-Jan-08
2693 ---------------------
2695 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2696 values in parens when parsing an RE using the C++ wrapper."
2698 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2699 Characters greater than 255 were excluded from the class instead of being
2700 included.
2702 3. The same bug as (2) above applied to negated POSIX classes such as
2703 [:^space:].
2705 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2706 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2707 I have changed it.
2709 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2710 first named subpattern) and a construct such as (?&a) would reference the
2711 first named subpattern whose name started with "a" (in other words, the
2712 length check was missing). Both these problems are fixed. "Subpattern name
2713 expected" is now given for (?&) (a zero-length name), and this patch also
2714 makes it give the same error for \k'' (previously it complained that that
2715 was a reference to a non-existent subpattern).
2717 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2718 this is right because (?- can be followed by option settings as well as by
2719 digits. I have, however, made the messages clearer.
2721 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2722 than the number used in the conditional) now cause a compile-time error.
2723 This is actually not compatible with Perl, which accepts such patterns, but
2724 treats the conditional as always being FALSE (as PCRE used to), but it
2725 seems to me that giving a diagnostic is better.
2727 8. Change "alphameric" to the more common word "alphanumeric" in comments
2728 and messages.
2730 9. Fix two occurrences of "backslash" in comments that should have been
2731 "backspace".
2733 10. Remove two redundant lines of code that can never be obeyed (their function
2734 was moved elsewhere).
2736 11. The program that makes PCRE's Unicode character property table had a bug
2737 which caused it to generate incorrect table entries for sequences of
2738 characters that have the same character type, but are in different scripts.
2739 It amalgamated them into a single range, with the script of the first of
2740 them. In other words, some characters were in the wrong script. There were
2741 thirteen such cases, affecting characters in the following ranges:
2743 U+002b0 - U+002c1
2744 U+0060c - U+0060d
2745 U+0061e - U+00612
2746 U+0064b - U+0065e
2747 U+0074d - U+0076d
2748 U+01800 - U+01805
2749 U+01d00 - U+01d77
2750 U+01d9b - U+01dbf
2751 U+0200b - U+0200f
2752 U+030fc - U+030fe
2753 U+03260 - U+0327f
2754 U+0fb46 - U+0fbb1
2755 U+10450 - U+1049d
2757 12. The -o option (show only the matching part of a line) for pcregrep was not
2758 compatible with GNU grep in that, if there was more than one match in a
2759 line, it showed only the first of them. It now behaves in the same way as
2760 GNU grep.
2762 13. If the -o and -v options were combined for pcregrep, it printed a blank
2763 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2764 does the same. The return code can be used to tell if there were any
2765 non-matching lines.
2767 14. Added --file-offsets and --line-offsets to pcregrep.
2769 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2770 infinitely looping recursion. The bug was that positive lookaheads were not
2771 being skipped when checking for a possible empty match (negative lookaheads
2772 and both kinds of lookbehind were skipped).
2774 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2775 inclusion of <windows.h> to before rather than after the definition of
2776 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2778 17. Specifying a possessive quantifier with a specific limit for a Unicode
2779 character property caused pcre_compile() to compile bad code, which led at
2780 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2781 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2782 caused the error; without that there was no problem.
2784 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2786 19. Added --enable-pcretest-libreadline.
2788 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2789 RE::GlobalReplace(). As a result, the number of replacements returned was
2790 double what it should be. I removed one of the increments, but Craig sent a
2791 later patch that removed the other one (the right fix) and added unit tests
2792 that check the return values (which was not done before).
2794 21. Several CMake things:
2796 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2797 the names libpcre and libpcreposix, not just pcre and pcreposix.
2799 (2) The above change means that pcretest and pcregrep are now correctly
2800 linked with the newly-built libraries, not previously installed ones.
2804 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2805 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2806 UTF-8 newline character). The key issue is that the pattern starts .*;
2807 this means that the match must be either at the beginning, or after a
2808 newline. The bug was in the code for advancing after a failed match and
2809 checking that the new position followed a newline. It was not taking
2810 account of UTF-8 characters correctly.
2812 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2813 character classes. PCRE was not treating the sequence [:...:] as a
2814 character class unless the ... were all letters. Perl, however, seems to
2815 allow any characters between [: and :], though of course it rejects as
2816 unknown any "names" that contain non-letters, because all the known class
2817 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2818 for example, whereas PCRE did not - it did not recognize a POSIX character
2819 class. This seemed a bit dangerous, so the code has been changed to be
2820 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2821 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2822 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2823 Perl does, and where it didn't before.
2825 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2826 Windows environments %n is disabled by default.
2829 Version 7.4 21-Sep-07
2830 ---------------------
2832 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2833 means that a class such as [\s] counted as "explicit reference to CR or
2834 LF". That isn't really right - the whole point of the change was to try to
2835 help when there was an actual mention of one of the two characters. So now
2836 the change happens only if \r or \n (or a literal CR or LF) character is
2837 encountered.
2839 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2840 of both had grown to the point where there were only 3 bits left.
2841 Fortunately, there was spare space in the data structure, and so I have
2842 moved the internal flags into a new 16-bit field to free up more option
2843 bits.
2845 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2846 but did not set the internal JCHANGED flag - either of these is enough to
2847 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2848 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2849 start sets both bits.
2851 4. Added options (at build time, compile time, exec time) to change \R from
2852 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2854 5. doc/pcresyntax.html was missing from the distribution.
2856 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2857 compatibility, even though it is no longer used.
2859 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2860 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2861 windows.h file is present (where different names are used). [This was
2862 reversed later after testing - see 16 below.]
2864 8. Changed all #include <config.h> to #include "config.h". There were also
2865 some further <pcre.h> cases that I changed to "pcre.h".
2867 9. When pcregrep was used with the --colour option, it missed the line ending
2868 sequence off the lines that it output.
2870 10. It was pointed out to me that arrays of string pointers cause lots of
2871 relocations when a shared library is dynamically loaded. A technique of
2872 using a single long string with a table of offsets can drastically reduce
2873 these. I have refactored PCRE in four places to do this. The result is
2874 dramatic:
2876 Originally: 290
2877 After changing UCP table: 187
2878 After changing error message table: 43
2879 After changing table of "verbs" 36
2880 After changing table of Posix names 22
2882 Thanks to the folks working on Gregex for glib for this insight.
2884 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2885 unicode-properties was also set.
2887 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2889 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2890 checked only for CRLF.
2892 14. Added casts to pcretest.c to avoid compiler warnings.
2894 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2896 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2897 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2898 entirely. This removes changes made in 7 above.
2900 17. The CMake files have been updated, and there is now more information about
2901 building with CMake in the NON-UNIX-USE document.
2904 Version 7.3 28-Aug-07
2905 ---------------------
2907 1. In the rejigging of the build system that eventually resulted in 7.1, the
2908 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2909 brackets there is not right, since it causes compilers to look for an
2910 installed pcre.h, not the version that is in the source that is being
2911 compiled (which of course may be different). I have changed it back to:
2913 #include "pcre.h"
2915 I have a vague recollection that the change was concerned with compiling in
2916 different directories, but in the new build system, that is taken care of
2917 by the VPATH setting the Makefile.
2919 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2920 when the subject happened to end in the byte 0x85 (e.g. if the last
2921 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2922 characters but of course it shouldn't be taken as a newline when it is part
2923 of another character. The bug was that, for an unlimited repeat of . in
2924 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2925 characters when looking for a newline.
2927 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2929 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2930 in debug output.
2932 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2933 long printing in the pcrecpp unittest when running under MinGW.
2935 6. ESC_K was left out of the EBCDIC table.
2937 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2938 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2939 limit also applies to "virtual nesting" when a pattern is recursive, and in
2940 this case 1000 isn't so big. I have been able to remove this limit at the
2941 expense of backing off one optimization in certain circumstances. Normally,
2942 when pcre_exec() would call its internal match() function recursively and
2943 immediately return the result unconditionally, it uses a "tail recursion"
2944 feature to save stack. However, when a subpattern that can match an empty
2945 string has an unlimited repetition quantifier, it no longer makes this
2946 optimization. That gives it a stack frame in which to save the data for
2947 checking that an empty string has been matched. Previously this was taken
2948 from the 1000-entry workspace that had been reserved. So now there is no
2949 explicit limit, but more stack is used.
2951 8. Applied Daniel's patches to solve problems with the import/export magic
2952 syntax that is required for Windows, and which was going wrong for the
2953 pcreposix and pcrecpp parts of the library. These were overlooked when this
2954 problem was solved for the main library.
2956 9. There were some crude static tests to avoid integer overflow when computing
2957 the size of patterns that contain repeated groups with explicit upper
2958 limits. As the maximum quantifier is 65535, the maximum group length was
2959 set at 30,000 so that the product of these two numbers did not overflow a
2960 32-bit integer. However, it turns out that people want to use groups that
2961 are longer than 30,000 bytes (though not repeat them that many times).
2962 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2963 made it possible to implement the integer overflow checks in a much more
2964 dynamic way, which I have now done. The artificial limitation on group
2965 length has been removed - we now have only the limit on the total length of
2966 the compiled pattern, which depends on the LINK_SIZE setting.
2968 10. Fixed a bug in the documentation for get/copy named substring when
2969 duplicate names are permitted. If none of the named substrings are set, the
2970 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2971 empty string.
2973 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2974 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2975 because the ] is interpreted as the first data character and the
2976 terminating ] is not found. PCRE has been made compatible with Perl in this
2977 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2978 cause memory overwriting.
2980 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2981 string has been matched (to stop an infinite loop). It was not recognizing
2982 a conditional subpattern that could match an empty string if that
2983 subpattern was within another subpattern. For example, it looped when
2984 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2985 condition was not nested. This bug has been fixed.
2987 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2988 past the start of the subject in the presence of bytes with the top bit
2989 set, for example "\x8aBCD".
2991 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2992 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2994 14. Optimized (?!) to (*FAIL).
2996 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2997 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2998 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2999 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3000 does: it's just the validity check that is more restrictive.
3002 16. Inserted checks for integer overflows during escape sequence (backslash)
3003 processing, and also fixed erroneous offset values for syntax errors during
3004 backslash processing.
3006 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3007 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3009 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3010 caused an overrun.
3012 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3013 something other than just ASCII characters) inside a group that had an
3014 unlimited repeat caused a loop at compile time (while checking to see
3015 whether the group could match an empty string).
3017 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3018 [\P{Any}] did so. (Error in the code for printing property names.)
3020 21. An orphan \E inside a character class could cause a crash.
3022 22. A repeated capturing bracket such as (A)? could cause a wild memory
3023 reference during compilation.
3025 23. There are several functions in pcre_compile() that scan along a compiled
3026 expression for various reasons (e.g. to see if it's fixed length for look
3027 behind). There were bugs in these functions when a repeated \p or \P was
3028 present in the pattern. These operators have additional parameters compared
3029 with \d, etc, and these were not being taken into account when moving along
3030 the compiled data. Specifically:
3032 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3033 length.
3035 (b) An item such as \pL+ within a repeated group could cause crashes or
3036 loops.
3038 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3039 "reference to non-existent subpattern" error.
3041 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3043 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3044 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3046 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3048 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3049 character were causing crashes (broken optimization).
3051 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3052 \p or \P) caused a compile-time loop.
3054 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3055 break. For example, the unstudied pattern [\r\n]A does not match the string
3056 "\r\nA" because change 7.0/46 below moves the current point on by two
3057 characters after failing to match at the start. However, the pattern \nA
3058 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3059 the same is true. There doesn't seem any very clean way out of this, but
3060 what I have chosen to do makes the common cases work: PCRE now takes note
3061 of whether there can be an explicit match for \r or \n anywhere in the
3062 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3063 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3064 pattern has explicit CR or LF references.
3066 29. Added (*CR) etc for changing newline setting at start of pattern.
3069 Version 7.2 19-Jun-07
3070 ---------------------
3072 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3073 which is apparently normally available under Windows.
3075 2. Re-jig the pcregrep tests with different newline settings in an attempt
3076 to make them independent of the local environment's newline setting.
3078 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3080 4. Some of the "internals" tests were previously cut out when the link size
3081 was not 2, because the output contained actual offsets. The recent new
3082 "Z" feature of pcretest means that these can be cut out, making the tests
3083 usable with all link sizes.
3085 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3086 stack recursion. This gives a massive performance boost under BSD, but just
3087 a small improvement under Linux. However, it saves one field in the frame
3088 in all cases.
3090 6. Added more features from the forthcoming Perl 5.10:
3092 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3093 recursion call. It refers to the nth most recently opened parentheses.
3095 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3096 to be opened parentheses.
3098 (c) Conditions that refer to capturing parentheses can be specified
3099 relatively, for example, (?(-2)... or (?(+3)...
3101 (d) \K resets the start of the current match so that everything before
3102 is not part of it.
3104 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3106 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3107 reference syntax.
3109 (g) (?| introduces a group in which the numbering of parentheses in each
3110 alternative starts with the same number.
3112 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3114 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3117 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3118 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3119 for detecting groups that can match an empty string.
3121 9. A pattern with a very large number of alternatives (more than several
3122 hundred) was running out of internal workspace during the pre-compile
3123 phase, where pcre_compile() figures out how much memory will be needed. A
3124 bit of new cunning has reduced the workspace needed for groups with
3125 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3126 workspace instead of running out of the 4096 that are available.
3128 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3130 11. Applied patch from Google to remove an optimization that didn't quite work.
3131 The report of the bug said:
3133 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3134 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3135 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3137 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3138 it matched the wrong number of bytes.
3141 Version 7.1 24-Apr-07
3142 ---------------------
3144 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3145 that is more "standard", making use of automake and other Autotools. There
3146 is some re-arrangement of the files and adjustment of comments consequent
3147 on this.
3149 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3150 for recursive directory scanning broke on some systems because the files
3151 are not scanned in any specific order and on different systems the order
3152 was different. A call to "sort" has been inserted into RunGrepTest for the
3153 approprate test as a short-term fix. In the longer term there may be an
3154 alternative.
3156 3. I had an email from Eric Raymond about problems translating some of PCRE's
3157 man pages to HTML (despite the fact that I distribute HTML pages, some
3158 people do their own conversions for various reasons). The problems
3159 concerned the use of low-level troff macros .br and .in. I have therefore
3160 removed all such uses from the man pages (some were redundant, some could
3161 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3162 HTML has been updated to handle .nf/.fi and to complain if it encounters
3163 .br or .in.
3165 4. Updated comments in configure.ac that get placed in config.h.in and also
3166 arranged for config.h to be included in the distribution, with the name
3167 config.h.generic, for the benefit of those who have to compile without
3168 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3170 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3171 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3172 makevp.bat for the current PCRE, using the additional files
3173 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3175 6. A Windows user reported a minor discrepancy with test 2, which turned out
3176 to be caused by a trailing space on an input line that had got lost in his
3177 copy. The trailing space was an accident, so I've just removed it.
3179 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3180 that is needed.
3182 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3183 as "const" (a) because they are and (b) because it helps the PHP
3184 maintainers who have recently made a script to detect big data structures
3185 in the php code that should be moved to the .rodata section. I remembered
3186 to update Builducptable as well, so it won't revert if ucptable.h is ever
3187 re-created.
3189 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3190 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3191 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3192 support is not required. This saves 1.5-2K of code, which is important in
3193 some applications.
3195 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3196 so as not to refer to the tables, even though these functions will never be
3197 called when UTF-8 support is disabled. Otherwise there are problems with a
3198 shared library.
3200 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3202 (a) It was defining its arguments as char * instead of void *.
3204 (b) It was assuming that all moves were upwards in memory; this was true
3205 a long time ago when I wrote it, but is no longer the case.
3207 The emulated memove() is provided for those environments that have neither
3208 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3209 is clearly not the case, as these two bugs were recently reported.
3211 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3212 and Detrail to create the HTML documentation, the .txt form of the man
3213 pages, and it removes trailing spaces from listed files. It also creates
3214 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3215 case, it wraps all the #defines with #ifndefs. This script should be run
3216 before "make dist".
3218 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3219 with Unicode property support.
3221 (a) For a maximizing quantifier, if the two different cases of the
3222 character were of different lengths in their UTF-8 codings (there are
3223 some cases like this - I found 11), and the matching function had to
3224 back up over a mixture of the two cases, it incorrectly assumed they
3225 were both the same length.
3227 (b) When PCRE was configured to use the heap rather than the stack for
3228 recursion during matching, it was not correctly preserving the data for
3229 the other case of a UTF-8 character when checking ahead for a match
3230 while processing a minimizing repeat. If the check also involved
3231 matching a wide character, but failed, corruption could cause an
3232 erroneous result when trying to check for a repeat of the original
3233 character.
3235 13. Some tidying changes to the testing mechanism:
3237 (a) The RunTest script now detects the internal link size and whether there
3238 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3239 values substituted by "configure". (The RunGrepTest script already did
3240 this for UTF-8.) The configure.ac script no longer substitutes the
3241 relevant variables.
3243 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3244 with length and offset values. This means that the output is different
3245 for different internal link sizes. Test 2 is skipped for link sizes
3246 other than 2 because of this, bypassing the problem. Unfortunately,
3247 there was also a test in test 3 (the locale tests) that used /B and
3248 failed for link sizes other than 2. Rather than cut the whole test out,
3249 I have added a new /Z option to pcretest that replaces the length and
3250 offset values with spaces. This is now used to make test 3 independent
3251 of link size. (Test 2 will be tidied up later.)
3253 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3254 segmentation fault instead of returning the appropriate error message.
3256 15. In multiline mode when the newline sequence was set to "any", the pattern
3257 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3258 This doesn't seem right; it now treats the CRLF combination as the line
3259 ending, and so does not match in that case. It's only a pattern such as ^$
3260 that would hit this one: something like ^ABC$ would have failed after \r
3261 and then tried again after \r\n.
3263 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3264 in an attempt to make files that differ only in their line terminators
3265 compare equal. This works on Linux.
3267 17. Under certain error circumstances pcregrep might try to free random memory
3268 as it exited. This is now fixed, thanks to valgrind.
3270 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3271 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3272 was because its rules for how to advance for /g after matching an empty
3273 string at the end of a line did not allow for this case. They now check for
3274 it specially.
3276 20. pcretest is supposed to handle patterns and data of any length, by
3277 extending its buffers when necessary. It was getting this wrong when the
3278 buffer for a data line had to be extended.
3280 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3281 CRLF as a newline sequence.
3283 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3284 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3285 I have nevertheless tidied it up.
3287 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3289 24. Added a man page for pcre-config.
3292 Version 7.0 19-Dec-06
3293 ---------------------
3295 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3296 moving to gcc 4.1.1.
3298 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3299 sys/time.h, which is documented as needed for this function. It doesn't
3300 seem to matter on Linux, but it showed up on some releases of OS X.
3302 3. It seems that there are systems where bytes whose values are greater than
3303 127 match isprint() in the "C" locale. The "C" locale should be the
3304 default when a C program starts up. In most systems, only ASCII printing
3305 characters match isprint(). This difference caused the output from pcretest
3306 to vary, making some of the tests fail. I have changed pcretest so that:
3308 (a) When it is outputting text in the compiled version of a pattern, bytes
3309 other than 32-126 are always shown as hex escapes.
3311 (b) When it is outputting text that is a matched part of a subject string,
3312 it does the same, unless a different locale has been set for the match
3313 (using the /L modifier). In this case, it uses isprint() to decide.
3315 4. Fixed a major bug that caused incorrect computation of the amount of memory
3316 required for a compiled pattern when options that changed within the
3317 pattern affected the logic of the preliminary scan that determines the
3318 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3319 that the computed length was too small. The symptoms of this bug were
3320 either the PCRE error "internal error: code overflow" from pcre_compile(),
3321 or a glibc crash with a message such as "pcretest: free(): invalid next
3322 size (fast)". Examples of patterns that provoked this bug (shown in
3323 pcretest format) are:
3325 /(?-x: )/x
3326 /(?x)(?-x: \s*#\s*)/
3327 /((?i)[\x{c0}])/8
3328 /(?i:[\x{c0}])/8
3330 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3331 is now done differently.
3333 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3334 wrapper classes; (b) implement a new function in the C++ scanner that is
3335 more efficient than the old way of doing things because it avoids levels of
3336 recursion in the regex matching; (c) add a paragraph to the documentation
3337 for the FullMatch() function.
3339 6. The escape sequence \n was being treated as whatever was defined as
3340 "newline". Not only was this contrary to the documentation, which states
3341 that \n is character 10 (hex 0A), but it also went horribly wrong when
3342 "newline" was defined as CRLF. This has been fixed.
3344 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3345 was being set to -1 for the "end of line" case (supposedly a value that no
3346 character can have). Though this value is never used (the check for end of
3347 line is "zero bytes in current character"), it caused compiler complaints.
3348 I've changed it to 0xffffffff.
3350 8. In pcre_version.c, the version string was being built by a sequence of
3351 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3352 string (as it is for production releases) called a macro with an empty
3353 argument. The C standard says the result of this is undefined. The gcc
3354 compiler treats it as an empty string (which was what was wanted) but it is
3355 reported that Visual C gives an error. The source has been hacked around to
3356 avoid this problem.
3358 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3359 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3360 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3361 of them did).
3363 10. Originally, pcretest opened its input and output without "b"; then I was
3364 told that "b" was needed in some environments, so it was added for release
3365 5.0 to both the input and output. (It makes no difference on Unix-like
3366 systems.) Later I was told that it is wrong for the input on Windows. I've
3367 now abstracted the modes into two macros, to make it easier to fiddle with
3368 them, and removed "b" from the input mode under Windows.
3370 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3372 12. Added -help and --help to pcretest as an official way of being reminded
3373 of the options.
3375 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3376 and pcrecpp.cc because they annoy compilers at high warning levels.
3378 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3380 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3381 scripts are not C programs :-) and which was not noticed because it works
3382 on Linux.
3384 16. pcretest is supposed to handle any length of pattern and data line (as one
3385 line or as a continued sequence of lines) by extending its input buffer if
3386 necessary. This feature was broken for very long pattern lines, leading to
3387 a string of junk being passed to pcre_compile() if the pattern was longer
3388 than about 50K.
3390 17. I have done a major re-factoring of the way pcre_compile() computes the
3391 amount of memory needed for a compiled pattern. Previously, there was code
3392 that made a preliminary scan of the pattern in order to do this. That was
3393 OK when PCRE was new, but as the facilities have expanded, it has become
3394 harder and harder to keep it in step with the real compile phase, and there
3395 have been a number of bugs (see for example, 4 above). I have now found a
3396 cunning way of running the real compile function in a "fake" mode that
3397 enables it to compute how much memory it would need, while actually only
3398 ever using a few hundred bytes of working memory and without too many
3399 tests of the mode. This should make future maintenance and development
3400 easier. A side effect of this work is that the limit of 200 on the nesting
3401 depth of parentheses has been removed (though this was never a serious
3402 limitation, I suspect). However, there is a downside: pcre_compile() now
3403 runs more slowly than before (30% or more, depending on the pattern). I
3404 hope this isn't a big issue. There is no effect on runtime performance.
3406 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3407 newline (only possible for the last line of a file) and it was a
3408 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3410 19. Added additional timing features to pcretest. (1) The -tm option now times
3411 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3412 separate command line item, by a number that specifies the number of
3413 repeats to use when timing. The default is 50000; this gives better
3414 precision, but takes uncomfortably long for very large patterns.
3416 20. Extended pcre_study() to be more clever in cases where a branch of a
3417 subpattern has no definite first character. For example, (a*|b*)[cd] would
3418 previously give no result from pcre_study(). Now it recognizes that the
3419 first character must be a, b, c, or d.
3421 21. There was an incorrect error "recursive call could loop indefinitely" if
3422 a subpattern (or the entire pattern) that was being tested for matching an
3423 empty string contained only one non-empty item after a nested subpattern.
3424 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3425 incorrectly, because the \d was being skipped in the check.
3427 22. The pcretest program now has a new pattern option /B and a command line
3428 option -b, which is equivalent to adding /B to every pattern. This causes
3429 it to show the compiled bytecode, without the additional information that
3430 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3431 is the same as /B/I).
3433 23. A new optimization is now able automatically to treat some sequences such
3434 as a*b as a*+b. More specifically, if something simple (such as a character
3435 or a simple class like \d) has an unlimited quantifier, and is followed by
3436 something that cannot possibly match the quantified thing, the quantifier
3437 is automatically "possessified".
3439 24. A recursive reference to a subpattern whose number was greater than 39
3440 went wrong under certain circumstances in UTF-8 mode. This bug could also
3441 have affected the operation of pcre_study().
3443 25. Realized that a little bit of performance could be had by replacing
3444 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3446 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3448 27. Possessive quantifiers such as a++ were previously implemented by turning
3449 them into atomic groups such as ($>a+). Now they have their own opcodes,
3450 which improves performance. This includes the automatically created ones
3451 from 23 above.
3453 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3454 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3455 the first matched character to be a colon. This applied both to named and
3456 numbered groups.
3458 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3460 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3461 building PCRE on Windows easier, so I have included it in the distribution.
3463 31. There is now a check in pcretest against a ridiculously large number being
3464 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3465 loop, the loop is abandoned.
3467 32. Forward references to subpatterns in conditions such as (?(2)...) where
3468 subpattern 2 is defined later cause pcre_compile() to search forwards in
3469 the pattern for the relevant set of parentheses. This search went wrong
3470 when there were unescaped parentheses in a character class, parentheses
3471 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3473 33. "Subroutine" calls and backreferences were previously restricted to
3474 referencing subpatterns earlier in the regex. This restriction has now
3475 been removed.
3477 34. Added a number of extra features that are going to be in Perl 5.10. On the
3478 whole, these are just syntactic alternatives for features that PCRE had
3479 previously implemented using the Python syntax or my own invention. The
3480 other formats are all retained for compatibility.
3482 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3483 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3484 also .NET compatible.
3486 (b) A recursion or subroutine call to a named group can now be defined as
3487 (?&name) as well as (?P>name).
3489 (c) A backreference to a named group can now be defined as \k<name> or
3490 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3491 5.10, are also .NET compatible.
3493 (d) A conditional reference to a named group can now use the syntax
3494 (?(<name>) or (?('name') as well as (?(name).
3496 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3497 groups (named and numbered) that are never evaluated inline, but can be
3498 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3499 is always false. There may be only one alternative in such a group.
3501 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3502 as the simple (?(R). The condition is true only if the most recent
3503 recursion is that of the given number or name. It does not search out
3504 through the entire recursion stack.
3506 (g) The escape \gN or \g{N} has been added, where N is a positive or
3507 negative number, specifying an absolute or relative reference.
3509 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3510 some "unreachable code" warnings.
3512 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3513 things, this adds five new scripts.
3515 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3516 There were also incompatibilities regarding the handling of \Q..\E inside
3517 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3518 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3520 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3521 matches an empty string, and forcibly breaks the loop. There were bugs in
3522 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3523 against aaaa the result was just "a" rather than "aaaa", for example. Two
3524 separate and independent bugs (that affected different cases) have been
3525 fixed.
3527 39. Refactored the code to abolish the use of different opcodes for small
3528 capturing bracket numbers. This is a tidy that I avoided doing when I
3529 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3530 The new approach is not only tidier, it makes it possible to reduce the
3531 memory needed to fix the previous bug (38).
3533 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3534 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3535 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3536 mode.
3538 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3539 report.
3541 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3542 copy construction and assignment in the C++ wrapper.
3544 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3545 couple of bugs that could have given wrong results in the "--newline=crlf"
3546 case.
3548 44. Added a number of casts and did some reorganization of signed/unsigned int
3549 variables following suggestions from Dair Grant. Also renamed the variable
3550 "this" as "item" because it is a C++ keyword.
3552 45. Arranged for dftables to add
3554 #include "pcre_internal.h"
3556 to pcre_chartables.c because without it, gcc 4.x may remove the array
3557 definition from the final binary if PCRE is built into a static library and
3558 dead code stripping is activated.
3560 46. For an unanchored pattern, if a match attempt fails at the start of a
3561 newline sequence, and the newline setting is CRLF or ANY, and the next two
3562 characters are CRLF, advance by two characters instead of one.
3565 Version 6.7 04-Jul-06
3566 ---------------------
3568 1. In order to handle tests when input lines are enormously long, pcretest has
3569 been re-factored so that it automatically extends its buffers when
3570 necessary. The code is crude, but this _is_ just a test program. The
3571 default size has been increased from 32K to 50K.
3573 2. The code in pcre_study() was using the value of the re argument before
3574 testing it for NULL. (Of course, in any sensible call of the function, it
3575 won't be NULL.)
3577 3. The memmove() emulation function in pcre_internal.h, which is used on
3578 systems that lack both memmove() and bcopy() - that is, hardly ever -
3579 was missing a "static" storage class specifier.
3581 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3582 containing an extended class (one that cannot be represented by a bitmap
3583 because it contains high-valued characters or Unicode property items, e.g.
3584 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3585 pattern, but PCRE should not loop if you do not (it no longer does).
3586 [Detail: two cases were found: (a) a repeated subpattern containing an
3587 extended class; (b) a recursive reference to a subpattern that followed a
3588 previous extended class. It wasn't skipping over the extended class
3589 correctly when UTF-8 mode was not set.]
3591 5. A negated single-character class was not being recognized as fixed-length
3592 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3593 compile error "lookbehind assertion is not fixed length".
3595 6. The RunPerlTest auxiliary script was showing an unexpected difference
3596 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3597 write a Perl script that can interpret lines of an input file either as
3598 byte characters or as UTF-8, which is what "perltest" was being required to
3599 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3600 can't do is switch easily at run time between having the "use utf8;" pragma
3601 or not. In the end, I fudged it by using the RunPerlTest script to insert
3602 "use utf8;" explicitly for the UTF-8 tests.
3604 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3605 the end of the subject string, contrary to the documentation and to what
3606 Perl does. This was true of both matching functions. Now it matches only at
3607 the start of the subject and immediately after *internal* newlines.
3609 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3610 a pointer to an int instead of a pointer to an unsigned long int. This
3611 caused problems on 64-bit systems.
3613 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3614 instance of the 'standard' template library not being so standard".
3616 10. There was no check on the number of named subpatterns nor the maximum
3617 length of a subpattern name. The product of these values is used to compute
3618 the size of the memory block for a compiled pattern. By supplying a very
3619 long subpattern name and a large number of named subpatterns, the size
3620 computation could be caused to overflow. This is now prevented by limiting
3621 the length of names to 32 characters, and the number of named subpatterns
3622 to 10,000.
3624 11. Subpatterns that are repeated with specific counts have to be replicated in
3625 the compiled pattern. The size of memory for this was computed from the
3626 length of the subpattern and the repeat count. The latter is limited to
3627 65535, but there was no limit on the former, meaning that integer overflow
3628 could in principle occur. The compiled length of a repeated subpattern is
3629 now limited to 30,000 bytes in order to prevent this.
3631 12. Added the optional facility to have named substrings with the same name.
3633 13. Added the ability to use a named substring as a condition, using the
3634 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3635 are numbers (not recommended). Forward references are permitted.
3637 14. Added forward references in named backreferences (if you see what I mean).
3639 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3640 pattern could run off the end of the subject. For example, the pattern
3641 "(?s)(.{1,5})"8 did this with the subject "ab".
3643 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3644 PCRE_CASELESS was set when matching characters that were quantified with ?
3645 or *.
3647 17. A character class other than a single negated character that had a minimum
3648 but no maximum quantifier - for example [ab]{6,} - was not handled
3649 correctly by pce_dfa_exec(). It would match only one character.
3651 18. A valid (though odd) pattern that looked like a POSIX character
3652 class but used an invalid character after [ (for example [[,abc,]]) caused
3653 pcre_compile() to give the error "Failed: internal error: code overflow" or
3654 in some cases to crash with a glibc free() error. This could even happen if
3655 the pattern terminated after [[ but there just happened to be a sequence of
3656 letters, a binary zero, and a closing ] in the memory that followed.
3658 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3659 over the years. Originally (before any Unicode support), just the bottom 8
3660 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3661 output from "man perlunicode" includes this:
3663 The regular expression compiler produces polymorphic opcodes. That
3664 is, the pattern adapts to the data and automatically switches to
3665 the Unicode character scheme when presented with Unicode data--or
3666 instead uses a traditional byte scheme when presented with byte
3667 data.
3669 Sadly, a wide octal escape does not cause a switch, and in a string with
3670 no other multibyte characters, these octal escapes are treated as before.
3671 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3672 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3673 Unicode string.
3675 I have not perpetrated such confusion in PCRE. Up till now, it took just
3676 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3677 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3678 translate to the appropriate multibyte character.
3680 29. Applied some refactoring to reduce the number of warnings from Microsoft
3681 and Borland compilers. This has included removing the fudge introduced
3682 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3683 a warning about an unused variable.
3685 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3686 characters since release 4.0, because Perl (from release 5.004) does not.
3687 [Or at least, is documented not to: some releases seem to be in conflict
3688 with the documentation.] However, when a pattern was studied with
3689 pcre_study() and all its branches started with \s, PCRE still included VT
3690 as a possible starting character. Of course, this did no harm; it just
3691 caused an unnecessary match attempt.
3693 22. Removed a now-redundant internal flag bit that recorded the fact that case
3694 dependency changed within the pattern. This was once needed for "required
3695 byte" processing, but is no longer used. This recovers a now-scarce options
3696 bit. Also moved the least significant internal flag bit to the most-
3697 significant bit of the word, which was not previously used (hangover from
3698 the days when it was an int rather than a uint) to free up another bit for
3699 the future.
3701 23. Added support for CRLF line endings as well as CR and LF. As well as the
3702 default being selectable at build time, it can now be changed at runtime
3703 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3704 specify that it is scanning data with non-default line endings.
3706 24. Changed the definition of CXXLINK to make it agree with the definition of
3707 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3709 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3710 recursions. This makes a big different to stack usage for some patterns.
3712 26. If a subpattern containing a named recursion or subroutine reference such
3713 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3714 the space required for the compiled pattern went wrong and gave too small a
3715 value. Depending on the environment, this could lead to "Failed: internal
3716 error: code overflow at offset 49" or "glibc detected double free or
3717 corruption" errors.
3719 27. Applied patches from Google (a) to support the new newline modes and (b) to
3720 advance over multibyte UTF-8 characters in GlobalReplace.
3722 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3723 difference for some implementation of PCRE in some Windows version.
3725 29. Added some extra testing facilities to pcretest:
3727 \q<number> in a data line sets the "match limit" value
3728 \Q<number> in a data line sets the "match recursion limt" value
3729 -S <number> sets the stack size, where <number> is in megabytes
3731 The -S option isn't available for Windows.
3734 Version 6.6 06-Feb-06
3735 ---------------------
3737 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3738 in pcreposix.h. I have copied the definition from pcre.h.
3740 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3741 because pcre.h is no longer a built file.
3743 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3744 not normally included in the compiled code.
3747 Version 6.5 01-Feb-06
3748 ---------------------
3750 1. When using the partial match feature with pcre_dfa_exec(), it was not
3751 anchoring the second and subsequent partial matches at the new starting
3752 point. This could lead to incorrect results. For example, with the pattern
3753 /1234/, partially matching against "123" and then "a4" gave a match.
3755 2. Changes to pcregrep:
3757 (a) All non-match returns from pcre_exec() were being treated as failures
3758 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3759 error message is output. Some extra information is given for the
3761 probably the only errors that are likely to be caused by users (by
3762 specifying a regex that has nested indefinite repeats, for instance).
3763 If there are more than 20 of these errors, pcregrep is abandoned.
3765 (b) A binary zero was treated as data while matching, but terminated the
3766 output line if it was written out. This has been fixed: binary zeroes
3767 are now no different to any other data bytes.
3769 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3770 used to set a locale for matching. The --locale=xxxx long option has
3771 been added (no short equivalent) to specify a locale explicitly on the
3772 pcregrep command, overriding the environment variables.
3774 (d) When -B was used with -n, some line numbers in the output were one less
3775 than they should have been.
3777 (e) Added the -o (--only-matching) option.
3779 (f) If -A or -C was used with -c (count only), some lines of context were
3780 accidentally printed for the final match.
3782 (g) Added the -H (--with-filename) option.
3784 (h) The combination of options -rh failed to suppress file names for files
3785 that were found from directory arguments.
3787 (i) Added the -D (--devices) and -d (--directories) options.
3789 (j) Added the -F (--fixed-strings) option.
3791 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3793 (l) Added the --colo(u)r option.
3795 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3796 is not present by default.
3798 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3799 items such as (?R) or (?1), when the recursion could match a number of
3800 alternatives. If it matched one of the alternatives, but subsequently,
3801 outside the recursion, there was a failure, the code tried to back up into
3802 the recursion. However, because of the way PCRE is implemented, this is not
3803 possible, and the result was an incorrect result from the match.
3805 In order to prevent this happening, the specification of recursion has
3806 been changed so that all such subpatterns are automatically treated as
3807 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3809 4. I had overlooked the fact that, in some locales, there are characters for
3810 which isalpha() is true but neither isupper() nor islower() are true. In
3811 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3812 and ordfeminine) are like this. This affected the treatment of \w and \W
3813 when they appeared in character classes, but not when they appeared outside
3814 a character class. The bit map for "word" characters is now created
3815 separately from the results of isalnum() instead of just taking it from the
3816 upper, lower, and digit maps. (Plus the underscore character, of course.)
3818 5. The above bug also affected the handling of POSIX character classes such as
3819 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3820 permanent tables. Instead, the bit maps for such a class were previously
3821 created as the appropriate unions of the upper, lower, and digit bitmaps.
3822 Now they are created by subtraction from the [[:word:]] class, which has
3823 its own bitmap.
3825 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3826 It is created by subtracting the vertical space characters (\x09, \x0a,
3827 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3828 subtraction was done in the overall bitmap for a character class, meaning
3829 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3830 be recognized. This bug has been fixed.
3832 7. Patches from the folks at Google:
3834 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3835 real life, but is still worth protecting against".
3837 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3838 regular expressions".
3840 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3841 have it.
3843 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3844 "configure" and the latter not, in order to fix a problem somebody had
3845 with compiling the Arg class on HP-UX.
3847 (e) Improve the error-handling of the C++ wrapper a little bit.
3849 (f) New tests for checking recursion limiting.
3851 8. The pcre_memmove() function, which is used only if the environment does not
3852 have a standard memmove() function (and is therefore rarely compiled),
3853 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3854 returning a result (though PCRE never actually uses the result).
3856 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3857 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3858 returned instead of calling malloc() with an overflowing number that would
3859 most likely cause subsequent chaos.
3861 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3863 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3864 with this option is matched, the nmatch and pmatch options of regexec() are
3865 ignored.
3867 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3868 provided in case anyone wants to the the POSIX interface with UTF-8
3869 strings.
3871 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3872 C++ linking (needed for some HP-UX environments).
3874 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3875 (unused parameter) and in the pcre_printint() function (omitted "default"
3876 switch label when the default is to do nothing).
3878 15. Added some code to make it possible, when PCRE is compiled as a C++
3879 library, to replace subject pointers for pcre_exec() with a smart pointer
3880 class, thus making it possible to process discontinuous strings.
3882 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3883 much the same function. They were added by different people who were trying
3884 to make PCRE easy to compile on non-Unix systems. It has been suggested
3885 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3886 for compiling on Windows systems. I have therefore replaced it with
3887 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3888 defaults to "extern" for C or "extern C" for C++, which works fine on
3889 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3890 SCOPE with something explicit in config.h. In addition:
3892 (a) pcreposix.h still had just "extern" instead of either of these macros;
3893 I have replaced it with PCRE_DATA_SCOPE.
3895 (b) Functions such as _pcre_xclass(), which are internal to the library,
3896 but external in the C sense, all had PCRE_EXPORT in their definitions.
3897 This is apparently wrong for the Windows case, so I have removed it.
3898 (It makes no difference on Unix-like systems.)
3900 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3901 of recursive calls to match(). This is different to MATCH_LIMIT because
3902 that limits the total number of calls to match(), not all of which increase
3903 the depth of recursion. Limiting the recursion depth limits the amount of
3904 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3905 when PCRE is compiled, and changed at run time. A patch from Google adds
3906 this functionality to the C++ interface.
3908 18. Changes to the handling of Unicode character properties:
3910 (a) Updated the table to Unicode 4.1.0.
3912 (b) Recognize characters that are not in the table as "Cn" (undefined).
3914 (c) I revised the way the table is implemented to a much improved format
3915 which includes recognition of ranges. It now supports the ranges that
3916 are defined in UnicodeData.txt, and it also amalgamates other
3917 characters into ranges. This has reduced the number of entries in the
3918 table from around 16,000 to around 3,000, thus reducing its size
3919 considerably. I realized I did not need to use a tree structure after
3920 all - a binary chop search is just as efficient. Having reduced the
3921 number of entries, I extended their size from 6 bytes to 8 bytes to
3922 allow for more data.
3924 (d) Added support for Unicode script names via properties such as \p{Han}.
3926 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3927 matching that character.
3929 20. When matching a repeated Unicode property with a minimum greater than zero,
3930 (for example \pL{2,}), PCRE could look past the end of the subject if it
3931 reached it while seeking the minimum number of characters. This could
3932 happen only if some of the characters were more than one byte long, because
3933 there is a check for at least the minimum number of bytes.
3935 21. Refactored the implementation of \p and \P so as to be more general, to
3936 allow for more different types of property in future. This has changed the
3937 compiled form incompatibly. Anybody with saved compiled patterns that use
3938 \p or \P will have to recompile them.
3940 22. Added "Any" and "L&" to the supported property types.
3942 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3943 but give a compile time error if the value is greater than 0xff.
3945 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3946 accidentally not being installed or uninstalled.
3948 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3949 made were to insert the current release number. This seemed silly, because
3950 it made things harder for people building PCRE on systems that don't run
3951 "configure". I have turned pcre.h into a distributed file, no longer built
3952 by "configure", with the version identification directly included. There is
3953 no longer a pcre.h.in file.
3955 However, this change necessitated a change to the pcre-config script as
3956 well. It is built from pcre-config.in, and one of the substitutions was the
3957 release number. I have updated configure.ac so that ./configure now finds
3958 the release number by grepping pcre.h.
3960 26. Added the ability to run the tests under valgrind.
3963 Version 6.4 05-Sep-05
3964 ---------------------
3966 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3967 "--" to be printed when multiple files were scanned, even when none of the
3968 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3969 consider it to be a bug, and have restored the previous behaviour.
3971 2. A couple of code tidies to get rid of compiler warnings.
3973 3. The pcretest program used to cheat by referring to symbols in the library
3974 whose names begin with _pcre_. These are internal symbols that are not
3975 really supposed to be visible externally, and in some environments it is
3976 possible to suppress them. The cheating is now confined to including
3977 certain files from the library's source, which is a bit cleaner.
3979 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3980 file's purpose clearer.
3982 5. Reorganized pcre_ucp_findchar().
3985 Version 6.3 15-Aug-05
3986 ---------------------
3988 1. The file libpcre.pc.in did not have general read permission in the tarball.
3990 2. There were some problems when building without C++ support:
3992 (a) If C++ support was not built, "make install" and "make test" still
3993 tried to test it.
3995 (b) There were problems when the value of CXX was explicitly set. Some
3996 changes have been made to try to fix these, and ...
3998 (c) --disable-cpp can now be used to explicitly disable C++ support.
4000 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4001 backslash in a target when C++ was disabled. This confuses some
4002 versions of "make", apparently. Using an intermediate variable solves
4003 this. (Same for CPP_LOBJ.)
4005 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4006 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4007 necessary on certain architectures.
4009 4. Added a setting of -export-symbols-regex to the link command to remove
4010 those symbols that are exported in the C sense, but actually are local
4011 within the library, and not documented. Their names all begin with
4012 "_pcre_". This is not a perfect job, because (a) we have to except some
4013 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4014 available (and never for static libraries). I have made a note to try to
4015 find a way round (a) in the future.
4018 Version 6.2 01-Aug-05
4019 ---------------------
4021 1. There was no test for integer overflow of quantifier values. A construction
4022 such as {1111111111111111} would give undefined results. What is worse, if
4023 a minimum quantifier for a parenthesized subpattern overflowed and became
4024 negative, the calculation of the memory size went wrong. This could have
4025 led to memory overwriting.
4027 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4029 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4030 operating environments where this matters.
4032 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4033 PCRE options from within the C++ wrapper.
4035 5. Named capturing subpatterns were not being correctly counted when a pattern
4036 was compiled. This caused two problems: (a) If there were more than 100
4037 such subpatterns, the calculation of the memory needed for the whole
4038 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4039 back references of the form \12, where the number was greater than 9, were
4040 not recognized as back references, even though there were sufficient
4041 previous subpatterns.
4043 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4044 versions of gcc, e.g. 2.95.4.
4047 Version 6.1 21-Jun-05
4048 ---------------------
4050 1. There was one reference to the variable "posix" in pcretest.c that was not
4051 surrounded by "#if !defined NOPOSIX".
4053 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4054 the cross-check on the old pcre_info() function, for the benefit of the
4055 cut-down version of PCRE that is currently imported into Exim.
4057 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4058 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4059 patterns that start (?i) but I don't think that matters. The use of (?i) is
4060 just an example; this all applies to the other options as well.
4062 4. Since libtool seems to echo the compile commands it is issuing, the output
4063 from "make" can be reduced a bit by putting "@" in front of each libtool
4064 compile command.
4066 5. Patch from the folks at Google for configure.in to be a bit more thorough
4067 in checking for a suitable C++ installation before trying to compile the
4068 C++ stuff. This should fix a reported problem when a compiler was present,
4069 but no suitable headers.
4071 6. The man pages all had just "PCRE" as their title. I have changed them to
4072 be the relevant file name. I have also arranged that these names are
4073 retained in the file doc/pcre.txt, which is a concatenation in text format
4074 of all the man pages except the little individual ones for each function.
4076 7. The NON-UNIX-USE file had not been updated for the different set of source
4077 files that come with release 6. I also added a few comments about the C++
4078 wrapper.
4081 Version 6.0 07-Jun-05
4082 ---------------------
4084 1. Some minor internal re-organization to help with my DFA experiments.
4086 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4087 didn't matter for the library itself when fully configured, but did matter
4088 when compiling without UCP support, or within Exim, where the ucp files are
4089 not imported.
4091 3. Refactoring of the library code to split up the various functions into
4092 different source modules. The addition of the new DFA matching code (see
4093 below) to a single monolithic source would have made it really too
4094 unwieldy, quite apart from causing all the code to be include in a
4095 statically linked application, when only some functions are used. This is
4096 relevant even without the DFA addition now that patterns can be compiled in
4097 one application and matched in another.
4099 The downside of splitting up is that there have to be some external
4100 functions and data tables that are used internally in different modules of
4101 the library but which are not part of the API. These have all had their
4102 names changed to start with "_pcre_" so that they are unlikely to clash
4103 with other external names.
4105 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4106 a different (DFA) algorithm. Although it is slower than the original
4107 function, it does have some advantages for certain types of matching
4108 problem.
4110 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4111 including restarting after a partial match.
4113 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4114 defined when compiling for Windows was sent to me. I have put it into the
4115 code, though I have no means of testing or verifying it.
4117 7. Added the pcre_refcount() auxiliary function.
4119 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4120 match before or at the first newline in the subject string. In pcretest,
4121 the /f option on a pattern can be used to set this.
4123 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4124 would behave wrongly. This has been present in PCRE since release 4.0.
4126 10. A number of changes to the pcregrep command:
4128 (a) Refactored how -x works; insert ^(...)$ instead of setting
4129 PCRE_ANCHORED and checking the length, in preparation for adding
4130 something similar for -w.
4132 (b) Added the -w (match as a word) option.
4134 (c) Refactored the way lines are read and buffered so as to have more
4135 than one at a time available.
4137 (d) Implemented a pcregrep test script.
4139 (e) Added the -M (multiline match) option. This allows patterns to match
4140 over several lines of the subject. The buffering ensures that at least
4141 8K, or the rest of the document (whichever is the shorter) is available
4142 for matching (and similarly the previous 8K for lookbehind assertions).
4144 (f) Changed the --help output so that it now says
4146 -w, --word-regex(p)
4148 instead of two lines, one with "regex" and the other with "regexp"
4149 because that confused at least one person since the short forms are the
4150 same. (This required a bit of code, as the output is generated
4151 automatically from a table. It wasn't just a text change.)
4153 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4154 option but starts with a hyphen. Could be a pattern or a path name
4155 starting with a hyphen, for instance.
4157 (h) "-" can be given as a file name to represent stdin.
4159 (i) When file names are being printed, "(standard input)" is used for
4160 the standard input, for compatibility with GNU grep. Previously
4161 "<stdin>" was used.
4163 (j) The option --label=xxx can be used to supply a name to be used for
4164 stdin when file names are being printed. There is no short form.
4166 (k) Re-factored the options decoding logic because we are going to add
4167 two more options that take data. Such options can now be given in four
4168 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4170 (l) Added the -A, -B, and -C options for requesting that lines of context
4171 around matches be printed.
4173 (m) Added the -L option to print the names of files that do not contain
4174 any matching lines, that is, the complement of -l.
4176 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4177 continue to scan other files.
4179 (o) The -s option was incorrectly implemented. For compatibility with other
4180 greps, it now suppresses the error message for a non-existent or non-
4181 accessible file (but not the return code). There is a new option called
4182 -q that suppresses the output of matching lines, which was what -s was
4183 previously doing.
4185 (p) Added --include and --exclude options to specify files for inclusion
4186 and exclusion when recursing.
4188 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4189 Hopefully, it now does.
4191 12. Missing cast in pcre_study().
4193 13. Added an "uninstall" target to the makefile.
4195 14. Replaced "extern" in the function prototypes in Makefile.in with
4196 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4197 world, but is set differently for Windows.
4199 15. Added a second compiling function called pcre_compile2(). The only
4200 difference is that it has an extra argument, which is a pointer to an
4201 integer error code. When there is a compile-time failure, this is set
4202 non-zero, in addition to the error test pointer being set to point to an
4203 error message. The new argument may be NULL if no error number is required
4204 (but then you may as well call pcre_compile(), which is now just a
4205 wrapper). This facility is provided because some applications need a
4206 numeric error indication, but it has also enabled me to tidy up the way
4207 compile-time errors are handled in the POSIX wrapper.
4209 16. Added VPATH=.libs to the makefile; this should help when building with one
4210 prefix path and installing with another. (Or so I'm told by someone who
4211 knows more about this stuff than I do.)
4213 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4214 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4215 match everything, including newlines. This is not POSIX-compatible, but
4216 somebody wanted the feature. From pcretest it can be activated by using
4217 both the P and the s flags.
4219 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4221 19. libpcre.pc was being incorrectly installed as executable.
4223 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4224 it now also looks for '\r' so that it will work unmodified on Windows.
4226 21. Added Google's contributed C++ wrapper to the distribution.
4228 22. Added some untidy missing memory free() calls in pcretest, to keep
4229 Electric Fence happy when testing.
4233 Version 5.0 13-Sep-04
4234 ---------------------
4236 1. Internal change: literal characters are no longer packed up into items
4237 containing multiple characters in a single byte-string. Each character
4238 is now matched using a separate opcode. However, there may be more than one
4239 byte in the character in UTF-8 mode.
4241 2. The pcre_callout_block structure has two new fields: pattern_position and
4242 next_item_length. These contain the offset in the pattern to the next match
4243 item, and its length, respectively.
4245 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4246 insertion of callouts before each pattern item. Added the /C option to
4247 pcretest to make use of this.
4249 4. On the advice of a Windows user, the lines
4251 #if defined(_WIN32) || defined(WIN32)
4252 _setmode( _fileno( stdout ), 0x8000 );
4253 #endif /* defined(_WIN32) || defined(WIN32) */
4255 have been added to the source of pcretest. This apparently does useful
4256 magic in relation to line terminators.
4258 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4259 for the benefit of those environments where the "b" makes a difference.
4261 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4262 to know about it. I have put a hack into configure.in that adds in code
4263 to set GCC=yes if CC=icc. This seems to end up at a point in the
4264 generated configure script that is early enough to affect the setting of
4265 compiler options, which is what is needed, but I have no means of testing
4266 whether it really works. (The user who reported this had patched the
4267 generated configure script, which of course I cannot do.)
4269 LATER: After change 22 below (new libtool files), the configure script
4270 seems to know about icc (and also ecc). Therefore, I have commented out
4271 this hack in configure.in.
4273 7. Added support for pkg-config (2 patches were sent in).
4275 8. Negated POSIX character classes that used a combination of internal tables
4276 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4277 [[:^ascii]]. Typically, they would match almost any characters. The other
4278 POSIX classes were not broken in this way.
4280 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4281 to find the match, as PCRE was deluded into thinking that the match had to
4282 start at the start point or following a newline. The same bug applied to
4283 patterns with negative forward assertions or any backward assertions
4284 preceding ".*" at the start, unless the pattern required a fixed first
4285 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4287 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4288 starting at the last subject character, bytes beyond the end of the subject
4289 string were read.
4291 11. Renamed the variable "class" as "classbits" to make life easier for C++
4292 users. (Previously there was a macro definition, but it apparently wasn't
4293 enough.)
4295 12. Added the new field "tables" to the extra data so that tables can be passed
4296 in at exec time, or the internal tables can be re-selected. This allows
4297 a compiled regex to be saved and re-used at a later time by a different
4298 program that might have everything at different addresses.
4300 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4301 -R library as well as a -L library.
4303 14. The debugging options of pcretest (-d on the command line or D on a
4304 pattern) showed incorrect output for anything following an extended class
4305 that contained multibyte characters and which was followed by a quantifier.
4307 15. Added optional support for general category Unicode character properties
4308 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4309 support. It adds about 90K to the size of the library. The meanings of the
4310 inbuilt class escapes such as \d and \s have NOT been changed.
4312 16. Updated pcredemo.c to include calls to free() to release the memory for the
4313 compiled pattern.
4315 17. The generated file chartables.c was being created in the source directory
4316 instead of in the building directory. This caused the build to fail if the
4317 source directory was different from the building directory, and was
4318 read-only.
4320 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4321 file. No doubt somebody will tell me if they don't make sense... Also added
4322 Dan Mooney's comments about building on OpenVMS.
4324 19. Added support for partial matching via the PCRE_PARTIAL option for
4325 pcre_exec() and the \P data escape in pcretest.
4327 20. Extended pcretest with 3 new pattern features:
4329 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4330 write the compiled pattern to the file whose name is "rest-of-line".
4331 This is a straight binary dump of the data, with the saved pointer to
4332 the character tables forced to be NULL. The study data, if any, is
4333 written too. After writing, pcretest reads a new pattern.
4335 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4336 compiled pattern from the given file. There must not be any
4337 occurrences of "<" in the file name (pretty unlikely); if there are,
4338 pcretest will instead treat the initial "<" as a pattern delimiter.
4339 After reading in the pattern, pcretest goes on to read data lines as
4340 usual.
4342 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4343 and 16-bit fields in a compiled pattern, to simulate a pattern that
4344 was compiled on a host of opposite endianness.
4346 21. The pcre-exec() function can now cope with patterns that were compiled on
4347 hosts of opposite endianness, with this restriction:
4349 As for any compiled expression that is saved and used later, the tables
4350 pointer field cannot be preserved; the extra_data field in the arguments
4351 to pcre_exec() should be used to pass in a tables address if a value
4352 other than the default internal tables were used at compile time.
4354 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4355 now diagnosed as an error. Previously, most of the time, a negative number
4356 would have been treated as zero, but if in addition "ovector" was passed as
4357 NULL, a crash could occur.
4359 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4360 new versions from the libtool 1.5 distribution (the last one is a copy of
4361 a file called libtool.m4). This seems to have fixed the need to patch
4362 "configure" to support Darwin 1.3 (which I used to do). However, I still
4363 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4364 workstation).
4366 24. Changed the PCRE licence to be the more standard "BSD" licence.
4369 Version 4.5 01-Dec-03
4370 ---------------------
4372 1. There has been some re-arrangement of the code for the match() function so
4373 that it can be compiled in a version that does not call itself recursively.
4374 Instead, it keeps those local variables that need separate instances for
4375 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4376 needs to "recurse". Keeping track of where control must go is done by means
4377 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4378 hide most of the details from the main code, and operates only if
4379 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4380 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4381 operating.
4383 To make it easier for callers to provide specially tailored get/free
4384 functions for this usage, two new functions, pcre_stack_malloc, and
4385 pcre_stack_free, are used. They are always called in strict stacking order,
4386 and the size of block requested is always the same.
4388 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4389 PCRE has been compiled to use the stack or the heap for recursion. The
4390 -C option of pcretest uses this to show which version is compiled.
4392 A new data escape \S, is added to pcretest; it causes the amounts of store
4393 obtained and freed by both kinds of malloc/free at match time to be added
4394 to the output.
4396 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4397 what's available on my current Linux desktop machine.
4399 3. When matching a UTF-8 string, the test for a valid string at the start has
4400 been extended. If start_offset is not zero, PCRE now checks that it points
4401 to a byte that is the start of a UTF-8 character. If not, it returns
4402 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4403 this is necessary because there may be backward assertions in the pattern.
4404 When matching the same subject several times, it may save resources to use
4405 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4407 4. The code for checking the validity of UTF-8 strings has been tightened so
4408 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4409 containing "overlong sequences".
4411 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4412 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4413 should have been "&", but it just so happened that all the cases this let
4414 through by mistake were picked up later in the function.
4416 6. I had used a variable called "isblank" - this is a C99 function, causing
4417 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4419 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4420 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4421 so the version is shown; (c) stop "make test" echoing "./RunTest".
4423 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4425 9. The prototype for memmove() for systems that don't have it was using
4426 size_t, but the inclusion of the header that defines size_t was later. I've
4427 moved the #includes for the C headers earlier to avoid this.
4429 10. Added some adjustments to the code to make it easier to compiler on certain
4430 special systems:
4432 (a) Some "const" qualifiers were missing.
4433 (b) Added the macro EXPORT before all exported functions; by default this
4434 is defined to be empty.
4435 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4436 that it reads its output file name as an argument instead of writing
4437 to the standard output and assuming this can be redirected.
4439 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4440 class containing characters with values greater than 255, PCRE compilation
4441 went into a loop.
4443 12. A recursive reference to a subpattern that was within another subpattern
4444 that had a minimum quantifier of zero caused PCRE to crash. For example,
4445 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4446 recursion. If the recursively-called subpattern itself had a zero repeat,
4447 that was OK.
4449 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4450 buffer into which it was copied (for escape processing) was still set at
4451 1024, so long lines caused crashes.
4453 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4454 "internal error: code overflow...". This applied to any character class
4455 that was followed by a possessive quantifier.
4457 15. Modified the Makefile to add libpcre.la as a prerequisite for
4458 libpcreposix.la because I was told this is needed for a parallel build to
4459 work.
4461 16. If a pattern that contained .* following optional items at the start was
4462 studied, the wrong optimizing data was generated, leading to matching
4463 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4464 matching string must start with a or b or c. The correct conclusion for
4465 this pattern is that a match can start with any character.
4468 Version 4.4 13-Aug-03
4469 ---------------------
4471 1. In UTF-8 mode, a character class containing characters with values between
4472 127 and 255 was not handled correctly if the compiled pattern was studied.
4473 In fixing this, I have also improved the studying algorithm for such
4474 classes (slightly).
4476 2. Three internal functions had redundant arguments passed to them. Removal
4477 might give a very teeny performance improvement.
4479 3. Documentation bug: the value of the capture_top field in a callout is *one
4480 more than* the number of the hightest numbered captured substring.
4482 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4483 in incorrectly linking with a previously installed version. They now link
4484 explicitly with libpcre.la.
4486 5. configure.in no longer needs to recognize Cygwin specially.
4488 6. A problem in pcre.in for Windows platforms is fixed.
4490 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4491 pcretest, it used to include the size of the study block as part of its
4492 output. Unfortunately, the structure contains a field that has a different
4493 size on different hardware architectures. This meant that the tests that
4494 showed this size failed. As the block is currently always of a fixed size,
4495 this information isn't actually particularly useful in pcretest output, so
4496 I have just removed it.
4498 8. Three pre-processor statements accidentally did not start in column 1.
4499 Sadly, there are *still* compilers around that complain, even though
4500 standard C has not required this for well over a decade. Sigh.
4502 9. In pcretest, the code for checking callouts passed small integers in the
4503 callout_data field, which is a void * field. However, some picky compilers
4504 complained about the casts involved for this on 64-bit systems. Now
4505 pcretest passes the address of the small integer instead, which should get
4506 rid of the warnings.
4508 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4509 both compile and run time, and gives an error if an invalid UTF-8 sequence
4510 is found. There is a option for disabling this check in cases where the
4511 string is known to be correct and/or the maximum performance is wanted.
4513 11. In response to a bug report, I changed one line in Makefile.in from
4515 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4516 to
4517 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4519 to look similar to other lines, but I have no way of telling whether this
4520 is the right thing to do, as I do not use Windows. No doubt I'll get told
4521 if it's wrong...
4524 Version 4.3 21-May-03
4525 ---------------------
4527 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4528 Makefile.
4530 2. Some refactoring to improve the quality of the code:
4532 (i) The utf8_table... variables are now declared "const".
4534 (ii) The code for \cx, which used the "case flipping" table to upper case
4535 lower case letters, now just substracts 32. This is ASCII-specific,
4536 but the whole concept of \cx is ASCII-specific, so it seems
4537 reasonable.
4539 (iii) PCRE was using its character types table to recognize decimal and
4540 hexadecimal digits in the pattern. This is silly, because it handles
4541 only 0-9, a-f, and A-F, but the character types table is locale-
4542 specific, which means strange things might happen. A private
4543 table is now used for this - though it costs 256 bytes, a table is
4544 much faster than multiple explicit tests. Of course, the standard
4545 character types table is still used for matching digits in subject
4546 strings against \d.
4548 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4549 ending in _t are). So I've renamed it as ESC_tee.
4551 3. The first argument for regexec() in the POSIX wrapper should have been
4552 defined as "const".
4554 4. Changed pcretest to use malloc() for its buffers so that they can be
4555 Electric Fenced for debugging.
4557 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4558 to read one or more bytes before the start of the subject string. Often this
4559 had no effect on PCRE's behaviour, but in some circumstances it could
4560 provoke a segmentation fault.
4562 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4563 to try to read one or more bytes before the start of the subject string.
4565 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4566 UTF-8 support could misbehave in various ways if the subject string
4567 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4568 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4569 back over UTF-8 characters.)
4572 Version 4.2 14-Apr-03
4573 ---------------------
4575 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4577 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4578 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4579 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4580 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4581 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4583 Note: automatic setting of the BUILD variables is not yet working
4584 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4585 compile-time but not at link-time
4586 [LINK]: use for linking executables only
4587 make different versions for Windows and non-Windows
4588 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4589 libraries
4590 [LINK_FOR_BUILD]: new variable
4591 [OBJEXT]: use throughout
4592 [EXEEXT]: use throughout
4593 <winshared>: new target
4594 <wininstall>: new target
4595 <dftables.o>: use native compiler
4596 <dftables>: use native linker
4597 <install>: handle Windows platform correctly
4598 <clean>: ditto
4599 <check>: ditto
4600 copy DLL to top builddir before testing
4602 As part of these changes, -no-undefined was removed again. This was reported
4603 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4604 in any case.
4606 3. Some tidies to get rid of compiler warnings:
4608 . In the match_data structure, match_limit was an unsigned long int, whereas
4609 match_call_count was an int. I've made them both unsigned long ints.
4611 . In pcretest the fact that a const uschar * doesn't automatically cast to
4612 a void * provoked a warning.
4614 . Turning on some more compiler warnings threw up some "shadow" variables
4615 and a few more missing casts.
4617 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4618 option, a class that contained a single character with a value between 128
4619 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4621 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4622 option, a class that contained several characters, but with at least one
4623 whose value was between 128 and 255 caused PCRE to crash.
4626 Version 4.1 12-Mar-03
4627 ---------------------
4629 1. Compiling with gcc -pedantic found a couple of places where casts were
4630 needed, and a string in dftables.c that was longer than standard compilers are
4631 required to support.
4633 2. Compiling with Sun's compiler found a few more places where the code could
4634 be tidied up in order to avoid warnings.
4636 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4637 first of these names is deprecated in the latest Autoconf in favour of the name
4638 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4639 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4640 analogy I have changed it to CFLAGS_FOR_BUILD.
4642 4. Added -no-undefined to the linking command in the Makefile, because this is
4643 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4644 linking step for the pcreposix library.
4646 5. PCRE was failing to diagnose the case of two named groups with the same
4647 name.
4649 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4650 literal character that is needed in the subject for a match, and scans along to
4651 ensure that it is present before embarking on the full matching process. This
4652 saves time in cases of nested unlimited repeats that are never going to match.
4653 Problem: the scan can take a lot of time if the subject is very long (e.g.
4654 megabytes), thus penalizing straightforward matches. It is now done only if the
4655 amount of subject to be scanned is less than 1000 bytes.
4657 7. A lesser problem with the same optimization is that it was recording the
4658 first character of an anchored pattern as "needed", thus provoking a search
4659 right along the subject, even when the first match of the pattern was going to
4660 fail. The "needed" character is now not set for anchored patterns, unless it
4661 follows something in the pattern that is of non-fixed length. Thus, it still
4662 fulfils its original purpose of finding quick non-matches in cases of nested
4663 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4666 Version 4.0 17-Feb-03
4667 ---------------------
4669 1. If a comment in an extended regex that started immediately after a meta-item
4670 extended to the end of string, PCRE compiled incorrect data. This could lead to
4671 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4673 2. Moved to autoconf 2.53 and libtool 1.4.2.
4675 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4676 the special perltest8 script is no longer needed - all the tests can be run
4677 from a single perltest script.
4679 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4680 by \s. It has now been removed in PCRE. This means it isn't recognized as
4681 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4682 class [:space:] *does* include VT, thereby creating a mess.
4684 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4685 space and tab.
4687 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4688 its new features into the main test script, reducing the number of scripts.
4690 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4691 were backward compatible, and made the (?i) apply to the whole pattern, as if
4692 /i were given. Now it behaves more logically, and applies the option setting
4693 only to what follows. PCRE has been changed to follow suit. However, if it
4694 finds options settings right at the start of the pattern, it extracts them into
4695 the global options, as before. Thus, they show up in the info data.
4697 8. Added support for the \Q...\E escape sequence. Characters in between are
4698 treated as literals. This is slightly different from Perl in that $ and @ are
4699 also handled as literals inside the quotes. In Perl, they will cause variable
4700 interpolation. Note the following examples:
4702 Pattern PCRE matches Perl matches
4704 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4705 \Qabc\$xyz\E abc\$xyz abc\$xyz
4706 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4708 For compatibility with Perl, \Q...\E sequences are recognized inside character
4709 classes as well as outside them.
4711 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4712 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4713 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4714 signed/unsigned warnings.
4716 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4717 option for pcretest, so I've replaced it by a simple function that does just
4718 that job.
4720 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4721 "pcregrep -".
4723 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4724 Java package. This provides some syntactic sugar for simple cases of what my
4725 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4726 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4727 item, you can use this simplified notation. Note that only makes sense with
4728 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4729 greediness, whatever the setting of the PCRE_UNGREEDY option.
4731 13. A change of greediness default within a pattern was not taking effect at
4732 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4733 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4734 was abstracted outside.
4736 14. PCRE now supports the \G assertion. It is true when the current matching
4737 position is at the start point of the match. This differs from \A when the
4738 starting offset is non-zero. Used with the /g option of pcretest (or similar
4739 code), it works in the same way as it does for Perl's /g option. If all
4740 alternatives of a regex begin with \G, the expression is anchored to the start
4741 match position, and the "anchored" flag is set in the compiled expression.
4743 15. Some bugs concerning the handling of certain option changes within patterns
4744 have been fixed. These applied to options other than (?ims). For example,
4745 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4746 been the other way round. Some of this was related to change 7 above.
4748 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4749 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4750 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4751 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4753 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4754 mode. Unlike ".", it always matches newline, whatever the setting of
4755 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4756 assertions. Perl allows it, but it doesn't (in general) work because it can't
4757 calculate the length of the lookbehind. At least, that's the case for Perl
4758 5.8.0 - I've been told they are going to document that it doesn't work in
4759 future.
4761 18. Added an error diagnosis for escapes that PCRE does not support: these are
4762 \L, \l, \N, \P, \p, \U, \u, and \X.
4764 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4765 reading past the end of the pattern in cases such as /[abcd/.
4767 20. PCRE was getting more memory than necessary for patterns with classes that
4768 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4770 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4771 compiling PCRE for use with Virtual Pascal.
4773 22. Small fix to the Makefile to make it work properly if the build is done
4774 outside the source tree.
4776 23. Added a new extension: a condition to go with recursion. If a conditional
4777 subpattern starts with (?(R) the "true" branch is used if recursion has
4778 happened, whereas the "false" branch is used only at the top level.
4780 24. When there was a very long string of literal characters (over 255 bytes
4781 without UTF support, over 250 bytes with UTF support), the computation of how
4782 much memory was required could be incorrect, leading to segfaults or other
4783 strange effects.
4785 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4786 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4787 there was a subsequent back reference to those brackets. This meant that, for
4788 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4789 possible to check for precisely this case. All we can do is abandon the
4790 optimization if .* occurs inside capturing brackets when there are any back
4791 references whatsoever. (See below for a better fix that came later.)
4793 26. The handling of the optimization for finding the first character of a
4794 non-anchored pattern, and for finding a character that is required later in the
4795 match were failing in some cases. This didn't break the matching; it just
4796 failed to optimize when it could. The way this is done has been re-implemented.
4798 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4800 28. Added a new feature that provides some of the functionality that Perl
4801 provides with (?{...}). The facility is termed a "callout". The way it is done
4802 in PCRE is for the caller to provide an optional function, by setting
4803 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4804 global variable. By default it is unset, which disables all calling out. To get
4805 the function called, the regex must include (?C) at appropriate points. This
4806 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4807 This provides a means of identifying different callout points. When PCRE
4808 reaches such a point in the regex, if pcre_callout has been set, the external
4809 function is called. It is provided with data in a structure called
4810 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4811 matching continues; if it returns a non-zero value, the match at the current
4812 point fails. However, backtracking will occur if possible. [This was changed
4813 later and other features added - see item 49 below.]
4815 29. pcretest is upgraded to test the callout functionality. It provides a
4816 callout function that displays information. By default, it shows the start of
4817 the match and the current position in the text. There are some new data escapes
4818 to vary what happens:
4820 \C+ in addition, show current contents of captured substrings
4821 \C- do not supply a callout function
4822 \C!n return 1 when callout number n is reached
4823 \C!n!m return 1 when callout number n is reached for the mth time
4825 30. If pcregrep was called with the -l option and just a single file name, it
4826 output "<stdin>" if a match was found, instead of the file name.
4828 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4829 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4830 pcre_exec(). This saves a malloc/free per call. The default value of
4831 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4832 when configuring.
4834 32. The default maximum size of a compiled pattern is 64K. There have been a
4835 few cases of people hitting this limit. The code now uses macros to handle the
4836 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4837 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4838 configuring. Tests 2 and 5 work only with 2-byte links because they output
4839 debugging information about compiled patterns.
4841 33. Internal code re-arrangements:
4843 (a) Moved the debugging function for printing out a compiled regex into
4844 its own source file (printint.c) and used #include to pull it into
4845 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4846 separate copies.
4848 (b) Defined the list of op-code names for debugging as a macro in
4849 internal.h so that it is next to the definition of the opcodes.
4851 (c) Defined a table of op-code lengths for simpler skipping along compiled
4852 code. This is again a macro in internal.h so that it is next to the
4853 definition of the opcodes.
4855 34. Added support for recursive calls to individual subpatterns, along the
4856 lines of Robin Houston's patch (but implemented somewhat differently).
4858 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4859 allow it to read and process whole directories in Win32. This code was
4860 contributed by Lionel Fourquaux; it has not been tested by me.
4862 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4863 used to name a group. Names consist of alphanumerics and underscores, and must
4864 be unique. Back references use the syntax (?P=name) and recursive calls use
4865 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4866 numbers. The function pcre_fullinfo() can be used after compilation to extract
4867 a name/number map. There are three relevant calls:
4869 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4870 PCRE_INFO_NAMECOUNT yields the number of entries
4871 PCRE_INFO_NAMETABLE yields a pointer to the map.
4873 The map is a vector of fixed-size entries. The size of each entry depends on
4874 the length of the longest name used. The first two bytes of each entry are the
4875 group number, most significant byte first. There follows the corresponding
4876 name, zero terminated. The names are in alphabetical order.
4878 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4879 case instead of 255. Making it the same both with and without UTF-8 support
4880 means that the same test output works with both.
4882 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4883 calling malloc() with a zero argument.
4885 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4886 optimization. I've improved things by keeping a bitmap of backreferences with
4887 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4888 fact referenced, the optimization can be applied. It is unlikely that a
4889 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4890 the match to follow \n) will appear inside brackets with a number greater than
4891 31, but if it does, any back reference > 31 suppresses the optimization.
4893 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4894 of disabling numbered capturing parentheses. Any opening parenthesis that is
4895 not followed by ? behaves as if it were followed by ?: but named parentheses
4896 can still be used for capturing (and they will acquire numbers in the usual
4897 way).
4899 41. Redesigned the return codes from the match() function into yes/no/error so
4900 that errors can be passed back from deep inside the nested calls. A malloc
4901 failure while inside a recursive subpattern call now causes the
4902 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4904 42. It is now possible to set a limit on the number of times the match()
4905 function is called in a call to pcre_exec(). This facility makes it possible to
4906 limit the amount of recursion and backtracking, though not in a directly
4907 obvious way, because the match() function is used in a number of different
4908 circumstances. The count starts from zero for each position in the subject
4909 string (for non-anchored patterns). The default limit is, for compatibility, a
4910 large number, namely 10 000 000. You can change this in two ways:
4912 (a) When configuring PCRE before making, you can use --with-match-limit=n
4913 to set a default value for the compiled library.
4915 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4916 a different value is set. See 45 below.
4918 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4920 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4921 of things that can be changed at compile time. The first argument specifies
4922 what is wanted and the second points to where the information is to be placed.
4923 The current list of available information is:
4927 The output is an integer that is set to one if UTF-8 support is available;
4928 otherwise it is set to zero.
4932 The output is an integer that it set to the value of the code that is used for
4933 newline. It is either LF (10) or CR (13).
4937 The output is an integer that contains the number of bytes used for internal
4938 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4942 The output is an integer that contains the threshold above which the POSIX
4943 interface uses malloc() for output vectors. See item 31 above.
4947 The output is an unsigned integer that contains the default limit of the number
4948 of match() calls in a pcre_exec() execution. See 42 above.
4950 44. pcretest has been upgraded by the addition of the -C option. This causes it
4951 to extract all the available output from the new pcre_config() function, and to
4952 output it. The program then exits immediately.
4954 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4955 order to support additional features. One way would have been to define
4956 pcre_exec2() (for example) with extra arguments, but this would not have been
4957 extensible, and would also have required all calls to the original function to
4958 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4959 is used for passing in "extra" data from pcre_study().
4961 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4962 contains the following fields:
4964 flags a bitmap indicating which of the following fields are set
4965 study_data opaque data from pcre_study()
4966 match_limit a way of specifying a limit on match() calls for a specific
4967 call to pcre_exec()
4968 callout_data data for callouts (see 49 below)
4970 The flag bits are also defined in pcre.h, and are
4976 The pcre_study() function now returns one of these new pcre_extra blocks, with
4977 the actual study data pointed to by the study_data field, and the
4978 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4979 before. That is, this change is entirely upwards-compatible and requires no
4980 change to existing code.
4982 If you want to pass in additional data to pcre_exec(), you can either place it
4983 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4984 block.
4986 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4987 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4988 times with different match limits, until it finds the minimum value needed for
4989 pcre_exec() to complete. The value is then output. This can be instructive; for
4990 most simple matches the number is quite small, but for pathological cases it
4991 gets very large very quickly.
4993 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4994 returns the size of the data block pointed to by the study_data field in a
4995 pcre_extra block, that is, the value that was passed as the argument to
4996 pcre_malloc() when PCRE was getting memory in which to place the information
4997 created by pcre_study(). The fourth argument should point to a size_t variable.
4998 pcretest has been extended so that this information is shown after a successful
4999 pcre_study() call when information about the compiled regex is being displayed.
5001 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5002 because what follows is always an absolute path. (Later: it turns out that this
5003 is more than cosmetic for MinGW, because it doesn't like empty path
5004 components.)
5006 49. Some changes have been made to the callout feature (see 28 above):
5008 (i) A callout function now has three choices for what it returns:
5010 0 => success, carry on matching
5011 > 0 => failure at this point, but backtrack if possible
5012 < 0 => serious error, return this value from pcre_exec()
5014 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5015 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5016 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5017 use by callout functions. It will never be used by PCRE itself.
5019 (ii) The pcre_extra structure (see 45 above) has a void * field called
5020 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5021 pcre_callout_block structure has a field of the same name. The contents of
5022 the field passed in the pcre_extra structure are passed to the callout
5023 function in the corresponding field in the callout block. This makes it
5024 easier to use the same callout-containing regex from multiple threads. For
5025 testing, the pcretest program has a new data escape
5027 \C*n pass the number n (may be negative) as callout_data
5029 If the callout function in pcretest receives a non-zero value as
5030 callout_data, it returns that value.
5032 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5033 there were some redundant $(CFLAGS) in commands that are now specified as
5034 $(LINK), which already includes $(CFLAGS).
5036 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5037 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5038 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5039 one-byte characters throughout. Note that case-insensitive matching applies
5040 only to characters whose values are less than 256. PCRE doesn't support the
5041 notion of cases for higher-valued characters.
5043 (i) A character class whose characters are all within 0-255 is handled as
5044 a bit map, and the map is inverted for negative classes. Previously, a
5045 character > 255 always failed to match such a class; however it should
5046 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5048 (ii) A negated character class with a single character < 255 is coded as
5049 "not this character" (OP_NOT). This wasn't working properly when the test
5050 character was multibyte, either singly or repeated.
5052 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5053 mode, for example: \x{100}{2,3}.
5055 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5056 singly or repeated) now correctly test multibyte characters. However,
5057 PCRE doesn't recognize any characters with values greater than 255 as
5058 digits, spaces, or word characters. Such characters always match \D, \S,
5059 and \W, and never match \d, \s, or \w.
5061 (v) Classes may now contain characters and character ranges with values
5062 greater than 255. For example: [ab\x{100}-\x{400}].
5064 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5065 PCRE in UTF-8 mode.
5067 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5068 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5069 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5070 value.)
5072 53. The single man page has become too large. I have therefore split it up into
5073 a number of separate man pages. These also give rise to individual HTML pages;
5074 these are now put in a separate directory, and there is an index.html page that
5075 lists them all. Some hyperlinking between the pages has been installed.
5077 54. Added convenience functions for handling named capturing parentheses.
5079 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5080 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5081 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5082 are faulted.
5084 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5085 calling configure. These values are used when compiling the dftables.c program
5086 which is run to generate the source of the default character tables. They
5087 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5088 you will need to set these values.
5090 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5093 Version 3.9 02-Jan-02
5094 ---------------------
5096 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5098 2. If --disable-static was given, the building process failed when trying to
5099 build pcretest and pcregrep. (For some reason it was using libtool to compile
5100 them, which is not right, as they aren't part of the library.)
5103 Version 3.8 18-Dec-01
5104 ---------------------
5106 1. The experimental UTF-8 code was completely screwed up. It was packing the
5107 bytes in the wrong order. How dumb can you get?
5110 Version 3.7 29-Oct-01
5111 ---------------------
5113 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5114 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5115 this didn't happen under Solaris 8, where I normally test things.
5117 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5118 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5119 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5120 this fix an hour or so after the initial 3.7 release.)
5123 Version 3.6 23-Oct-01
5124 ---------------------
5126 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5127 offsets passed as NULL with zero offset count.
5129 2. The config.guess and config.sub files had not been updated when I moved to
5130 the latest autoconf.
5133 Version 3.5 15-Aug-01
5134 ---------------------
5136 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5137 had been forgotten.
5139 2. By using declared but undefined structures, we can avoid using "void"
5140 definitions in pcre.h while keeping the internal definitions of the structures
5141 private.
5143 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5144 user point of view, this means that both static and shared libraries are built
5145 by default, but this can be individually controlled. More of the work of
5146 handling this static/shared cases is now inside libtool instead of PCRE's make
5147 file.
5149 4. The pcretest utility is now installed along with pcregrep because it is
5150 useful for users (to test regexs) and by doing this, it automatically gets
5151 relinked by libtool. The documentation has been turned into a man page, so
5152 there are now .1, .txt, and .html versions in /doc.
5154 5. Upgrades to pcregrep:
5155 (i) Added long-form option names like gnu grep.
5156 (ii) Added --help to list all options with an explanatory phrase.
5157 (iii) Added -r, --recursive to recurse into sub-directories.
5158 (iv) Added -f, --file to read patterns from a file.
5160 6. pcre_exec() was referring to its "code" argument before testing that
5161 argument for NULL (and giving an error if it was NULL).
5163 7. Upgraded Makefile.in to allow for compiling in a different directory from
5164 the source directory.
5166 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5167 options bits, the pointer it was passed was to an int instead of to an unsigned
5168 long int. This mattered only on 64-bit systems.
5170 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5171 generated) instead of pcre.in, which it its source. Also made the same change
5172 in several of the .c files.
5174 10. A new release of gcc defines printf() as a macro, which broke pcretest
5175 because it had an ifdef in the middle of a string argument for printf(). Fixed
5176 by using separate calls to printf().
5178 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5179 script, to force use of CR or LF instead of \n in the source. On non-Unix
5180 systems, the value can be set in config.h.
5182 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5183 absolute limit. Changed the text of the error message to make this clear, and
5184 likewise updated the man page.
5186 13. The limit of 99 on the number of capturing subpatterns has been removed.
5187 The new limit is 65535, which I hope will not be a "real" limit.
5190 Version 3.4 22-Aug-00
5191 ---------------------
5193 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5195 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5198 Version 3.3 01-Aug-00
5199 ---------------------
5201 1. If an octal character was given, but the value was greater than \377, it
5202 was not getting masked to the least significant bits, as documented. This could
5203 lead to crashes in some systems.
5205 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5206 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5208 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5209 These just pass their arguments on to (pcre_free)(), but they are provided
5210 because some uses of PCRE bind it to non-C systems that can call its functions,
5211 but cannot call free() or pcre_free() directly.
5213 4. Add "make test" as a synonym for "make check". Corrected some comments in
5214 the Makefile.
5216 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5217 Makefile.
5219 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5220 command called pgrep for grepping around the active processes.
5222 7. Added the beginnings of support for UTF-8 character strings.
5224 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5225 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5226 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5227 out for the ar command.)
5230 Version 3.2 12-May-00
5231 ---------------------
5233 This is purely a bug fixing release.
5235 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5236 of ZA. This was just one example of several cases that could provoke this bug,
5237 which was introduced by change 9 of version 2.00. The code for breaking
5238 infinite loops after an iteration that matches an empty string was't working
5239 correctly.
5241 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5242 when matched against abbab (for example). After matching an empty string, it
5243 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5244 caused it to match further down the string than it should.
5246 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5247 was there because it doesn't seem to be needed, and it causes trouble on some
5248 systems, as it is not a Standard C header. It has been removed.
5250 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5251 were reported on the Macintosh. The changes were from
5253 while ((c = *(++ptr)) != 0 && c != '\n');
5254 to
5255 while ((c = *(++ptr)) != 0 && c != '\n') ;
5257 Totally extraordinary, but if that's what it takes...
5259 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5260 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5261 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5262 assumes the way PCRE uses memmove() (always moving upwards).
5264 6. PCRE is being used in one environment where strchr() is not available. There
5265 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5266 faster code anyway.
5269 Version 3.1 09-Feb-00
5270 ---------------------
5272 The only change in this release is the fixing of some bugs in Makefile.in for
5273 the "install" target:
5275 (1) It was failing to install pcreposix.h.
5277 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5280 Version 3.0 01-Feb-00
5281 ---------------------
5283 1. Add support for the /+ modifier to perltest (to output $` like it does in
5284 pcretest).
5286 2. Add support for the /g modifier to perltest.
5288 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5289 matches null strings.
5291 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5292 pattern. Perl treats empty patterns specially - it reuses the most recent
5293 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5294 effect.
5296 5. The POSIX interface was broken in that it was just handing over the POSIX
5297 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5298 required a bigger vector, with some working space on the end. This means that
5299 the POSIX wrapper now has to get and free some memory, and copy the results.
5301 6. Added some simple autoconf support, placing the test data and the
5302 documentation in separate directories, re-organizing some of the
5303 information files, and making it build pcre-config (a GNU standard). Also added
5304 libtool support for building PCRE as a shared library, which is now the
5305 default.
5307 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5308 09 are not valid octal constants. Single digits will be used for minor values
5309 less than 10.
5311 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5312 existing programs that set these in the POSIX interface can use PCRE without
5313 modification.
5315 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5316 return all that pcre_info() returns, plus additional data. The pcre_info()
5317 function is retained for compatibility, but is considered to be obsolete.
5319 10. Added experimental recursion feature (?R) to handle one common case that
5320 Perl 5.6 will be able to do with (?p{...}).
5322 11. Added support for POSIX character classes like [:alpha:], which Perl is
5323 adopting.
5326 Version 2.08 31-Aug-99
5327 ----------------------
5329 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5330 trying to match at the startoffset position, but instead was moving forward to
5331 the next newline as if a previous match had failed.
5333 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5334 and could get into a loop if a null string was matched other than at the start
5335 of the subject.
5337 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5338 be distinguished at compile time, and for completeness also added PCRE_DATE.
5340 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5341 in GnuWin32 environments.
5344 Version 2.07 29-Jul-99
5345 ----------------------
5347 1. The documentation is now supplied in plain text form and HTML as well as in
5348 the form of man page sources.
5350 2. C++ compilers don't like assigning (void *) values to other pointer types.
5351 In particular this affects malloc(). Although there is no problem in Standard
5352 C, I've put in casts to keep C++ compilers happy.
5354 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5355 should be (const char *).
5357 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5358 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5359 However, I haven't made this a standard facility. The documentation doesn't
5360 mention it, and the Makefile doesn't support it.
5362 5. The Makefile now contains an "install" target, with editable destinations at
5363 the top of the file. The pcretest program is not installed.
5365 6. pgrep -V now gives the PCRE version number and date.
5367 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5368 causing the entire string to be ignored, instead of just the last character.
5370 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5371 non-matching string, it can take a very, very long time, even for strings of
5372 quite modest length, because of the nested recursion. PCRE now does better in
5373 some of these cases. It does this by remembering the last required literal
5374 character in the pattern, and pre-searching the subject to ensure it is present
5375 before running the real match. In other words, it applies a heuristic to detect
5376 some types of certain failure quickly, and in the above example, if presented
5377 with a string that has no trailing " it gives "no match" very quickly.
5379 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5380 other alternatives are tried instead.
5383 Version 2.06 09-Jun-99
5384 ----------------------
5386 1. Change pcretest's output for amount of store used to show just the code
5387 space, because the remainder (the data block) varies in size between 32-bit and
5388 64-bit systems.
5390 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5391 start matching at. This allows lookbehinds to work when searching for multiple
5392 occurrences in a string.
5394 3. Added additional options to pcretest for testing multiple occurrences:
5396 /+ outputs the rest of the string that follows a match
5397 /g loops for multiple occurrences, using the new startoffset argument
5398 /G loops for multiple occurrences by passing an incremented pointer
5400 4. PCRE wasn't doing the "first character" optimization for patterns starting
5401 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5402 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5403 the letter 'x'. On long subject strings, this gives a significant speed-up.
5406 Version 2.05 21-Apr-99
5407 ----------------------
5409 1. Changed the type of magic_number from int to long int so that it works
5410 properly on 16-bit systems.
5412 2. Fixed a bug which caused patterns starting with .* not to work correctly
5413 when the subject string contained newline characters. PCRE was assuming
5414 anchoring for such patterns in all cases, which is not correct because .* will
5415 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5416 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5417 must be retried after every newline in the subject.
5420 Version 2.04 18-Feb-99
5421 ----------------------
5423 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5424 computation of the store needed to hold the pattern was incorrect (too large).
5425 If such patterns were nested a few deep, this could multiply and become a real
5426 problem.
5428 2. Added /M option to pcretest to show the memory requirement of a specific
5429 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5431 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5432 compiled in such a way that the backtracking after subsequent failure was
5433 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5434 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5437 Version 2.03 02-Feb-99
5438 ----------------------
5440 1. Fixed typo and small mistake in man page.
5442 2. Added 4th condition (GPL supersedes if conflict) and created separate
5443 LICENCE file containing the conditions.
5445 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5446 Perl, that is the internal \ allows the delimiter to be included in the
5447 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5448 the final delimiter, add \ to the end of the pattern (to test the error).
5450 4. Added the convenience functions for extracting substrings after a successful
5451 match. Updated pcretest to make it able to test these functions.
5454 Version 2.02 14-Jan-99
5455 ----------------------
5457 1. Initialized the working variables associated with each extraction so that
5458 their saving and restoring doesn't refer to uninitialized store.
5460 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5461 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5462 fix the problem.
5464 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5465 calls, and wasn't printing the correct value for compiling calls. Increased the
5466 default value of LOOPREPEAT, and the number of significant figures in the
5467 times.
5469 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5471 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5472 a building problem on Windows NT with a FAT file system.
5475 Version 2.01 21-Oct-98
5476 ----------------------
5478 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5479 to character tables built by pcre_maketables() in the current locale. If NULL
5480 is passed, the default tables are used.
5483 Version 2.00 24-Sep-98
5484 ----------------------
5486 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5487 it any more.
5489 2. Allow quantification of (?>) groups, and make it work correctly.
5491 3. The first character computation wasn't working for (?>) groups.
5493 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5494 end of the subject) and add 5.005's \z, which really does match only at the
5495 very end of the subject.
5497 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5499 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5500 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5501 localized options. All options to pcre_study() were also removed.
5503 7. Add other new features from 5.005:
5505 $(?<= positive lookbehind
5506 $(?<! negative lookbehind
5507 (?imsx-imsx) added the unsetting capability
5508 such a setting is global if at outer level; local otherwise
5509 (?imsx-imsx:) non-capturing groups with option setting
5510 (?(cond)re|re) conditional pattern matching
5512 A backreference to itself in a repeated group matches the previous
5513 captured string.
5515 8. General tidying up of studying (both automatic and via "study")
5516 consequential on the addition of new assertions.
5518 9. As in 5.005, unlimited repeated groups that could match an empty substring
5519 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5520 runtime if any iteration does actually match an empty substring.
5522 10. Include the RunTest script in the distribution.
5524 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5525 discrepancies, some of which were old and were also with respect to 5.004. They
5526 have now been fixed.
5529 Version 1.09 28-Apr-98
5530 ----------------------
5532 1. A negated single character class followed by a quantifier with a minimum
5533 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5534 program crashes, or just wrong answers. This did not apply to negated classes
5535 containing more than one character, or to minima other than one.
5538 Version 1.08 27-Mar-98
5539 ----------------------
5541 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5543 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5544 latter must appear before anything that relies on it in the pattern.
5547 Version 1.07 16-Feb-98
5548 ----------------------
5550 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5551 repeat of a potentially empty string).
5554 Version 1.06 23-Jan-98
5555 ----------------------
5557 1. Added Markus Oberhumer's little patches for C++.
5559 2. Literal strings longer than 255 characters were broken.
5562 Version 1.05 23-Dec-97
5563 ----------------------
5565 1. Negated character classes containing more than one character were failing if
5566 PCRE_CASELESS was set at run time.
5569 Version 1.04 19-Dec-97
5570 ----------------------
5572 1. Corrected the man page, where some "const" qualifiers had been omitted.
5574 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5575 input syntax.
5577 3. Fixed memory leak which occurred when a regex with back references was
5578 matched with an offsets vector that wasn't big enough. The temporary memory
5579 that is used in this case wasn't being freed if the match failed.
5581 4. Tidied pcretest to ensure it frees memory that it gets.
5583 5. Temporary memory was being obtained in the case where the passed offsets
5584 vector was exactly big enough.
5586 6. Corrected definition of offsetof() from change 5 below.
5588 7. I had screwed up change 6 below and broken the rules for the use of
5589 setjmp(). Now fixed.
5592 Version 1.03 18-Dec-97
5593 ----------------------
5595 1. A erroneous regex with a missing opening parenthesis was correctly
5596 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5597 on some systems.
5599 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5600 it was reported that one broken compiler failed on the former because "code" is
5601 also an independent variable.
5603 3. The erroneous regex a[]b caused an array overrun reference.
5605 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5606 fail on data ending with that character. (It was going on too far, and checking
5607 the next character, typically a binary zero.) This was specific to the
5608 optimized code for single-character negative classes.
5610 5. Added a contributed patch from the TIN world which does the following:
5612 + Add an undef for memmove, in case the the system defines a macro for it.
5614 + Add a definition of offsetof(), in case there isn't one. (I don't know
5615 the reason behind this - offsetof() is part of the ANSI standard - but
5616 it does no harm).
5618 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5619 most of the places where whitespace preceded '#'. I have given up and
5620 allowed the remaining 2 cases to be at the margin.
5622 + Rename some variables in pcre to eliminate shadowing. This seems very
5623 pedantic, but does no harm, of course.
5625 6. Moved the call to setjmp() into its own function, to get rid of warnings
5626 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5628 7. Constructs such as \d{8,} were compiling into the equivalent of
5629 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5630 outcome, but in this particular case used more store than had been allocated,
5631 which caused the bug to be discovered because it threw up an internal error.
5633 8. The debugging code in both pcre and pcretest for outputting the compiled
5634 form of a regex was going wrong in the case of back references followed by
5635 curly-bracketed repeats.
5638 Version 1.02 12-Dec-97
5639 ----------------------
5641 1. Typos in pcre.3 and comments in the source fixed.
5643 2. Applied a contributed patch to get rid of places where it used to remove
5644 'const' from variables, and fixed some signed/unsigned and uninitialized
5645 variable warnings.
5647 3. Added the "runtest" target to Makefile.
5649 4. Set default compiler flag to -O2 rather than just -O.
5652 Version 1.01 19-Nov-97
5653 ----------------------
5655 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5656 like /([ab]*)*/, that is, for classes with more than one character in them.
5658 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5659 as /((?>a*))*/ (a PCRE_EXTRA facility).
5662 Version 1.00 18-Nov-97
5663 ----------------------
5665 1. Added compile-time macros to support systems such as SunOS4 which don't have
5666 memmove() or strerror() but have other things that can be used instead.
5668 2. Arranged that "make clean" removes the executables.
5671 Version 0.99 27-Oct-97
5672 ----------------------
5674 1. Fixed bug in code for optimizing classes with only one character. It was
5675 initializing a 32-byte map regardless, which could cause it to run off the end
5676 of the memory it had got.
5678 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5681 Version 0.98 22-Oct-97
5682 ----------------------
5684 1. Fixed bug in code for handling temporary memory usage when there are more
5685 back references than supplied space in the ovector. This could cause segfaults.
5688 Version 0.97 21-Oct-97
5689 ----------------------
5691 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5693 2. Optimized negated single characters not to use a bit map.
5695 3. Brought error texts together as macro definitions; clarified some of them;
5696 fixed one that was wrong - it said "range out of order" when it meant "invalid
5697 escape sequence".
5699 4. Changed some char * arguments to const char *.
5701 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5703 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5704 pcretest.
5707 Version 0.96 16-Oct-97
5708 ----------------------
5710 1. Added a simple "pgrep" utility to the distribution.
5712 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5713 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5714 where "ddd" means "one or more decimal digits".
5716 3. Fixed serious bug. If a pattern had a back reference, but the call to
5717 pcre_exec() didn't supply a large enough ovector to record the related
5718 identifying subpattern, the match always failed. PCRE now remembers the number
5719 of the largest back reference, and gets some temporary memory in which to save
5720 the offsets during matching if necessary, in order to ensure that
5721 backreferences always work.
5723 4. Increased the compatibility with Perl in a number of ways:
5725 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5726 to request this handling. The option can be set at compile or exec time.
5728 (b) $ matches before a terminating newline by default; an option
5729 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5730 mode). The option can be set at compile or exec time.
5732 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5733 the same as Perl's. If the decimal number it represents is less than 10
5734 or there aren't that many previous left capturing parentheses, an octal
5735 escape is read. Inside a character class, it's always an octal escape,
5736 even if it is a single digit.
5738 (d) An escaped but undefined alphabetic character is taken as a literal,
5739 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5740 escapes.
5742 (e) {0} is now permitted. (The previous item is removed from the compiled
5743 pattern).
5745 5. Changed all the names of code files so that the basic parts are no longer
5746 than 10 characters, and abolished the teeny "globals.c" file.
5748 6. Changed the handling of character classes; they are now done with a 32-byte
5749 bit map always.
5751 7. Added the -d and /D options to pcretest to make it possible to look at the
5752 internals of compilation without having to recompile pcre.
5755 Version 0.95 23-Sep-97
5756 ----------------------
5758 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5759 \x20 at the start of a run of normal characters. These were being treated as
5760 real characters, instead of the source characters being re-checked.
5763 Version 0.94 18-Sep-97
5764 ----------------------
5766 1. The functions are now thread-safe, with the caveat that the global variables
5767 containing pointers to malloc() and free() or alternative functions are the
5768 same for all threads.
5770 2. Get pcre_study() to generate a bitmap of initial characters for non-
5771 anchored patterns when this is possible, and use it if passed to pcre_exec().
5774 Version 0.93 15-Sep-97
5775 ----------------------
5777 1. /(b)|(:+)/ was computing an incorrect first character.
5779 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5780 but not actually doing anything yet.
5782 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5783 as Perl does (e.g. [-az] or [az-]).
5785 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5786 all possible positions.
5788 5. Split up into different modules to avoid including unneeded functions in a
5789 compiled binary. However, compile and exec are still in one module. The "study"
5790 function is split off.
5792 6. The character tables are now in a separate module whose source is generated
5793 by an auxiliary program - but can then be edited by hand if required. There are
5794 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5795 toupper() in the code.
5797 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5798 make them global. Abolish the function for setting them, as the caller can now
5799 set them directly.
5802 Version 0.92 11-Sep-97
5803 ----------------------
5805 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5806 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5808 2. Caseless matching was not working in character classes if the characters in
5809 the pattern were in upper case.
5811 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5813 4. Make PCRE_ANCHORED public and accept as a compile option.
5815 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5816 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5817 pass them.
5819 6. Give an error if bad option bits passed at compile or run time.
5821 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5822 pcretest to cause it to pass that flag.
5824 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5825 options, and the first character, if set.
5827 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5830 Version 0.91 10-Sep-97
5831 ----------------------
5833 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5834 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5836 2. PCRE was looping on encountering an indefinitely repeated back reference to
5837 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5838 Perl does - treats the match as successful.
5840 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5