ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1589 - (show annotations)
Mon Aug 10 14:19:06 2015 UTC (4 years, 5 months ago) by ph10
File size: 282855 byte(s)
Add missing integer overflow checks.
1 ChangeLog for PCRE
2 ------------------
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
7 Version 8.38 xx-xxx-xxxx
8 ------------------------
10 1. If a group that contained a recursive back reference also contained a
11 forward reference subroutine call followed by a non-forward-reference
12 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13 compile correct code, leading to undefined behaviour or an internally
14 detected error. This bug was discovered by the LLVM fuzzer.
16 2. Quantification of certain items (e.g. atomic back references) could cause
17 incorrect code to be compiled when recursive forward references were
18 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19 This bug was discovered by the LLVM fuzzer.
21 3. A repeated conditional group whose condition was a reference by name caused
22 a buffer overflow if there was more than one group with the given name.
23 This bug was discovered by the LLVM fuzzer.
25 4. A recursive back reference by name within a group that had the same name as
26 another group caused a buffer overflow. For example:
27 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
29 5. A forward reference by name to a group whose number is the same as the
30 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31 a buffer overflow at compile time. This bug was discovered by the LLVM
32 fuzzer.
34 6. A lookbehind assertion within a set of mutually recursive subpatterns could
35 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
37 7. Another buffer overflow bug involved duplicate named groups with a
38 reference between their definition, with a group that reset capture
39 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40 fixed by always allowing for more memory, even if not needed. (A proper fix
41 is implemented in PCRE2, but it involves more refactoring.)
43 8. There was no check for integer overflow in subroutine calls such as (?123).
45 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
46 being treated as a literal 'l' instead of causing an error.
48 10. There was a buffer overflow if pcre_exec() was called with an ovector of
49 size 1. This bug was found by american fuzzy lop.
51 11. If a non-capturing group containing a conditional group that could match
52 an empty string was repeated, it was not identified as matching an empty
53 string itself. For example: /^(?:(?(1)x|)+)+$()/.
55 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
56 \a and \e in test subject lines.
58 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
59 instead of the EBCDIC value.
61 14. The handling of \c in an EBCDIC environment has been revised so that it is
62 now compatible with the specification in Perl's perlebcdic page.
64 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
65 ASCII/Unicode. This has now been added to the list of characters that are
66 recognized as white space in EBCDIC.
68 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
69 error (correctly) when used outside a class, but did not give an error
70 within a class.
72 17. \h within a class was incorrectly compiled in EBCDIC environments.
74 18. A pattern with an unmatched closing parenthesis that contained a backward
75 assertion which itself contained a forward reference caused buffer
76 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
78 19. JIT should return with error when the compiled pattern requires more stack
79 space than the maximum.
81 20. A possessively repeated conditional group that could match an empty string,
82 for example, /(?(R))*+/, was incorrectly compiled.
84 21. Fix infinite recursion in the JIT compiler when certain patterns such as
85 /(?:|a|){100}x/ are analysed.
87 22. Some patterns with character classes involving [: and \\ were incorrectly
88 compiled and could cause reading from uninitialized memory or an incorrect
89 error diagnosis.
91 23. Pathological patterns containing many nested occurrences of [: caused
92 pcre_compile() to run for a very long time.
94 24. A conditional group with only one branch has an implicit empty alternative
95 branch and must therefore be treated as potentially matching an empty
96 string.
98 25. If (?R was followed by - or + incorrect behaviour happened instead of a
99 diagnostic.
101 26. Arrange to give up on finding the minimum matching length for overly
102 complex patterns.
104 27. Similar to (4) above: in a pattern with duplicated named groups and an
105 occurrence of (?| it is possible for an apparently non-recursive back
106 reference to become recursive if a later named group with the relevant
107 number is encountered. This could lead to a buffer overflow. Wen Guanxing
108 from Venustech ADLAB discovered this bug.
110 28. If pcregrep was given the -q option with -c or -l, or when handling a
111 binary file, it incorrectly wrote output to stdout.
113 29. The JIT compiler did not restore the control verb head in case of *THEN
114 control verbs. This issue was found by Karl Skomski with a custom LLVM
115 fuzzer.
117 30. Error messages for syntax errors following \g and \k were giving inaccurate
118 offsets in the pattern.
120 31. Added a check for integer overflow in conditions (?(<digits>) and
121 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
122 fuzzer.
125 Version 8.37 28-April-2015
126 --------------------------
128 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
129 for those parentheses to be closed with whatever has been captured so far.
130 However, it was failing to mark any other groups between the hightest
131 capture so far and the currrent group as "unset". Thus, the ovector for
132 those groups contained whatever was previously there. An example is the
133 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
135 2. If an assertion condition was quantified with a minimum of zero (an odd
136 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
138 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
139 unrecognized modifier, a crash could occur.
141 4. An attempt to do global matching in pcretest with a zero-length ovector
142 caused a crash.
144 5. Fixed a memory leak during matching that could occur for a subpattern
145 subroutine call (recursive or otherwise) if the number of captured groups
146 that had to be saved was greater than ten.
148 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
149 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
150 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
152 7. A UTF pattern containing a "not" match of a non-ASCII character and a
153 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
155 8. When a pattern is compiled, it remembers the highest back reference so that
156 when matching, if the ovector is too small, extra memory can be obtained to
157 use instead. A conditional subpattern whose condition is a check on a
158 capture having happened, such as, for example in the pattern
159 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
160 setting the highest backreference number. This mattered only if pcre_exec()
161 was called with an ovector that was too small to hold the capture, and there
162 was no other kind of back reference (a situation which is probably quite
163 rare). The effect of the bug was that the condition was always treated as
164 FALSE when the capture could not be consulted, leading to a incorrect
165 behaviour by pcre_exec(). This bug has been fixed.
167 9. A reference to a duplicated named group (either a back reference or a test
168 for being set in a conditional) that occurred in a part of the pattern where
169 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
170 to be incorrectly calculated, leading to overwriting.
172 10. A mutually recursive set of back references such as (\2)(\1) caused a
173 segfault at study time (while trying to find the minimum matching length).
174 The infinite loop is now broken (with the minimum length unset, that is,
175 zero).
177 11. If an assertion that was used as a condition was quantified with a minimum
178 of zero, matching went wrong. In particular, if the whole group had
179 unlimited repetition and could match an empty string, a segfault was
180 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
181 assertions to be quantified, but not if they are being used as conditions,
182 so the above pattern is faulted by Perl. PCRE has now been changed so that
183 it also rejects such patterns.
185 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
186 failed to allow the zero-repeat case if pcre2_exec() was called with an
187 ovector too small to capture the group.
189 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
190 Red Hat Product Security:
192 (a) A crash if /K and /F were both set with the option to save the compiled
193 pattern.
195 (b) Another crash if the option to print captured substrings in a callout
196 was combined with setting a null ovector, for example \O\C+ as a subject
197 string.
199 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
200 forward reference repeated a large (but limited) number of times within a
201 repeated outer group that has a zero minimum quantifier, caused incorrect
202 code to be compiled, leading to the error "internal error:
203 previously-checked referenced subpattern not found" when an incorrect
204 memory address was read. This bug was reported as "heap overflow",
205 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
206 CVE-2015-2325.
208 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
209 call within a group that also contained a recursive back reference caused
210 incorrect code to be compiled. This bug was reported as "heap overflow",
211 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
212 number CVE-2015-2326.
214 24. Computing the size of the JIT read-only data in advance has been a source
215 of various issues, and new ones are still appear unfortunately. To fix
216 existing and future issues, size computation is eliminated from the code,
217 and replaced by on-demand memory allocation.
219 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
220 adjacent to the end of the range, and the range contained characters with
221 more than one other case, caused incorrect behaviour when compiled in UTF
222 mode. In that example, the range a-j was left out of the class.
224 26. Fix JIT compilation of conditional blocks, which assertion
225 is converted to (*FAIL). E.g: /(?(?!))/.
227 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
228 discovered by the LLVM fuzzer.
230 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
231 when this assertion was used as a condition, for example (?(?!)a|b). In
232 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
233 error about an unsupported item.
235 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
236 possessification code could take exponential time to complete. A recursion
237 depth limit of 1000 has been imposed to limit the resources used by this
238 optimization.
240 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
241 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
242 because \S ensures they are all in the class. The code for doing this was
243 interacting badly with the code for computing the amount of space needed to
244 compile the pattern, leading to a buffer overflow. This bug was discovered
245 by the LLVM fuzzer.
247 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
248 other kinds of group caused stack overflow at compile time. This bug was
249 discovered by the LLVM fuzzer.
251 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
252 between a subroutine call and its quantifier was incorrectly compiled,
253 leading to buffer overflow or other errors. This bug was discovered by the
254 LLVM fuzzer.
256 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
257 assertion after (?(. The code was failing to check the character after
258 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
259 was discovered by the LLVM fuzzer.
261 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
262 a fixed maximum following a group that contains a subroutine reference was
263 incorrectly compiled and could trigger buffer overflow. This bug was
264 discovered by the LLVM fuzzer.
266 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
267 caused a stack overflow instead of the diagnosis of a non-fixed length
268 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
270 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
271 (e.g. /(?<=\Ka)/) could make pcregrep loop.
273 37. There was a similar problem to 36 in pcretest for global matches.
275 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
276 and a subsequent item in the pattern caused a non-match, backtracking over
277 the repeated \X did not stop, but carried on past the start of the subject,
278 causing reference to random memory and/or a segfault. There were also some
279 other cases where backtracking after \C could crash. This set of bugs was
280 discovered by the LLVM fuzzer.
282 39. The function for finding the minimum length of a matching string could take
283 a very long time if mutual recursion was present many times in a pattern,
284 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
285 method has been implemented. This infelicity was discovered by the LLVM
286 fuzzer.
288 40. Static linking against the PCRE library using the pkg-config module was
289 failing on missing pthread symbols.
292 Version 8.36 26-September-2014
293 ------------------------------
295 1. Got rid of some compiler warnings in the C++ modules that were shown up by
296 -Wmissing-field-initializers and -Wunused-parameter.
298 2. The tests for quantifiers being too big (greater than 65535) were being
299 applied after reading the number, and stupidly assuming that integer
300 overflow would give a negative number. The tests are now applied as the
301 numbers are read.
303 3. Tidy code in pcre_exec.c where two branches that used to be different are
304 now the same.
306 4. The JIT compiler did not generate match limit checks for certain
307 bracketed expressions with quantifiers. This may lead to exponential
308 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
309 issue should be resolved now.
311 5. Fixed an issue, which occures when nested alternatives are optimized
312 with table jumps.
314 6. Inserted two casts and changed some ints to size_t in the light of some
315 reported 64-bit compiler warnings (Bugzilla 1477).
317 7. Fixed a bug concerned with zero-minimum possessive groups that could match
318 an empty string, which sometimes were behaving incorrectly in the
319 interpreter (though correctly in the JIT matcher). This pcretest input is
320 an example:
322 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
325 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
326 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
327 for an empty string was breaking the inner loop and carrying on at a lower
328 level, when possessive repeated groups should always return to a higher
329 level as they have no backtrack points in them. The empty string test now
330 occurs at the outer level.
332 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
333 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
335 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
336 doesn't).
338 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
339 to hold a pattern to be incorrectly computed (too small) when there were
340 named back references to duplicated names. This could cause "internal
341 error: code overflow" or "double free or corruption" or other memory
342 handling errors.
344 11. When named subpatterns had the same prefixes, back references could be
345 confused. For example, in this pattern:
347 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
349 the reference to 'Name' was incorrectly treated as a reference to a
350 duplicate name.
352 12. A pattern such as /^s?c/mi8 where the optional character has more than
353 one "other case" was incorrectly compiled such that it would only try to
354 match starting at "c".
356 13. When a pattern starting with \s was studied, VT was not included in the
357 list of possible starting characters; this should have been part of the
358 8.34/18 patch.
360 14. If a character class started [\Qx]... where x is any character, the class
361 was incorrectly terminated at the ].
363 15. If a pattern that started with a caseless match for a character with more
364 than one "other case" was studied, PCRE did not set up the starting code
365 unit bit map for the list of possible characters. Now it does. This is an
366 optimization improvement, not a bug fix.
368 16. The Unicode data tables have been updated to Unicode 7.0.0.
370 17. Fixed a number of memory leaks in pcregrep.
372 18. Avoid a compiler warning (from some compilers) for a function call with
373 a cast that removes "const" from an lvalue by using an intermediate
374 variable (to which the compiler does not object).
376 19. Incorrect code was compiled if a group that contained an internal recursive
377 back reference was optional (had quantifier with a minimum of zero). This
378 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
379 caused segmentation faults because of stack overflows at compile time.
381 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
382 group that is quantified with an indefinite repeat, caused a compile-time
383 loop which used up all the system stack and provoked a segmentation fault.
384 This was not the same bug as 19 above.
386 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
387 Patch by Mike Frysinger.
390 Version 8.35 04-April-2014
391 --------------------------
393 1. A new flag is set, when property checks are present in an XCLASS.
394 When this flag is not set, PCRE can perform certain optimizations
395 such as studying these XCLASS-es.
397 2. The auto-possessification of character sets were improved: a normal
398 and an extended character set can be compared now. Furthermore
399 the JIT compiler optimizes more character set checks.
401 3. Got rid of some compiler warnings for potentially uninitialized variables
402 that show up only when compiled with -O2.
404 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
405 of a match later then the end of the match. The pcretest program was not
406 handling the case sensibly - it was outputting from the start to the next
407 binary zero. It now reports this situation in a message, and outputs the
408 text from the end to the start.
410 5. Fast forward search is improved in JIT. Instead of the first three
411 characters, any three characters with fixed position can be searched.
412 Search order: first, last, middle.
414 6. Improve character range checks in JIT. Characters are read by an inprecise
415 function now, which returns with an unknown value if the character code is
416 above a certain threshold (e.g: 256). The only limitation is that the value
417 must be bigger than the threshold as well. This function is useful when
418 the characters above the threshold are handled in the same way.
420 7. The macros whose names start with RAWUCHAR are placeholders for a future
421 mode in which only the bottom 21 bits of 32-bit data items are used. To
422 make this more memorable for those maintaining the code, the names have
423 been changed to start with UCHAR21, and an extensive comment has been added
424 to their definition.
426 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
427 to the export list in Makefile.am (they were accidentally omitted from the
428 8.34 tarball).
430 9. The informational output from pcretest used the phrase "starting byte set"
431 which is inappropriate for the 16-bit and 32-bit libraries. As the output
432 for "first char" and "need char" really means "non-UTF-char", I've changed
433 "byte" to "char", and slightly reworded the output. The documentation about
434 these values has also been (I hope) clarified.
436 10. Another JIT related optimization: use table jumps for selecting the correct
437 backtracking path, when more than four alternatives are present inside a
438 bracket.
440 11. Empty match is not possible, when the minimum length is greater than zero,
441 and there is no \K in the pattern. JIT should avoid empty match checks in
442 such cases.
444 12. In a caseless character class with UCP support, when a character with more
445 than one alternative case was not the first character of a range, not all
446 the alternative cases were added to the class. For example, s and \x{17f}
447 are both alternative cases for S: the class [RST] was handled correctly,
448 but [R-T] was not.
450 13. The configure.ac file always checked for pthread support when JIT was
451 enabled. This is not used in Windows, so I have put this test inside a
452 check for the presence of windows.h (which was already tested for).
454 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
455 The algorithm provides a way to skip certain starting offsets, and usually
456 faster than linear prefix searches.
458 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
459 as for 'fr_FR' and 'french'. For some reason, however, it then used the
460 Windows-specific input and output files, which have 'french' screwed in.
461 So this could never have worked. One of the problems with locales is that
462 they aren't always the same. I have now updated RunTest so that it checks
463 the output of the locale test (test 3) against three different output
464 files, and it allows the test to pass if any one of them matches. With luck
465 this should make the test pass on some versions of Solaris where it was
466 failing. Because of the uncertainty, the script did not used to stop if
467 test 3 failed; it now does. If further versions of a French locale ever
468 come to light, they can now easily be added.
470 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
471 there was a message during ./configure, but it did not stop. This now
472 provokes an error. The invalid example in README has been corrected.
473 If a value less than the minimum is given, the minimum value has always
474 been used, but now a warning is given.
476 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
477 was a bug in the test system, which is now fixed. Also, the list of various
478 configurations that are tested for each release did not have one with both
479 16/32 bits and --enable-bar-anycrlf. It now does.
481 18. pcretest was missing "-C bsr" for displaying the \R default setting.
483 19. Little endian PowerPC systems are supported now by the JIT compiler.
485 20. The fast forward newline mechanism could enter to an infinite loop on
486 certain invalid UTF-8 input. Although we don't support these cases
487 this issue can be fixed by a performance optimization.
489 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
490 not take account if existing stack usage. There is now a new global
491 variable called pcre_stack_guard that can be set to point to an external
492 function to check stack availability. It is called at the start of
493 processing every parenthesized group.
495 22. A typo in the code meant that in ungreedy mode the max/min qualifier
496 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
497 not match "ab".
499 23. When UTF was disabled, the JIT program reported some incorrect compile
500 errors. These messages are silenced now.
502 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
503 compiler.
505 25. Change all the temporary files used in RunGrepTest to be different to those
506 used by RunTest so that the tests can be run simultaneously, for example by
507 "make -j check".
510 Version 8.34 15-December-2013
511 -----------------------------
513 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
514 executable memory. Patch inspired by Carsten Klein.
516 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
517 this macro is never tested and has no effect, because the work to support
518 coverage involves only compiling and linking options and special targets in
519 the Makefile. The comment in config.h implied that defining the macro would
520 enable coverage support, which is totally false. There was also support for
521 setting this macro in the CMake files (my fault, I just copied it from
522 configure). SUPPORT_GCOV has now been removed.
524 3. Make a small performance improvement in strlen16() and strlen32() in
525 pcretest.
527 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
528 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
529 these cases). There was also one in pcretest.c.
531 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
533 6. In UTF mode, the code for checking whether a group could match an empty
534 string (which is used for indefinitely repeated groups to allow for
535 breaking an infinite loop) was broken when the group contained a repeated
536 negated single-character class with a character that occupied more than one
537 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
538 UTF-8 mode). The effect was undefined: the group might or might not be
539 deemed as matching an empty string, or the program might have crashed.
541 7. The code for checking whether a group could match an empty string was not
542 recognizing that \h, \H, \v, \V, and \R must match a character.
544 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
545 an empty string. If it can, pcretest shows this in its information output.
547 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
548 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
549 matched by pcre_exec() without using JIT:
551 (a) If the rest of the pattern did not match after a maximal run of
552 grapheme clusters, the code for backing up to try with fewer of them
553 did not always back up over a full grapheme when characters that do not
554 have the modifier quality were involved, e.g. Hangul syllables.
556 (b) If the match point in a subject started with modifier character, and
557 there was no match, the code could incorrectly back up beyond the match
558 point, and potentially beyond the first character in the subject,
559 leading to a segfault or an incorrect match result.
561 10. A conditional group with an assertion condition could lead to PCRE
562 recording an incorrect first data item for a match if no other first data
563 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
564 first data item, and therefore matched "ca" after "c" instead of at the
565 start.
567 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
568 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
570 12. The source of pcregrep now includes z/OS-specific code so that it can be
571 compiled for z/OS as part of the special z/OS distribution.
573 13. Added the -T and -TM options to pcretest.
575 14. The code in pcre_compile.c for creating the table of named capturing groups
576 has been refactored. Instead of creating the table dynamically during the
577 actual compiling pass, the information is remembered during the pre-compile
578 pass (on the stack unless there are more than 20 named groups, in which
579 case malloc() is used) and the whole table is created before the actual
580 compile happens. This has simplified the code (it is now nearly 150 lines
581 shorter) and prepared the way for better handling of references to groups
582 with duplicate names.
584 15. A back reference to a named subpattern when there is more than one of the
585 same name now checks them in the order in which they appear in the pattern.
586 The first one that is set is used for the reference. Previously only the
587 first one was inspected. This change makes PCRE more compatible with Perl.
589 16. Unicode character properties were updated from Unicode 6.3.0.
591 17. The compile-time code for auto-possessification has been refactored, based
592 on a patch by Zoltan Herczeg. It now happens after instead of during
593 compilation. The code is cleaner, and more cases are handled. The option
594 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
595 options in pcretest are provided to set it. It can also be set by
596 (*NO_AUTO_POSSESS) at the start of a pattern.
598 18. The character VT has been added to the default ("C" locale) set of
599 characters that match \s and are generally treated as white space,
600 following this same change in Perl 5.18. There is now no difference between
601 "Perl space" and "POSIX space". Whether VT is treated as white space in
602 other locales depends on the locale.
604 19. The code for checking named groups as conditions, either for being set or
605 for being recursed, has been refactored (this is related to 14 and 15
606 above). Processing unduplicated named groups should now be as fast at
607 numerical groups, and processing duplicated groups should be faster than
608 before.
610 20. Two patches to the CMake build system, by Alexander Barkov:
612 (1) Replace the "source" command by "." in CMakeLists.txt because
613 "source" is a bash-ism.
615 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
616 without these the CMake build does not work on Solaris.
618 21. Perl has changed its handling of \8 and \9. If there is no previously
619 encountered capturing group of those numbers, they are treated as the
620 literal characters 8 and 9 instead of a binary zero followed by the
621 literals. PCRE now does the same.
623 22. Following Perl, added \o{} to specify codepoints in octal, making it
624 possible to specify values greater than 0777 and also making them
625 unambiguous.
627 23. Perl now gives an error for missing closing braces after \x{... instead of
628 treating the string as literal. PCRE now does the same.
630 24. RunTest used to grumble if an inappropriate test was selected explicitly,
631 but just skip it when running all tests. This make it awkward to run ranges
632 of tests when one of them was inappropriate. Now it just skips any
633 inappropriate tests, as it always did when running all tests.
635 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
636 character types such as \d or \w, too many callouts were inserted, and the
637 data that they returned was rubbish.
639 26. In UCP mode, \s was not matching two of the characters that Perl matches,
640 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
641 were matched by \h. The code has now been refactored so that the lists of
642 the horizontal and vertical whitespace characters used for \h and \v (which
643 are defined only in one place) are now also used for \s.
645 27. Add JIT support for the 64 bit TileGX architecture.
646 Patch by Jiong Wang (Tilera Corporation).
648 28. Possessive quantifiers for classes (both explicit and automatically
649 generated) now use special opcodes instead of wrapping in ONCE brackets.
651 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
652 (because it's meaningless), this was not happening when PCRE_CASELESS was
653 set. Not wrong, but inefficient.
655 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
656 properties for \w, \d, etc) is present in a test regex. Otherwise if the
657 test contains no characters greater than 255, Perl doesn't realise it
658 should be using Unicode semantics.
660 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
661 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
662 does in Unicode mode.
664 32. Added the "forbid" facility to pcretest so that putting tests into the
665 wrong test files can sometimes be quickly detected.
667 33. There is now a limit (default 250) on the depth of nesting of parentheses.
668 This limit is imposed to control the amount of system stack used at compile
669 time. It can be changed at build time by --with-parens-nest-limit=xxx or
670 the equivalent in CMake.
672 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
673 errors. Perl warns for these when in warning mode, but PCRE has no facility
674 for giving warnings.
676 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
677 However, this was not working for (?!) because it is optimized to (*FAIL),
678 for which PCRE does not allow quantifiers. The optimization is now disabled
679 when a quantifier follows (?!). I can't see any use for this, but it makes
680 things uniform.
682 36. Perl no longer allows group names to start with digits, so I have made this
683 change also in PCRE. It simplifies the code a bit.
685 37. In extended mode, Perl ignores spaces before a + that indicates a
686 possessive quantifier. PCRE allowed a space before the quantifier, but not
687 before the possessive +. It now does.
689 38. The use of \K (reset reported match start) within a repeated possessive
690 group such as (a\Kb)*+ was not working.
692 40. Document that the same character tables must be used at compile time and
693 run time, and that the facility to pass tables to pcre_exec() and
694 pcre_dfa_exec() is for use only with saved/restored patterns.
696 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
697 features for Builds with MSVC:
699 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
700 builds). The .rc files can be used to set FileDescription and many other
701 attributes.
703 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
704 This allows higher-level build scripts which want .pdb files to avoid
705 hard-coding the exact files needed."
707 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
708 mean "start of word" and "end of word", respectively, as a transition aid.
710 43. A minimizing repeat of a class containing codepoints greater than 255 in
711 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
712 compiled to use the heap for recursion.
714 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
715 is configured.
718 Version 8.33 28-May-2013
719 ------------------------
721 1. Added 'U' to some constants that are compared to unsigned integers, to
722 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
723 variables that are added to signed variables, to ensure the result is
724 signed and can be negated.
726 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
727 CMake config files.
729 3. Revise the creation of config.h.generic so that all boolean macros are
730 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
731 overriding via -D on the command line possible.
733 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
734 to unsigned int is reported to make a quite noticeable speed difference in
735 a specific Windows environment. Testing on Linux did also appear to show
736 some benefit (and it is clearly not harmful). Also fixed the definition of
737 Xop which should be unsigned.
739 5. Related to (4), changing the definition of the intermediate variable cc
740 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
741 improvements.
743 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
744 unnecessary spaces.
746 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
747 and later.
749 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
751 9. Optimizing fast_forward_start_bits in JIT.
753 10. Adding support for callouts in JIT, and fixing some issues revealed
754 during this work. Namely:
756 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
758 (b) Minimum length was not checked before the matching is started.
760 11. The value of capture_last that is passed to callouts was incorrect in some
761 cases when there was a capture on one path that was subsequently abandoned
762 after a backtrack. Also, the capture_last value is now reset after a
763 recursion, since all captures are also reset in this case.
765 12. The interpreter no longer returns the "too many substrings" error in the
766 case when an overflowing capture is in a branch that is subsequently
767 abandoned after a backtrack.
769 13. In the pathological case when an offset vector of size 2 is used, pcretest
770 now prints out the matched string after a yield of 0 or 1.
772 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
773 Only supported by the JIT compiler at the moment.
775 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
777 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
778 offset of the starting point of the matching process, provided the offsets
779 vector is large enough.
781 17. The \A escape now records a lookbehind value of 1, though its execution
782 does not actually inspect the previous character. This is to ensure that,
783 in partial multi-segment matching, at least one character from the old
784 segment is retained when a new segment is processed. Otherwise, if there
785 are no lookbehinds in the pattern, \A might match incorrectly at the start
786 of a new segment.
788 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
790 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
791 gives some modest performance improvement in 8-bit mode.
793 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
794 be expressed in certain programming languages using Universal Character
795 Names.
797 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
798 which points out that "non characters" are not "characters that may not
799 appear in Unicode strings" but rather "characters that are reserved for
800 internal use and have only local meaning".
802 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
803 there was a conditional group that depended on an assertion, if the
804 assertion was false, the callout that immediately followed the alternation
805 in the condition was skipped when pcre_exec() was used for matching.
807 23. Allow an explicit callout to be inserted before an assertion that is the
808 condition for a conditional group, for compatibility with automatic
809 callouts, which always insert a callout at this point.
811 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
812 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
814 25. (*PRUNE) is now supported by the JIT compiler.
816 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
818 27. Fix the case where there are two or more SKIPs with arguments that may be
819 ignored.
821 28. (*SKIP) is now supported by the JIT compiler.
823 29. (*THEN) is now supported by the JIT compiler.
825 30. Update RunTest with additional test selector options.
827 31. The way PCRE handles backtracking verbs has been changed in two ways.
829 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
830 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
831 backtracking. In some cases this makes it more Perl-compatible, but Perl's
832 rather obscure rules do not always do the same thing.
834 (2) Previously, backtracking verbs were confined within assertions. This is
835 no longer the case for positive assertions, except for (*ACCEPT). Again,
836 this sometimes improves Perl compatibility, and sometimes does not.
838 32. A number of tests that were in test 2 because Perl did things differently
839 have been moved to test 1, because either Perl or PCRE has changed, and
840 these tests are now compatible.
842 32. Backtracking control verbs are now handled in the same way in JIT and
843 interpreter.
845 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
846 contained a forward subroutine reference caused a compile error.
848 34. Auto-detect and optimize limited repetitions in JIT.
850 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
851 blocking (*UTF) etc.
853 36. In the interpreter, maximizing pattern repetitions for characters and
854 character types now use tail recursion, which reduces stack usage.
856 37. The value of the max lookbehind was not correctly preserved if a compiled
857 and saved regex was reloaded on a host of different endianness.
859 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
860 of the compiled pattern block, expand the flags field from 16 to 32 bits
861 because it was almost full.
863 39. Try madvise first before posix_madvise.
865 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
866 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
867 original change locked out all matches of empty strings. This has been
868 changed so that one match of an empty string per line is recognized.
869 Subsequent searches on the same line (for colouring or for --only-matching,
870 for example) do not recognize empty strings.
872 41. Applied a user patch to fix a number of spelling mistakes in comments.
874 42. Data lines longer than 65536 caused pcretest to crash.
876 43. Clarified the data type for length and startoffset arguments for pcre_exec
877 and pcre_dfa_exec in the function-specific man pages, where they were
878 explicitly stated to be in bytes, never having been updated. I also added
879 some clarification to the pcreapi man page.
881 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
882 a segmentation fault.
885 Version 8.32 30-November-2012
886 -----------------------------
888 1. Improved JIT compiler optimizations for first character search and single
889 character iterators.
891 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
892 Patch by Daniel Richard G.
894 3. Single character iterator optimizations in the JIT compiler.
896 4. Improved JIT compiler optimizations for character ranges.
898 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
899 Reported by Giuseppe D'Angelo.
901 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
902 of a line, was being set incorrectly in cases where .* appeared inside
903 atomic brackets at the start of a pattern, or where there was a subsequent
904 *PRUNE or *SKIP.
906 7. Improved instruction cache flush for POWER/PowerPC.
907 Patch by Daniel Richard G.
909 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
910 grep:
912 (a) There is now no limit to the number of patterns to be matched.
914 (b) An error is given if a pattern is too long.
916 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
917 are now supported.
919 (d) --exclude-from and --include-from (multiple use) have been added.
921 (e) Exclusions and inclusions now apply to all files and directories, not
922 just to those obtained from scanning a directory recursively.
924 (f) Multiple uses of -f and --file-list are now supported.
926 (g) In a Windows environment, the default for -d has been changed from
927 "read" (the GNU grep default) to "skip", because otherwise the presence
928 of a directory in the file list provokes an error.
930 (h) The documentation has been revised and clarified in places.
932 9. Improve the matching speed of capturing brackets.
934 10. Changed the meaning of \X so that it now matches a Unicode extended
935 grapheme cluster.
937 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
938 out POSIX threads when JIT support is configured.
940 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
942 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
943 structure could go wrong in environments where size_t is not the same size
944 as int.
946 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
947 set.
949 15. The EBCDIC support had decayed; later updates to the code had included
950 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
951 general tidy up of EBCDIC-related issues, and the documentation was also
952 not quite right. There is now a test that can be run on ASCII systems to
953 check some of the EBCDIC-related things (but is it not a full test).
955 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
956 in a small tidy to the code.
958 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
960 18. If the --only-matching (-o) option in pcregrep is specified multiple
961 times, each one causes appropriate output. For example, -o1 -o2 outputs the
962 substrings matched by the 1st and 2nd capturing parentheses. A separating
963 string can be specified by --om-separator (default empty).
965 19. Improving the first n character searches.
967 20. Turn case lists for horizontal and vertical white space into macros so that
968 they are defined only once.
970 21. This set of changes together give more compatible Unicode case-folding
971 behaviour for characters that have more than one other case when UCP
972 support is available.
974 (a) The Unicode property table now has offsets into a new table of sets of
975 three or more characters that are case-equivalent. The MultiStage2.py
976 script that generates these tables (the pcre_ucd.c file) now scans
977 CaseFolding.txt instead of UnicodeData.txt for character case
978 information.
980 (b) The code for adding characters or ranges of characters to a character
981 class has been abstracted into a generalized function that also handles
982 case-independence. In UTF-mode with UCP support, this uses the new data
983 to handle characters with more than one other case.
985 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
986 whose other case is greater than 256 are now correctly matched
987 caselessly. Previously, the high codepoint matched the low one, but not
988 vice versa.
990 (d) The processing of \h, \H, \v, and \ in character classes now makes use
991 of the new class addition function, using character lists defined as
992 macros alongside the case definitions of 20 above.
994 (e) Caseless back references now work with characters that have more than
995 one other case.
997 (f) General caseless matching of characters with more than one other case
998 is supported.
1000 22. Unicode character properties were updated from Unicode 6.2.0
1002 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1004 24. Add support for 32-bit character strings, and UTF-32
1006 25. Major JIT compiler update (code refactoring and bugfixing).
1007 Experimental Sparc 32 support is added.
1009 26. Applied a modified version of Daniel Richard G's patch to create
1010 pcre.h.generic and config.h.generic by "make" instead of in the
1011 PrepareRelease script.
1013 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1014 pcre_compile.c when checking for a zero character.
1016 28. Introducing a native interface for JIT. Through this interface, the compiled
1017 machine code can be directly executed. The purpose of this interface is to
1018 provide fast pattern matching, so several sanity checks are not performed.
1019 However, feature tests are still performed. The new interface provides
1020 1.4x speedup compared to the old one.
1022 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1023 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1024 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1026 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1027 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1028 the "old" RFC 2279). Instead, it ended up passing a negative length to
1029 pcre_exec().
1031 31. Add support for GCC's visibility feature to hide internal functions.
1033 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1034 "unknown -C option" after outputting 0 or 1.
1036 33. There is now support for generating a code coverage report for the test
1037 suite in environments where gcc is the compiler and lcov is installed. This
1038 is mainly for the benefit of the developers.
1040 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1041 unaddressable using valgrind annotations, allowing valgrind to detect
1042 invalid memory accesses. This is mainly for the benefit of the developers.
1044 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1046 26. Give configure error if --enable-cpp but no C++ compiler found.
1049 Version 8.31 06-July-2012
1050 -------------------------
1052 1. Fixing a wrong JIT test case and some compiler warnings.
1054 2. Removed a bashism from the RunTest script.
1056 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1057 to unsigned type, result still unsigned" that was given by an MS compiler
1058 on encountering the code "-sizeof(xxx)".
1060 4. Partial matching support is added to the JIT compiler.
1062 5. Fixed several bugs concerned with partial matching of items that consist
1063 of more than one character:
1065 (a) /^(..)\1/ did not partially match "aba" because checking references was
1066 done on an "all or nothing" basis. This also applied to repeated
1067 references.
1069 (b) \R did not give a hard partial match if \r was found at the end of the
1070 subject.
1072 (c) \X did not give a hard partial match after matching one or more
1073 characters at the end of the subject.
1075 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1076 a partial match for the string "\r".
1078 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1079 a partial match for a CR character at the end of the subject string.
1081 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1082 running pcretest, the text "(JIT)" added to the output whenever JIT is
1083 actually used to run the match.
1085 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1086 or /S+[+] with a digit between 1 and 7.
1088 8. OP_NOT now supports any UTF character not just single-byte ones.
1090 9. (*MARK) control verb is now supported by the JIT compiler.
1092 10. The command "./RunTest list" lists the available tests without actually
1093 running any of them. (Because I keep forgetting what they all are.)
1097 12. Applied a (slightly modified) user-supplied patch that improves performance
1098 when the heap is used for recursion (compiled with --disable-stack-for-
1099 recursion). Instead of malloc and free for each heap frame each time a
1100 logical recursion happens, frames are retained on a chain and re-used where
1101 possible. This sometimes gives as much as 30% improvement.
1103 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1104 call.
1106 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1108 15. It is now possible to link pcretest with libedit as an alternative to
1109 libreadline.
1111 16. (*COMMIT) control verb is now supported by the JIT compiler.
1113 17. The Unicode data tables have been updated to Unicode 6.1.0.
1115 18. Added --file-list option to pcregrep.
1117 19. Added binary file support to pcregrep, including the -a, --binary-files,
1118 -I, and --text options.
1120 20. The madvise function is renamed for posix_madvise for QNX compatibility
1121 reasons. Fixed by Giuseppe D'Angelo.
1123 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1125 22. Changed the option for creating symbolic links for 16-bit man pages from
1126 -s to -sf so that re-installing does not cause issues.
1128 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1130 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1131 matches in certain environments (the workspace was not being correctly
1132 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1133 some of the workspace data at the beginning of a restart.
1135 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1136 was not doing so when it should - probably a typo introduced by SVN 528
1137 (change 8.10/14).
1139 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1140 \w+ when the character tables indicated that \x{c4} was a word character.
1141 There were several related cases, all because the tests for doing a table
1142 lookup were testing for characters less than 127 instead of 255.
1144 27. If a pattern contains capturing parentheses that are not used in a match,
1145 their slots in the ovector are set to -1. For those that are higher than
1146 any matched groups, this happens at the end of processing. In the case when
1147 there were back references that the ovector was too small to contain
1148 (causing temporary malloc'd memory to be used during matching), and the
1149 highest capturing number was not used, memory off the end of the ovector
1150 was incorrectly being set to -1. (It was using the size of the temporary
1151 memory instead of the true size.)
1153 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1154 ovector size, it uses memory at the end of the block that it has got.
1156 29. Check for an overlong MARK name and give an error at compile time. The
1157 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1159 30. JIT compiler update.
1161 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1162 Rill for the patch.
1164 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1166 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1168 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1169 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1170 was enabled.
1172 35. Improve JIT code generation for greedy plus quantifier.
1174 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1175 1 to "aa" instead of to an empty string. The bug affected repeated groups
1176 that could potentially match an empty string.
1178 37. Optimizing single character iterators in JIT.
1180 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1181 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1182 codepoints that are too big for the mode are faulted, and in a UTF mode,
1183 disallowed codepoints are also faulted.
1185 39. If PCRE was compiled with UTF support, in three places in the DFA
1186 matcher there was code that should only have been obeyed in UTF mode, but
1187 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1188 processing when bytes with values greater than 127 were present. In 16-bit
1189 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1190 both cases the values are those that cannot be the first data item in a UTF
1191 character. The three items that might have provoked this were recursions,
1192 possessively repeated groups, and atomic groups.
1194 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1195 and pcregrep, because some OS require shared objects to be explicitly
1196 passed to ld, causing the link step to fail if they are not.
1198 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1199 mode, patterns that started with \h* or \R* might be incorrectly matched.
1202 Version 8.30 04-February-2012
1203 -----------------------------
1205 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1206 name is defined in ctype.h.
1208 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1209 only in quite long subpatterns.
1211 3. Removed the function pcre_info(), which has been obsolete and deprecated
1212 since it was replaced by pcre_fullinfo() in February 2000.
1214 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1215 match a (*MARK), and the match failed at the start of the subject, a
1216 reference to memory before the start of the subject could occur. This bug
1217 was introduced by fix 17 of release 8.21.
1219 5. A reference to an unset group with zero minimum repetition was giving
1220 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1221 /(another)?(\1?)test/ matched against "hello world test". This bug was
1222 introduced in release 8.13.
1224 6. Add support for 16-bit character strings (a large amount of work involving
1225 many changes and refactorings).
1227 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1228 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1229 from a file.
1231 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1232 rounding is not applied in this particular case).
1234 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1235 if they appear, or are escaped, in patterns.
1237 10. Get rid of a number of -Wunused-but-set-variable warnings.
1239 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1240 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1241 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1242 also returns the mark "x". This bug applied to capturing parentheses,
1243 non-capturing parentheses, and atomic parentheses. It also applied to some
1244 assertions.
1246 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1247 information out of configure.ac instead of relying on pcre.h.generic, which
1248 is not stored in the repository.
1250 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1251 -lreadline.
1253 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1255 15. Applied Graycode's patch to put the top-level frame on the stack rather
1256 than the heap when not using the stack for recursion. This gives a
1257 performance improvement in many cases when recursion is not deep.
1259 16. Experimental code added to "pcretest -C" to output the stack frame size.
1262 Version 8.21 12-Dec-2011
1263 ------------------------
1265 1. Updating the JIT compiler.
1267 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1268 are added as well.
1270 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1271 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1272 calling _pcre_jit_exec. Some extra comments are added.
1274 4. (*MARK) settings inside atomic groups that do not contain any capturing
1275 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1276 was introduced by change 18 for 8.20.
1278 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1279 ECMA-262 standard.
1281 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1282 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1283 This bug was probably introduced by change 9 of 8.13.
1285 7. While fixing 6 above, I noticed that a number of other items were being
1286 incorrectly rejected as "not fixed length". This arose partly because newer
1287 opcodes had not been added to the fixed-length checking code. I have (a)
1288 corrected the bug and added tests for these items, and (b) arranged for an
1289 error to occur if an unknown opcode is encountered while checking for fixed
1290 length instead of just assuming "not fixed length". The items that were
1291 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1292 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1293 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1295 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1296 being incorrectly compiled and would have given unpredicatble results.
1298 9. A possessively repeated subpattern with minimum repeat count greater than
1299 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1300 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1301 the first (A) could occur when it should not.
1303 10. Add a cast and remove a redundant test from the code.
1305 11. JIT should use pcre_malloc/pcre_free for allocation.
1307 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1308 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1309 is anything other than /usr, -L is still shown).
1311 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1313 14. Perl does not support \N without a following name in a [] class; PCRE now
1314 also gives an error.
1316 15. If a forward reference was repeated with an upper limit of around 2000,
1317 it caused the error "internal error: overran compiling workspace". The
1318 maximum number of forward references (including repeats) was limited by the
1319 internal workspace, and dependent on the LINK_SIZE. The code has been
1320 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1321 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1322 of around 200,000 forward references. While doing this, I also speeded up
1323 the filling in of repeated forward references.
1325 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1326 incorrectly expecting the subject to contain another "a" after the start.
1328 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1329 in the match, the SKIP should be ignored. This was not happening; instead
1330 the SKIP was being treated as NOMATCH. For patterns such as
1331 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1332 tested.
1334 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1335 now much more compatible with Perl, in particular in cases where the result
1336 is a non-match for a non-anchored pattern. For example, if
1337 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1338 "m", where previously it did not return a name. A side effect of this
1339 change is that for partial matches, the last encountered mark name is
1340 returned, as for non matches. A number of tests that were previously not
1341 Perl-compatible have been moved into the Perl-compatible test files. The
1342 refactoring has had the pleasing side effect of removing one argument from
1343 the match() function, thus reducing its stack requirements.
1345 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1346 subsequent uses of /S (without +) incorrectly behaved like /S+.
1348 21. Retrieve executable code size support for the JIT compiler and fixing
1349 some warnings.
1351 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1352 not work when the shorter character appeared right at the end of the
1353 subject string.
1355 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1356 systems.
1358 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1359 output it when the /M option is used in pcretest.
1361 25. The CheckMan script was not being included in the distribution. Also, added
1362 an explicit "perl" to run Perl scripts from the PrepareRelease script
1363 because this is reportedly needed in Windows.
1365 26. If study data was being save in a file and studying had not found a set of
1366 "starts with" bytes for the pattern, the data written to the file (though
1367 never used) was taken from uninitialized memory and so caused valgrind to
1368 complain.
1370 27. Updated RunTest.bat as provided by Sheri Pierce.
1372 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1374 29. Computation of memory usage for the table of capturing group names was
1375 giving an unnecessarily large value.
1378 Version 8.20 21-Oct-2011
1379 ------------------------
1381 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1382 a POSIX class. After further experiments with Perl, which convinced me that
1383 Perl has bugs and confusions, a closing square bracket is no longer allowed
1384 in a POSIX name. This bug also affected patterns with classes that started
1385 with full stops.
1387 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1388 captured substring, but while checking the failing first alternative,
1389 substring 1 is temporarily captured. If the output vector supplied to
1390 pcre_exec() was not big enough for this capture, the yield of the function
1391 was still zero ("insufficient space for captured substrings"). This cannot
1392 be totally fixed without adding another stack variable, which seems a lot
1393 of expense for a edge case. However, I have improved the situation in cases
1394 such as /(a)(b)x|abc/ matched against "abc", where the return code
1395 indicates that fewer than the maximum number of slots in the ovector have
1396 been set.
1398 3. Related to (2) above: when there are more back references in a pattern than
1399 slots in the output vector, pcre_exec() uses temporary memory during
1400 matching, and copies in the captures as far as possible afterwards. It was
1401 using the entire output vector, but this conflicts with the specification
1402 that only 2/3 is used for passing back captured substrings. Now it uses
1403 only the first 2/3, for compatibility. This is, of course, another edge
1404 case.
1406 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1407 main code base, and can be used by building with --enable-jit. When this is
1408 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1409 runtime --no-jit option is given.
1411 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1412 ovector, the return from the function was zero, implying that there were
1413 other matches that did not fit. The correct "exactly full" value is now
1414 returned.
1416 6. If a subpattern that was called recursively or as a subroutine contained
1417 (*PRUNE) or any other control that caused it to give a non-standard return,
1418 invalid errors such as "Error -26 (nested recursion at the same subject
1419 position)" or even infinite loops could occur.
1421 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1422 computing the minimum length on reaching *ACCEPT, and so ended up with the
1423 wrong value of 1 rather than 0. Further investigation indicates that
1424 computing a minimum subject length in the presence of *ACCEPT is difficult
1425 (think back references, subroutine calls), and so I have changed the code
1426 so that no minimum is registered for a pattern that contains *ACCEPT.
1428 8. If (*THEN) was present in the first (true) branch of a conditional group,
1429 it was not handled as intended. [But see 16 below.]
1431 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1432 Sheri Pierce.
1434 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1435 the first byte in a match must be "a".
1437 11. Change 17 for 8.13 increased the recursion depth for patterns like
1438 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1439 pattern contains any instances of (*THEN). If it does not, the old
1440 optimizations are restored. It would be nice to do this on a per-group
1441 basis, but at the moment that is not feasible.
1443 12. In some environments, the output of pcretest -C is CRLF terminated. This
1444 broke RunTest's code that checks for the link size. A single white space
1445 character after the value is now allowed for.
1447 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1448 For "fr", it uses the Windows-specific input and output files.
1450 14. If (*THEN) appeared in a group that was called recursively or as a
1451 subroutine, it did not work as intended. [But see next item.]
1453 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1454 pattern fragments (but not containing any | characters). If A and B are
1455 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1456 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1457 D. In other words, Perl considers parentheses that do not contain any |
1458 characters to be part of a surrounding alternative, whereas PCRE was
1459 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1460 differently. PCRE now behaves in the same way as Perl, except in the case
1461 of subroutine/recursion calls such as (?1) which have in any case always
1462 been different (but PCRE had them first :-).
1464 16. Related to 15 above: Perl does not treat the | in a conditional group as
1465 creating alternatives. Such a group is treated in the same way as an
1466 ordinary group without any | characters when processing (*THEN). PCRE has
1467 been changed to match Perl's behaviour.
1469 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1470 RunGrepTest script failed.
1472 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1473 inevitable for groups that contain captures, but it can lead to a lot of
1474 stack use in large patterns. The old behaviour has been restored for atomic
1475 groups that do not contain any capturing parentheses.
1477 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1478 suppress the check for a minimum subject length at run time. (If it was
1479 given to pcre_exec() or pcre_dfa_exec() it did work.)
1481 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1482 fail to work when decoding hex characters in data strings in EBCDIC
1483 environments.
1485 21. It appears that in at least one Mac OS environment, the isxdigit() function
1486 is implemented as a macro that evaluates to its argument more than once,
1487 contravening the C 90 Standard (I haven't checked a later standard). There
1488 was an instance in pcretest which caused it to go wrong when processing
1489 \x{...} escapes in subject strings. The has been rewritten to avoid using
1490 things like p++ in the argument of isxdigit().
1493 Version 8.13 16-Aug-2011
1494 ------------------------
1496 1. The Unicode data tables have been updated to Unicode 6.0.0.
1498 2. Two minor typos in pcre_internal.h have been fixed.
1500 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1501 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1502 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1504 4. There were a number of related bugs in the code for matching backrefences
1505 caselessly in UTF-8 mode when codes for the characters concerned were
1506 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1507 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1508 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1509 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1510 2-byte code at the end of the subject (it thought there wasn't enough data
1511 left).
1513 5. Comprehensive information about what went wrong is now returned by
1514 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1515 as the output vector has at least 2 elements. The offset of the start of
1516 the failing character and a reason code are placed in the vector.
1518 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1519 now returned is for the first byte of the failing character, instead of the
1520 last byte inspected. This is an incompatible change, but I hope it is small
1521 enough not to be a problem. It makes the returned offset consistent with
1522 pcre_exec() and pcre_dfa_exec().
1524 7. pcretest now gives a text phrase as well as the error number when
1525 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1526 failure, the offset and reason code are output.
1528 8. When \R was used with a maximizing quantifier it failed to skip backwards
1529 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1530 back over a single character (\n). This seems wrong (because it treated the
1531 two characters as a single entity when going forwards), conflicts with the
1532 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1533 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1534 has been changed.
1536 9. Some internal refactoring has changed the processing so that the handling
1537 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1538 time (the PCRE_DOTALL option was changed this way some time ago: version
1539 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1540 which was always a bit of a fudge. It also means that there is one less
1541 argument for the match() function, which reduces its stack requirements
1542 slightly. This change also fixes an incompatibility with Perl: the pattern
1543 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1545 10. More internal refactoring has drastically reduced the number of recursive
1546 calls to match() for possessively repeated groups such as (abc)++ when
1547 using pcre_exec().
1549 11. While implementing 10, a number of bugs in the handling of groups were
1550 discovered and fixed:
1552 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1553 (a|)*(?1) gave a compile-time internal error.
1554 ((a|)+)+ did not notice that the outer group could match an empty string.
1555 (^a|^)+ was not marked as anchored.
1556 (.*a|.*)+ was not marked as matching at start or after a newline.
1558 12. Yet more internal refactoring has removed another argument from the match()
1559 function. Special calls to this function are now indicated by setting a
1560 value in a variable in the "match data" data block.
1562 13. Be more explicit in pcre_study() instead of relying on "default" for
1563 opcodes that mean there is no starting character; this means that when new
1564 ones are added and accidentally left out of pcre_study(), testing should
1565 pick them up.
1567 14. The -s option of pcretest has been documented for ages as being an old
1568 synonym of -m (show memory usage). I have changed it to mean "force study
1569 for every regex", that is, assume /S for every regex. This is similar to -i
1570 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1571 using it. It makes it easier to run collections of tests with and without
1572 study enabled, and thereby test pcre_study() more easily. All the standard
1573 tests are now run with and without -s (but some patterns can be marked as
1574 "never study" - see 20 below).
1576 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1577 restoration of the capturing data to the outer values was not happening
1578 correctly.
1580 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1581 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1582 pattern had matched an empty string, and so incorrectly returned a no
1583 match.
1585 17. There was optimizing code for the last branch of non-capturing parentheses,
1586 and also for the obeyed branch of a conditional subexpression, which used
1587 tail recursion to cut down on stack usage. Unfortunately, now that there is
1588 the possibility of (*THEN) occurring in these branches, tail recursion is
1589 no longer possible because the return has to be checked for (*THEN). These
1590 two optimizations have therefore been removed. [But see 8.20/11 above.]
1592 18. If a pattern containing \R was studied, it was assumed that \R always
1593 matched two bytes, thus causing the minimum subject length to be
1594 incorrectly computed because \R can also match just one byte.
1596 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1597 was incorrectly computed.
1599 20. If /S is present twice on a test pattern in pcretest input, it now
1600 *disables* studying, thereby overriding the use of -s on the command line
1601 (see 14 above). This is necessary for one or two tests to keep the output
1602 identical in both cases.
1604 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1605 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1607 22. When an atomic group that contained a capturing parenthesis was
1608 successfully matched, but the branch in which it appeared failed, the
1609 capturing was not being forgotten if a higher numbered group was later
1610 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1611 group 1 to "a", when in fact it should be unset. This applied to multi-
1612 branched capturing and non-capturing groups, repeated or not, and also to
1613 positive assertions (capturing in negative assertions does not happen
1614 in PCRE) and also to nested atomic groups.
1616 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1617 subject after a captured substring, to make it easier to tell which of a
1618 number of identical substrings has been captured.
1620 24. The way atomic groups are processed by pcre_exec() has been changed so that
1621 if they are repeated, backtracking one repetition now resets captured
1622 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1623 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1624 "aaa". Previously, it would have been "a". As part of this code
1625 refactoring, the way recursive calls are handled has also been changed.
1627 25. If an assertion condition captured any substrings, they were not passed
1628 back unless some other capturing happened later. For example, if
1629 (?(?=(a))a) was matched against "a", no capturing was returned.
1631 26. When studying a pattern that contained subroutine calls or assertions,
1632 the code for finding the minimum length of a possible match was handling
1633 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1634 group 1 called group 2 while simultaneously a separate group 2 called group
1635 1). A stack overflow occurred in this case. I have fixed this by limiting
1636 the recursion depth to 10.
1638 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1639 Fortmann. This supports explicit test numbers on the command line, and has
1640 argument validation and error reporting.
1642 28. An instance of \X with an unlimited repeat could fail if at any point the
1643 first character it looked at was a mark character.
1645 29. Some minor code refactoring concerning Unicode properties and scripts
1646 should reduce the stack requirement of match() slightly.
1648 30. Added the '=' option to pcretest to check the setting of unused capturing
1649 slots at the end of the pattern, which are documented as being -1, but are
1650 not included in the return count.
1652 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1653 compiled something random. Now it gives a compile-time error (as does
1654 Perl).
1656 32. A *MARK encountered during the processing of a positive assertion is now
1657 recorded and passed back (compatible with Perl).
1659 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1660 had alternative anchored branches, the search for a second match in a line
1661 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1662 matched the line "0102" twice. The same bug affected patterns that started
1663 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1664 twice.
1666 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1667 does, and because of capturing effects, quantifying parenthesized
1668 assertions may at times be useful. Quantifiers are now allowed for
1669 parenthesized assertions.
1671 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1673 36. \g was being checked for fancy things in a character class, when it should
1674 just be a literal "g".
1676 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1677 appearance of a nested POSIX class supersedes an apparent external class.
1678 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1679 unescaped square brackets may also appear as part of class names. For
1680 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1681 more like Perl. (But see 8.20/1 above.)
1683 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1684 was because it thought it was \N{name}, which is not supported).
1686 39. Add minix to OS list not supporting the -S option in pcretest.
1688 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1689 cannot analyze patterns in sufficient detail to catch mutual recursions
1690 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1691 subgroup is called recursively as a subpattern for a second time at the
1692 same position in the subject string. In previous releases this might have
1693 been caught by the recursion limit, or it might have run out of stack.
1695 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1696 happen only once. PCRE was, however incorrectly giving a compile time error
1697 "recursive call could loop indefinitely" because it cannot analyze the
1698 pattern in sufficient detail. The compile time test no longer happens when
1699 PCRE is compiling a conditional subpattern, but actual runaway loops are
1700 now caught at runtime (see 40 above).
1702 42. It seems that Perl allows any characters other than a closing parenthesis
1703 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1704 has been changed to be the same.
1706 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1707 as not to get warnings when autogen.sh is called. Also changed
1708 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1710 44. To help people who use pcregrep to scan files containing exceedingly long
1711 lines, the following changes have been made:
1713 (a) The default value of the buffer size parameter has been increased from
1714 8K to 20K. (The actual buffer used is three times this size.)
1716 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1717 PCRE is built.
1719 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1720 to be set at run time.
1722 (d) Numerical values in pcregrep options can be followed by K or M, for
1723 example --buffer-size=50K.
1725 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1726 given and the return code is set to 2.
1728 45. Add a pointer to the latest mark to the callout data block.
1730 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1731 partial match of an empty string instead of no match. This was specific to
1732 the use of ".".
1734 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1735 complete match instead of a partial match. This bug was dependent on both
1736 the PCRE_UTF8 and PCRE_DOTALL options being set.
1738 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1739 starting byte set, because \b was not being ignored.
1742 Version 8.12 15-Jan-2011
1743 ------------------------
1745 1. Fixed some typos in the markup of the man pages, and wrote a script that
1746 checks for such things as part of the documentation building process.
1748 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1749 --match-limit and --recursion-limit options (added for 8.11). In
1750 particular, this made one of the standard tests fail. (The integer value
1751 went into the wrong half of a long int.)
1753 3. If the --colour option was given to pcregrep with -v (invert match), it
1754 did strange things, either producing crazy output, or crashing. It should,
1755 of course, ignore a request for colour when reporting lines that do not
1756 match.
1758 4. Another pcregrep bug caused similar problems if --colour was specified with
1759 -M (multiline) and the pattern match finished with a line ending.
1761 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1762 matched in multiline mode, the following line was shown as part of the
1763 match. This seems wrong, so I have changed it.
1765 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1766 the check for further matches in the same line (so they could be coloured)
1767 to overrun the end of the current line. If another match was found, it was
1768 incorrectly shown (and then shown again when found in the next line).
1770 7. If pcregrep was compiled under Windows, there was a reference to the
1771 function pcregrep_exit() before it was defined. I am assuming this was
1772 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1773 reported by a user. I've moved the definition above the reference.
1776 Version 8.11 10-Dec-2010
1777 ------------------------
1779 1. (*THEN) was not working properly if there were untried alternatives prior
1780 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1781 backtracked to try for "b" instead of moving to the next alternative branch
1782 at the same level (in this case, to look for "c"). The Perl documentation
1783 is clear that when (*THEN) is backtracked onto, it goes to the "next
1784 alternative in the innermost enclosing group".
1786 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1787 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1788 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1789 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1790 (*THEN).
1792 3. If \s appeared in a character class, it removed the VT character from
1793 the class, even if it had been included by some previous item, for example
1794 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1795 of \s, but is part of the POSIX "space" class.)
1797 4. A partial match never returns an empty string (because you can always
1798 match an empty string at the end of the subject); however the checking for
1799 an empty string was starting at the "start of match" point. This has been
1800 changed to the "earliest inspected character" point, because the returned
1801 data for a partial match starts at this character. This means that, for
1802 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1803 (previously it gave "no match").
1805 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1806 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1807 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1808 has an implication that the given string is incomplete (because a partial
1809 match is preferred over a full match). For this reason, these items now
1810 give a partial match in this situation. [Aside: previously, the one case
1811 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1812 match rather than a full match, which was wrong by the old rules, but is
1813 now correct.]
1815 6. There was a bug in the handling of #-introduced comments, recognized when
1816 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1817 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1818 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1819 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1820 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1821 places in pcre_compile().
1823 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1824 comments when looking ahead for named forward references to subpatterns,
1825 the only newline sequence it recognized was NL. It now handles newlines
1826 according to the set newline convention.
1828 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1829 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1830 cater for a lack of strerror(). These oversights have been fixed.
1832 9. Added --match-limit and --recursion-limit to pcregrep.
1834 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1836 11. When the -o option was used, pcregrep was setting a return code of 1, even
1837 when matches were found, and --line-buffered was not being honoured.
1839 12. Added an optional parentheses number to the -o and --only-matching options
1840 of pcregrep.
1842 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1843 can match an empty string. The code to do it in pcretest and pcredemo
1844 needed fixing:
1846 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1847 only one byte after an empty string match just before CRLF (this case
1848 just got forgotten; "any" and "anycrlf" were OK).
1850 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1851 mode when an empty string match preceded an ASCII character followed by
1852 a non-ASCII character. (The code for advancing by one character rather
1853 than one byte was nonsense.)
1855 (c) The pcredemo.c sample program did not have any code at all to handle
1856 the cases when CRLF is a valid newline sequence.
1858 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1859 as a starting offset was within the subject string. There is now a new
1860 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1861 negative or greater than the length of the string. In order to test this,
1862 pcretest is extended to allow the setting of negative starting offsets.
1864 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1865 starting offset points to the beginning of a UTF-8 character was
1866 unnecessarily clumsy. I tidied it up.
1868 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1869 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1871 17. Nobody had reported that the --include_dir option, which was added in
1872 release 7.7 should have been called --include-dir (hyphen, not underscore)
1873 for compatibility with GNU grep. I have changed it to --include-dir, but
1874 left --include_dir as an undocumented synonym, and the same for
1875 --exclude-dir, though that is not available in GNU grep, at least as of
1876 release 2.5.4.
1878 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1879 characters from a string of bytes) have been redefined so as not to use
1880 loops, in order to improve performance in some environments. At the same
1881 time, I abstracted some of the common code into auxiliary macros to save
1882 repetition (this should not affect the compiled code).
1884 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1885 compile-time error is now given if \c is not followed by an ASCII
1886 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1887 different, and any byte value is allowed.)
1889 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1890 START_OPTIMIZE option, which is now allowed at compile time - but just
1891 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1892 to pcregrep and other applications that have no direct access to PCRE
1893 options. The new /Y option in pcretest sets this option when calling
1894 pcre_compile().
1896 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1897 back references. Groups containing recursive back references were forced to
1898 be atomic by that change, but in the case of named groups, the amount of
1899 memory required was incorrectly computed, leading to "Failed: internal
1900 error: code overflow". This has been fixed.
1902 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1903 pcretest.c, to avoid build problems in some Borland environments.
1906 Version 8.10 25-Jun-2010
1907 ------------------------
1909 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1910 THEN.
1912 2. (*ACCEPT) was not working when inside an atomic group.
1914 3. Inside a character class, \B is treated as a literal by default, but
1915 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1916 causes the error). The code is unchanged, but I tidied the documentation.
1918 4. Inside a character class, PCRE always treated \R and \X as literals,
1919 whereas Perl faults them if its -w option is set. I have changed PCRE so
1920 that it faults them when PCRE_EXTRA is set.
1922 5. Added support for \N, which always matches any character other than
1923 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1925 6. When compiling pcregrep with newer versions of gcc which may have
1926 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1927 declared with attribute warn_unused_result" were given. Just casting the
1928 result to (void) does not stop the warnings; a more elaborate fudge is
1929 needed. I've used a macro to implement this.
1931 7. Minor change to pcretest.c to avoid a compiler warning.
1933 8. Added four artifical Unicode properties to help with an option to make
1934 \s etc use properties (see next item). The new properties are: Xan
1935 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1937 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1938 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1939 this option. Modified pcretest to add /W to test this facility. Added
1940 REG_UCP to make it available via the POSIX interface.
1942 10. Added --line-buffered to pcregrep.
1944 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1945 studied, and the match started with a letter with a code point greater than
1946 127 whose first byte was different to the first byte of the other case of
1947 the letter, the other case of this starting letter was not recognized
1948 (#976).
1950 12. If a pattern that was studied started with a repeated Unicode property
1951 test, for example, \p{Nd}+, there was the theoretical possibility of
1952 setting up an incorrect bitmap of starting bytes, but fortunately it could
1953 not have actually happened in practice until change 8 above was made (it
1954 added property types that matched character-matching opcodes).
1956 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1957 possible starting bytes for non-anchored patterns.
1959 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1960 \R, and also a number of cases that involve Unicode properties, both
1961 explicit and implicit when PCRE_UCP is set.
1963 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1964 input, it could crash or give wrong results if characters with values
1965 greater than 0xc0 were present in the subject string. (Detail: it assumed
1966 UTF-8 input when processing these items.)
1968 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1969 size_t is 64-bit (#991).
1971 17. Added a check for running out of memory when PCRE is compiled with
1972 --disable-stack-for-recursion (#990).
1974 18. If the last data line in a file for pcretest does not have a newline on
1975 the end, a newline was missing in the output.
1977 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1978 less than 128) in its various bitmaps. However, there is a facility for
1979 generating tables according to the current locale when PCRE is compiled. It
1980 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1981 characters, are recognized by isspace() and therefore were getting set in
1982 these tables, and indeed these tables seem to approximate to ISO 8859. This
1983 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1984 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1985 which of course cannot start UTF-8 characters. I have changed the code so
1986 that only real ASCII characters (less than 128) and the correct starting
1987 bytes for UTF-8 encodings are set for characters greater than 127 when in
1988 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1989 altogether.)
1991 20. Added the /T option to pcretest so as to be able to run tests with non-
1992 standard character tables, thus making it possible to include the tests
1993 used for 19 above in the standard set of tests.
1995 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1996 reference to a subpattern the other side of a comment that contains an
1997 opening parenthesis caused either an internal compiling error, or a
1998 reference to the wrong subpattern.
2001 Version 8.02 19-Mar-2010
2002 ------------------------
2004 1. The Unicode data tables have been updated to Unicode 5.2.0.
2006 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2007 configured.
2009 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2010 original author of that file, following a query about its status.
2012 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2013 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2015 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2016 quantifier applied to a forward-referencing subroutine call, could compile
2017 incorrect code or give the error "internal error: previously-checked
2018 referenced subpattern not found".
2020 6. Both MS Visual Studio and Symbian OS have problems with initializing
2021 variables to point to external functions. For these systems, therefore,
2022 pcre_malloc etc. are now initialized to local functions that call the
2023 relevant global functions.
2025 7. There were two entries missing in the vectors called coptable and poptable
2026 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2027 I've fixed the data, and added a kludgy way of testing at compile time that
2028 the lengths are correct (equal to the number of opcodes).
2030 8. Following on from 7, I added a similar kludge to check the length of the
2031 eint vector in pcreposix.c.
2033 9. Error texts for pcre_compile() are held as one long string to avoid too
2034 much relocation at load time. To find a text, the string is searched,
2035 counting zeros. There was no check for running off the end of the string,
2036 which could happen if a new error number was added without updating the
2037 string.
2039 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2041 11. \K was not working if it appeared in an atomic group or in a group that
2042 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2043 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2044 the assertion is positive, but not if it is negative.
2046 12. Change 11 fortuitously reduced the size of the stack frame used in the
2047 "match()" function of pcre_exec.c by one pointer. Forthcoming
2048 implementation of support for (*MARK) will need an extra pointer on the
2049 stack; I have reserved it now, so that the stack frame size does not
2050 decrease.
2052 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2053 item in branch that calls a recursion is a subroutine call - as in the
2054 second branch in the above example - was incorrectly given the compile-
2055 time error "recursive call could loop indefinitely" because pcre_compile()
2056 was not correctly checking the subroutine for matching a non-empty string.
2058 14. The checks for overrunning compiling workspace could trigger after an
2059 overrun had occurred. This is a "should never occur" error, but it can be
2060 triggered by pathological patterns such as hundreds of nested parentheses.
2061 The checks now trigger 100 bytes before the end of the workspace.
2063 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2066 Version 8.01 19-Jan-2010
2067 ------------------------
2069 1. If a pattern contained a conditional subpattern with only one branch (in
2070 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2071 computed the wrong minimum data length (which is of course zero for such
2072 subpatterns). This could cause incorrect "no match" results.
2074 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2075 the pattern is reset in the first branch, pcre_compile() failed with
2076 "internal error: code overflow at offset...". This happened only when
2077 the reset was to the original external option setting. (An optimization
2078 abstracts leading options settings into an external setting, which was the
2079 cause of this.)
2081 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2082 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2083 assertion pattern did not match (meaning that the assertion was true), it
2084 was incorrectly treated as false if the SKIP had been reached during the
2085 matching. This also applied to assertions used as conditions.
2087 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2088 assertion subpattern, including such a pattern used as a condition,
2089 unpredictable results occurred, instead of the error return
2092 5. The C++ GlobalReplace function was not working like Perl for the special
2093 situation when an empty string is matched. It now does the fancy magic
2094 stuff that is necessary.
2096 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2097 removed. (These were left over from very, very early versions of PCRE.)
2099 7. Some cosmetic changes to the code to make life easier when compiling it
2100 as part of something else:
2102 (a) Change DEBUG to PCRE_DEBUG.
2104 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2105 called "current" as "current_branch", to prevent a collision with the
2106 Linux macro when compiled as a kernel module.
2108 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2109 prevent a collision with the Linux macro when compiled as a kernel
2110 module.
2112 8. In pcre_compile() there are some checks for integer overflows that used to
2113 cast potentially large values to (double). This has been changed to that
2114 when building, a check for int64_t is made, and if it is found, it is used
2115 instead, thus avoiding the use of floating point arithmetic. (There is no
2116 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2117 double.
2119 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2120 2005 (difference between two addresses compared to an unsigned value).
2122 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2123 custom one, because of the following reported problem in Windows:
2125 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2126 under Win32.
2127 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2128 therefore missing the function definition.
2129 - The compiler thus generates a "C" signature for the test function.
2130 - The linker fails to find the "C" function.
2131 - PCRE fails to configure if asked to do so against libbz2.
2133 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2134 messages were output:
2136 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2137 rerunning libtoolize, to keep the correct libtool macros in-tree.
2138 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2140 I have done both of these things.
2142 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2143 most of the time, it *can* run out if it is given a pattern that contains a
2144 runaway infinite recursion. I updated the discussion in the pcrestack man
2145 page.
2147 13. Now that we have gone to the x.xx style of version numbers, the minor
2148 version may start with zero. Using 08 or 09 is a bad idea because users
2149 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2150 interpreted as invalid octal numbers. I've updated the previous comment in
2151 configure.ac, and also added a check that gives an error if 08 or 09 are
2152 used.
2154 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2155 causing partial matching to fail when the end of the subject matched \W
2156 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2158 15. There were some discrepancies between the declarations in pcre_internal.h
2159 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2160 their definitions. The declarations used "const uschar *" and the
2161 definitions used USPTR. Even though USPTR is normally defined as "const
2162 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2163 reported that: "This difference in casting confuses some C++ compilers, for
2164 example, SunCC recognizes above declarations as different functions and
2165 generates broken code for hbpcre." I have changed the declarations to use
2166 USPTR.
2168 16. GNU libtool is named differently on some systems. The autogen.sh script now
2169 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2170 (FreeBSD).
2172 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2173 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2174 comment: "Figure out how to create a longlong from a string: strtoll and
2175 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2176 instance, but it only takes 2 args instead of 3!"
2178 18. A subtle bug concerned with back references has been fixed by a change of
2179 specification, with a corresponding code fix. A pattern such as
2180 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2181 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2182 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2183 same bug. Such groups have to be quantified to be useful, or contained
2184 inside another quantified group. (If there's no repetition, the reference
2185 can never match.) The problem arises because, having left the group and
2186 moved on to the rest of the pattern, a later failure that backtracks into
2187 the group uses the captured value from the final iteration of the group
2188 rather than the correct earlier one. I have fixed this in PCRE by forcing
2189 any group that contains a reference to itself to be an atomic group; that
2190 is, there cannot be any backtracking into it once it has completed. This is
2191 similar to recursive and subroutine calls.
2194 Version 8.00 19-Oct-09
2195 ----------------------
2197 1. The table for translating pcre_compile() error codes into POSIX error codes
2198 was out-of-date, and there was no check on the pcre_compile() error code
2199 being within the table. This could lead to an OK return being given in
2200 error.
2202 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2203 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2204 in a Windows environment.
2206 3. The pcregrep --count option prints the count for each file even when it is
2207 zero, as does GNU grep. However, pcregrep was also printing all files when
2208 --files-with-matches was added. Now, when both options are given, it prints
2209 counts only for those files that have at least one match. (GNU grep just
2210 prints the file name in this circumstance, but including the count seems
2211 more useful - otherwise, why use --count?) Also ensured that the
2212 combination -clh just lists non-zero counts, with no names.
2214 4. The long form of the pcregrep -F option was incorrectly implemented as
2215 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2216 but it seems right to fix it, and I didn't think it was worth preserving
2217 the old behaviour.
2219 5. The command line items --regex=pattern and --regexp=pattern were not
2220 recognized by pcregrep, which required --regex pattern or --regexp pattern
2221 (with a space rather than an '='). The man page documented the '=' forms,
2222 which are compatible with GNU grep; these now work.
2224 6. No libpcreposix.pc file was created for pkg-config; there was just
2225 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2227 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2228 when UCP support is not needed, by modifying the Python script that
2229 generates it from Unicode data files. This should not matter if the module
2230 is correctly used as a library, but I received one complaint about 50K of
2231 unwanted data. My guess is that the person linked everything into his
2232 program rather than using a library. Anyway, it does no harm.
2234 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2235 was a minimum greater than 1 for a wide character in a possessive
2236 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2237 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2238 character. Chaos in the form of incorrect output or a compiling loop could
2239 result.
2241 9. The restrictions on what a pattern can contain when partial matching is
2242 requested for pcre_exec() have been removed. All patterns can now be
2243 partially matched by this function. In addition, if there are at least two
2244 slots in the offset vector, the offset of the earliest inspected character
2245 for the match and the offset of the end of the subject are set in them when
2246 PCRE_ERROR_PARTIAL is returned.
2248 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2249 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2250 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2251 and may be more useful for multi-segment matching.
2253 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2254 used to be given if ever the end of the subject was reached; now it is
2255 given only if matching could not proceed because another character was
2256 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2257 string "Z", which now yields "no match" instead of "partial match". In the
2258 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2259 final character ended with (*FAIL).
2261 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2262 if the pattern had a "must contain" character that was already found in the
2263 earlier partial match, unless partial matching was again requested. For
2264 example, with the pattern /dog.(body)?/, the "must contain" character is
2265 "g". If the first part-match was for the string "dog", restarting with
2266 "sbody" failed. This bug has been fixed.
2268 13. The string returned by pcre_dfa_exec() after a partial match has been
2269 changed so that it starts at the first inspected character rather than the
2270 first character of the match. This makes a difference only if the pattern
2271 starts with a lookbehind assertion or \b or \B (\K is not supported by
2272 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2273 matching functions compatible, and I think it's the right thing to do.
2275 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2276 so that the demonstration program is easily available in environments where
2277 PCRE has not been installed from source.
2279 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2280 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2281 library.
2283 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2284 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2285 is not the first non-POSIX option to be added. Clearly some people find
2286 these options useful.
2288 17. If a caller to the POSIX matching function regexec() passes a non-zero
2289 value for nmatch with a NULL value for pmatch, the value of
2290 nmatch is forced to zero.
2292 18. RunGrepTest did not have a test for the availability of the -u option of
2293 the diff command, as RunTest does. It now checks in the same way as
2294 RunTest, and also checks for the -b option.
2296 19. If an odd number of negated classes containing just a single character
2297 interposed, within parentheses, between a forward reference to a named
2298 subpattern and the definition of the subpattern, compilation crashed with
2299 an internal error, complaining that it could not find the referenced
2300 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2301 [The bug was that it was starting one character too far in when skipping
2302 over the character class, thus treating the ] as data rather than
2303 terminating the class. This meant it could skip too much.]
2305 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2306 /g option in pcretest when the pattern contains \K, which makes it possible
2307 to have an empty string match not at the start, even when the pattern is
2308 anchored. Updated pcretest and pcredemo to use this option.
2310 21. If the maximum number of capturing subpatterns in a recursion was greater
2311 than the maximum at the outer level, the higher number was returned, but
2312 with unset values at the outer level. The correct (outer level) value is
2313 now given.
2315 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2316 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2317 make it do so. The string so far is captured, making this feature
2318 compatible with Perl.
2320 23. The tests have been re-organized, adding tests 11 and 12, to make it
2321 possible to check the Perl 5.10 features against Perl 5.10.
2323 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2324 pattern matches a fixed length string. PCRE did not allow this; now it
2325 does. Neither allows recursion.
2327 25. I finally figured out how to implement a request to provide the minimum
2328 length of subject string that was needed in order to match a given pattern.
2329 (It was back references and recursion that I had previously got hung up
2330 on.) This code has now been added to pcre_study(); it finds a lower bound
2331 to the length of subject needed. It is not necessarily the greatest lower
2332 bound, but using it to avoid searching strings that are too short does give
2333 some useful speed-ups. The value is available to calling programs via
2334 pcre_fullinfo().
2336 26. While implementing 25, I discovered to my embarrassment that pcretest had
2337 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2338 study optimizations had never been tested with that matching function.
2339 Oops. What is worse, even when it was passed study data, there was a bug in
2340 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2341 were also very few tests of studied patterns with pcre_dfa_exec().
2343 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2344 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2345 on the other side of the coin, they are no longer allowed to have different
2346 names, because these cannot be distinguished in PCRE, and this has caused
2347 confusion. (This is a difference from Perl.)
2349 28. When duplicate subpattern names are present (necessarily with different
2350 numbers, as required by 27 above), and a test is made by name in a
2351 conditional pattern, either for a subpattern having been matched, or for
2352 recursion in such a pattern, all the associated numbered subpatterns are
2353 tested, and the overall condition is true if the condition is true for any
2354 one of them. This is the way Perl works, and is also more like the way
2355 testing by number works.
2358 Version 7.9 11-Apr-09
2359 ---------------------
2361 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2362 (pcretest), all targets were linked against these libraries. This included
2363 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2364 libraries. This caused unwanted dependencies to be created. This problem
2365 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2366 pcretest is linked with readline.
2368 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2369 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2370 moved outside it again, because FALSE and TRUE are already defined in AIX,
2371 but BOOL is not.
2373 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2374 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2376 4. The pcregrep documentation said spaces were inserted as well as colons (or
2377 hyphens) following file names and line numbers when outputting matching
2378 lines. This is not true; no spaces are inserted. I have also clarified the
2379 wording for the --colour (or --color) option.
2381 5. In pcregrep, when --colour was used with -o, the list of matching strings
2382 was not coloured; this is different to GNU grep, so I have changed it to be
2383 the same.
2385 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2386 each matching line was coloured. Now it goes on to look for further matches
2387 of any of the test patterns, which is the same behaviour as GNU grep.
2389 7. A pattern that could match an empty string could cause pcregrep to loop; it
2390 doesn't make sense to accept an empty string match in pcregrep, so I have
2391 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2392 seems to be how GNU grep behaves. [But see later change 40 for release
2393 8.33.]
2395 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2396 start or after a newline", because the conditional assertion was not being
2397 correctly handled. The rule now is that both the assertion and what follows
2398 in the first alternative must satisfy the test.
2400 9. If auto-callout was enabled in a pattern with a conditional group whose
2401 condition was an assertion, PCRE could crash during matching, both with
2402 pcre_exec() and pcre_dfa_exec().
2404 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2405 used for matching.
2407 11. Unicode property support in character classes was not working for
2408 characters (bytes) greater than 127 when not in UTF-8 mode.
2410 12. Added the -M command line option to pcretest.
2412 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2414 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2416 16. Added comments and documentation about mis-use of no_arg in the C++
2417 wrapper.
2419 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2420 from Martin Jerabek that uses macro names for all relevant character and
2421 string constants.
2423 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2424 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2425 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2426 these, but not everybody uses configure.
2428 19. A conditional group that had only one branch was not being correctly
2429 recognized as an item that could match an empty string. This meant that an
2430 enclosing group might also not be so recognized, causing infinite looping
2431 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2432 with the subject "ab", where knowledge that the repeated group can match
2433 nothing is needed in order to break the loop.
2435 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2436 exec(), but without supplying a callout function, matching went wrong.
2438 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2439 leak if the size of the offset vector was greater than 30. When the vector
2440 is smaller, the saved offsets during recursion go onto a local stack
2441 vector, but for larger vectors malloc() is used. It was failing to free
2442 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2443 error, in fact).
2445 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2446 heapframe that is used only when UTF-8 support is enabled. This caused no
2447 problem, but was untidy.
2449 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2450 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2451 included within another project.
2453 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2454 slightly modified by me:
2456 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2457 not building pcregrep.
2459 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2460 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2462 25. Forward references, both numeric and by name, in patterns that made use of
2463 duplicate group numbers, could behave incorrectly or give incorrect errors,
2464 because when scanning forward to find the reference group, PCRE was not
2465 taking into account the duplicate group numbers. A pattern such as
2466 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2468 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2469 the feature of a custom pointer more persuasive (as requested by a user).
2471 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2472 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2473 pre-defined.
2475 28. Added support for (*UTF8) at the start of a pattern.
2477 29. Arrange for flags added by the "release type" setting in CMake to be shown
2478 in the configuration summary.
2481 Version 7.8 05-Sep-08
2482 ---------------------
2484 1. Replaced UCP searching code with optimized version as implemented for Ad
2485 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2486 stage table and inline lookup instead of a function, giving speed ups of 2
2487 to 5 times on some simple patterns that I tested. Permission was given to
2488 distribute the MultiStage2.py script that generates the tables (it's not in
2489 the tarball, but is in the Subversion repository).
2491 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2492 scripts.
2494 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2495 a group with a zero qualifier. The result of the study could be incorrect,
2496 or the function might crash, depending on the pattern.
2498 4. Caseless matching was not working for non-ASCII characters in back
2499 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2500 It now works when Unicode Property Support is available.
2502 5. In pcretest, an escape such as \x{de} in the data was always generating
2503 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2504 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2505 truncation.
2507 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2509 7. Added two (int) casts to pcregrep when printing the difference of two
2510 pointers, in case they are 64-bit values.
2512 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2513 test 2 if it fails.
2515 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2516 and a #define of that name to empty if it is not externally set. This is to
2517 allow users of MSVC to set it if necessary.
2519 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2520 the convenience functions in the pcre_get.c source file.
2522 11. An option change at the start of a pattern that had top-level alternatives
2523 could cause overwriting and/or a crash. This command provoked a crash in
2524 some environments:
2526 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2528 This potential security problem was recorded as CVE-2008-2371.
2530 12. For a pattern where the match had to start at the beginning or immediately
2531 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2532 pcre_dfa_exec() could read past the end of the passed subject if there was
2533 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2534 pcretest so that it places the subject at the end of its malloc-ed buffer.
2536 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2537 exec() might read past the end of the data buffer in UTF-8 mode.
2539 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2540 the data contained the byte 0x85 as part of a UTF-8 character within its
2541 first line. This applied both to normal and DFA matching.
2543 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2544 /^[^d]*?$/8 failed to match "abc".
2546 16. Added a missing copyright notice to pcrecpp_internal.h.
2548 17. Make it more clear in the documentation that values returned from
2549 pcre_exec() in ovector are byte offsets, not character counts.
2551 18. Tidied a few places to stop certain compilers from issuing warnings.
2553 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2554 supplied by Stefan Weber. I made a further small update for 7.8 because
2555 there is a change of source arrangements: the pcre_searchfuncs.c module is
2556 replaced by pcre_ucd.c.
2559 Version 7.7 07-May-08
2560 ---------------------
2562 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2563 a string to a long long, pretend we don't even have a long long." This is
2564 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2566 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2567 pre-7.6 versions, which defined a global no_arg variable instead of putting
2568 it in the RE class. (See also #8 below.)
2570 3. Remove a line of dead code, identified by coverity and reported by Nuno
2571 Lopes.
2573 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2575 (1) The include/exclude patterns were being applied to the whole pathnames
2576 of files, instead of just to the final components.
2578 (2) If there was more than one level of directory, the subdirectories were
2579 skipped unless they satisfied the include/exclude conditions. This is
2580 inconsistent with GNU grep (and could even be seen as contrary to the
2581 pcregrep specification - which I improved to make it absolutely clear).
2582 The action now is always to scan all levels of directory, and just
2583 apply the include/exclude patterns to regular files.
2585 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2586 --exclude_dir in the tests to avoid scanning .svn directories.
2588 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2589 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2590 doesn't support NULs in patterns.
2592 7. Added some missing "const"s to declarations of static tables in
2593 pcre_compile.c and pcre_dfa_exec.c.
2595 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2596 caused by fix #2 above. (Subsequently also a second patch to fix the
2597 first patch. And a third patch - this was a messy problem.)
2599 9. Applied Craig's patch to remove the use of push_back().
2601 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2602 matching function regexec().
2604 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2605 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2606 references. PCRE supports relative numbers with this syntax (I don't think
2607 Oniguruma does).
2609 12. Previously, a group with a zero repeat such as (...){0} was completely
2610 omitted from the compiled regex. However, this means that if the group
2611 was called as a subroutine from elsewhere in the pattern, things went wrong
2612 (an internal error was given). Such groups are now left in the compiled
2613 pattern, with a new opcode that causes them to be skipped at execution
2614 time.
2616 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2617 to the way PCRE behaves:
2619 (a) A lone ] character is dis-allowed (Perl treats it as data).
2621 (b) A back reference to an unmatched subpattern matches an empty string
2622 (Perl fails the current match path).
2624 (c) A data ] in a character class must be notated as \] because if the
2625 first data character in a class is ], it defines an empty class. (In
2626 Perl it is not possible to have an empty class.) The empty class []
2627 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2628 The negative empty class [^] matches any one character, independently
2629 of the DOTALL setting.
2631 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2632 non-existent subpattern following a character class starting with ']' and
2633 containing () gave an internal compiling error instead of "reference to
2634 non-existent subpattern". Fortunately, when the pattern did exist, the
2635 compiled code was correct. (When scanning forwards to check for the
2636 existence of the subpattern, it was treating the data ']' as terminating
2637 the class, so got the count wrong. When actually compiling, the reference
2638 was subsequently set up correctly.)
2640 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2641 it was being rejected as not supported by pcre_dfa_exec(), even though
2642 other assertions are supported. I have made pcre_dfa_exec() support
2643 (*FAIL).
2645 16. The implementation of 13c above involved the invention of a new opcode,
2646 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2647 cannot be changed at match time, I realized I could make a small
2648 improvement to matching performance by compiling OP_ALLANY instead of
2649 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2650 on the OP_ANY path.
2652 17. Compiling pcretest on Windows with readline support failed without the
2653 following two fixes: (1) Make the unistd.h include conditional on
2654 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2656 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2657 ncurses library to be included for pcretest when ReadLine support is
2658 requested, but also to allow for it to be overridden. This patch came from
2659 Daniel Bergström.
2661 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2662 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2663 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2664 spotting this.
2667 Version 7.6 28-Jan-08
2668 ---------------------
2670 1. A character class containing a very large number of characters with
2671 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2672 overflow.
2674 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2675 HAVE_LONG_LONG is not defined.
2677 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2678 bring them up to date and include new features. This patch includes:
2680 - Fixed PH's badly added libz and libbz2 support.
2681 - Fixed a problem with static linking.
2682 - Added pcredemo. [But later removed - see 7 below.]
2683 - Fixed dftables problem and added an option.
2684 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2686 - Added readline support for pcretest.
2687 - Added an listing of the option settings after cmake has run.
2689 4. A user submitted a patch to Makefile that makes it easy to create
2690 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2691 Makefile.am that cause it to include this special target, without
2692 affecting anything else. Note that the same mingw target plus all
2693 the other distribution libraries and programs are now supported
2694 when configuring with CMake (see 6 below) instead of with
2695 Configure/Make.
2697 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2698 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2699 exported in the Windows port". It has not yet been confirmed that the patch
2700 solves the problem, but it does no harm.
2702 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2703 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2704 with CMake, and also correct the comment about stack recursion.
2706 7. Remove the automatic building of pcredemo from the ./configure system and
2707 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2708 of a program that users should build themselves after PCRE is installed, so
2709 building it automatically is not really right. What is more, it gave
2710 trouble in some build environments.
2712 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2715 Version 7.5 10-Jan-08
2716 ---------------------
2718 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2719 values in parens when parsing an RE using the C++ wrapper."
2721 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2722 Characters greater than 255 were excluded from the class instead of being
2723 included.
2725 3. The same bug as (2) above applied to negated POSIX classes such as
2726 [:^space:].
2728 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2729 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2730 I have changed it.
2732 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2733 first named subpattern) and a construct such as (?&a) would reference the
2734 first named subpattern whose name started with "a" (in other words, the
2735 length check was missing). Both these problems are fixed. "Subpattern name
2736 expected" is now given for (?&) (a zero-length name), and this patch also
2737 makes it give the same error for \k'' (previously it complained that that
2738 was a reference to a non-existent subpattern).
2740 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2741 this is right because (?- can be followed by option settings as well as by
2742 digits. I have, however, made the messages clearer.
2744 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2745 than the number used in the conditional) now cause a compile-time error.
2746 This is actually not compatible with Perl, which accepts such patterns, but
2747 treats the conditional as always being FALSE (as PCRE used to), but it
2748 seems to me that giving a diagnostic is better.
2750 8. Change "alphameric" to the more common word "alphanumeric" in comments
2751 and messages.
2753 9. Fix two occurrences of "backslash" in comments that should have been
2754 "backspace".
2756 10. Remove two redundant lines of code that can never be obeyed (their function
2757 was moved elsewhere).
2759 11. The program that makes PCRE's Unicode character property table had a bug
2760 which caused it to generate incorrect table entries for sequences of
2761 characters that have the same character type, but are in different scripts.
2762 It amalgamated them into a single range, with the script of the first of
2763 them. In other words, some characters were in the wrong script. There were
2764 thirteen such cases, affecting characters in the following ranges:
2766 U+002b0 - U+002c1
2767 U+0060c - U+0060d
2768 U+0061e - U+00612
2769 U+0064b - U+0065e
2770 U+0074d - U+0076d
2771 U+01800 - U+01805
2772 U+01d00 - U+01d77
2773 U+01d9b - U+01dbf
2774 U+0200b - U+0200f
2775 U+030fc - U+030fe
2776 U+03260 - U+0327f
2777 U+0fb46 - U+0fbb1
2778 U+10450 - U+1049d
2780 12. The -o option (show only the matching part of a line) for pcregrep was not
2781 compatible with GNU grep in that, if there was more than one match in a
2782 line, it showed only the first of them. It now behaves in the same way as
2783 GNU grep.
2785 13. If the -o and -v options were combined for pcregrep, it printed a blank
2786 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2787 does the same. The return code can be used to tell if there were any
2788 non-matching lines.
2790 14. Added --file-offsets and --line-offsets to pcregrep.
2792 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2793 infinitely looping recursion. The bug was that positive lookaheads were not
2794 being skipped when checking for a possible empty match (negative lookaheads
2795 and both kinds of lookbehind were skipped).
2797 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2798 inclusion of <windows.h> to before rather than after the definition of
2799 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2801 17. Specifying a possessive quantifier with a specific limit for a Unicode
2802 character property caused pcre_compile() to compile bad code, which led at
2803 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2804 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2805 caused the error; without that there was no problem.
2807 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2809 19. Added --enable-pcretest-libreadline.
2811 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2812 RE::GlobalReplace(). As a result, the number of replacements returned was
2813 double what it should be. I removed one of the increments, but Craig sent a
2814 later patch that removed the other one (the right fix) and added unit tests
2815 that check the return values (which was not done before).
2817 21. Several CMake things:
2819 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2820 the names libpcre and libpcreposix, not just pcre and pcreposix.
2822 (2) The above change means that pcretest and pcregrep are now correctly
2823 linked with the newly-built libraries, not previously installed ones.
2827 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2828 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2829 UTF-8 newline character). The key issue is that the pattern starts .*;
2830 this means that the match must be either at the beginning, or after a
2831 newline. The bug was in the code for advancing after a failed match and
2832 checking that the new position followed a newline. It was not taking
2833 account of UTF-8 characters correctly.
2835 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2836 character classes. PCRE was not treating the sequence [:...:] as a
2837 character class unless the ... were all letters. Perl, however, seems to
2838 allow any characters between [: and :], though of course it rejects as
2839 unknown any "names" that contain non-letters, because all the known class
2840 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2841 for example, whereas PCRE did not - it did not recognize a POSIX character
2842 class. This seemed a bit dangerous, so the code has been changed to be
2843 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2844 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2845 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2846 Perl does, and where it didn't before.
2848 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2849 Windows environments %n is disabled by default.
2852 Version 7.4 21-Sep-07
2853 ---------------------
2855 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2856 means that a class such as [\s] counted as "explicit reference to CR or
2857 LF". That isn't really right - the whole point of the change was to try to
2858 help when there was an actual mention of one of the two characters. So now
2859 the change happens only if \r or \n (or a literal CR or LF) character is
2860 encountered.
2862 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2863 of both had grown to the point where there were only 3 bits left.
2864 Fortunately, there was spare space in the data structure, and so I have
2865 moved the internal flags into a new 16-bit field to free up more option
2866 bits.
2868 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2869 but did not set the internal JCHANGED flag - either of these is enough to
2870 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2871 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2872 start sets both bits.
2874 4. Added options (at build time, compile time, exec time) to change \R from
2875 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2877 5. doc/pcresyntax.html was missing from the distribution.
2879 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2880 compatibility, even though it is no longer used.
2882 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2883 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2884 windows.h file is present (where different names are used). [This was
2885 reversed later after testing - see 16 below.]
2887 8. Changed all #include <config.h> to #include "config.h". There were also
2888 some further <pcre.h> cases that I changed to "pcre.h".
2890 9. When pcregrep was used with the --colour option, it missed the line ending
2891 sequence off the lines that it output.
2893 10. It was pointed out to me that arrays of string pointers cause lots of
2894 relocations when a shared library is dynamically loaded. A technique of
2895 using a single long string with a table of offsets can drastically reduce
2896 these. I have refactored PCRE in four places to do this. The result is
2897 dramatic:
2899 Originally: 290
2900 After changing UCP table: 187
2901 After changing error message table: 43
2902 After changing table of "verbs" 36
2903 After changing table of Posix names 22
2905 Thanks to the folks working on Gregex for glib for this insight.
2907 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2908 unicode-properties was also set.
2910 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2912 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2913 checked only for CRLF.
2915 14. Added casts to pcretest.c to avoid compiler warnings.
2917 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2919 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2920 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2921 entirely. This removes changes made in 7 above.
2923 17. The CMake files have been updated, and there is now more information about
2924 building with CMake in the NON-UNIX-USE document.
2927 Version 7.3 28-Aug-07
2928 ---------------------
2930 1. In the rejigging of the build system that eventually resulted in 7.1, the
2931 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2932 brackets there is not right, since it causes compilers to look for an
2933 installed pcre.h, not the version that is in the source that is being
2934 compiled (which of course may be different). I have changed it back to:
2936 #include "pcre.h"
2938 I have a vague recollection that the change was concerned with compiling in
2939 different directories, but in the new build system, that is taken care of
2940 by the VPATH setting the Makefile.
2942 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2943 when the subject happened to end in the byte 0x85 (e.g. if the last
2944 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2945 characters but of course it shouldn't be taken as a newline when it is part
2946 of another character. The bug was that, for an unlimited repeat of . in
2947 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2948 characters when looking for a newline.
2950 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2952 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2953 in debug output.
2955 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2956 long printing in the pcrecpp unittest when running under MinGW.
2958 6. ESC_K was left out of the EBCDIC table.
2960 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2961 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2962 limit also applies to "virtual nesting" when a pattern is recursive, and in
2963 this case 1000 isn't so big. I have been able to remove this limit at the
2964 expense of backing off one optimization in certain circumstances. Normally,
2965 when pcre_exec() would call its internal match() function recursively and
2966 immediately return the result unconditionally, it uses a "tail recursion"
2967 feature to save stack. However, when a subpattern that can match an empty
2968 string has an unlimited repetition quantifier, it no longer makes this
2969 optimization. That gives it a stack frame in which to save the data for
2970 checking that an empty string has been matched. Previously this was taken
2971 from the 1000-entry workspace that had been reserved. So now there is no
2972 explicit limit, but more stack is used.
2974 8. Applied Daniel's patches to solve problems with the import/export magic
2975 syntax that is required for Windows, and which was going wrong for the
2976 pcreposix and pcrecpp parts of the library. These were overlooked when this
2977 problem was solved for the main library.
2979 9. There were some crude static tests to avoid integer overflow when computing
2980 the size of patterns that contain repeated groups with explicit upper
2981 limits. As the maximum quantifier is 65535, the maximum group length was
2982 set at 30,000 so that the product of these two numbers did not overflow a
2983 32-bit integer. However, it turns out that people want to use groups that
2984 are longer than 30,000 bytes (though not repeat them that many times).
2985 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2986 made it possible to implement the integer overflow checks in a much more
2987 dynamic way, which I have now done. The artificial limitation on group
2988 length has been removed - we now have only the limit on the total length of
2989 the compiled pattern, which depends on the LINK_SIZE setting.
2991 10. Fixed a bug in the documentation for get/copy named substring when
2992 duplicate names are permitted. If none of the named substrings are set, the
2993 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2994 empty string.
2996 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2997 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2998 because the ] is interpreted as the first data character and the
2999 terminating ] is not found. PCRE has been made compatible with Perl in this
3000 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3001 cause memory overwriting.
3003 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3004 string has been matched (to stop an infinite loop). It was not recognizing
3005 a conditional subpattern that could match an empty string if that
3006 subpattern was within another subpattern. For example, it looped when
3007 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3008 condition was not nested. This bug has been fixed.
3010 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3011 past the start of the subject in the presence of bytes with the top bit
3012 set, for example "\x8aBCD".
3014 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3015 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3017 14. Optimized (?!) to (*FAIL).
3019 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3020 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3021 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3022 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3023 does: it's just the validity check that is more restrictive.
3025 16. Inserted checks for integer overflows during escape sequence (backslash)
3026 processing, and also fixed erroneous offset values for syntax errors during
3027 backslash processing.
3029 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3030 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3032 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3033 caused an overrun.
3035 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3036 something other than just ASCII characters) inside a group that had an
3037 unlimited repeat caused a loop at compile time (while checking to see
3038 whether the group could match an empty string).
3040 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3041 [\P{Any}] did so. (Error in the code for printing property names.)
3043 21. An orphan \E inside a character class could cause a crash.
3045 22. A repeated capturing bracket such as (A)? could cause a wild memory
3046 reference during compilation.
3048 23. There are several functions in pcre_compile() that scan along a compiled
3049 expression for various reasons (e.g. to see if it's fixed length for look
3050 behind). There were bugs in these functions when a repeated \p or \P was
3051 present in the pattern. These operators have additional parameters compared
3052 with \d, etc, and these were not being taken into account when moving along
3053 the compiled data. Specifically:
3055 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3056 length.
3058 (b) An item such as \pL+ within a repeated group could cause crashes or
3059 loops.
3061 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3062 "reference to non-existent subpattern" error.
3064 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3066 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3067 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3069 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3071 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3072 character were causing crashes (broken optimization).
3074 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3075 \p or \P) caused a compile-time loop.
3077 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3078 break. For example, the unstudied pattern [\r\n]A does not match the string
3079 "\r\nA" because change 7.0/46 below moves the current point on by two
3080 characters after failing to match at the start. However, the pattern \nA
3081 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3082 the same is true. There doesn't seem any very clean way out of this, but
3083 what I have chosen to do makes the common cases work: PCRE now takes note
3084 of whether there can be an explicit match for \r or \n anywhere in the
3085 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3086 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3087 pattern has explicit CR or LF references.
3089 29. Added (*CR) etc for changing newline setting at start of pattern.
3092 Version 7.2 19-Jun-07
3093 ---------------------
3095 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3096 which is apparently normally available under Windows.
3098 2. Re-jig the pcregrep tests with different newline settings in an attempt
3099 to make them independent of the local environment's newline setting.
3101 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3103 4. Some of the "internals" tests were previously cut out when the link size
3104 was not 2, because the output contained actual offsets. The recent new
3105 "Z" feature of pcretest means that these can be cut out, making the tests
3106 usable with all link sizes.
3108 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3109 stack recursion. This gives a massive performance boost under BSD, but just
3110 a small improvement under Linux. However, it saves one field in the frame
3111 in all cases.
3113 6. Added more features from the forthcoming Perl 5.10:
3115 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3116 recursion call. It refers to the nth most recently opened parentheses.
3118 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3119 to be opened parentheses.
3121 (c) Conditions that refer to capturing parentheses can be specified
3122 relatively, for example, (?(-2)... or (?(+3)...
3124 (d) \K resets the start of the current match so that everything before
3125 is not part of it.
3127 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3129 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3130 reference syntax.
3132 (g) (?| introduces a group in which the numbering of parentheses in each
3133 alternative starts with the same number.
3135 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3137 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3140 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3141 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3142 for detecting groups that can match an empty string.
3144 9. A pattern with a very large number of alternatives (more than several
3145 hundred) was running out of internal workspace during the pre-compile
3146 phase, where pcre_compile() figures out how much memory will be needed. A
3147 bit of new cunning has reduced the workspace needed for groups with
3148 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3149 workspace instead of running out of the 4096 that are available.
3151 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3153 11. Applied patch from Google to remove an optimization that didn't quite work.
3154 The report of the bug said:
3156 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3157 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3158 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3160 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3161 it matched the wrong number of bytes.
3164 Version 7.1 24-Apr-07
3165 ---------------------
3167 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3168 that is more "standard", making use of automake and other Autotools. There
3169 is some re-arrangement of the files and adjustment of comments consequent
3170 on this.
3172 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3173 for recursive directory scanning broke on some systems because the files
3174 are not scanned in any specific order and on different systems the order
3175 was different. A call to "sort" has been inserted into RunGrepTest for the
3176 approprate test as a short-term fix. In the longer term there may be an
3177 alternative.
3179 3. I had an email from Eric Raymond about problems translating some of PCRE's
3180 man pages to HTML (despite the fact that I distribute HTML pages, some
3181 people do their own conversions for various reasons). The problems
3182 concerned the use of low-level troff macros .br and .in. I have therefore
3183 removed all such uses from the man pages (some were redundant, some could
3184 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3185 HTML has been updated to handle .nf/.fi and to complain if it encounters
3186 .br or .in.
3188 4. Updated comments in configure.ac that get placed in config.h.in and also
3189 arranged for config.h to be included in the distribution, with the name
3190 config.h.generic, for the benefit of those who have to compile without
3191 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3193 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3194 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3195 makevp.bat for the current PCRE, using the additional files
3196 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3198 6. A Windows user reported a minor discrepancy with test 2, which turned out
3199 to be caused by a trailing space on an input line that had got lost in his
3200 copy. The trailing space was an accident, so I've just removed it.
3202 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3203 that is needed.
3205 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3206 as "const" (a) because they are and (b) because it helps the PHP
3207 maintainers who have recently made a script to detect big data structures
3208 in the php code that should be moved to the .rodata section. I remembered
3209 to update Builducptable as well, so it won't revert if ucptable.h is ever
3210 re-created.
3212 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3213 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3214 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3215 support is not required. This saves 1.5-2K of code, which is important in
3216 some applications.
3218 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3219 so as not to refer to the tables, even though these functions will never be
3220 called when UTF-8 support is disabled. Otherwise there are problems with a
3221 shared library.
3223 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3225 (a) It was defining its arguments as char * instead of void *.
3227 (b) It was assuming that all moves were upwards in memory; this was true
3228 a long time ago when I wrote it, but is no longer the case.
3230 The emulated memove() is provided for those environments that have neither
3231 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3232 is clearly not the case, as these two bugs were recently reported.
3234 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3235 and Detrail to create the HTML documentation, the .txt form of the man
3236 pages, and it removes trailing spaces from listed files. It also creates
3237 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3238 case, it wraps all the #defines with #ifndefs. This script should be run
3239 before "make dist".
3241 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3242 with Unicode property support.
3244 (a) For a maximizing quantifier, if the two different cases of the
3245 character were of different lengths in their UTF-8 codings (there are
3246 some cases like this - I found 11), and the matching function had to
3247 back up over a mixture of the two cases, it incorrectly assumed they
3248 were both the same length.
3250 (b) When PCRE was configured to use the heap rather than the stack for
3251 recursion during matching, it was not correctly preserving the data for
3252 the other case of a UTF-8 character when checking ahead for a match
3253 while processing a minimizing repeat. If the check also involved
3254 matching a wide character, but failed, corruption could cause an
3255 erroneous result when trying to check for a repeat of the original
3256 character.
3258 13. Some tidying changes to the testing mechanism:
3260 (a) The RunTest script now detects the internal link size and whether there
3261 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3262 values substituted by "configure". (The RunGrepTest script already did
3263 this for UTF-8.) The configure.ac script no longer substitutes the
3264 relevant variables.
3266 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3267 with length and offset values. This means that the output is different
3268 for different internal link sizes. Test 2 is skipped for link sizes
3269 other than 2 because of this, bypassing the problem. Unfortunately,
3270 there was also a test in test 3 (the locale tests) that used /B and
3271 failed for link sizes other than 2. Rather than cut the whole test out,
3272 I have added a new /Z option to pcretest that replaces the length and
3273 offset values with spaces. This is now used to make test 3 independent
3274 of link size. (Test 2 will be tidied up later.)
3276 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3277 segmentation fault instead of returning the appropriate error message.
3279 15. In multiline mode when the newline sequence was set to "any", the pattern
3280 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3281 This doesn't seem right; it now treats the CRLF combination as the line
3282 ending, and so does not match in that case. It's only a pattern such as ^$
3283 that would hit this one: something like ^ABC$ would have failed after \r
3284 and then tried again after \r\n.
3286 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3287 in an attempt to make files that differ only in their line terminators
3288 compare equal. This works on Linux.
3290 17. Under certain error circumstances pcregrep might try to free random memory
3291 as it exited. This is now fixed, thanks to valgrind.
3293 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3294 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3295 was because its rules for how to advance for /g after matching an empty
3296 string at the end of a line did not allow for this case. They now check for
3297 it specially.
3299 20. pcretest is supposed to handle patterns and data of any length, by
3300 extending its buffers when necessary. It was getting this wrong when the
3301 buffer for a data line had to be extended.
3303 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3304 CRLF as a newline sequence.
3306 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3307 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3308 I have nevertheless tidied it up.
3310 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3312 24. Added a man page for pcre-config.
3315 Version 7.0 19-Dec-06
3316 ---------------------
3318 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3319 moving to gcc 4.1.1.
3321 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3322 sys/time.h, which is documented as needed for this function. It doesn't
3323 seem to matter on Linux, but it showed up on some releases of OS X.
3325 3. It seems that there are systems where bytes whose values are greater than
3326 127 match isprint() in the "C" locale. The "C" locale should be the
3327 default when a C program starts up. In most systems, only ASCII printing
3328 characters match isprint(). This difference caused the output from pcretest
3329 to vary, making some of the tests fail. I have changed pcretest so that:
3331 (a) When it is outputting text in the compiled version of a pattern, bytes
3332 other than 32-126 are always shown as hex escapes.
3334 (b) When it is outputting text that is a matched part of a subject string,
3335 it does the same, unless a different locale has been set for the match
3336 (using the /L modifier). In this case, it uses isprint() to decide.
3338 4. Fixed a major bug that caused incorrect computation of the amount of memory
3339 required for a compiled pattern when options that changed within the
3340 pattern affected the logic of the preliminary scan that determines the
3341 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3342 that the computed length was too small. The symptoms of this bug were
3343 either the PCRE error "internal error: code overflow" from pcre_compile(),
3344 or a glibc crash with a message such as "pcretest: free(): invalid next
3345 size (fast)". Examples of patterns that provoked this bug (shown in
3346 pcretest format) are:
3348 /(?-x: )/x
3349 /(?x)(?-x: \s*#\s*)/
3350 /((?i)[\x{c0}])/8
3351 /(?i:[\x{c0}])/8
3353 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3354 is now done differently.
3356 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3357 wrapper classes; (b) implement a new function in the C++ scanner that is
3358 more efficient than the old way of doing things because it avoids levels of
3359 recursion in the regex matching; (c) add a paragraph to the documentation
3360 for the FullMatch() function.
3362 6. The escape sequence \n was being treated as whatever was defined as
3363 "newline". Not only was this contrary to the documentation, which states
3364 that \n is character 10 (hex 0A), but it also went horribly wrong when
3365 "newline" was defined as CRLF. This has been fixed.
3367 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3368 was being set to -1 for the "end of line" case (supposedly a value that no
3369 character can have). Though this value is never used (the check for end of
3370 line is "zero bytes in current character"), it caused compiler complaints.
3371 I've changed it to 0xffffffff.
3373 8. In pcre_version.c, the version string was being built by a sequence of
3374 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3375 string (as it is for production releases) called a macro with an empty
3376 argument. The C standard says the result of this is undefined. The gcc
3377 compiler treats it as an empty string (which was what was wanted) but it is
3378 reported that Visual C gives an error. The source has been hacked around to
3379 avoid this problem.
3381 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3382 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3383 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3384 of them did).
3386 10. Originally, pcretest opened its input and output without "b"; then I was
3387 told that "b" was needed in some environments, so it was added for release
3388 5.0 to both the input and output. (It makes no difference on Unix-like
3389 systems.) Later I was told that it is wrong for the input on Windows. I've
3390 now abstracted the modes into two macros, to make it easier to fiddle with
3391 them, and removed "b" from the input mode under Windows.
3393 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3395 12. Added -help and --help to pcretest as an official way of being reminded
3396 of the options.
3398 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3399 and pcrecpp.cc because they annoy compilers at high warning levels.
3401 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3403 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3404 scripts are not C programs :-) and which was not noticed because it works
3405 on Linux.
3407 16. pcretest is supposed to handle any length of pattern and data line (as one
3408 line or as a continued sequence of lines) by extending its input buffer if
3409 necessary. This feature was broken for very long pattern lines, leading to
3410 a string of junk being passed to pcre_compile() if the pattern was longer
3411 than about 50K.
3413 17. I have done a major re-factoring of the way pcre_compile() computes the
3414 amount of memory needed for a compiled pattern. Previously, there was code
3415 that made a preliminary scan of the pattern in order to do this. That was
3416 OK when PCRE was new, but as the facilities have expanded, it has become
3417 harder and harder to keep it in step with the real compile phase, and there
3418 have been a number of bugs (see for example, 4 above). I have now found a
3419 cunning way of running the real compile function in a "fake" mode that
3420 enables it to compute how much memory it would need, while actually only
3421 ever using a few hundred bytes of working memory and without too many
3422 tests of the mode. This should make future maintenance and development
3423 easier. A side effect of this work is that the limit of 200 on the nesting
3424 depth of parentheses has been removed (though this was never a serious
3425 limitation, I suspect). However, there is a downside: pcre_compile() now
3426 runs more slowly than before (30% or more, depending on the pattern). I
3427 hope this isn't a big issue. There is no effect on runtime performance.
3429 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3430 newline (only possible for the last line of a file) and it was a
3431 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3433 19. Added additional timing features to pcretest. (1) The -tm option now times
3434 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3435 separate command line item, by a number that specifies the number of
3436 repeats to use when timing. The default is 50000; this gives better
3437 precision, but takes uncomfortably long for very large patterns.
3439 20. Extended pcre_study() to be more clever in cases where a branch of a
3440 subpattern has no definite first character. For example, (a*|b*)[cd] would
3441 previously give no result from pcre_study(). Now it recognizes that the
3442 first character must be a, b, c, or d.
3444 21. There was an incorrect error "recursive call could loop indefinitely" if
3445 a subpattern (or the entire pattern) that was being tested for matching an
3446 empty string contained only one non-empty item after a nested subpattern.
3447 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3448 incorrectly, because the \d was being skipped in the check.
3450 22. The pcretest program now has a new pattern option /B and a command line
3451 option -b, which is equivalent to adding /B to every pattern. This causes
3452 it to show the compiled bytecode, without the additional information that
3453 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3454 is the same as /B/I).
3456 23. A new optimization is now able automatically to treat some sequences such
3457 as a*b as a*+b. More specifically, if something simple (such as a character
3458 or a simple class like \d) has an unlimited quantifier, and is followed by
3459 something that cannot possibly match the quantified thing, the quantifier
3460 is automatically "possessified".
3462 24. A recursive reference to a subpattern whose number was greater than 39
3463 went wrong under certain circumstances in UTF-8 mode. This bug could also
3464 have affected the operation of pcre_study().
3466 25. Realized that a little bit of performance could be had by replacing
3467 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3469 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3471 27. Possessive quantifiers such as a++ were previously implemented by turning
3472 them into atomic groups such as ($>a+). Now they have their own opcodes,
3473 which improves performance. This includes the automatically created ones
3474 from 23 above.
3476 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3477 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3478 the first matched character to be a colon. This applied both to named and
3479 numbered groups.
3481 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3483 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3484 building PCRE on Windows easier, so I have included it in the distribution.
3486 31. There is now a check in pcretest against a ridiculously large number being
3487 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3488 loop, the loop is abandoned.
3490 32. Forward references to subpatterns in conditions such as (?(2)...) where
3491 subpattern 2 is defined later cause pcre_compile() to search forwards in
3492 the pattern for the relevant set of parentheses. This search went wrong
3493 when there were unescaped parentheses in a character class, parentheses
3494 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3496 33. "Subroutine" calls and backreferences were previously restricted to
3497 referencing subpatterns earlier in the regex. This restriction has now
3498 been removed.
3500 34. Added a number of extra features that are going to be in Perl 5.10. On the
3501 whole, these are just syntactic alternatives for features that PCRE had
3502 previously implemented using the Python syntax or my own invention. The
3503 other formats are all retained for compatibility.
3505 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3506 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3507 also .NET compatible.
3509 (b) A recursion or subroutine call to a named group can now be defined as
3510 (?&name) as well as (?P>name).
3512 (c) A backreference to a named group can now be defined as \k<name> or
3513 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3514 5.10, are also .NET compatible.
3516 (d) A conditional reference to a named group can now use the syntax
3517 (?(<name>) or (?('name') as well as (?(name).
3519 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3520 groups (named and numbered) that are never evaluated inline, but can be
3521 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3522 is always false. There may be only one alternative in such a group.
3524 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3525 as the simple (?(R). The condition is true only if the most recent
3526 recursion is that of the given number or name. It does not search out
3527 through the entire recursion stack.
3529 (g) The escape \gN or \g{N} has been added, where N is a positive or
3530 negative number, specifying an absolute or relative reference.
3532 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3533 some "unreachable code" warnings.
3535 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3536 things, this adds five new scripts.
3538 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3539 There were also incompatibilities regarding the handling of \Q..\E inside
3540 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3541 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3543 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3544 matches an empty string, and forcibly breaks the loop. There were bugs in
3545 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3546 against aaaa the result was just "a" rather than "aaaa", for example. Two
3547 separate and independent bugs (that affected different cases) have been
3548 fixed.
3550 39. Refactored the code to abolish the use of different opcodes for small
3551 capturing bracket numbers. This is a tidy that I avoided doing when I
3552 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3553 The new approach is not only tidier, it makes it possible to reduce the
3554 memory needed to fix the previous bug (38).
3556 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3557 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3558 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3559 mode.
3561 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3562 report.
3564 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3565 copy construction and assignment in the C++ wrapper.
3567 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3568 couple of bugs that could have given wrong results in the "--newline=crlf"
3569 case.
3571 44. Added a number of casts and did some reorganization of signed/unsigned int
3572 variables following suggestions from Dair Grant. Also renamed the variable
3573 "this" as "item" because it is a C++ keyword.
3575 45. Arranged for dftables to add
3577 #include "pcre_internal.h"
3579 to pcre_chartables.c because without it, gcc 4.x may remove the array
3580 definition from the final binary if PCRE is built into a static library and
3581 dead code stripping is activated.
3583 46. For an unanchored pattern, if a match attempt fails at the start of a
3584 newline sequence, and the newline setting is CRLF or ANY, and the next two
3585 characters are CRLF, advance by two characters instead of one.
3588 Version 6.7 04-Jul-06
3589 ---------------------
3591 1. In order to handle tests when input lines are enormously long, pcretest has
3592 been re-factored so that it automatically extends its buffers when
3593 necessary. The code is crude, but this _is_ just a test program. The
3594 default size has been increased from 32K to 50K.
3596 2. The code in pcre_study() was using the value of the re argument before
3597 testing it for NULL. (Of course, in any sensible call of the function, it
3598 won't be NULL.)
3600 3. The memmove() emulation function in pcre_internal.h, which is used on
3601 systems that lack both memmove() and bcopy() - that is, hardly ever -
3602 was missing a "static" storage class specifier.
3604 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3605 containing an extended class (one that cannot be represented by a bitmap
3606 because it contains high-valued characters or Unicode property items, e.g.
3607 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3608 pattern, but PCRE should not loop if you do not (it no longer does).
3609 [Detail: two cases were found: (a) a repeated subpattern containing an
3610 extended class; (b) a recursive reference to a subpattern that followed a
3611 previous extended class. It wasn't skipping over the extended class
3612 correctly when UTF-8 mode was not set.]
3614 5. A negated single-character class was not being recognized as fixed-length
3615 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3616 compile error "lookbehind assertion is not fixed length".
3618 6. The RunPerlTest auxiliary script was showing an unexpected difference
3619 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3620 write a Perl script that can interpret lines of an input file either as
3621 byte characters or as UTF-8, which is what "perltest" was being required to
3622 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3623 can't do is switch easily at run time between having the "use utf8;" pragma
3624 or not. In the end, I fudged it by using the RunPerlTest script to insert
3625 "use utf8;" explicitly for the UTF-8 tests.
3627 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3628 the end of the subject string, contrary to the documentation and to what
3629 Perl does. This was true of both matching functions. Now it matches only at
3630 the start of the subject and immediately after *internal* newlines.
3632 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3633 a pointer to an int instead of a pointer to an unsigned long int. This
3634 caused problems on 64-bit systems.
3636 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3637 instance of the 'standard' template library not being so standard".
3639 10. There was no check on the number of named subpatterns nor the maximum
3640 length of a subpattern name. The product of these values is used to compute
3641 the size of the memory block for a compiled pattern. By supplying a very
3642 long subpattern name and a large number of named subpatterns, the size
3643 computation could be caused to overflow. This is now prevented by limiting
3644 the length of names to 32 characters, and the number of named subpatterns
3645 to 10,000.
3647 11. Subpatterns that are repeated with specific counts have to be replicated in
3648 the compiled pattern. The size of memory for this was computed from the
3649 length of the subpattern and the repeat count. The latter is limited to
3650 65535, but there was no limit on the former, meaning that integer overflow
3651 could in principle occur. The compiled length of a repeated subpattern is
3652 now limited to 30,000 bytes in order to prevent this.
3654 12. Added the optional facility to have named substrings with the same name.
3656 13. Added the ability to use a named substring as a condition, using the
3657 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3658 are numbers (not recommended). Forward references are permitted.
3660 14. Added forward references in named backreferences (if you see what I mean).
3662 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3663 pattern could run off the end of the subject. For example, the pattern
3664 "(?s)(.{1,5})"8 did this with the subject "ab".
3666 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3667 PCRE_CASELESS was set when matching characters that were quantified with ?
3668 or *.
3670 17. A character class other than a single negated character that had a minimum
3671 but no maximum quantifier - for example [ab]{6,} - was not handled
3672 correctly by pce_dfa_exec(). It would match only one character.
3674 18. A valid (though odd) pattern that looked like a POSIX character
3675 class but used an invalid character after [ (for example [[,abc,]]) caused
3676 pcre_compile() to give the error "Failed: internal error: code overflow" or
3677 in some cases to crash with a glibc free() error. This could even happen if
3678 the pattern terminated after [[ but there just happened to be a sequence of
3679 letters, a binary zero, and a closing ] in the memory that followed.
3681 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3682 over the years. Originally (before any Unicode support), just the bottom 8
3683 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3684 output from "man perlunicode" includes this:
3686 The regular expression compiler produces polymorphic opcodes. That
3687 is, the pattern adapts to the data and automatically switches to
3688 the Unicode character scheme when presented with Unicode data--or
3689 instead uses a traditional byte scheme when presented with byte
3690 data.
3692 Sadly, a wide octal escape does not cause a switch, and in a string with
3693 no other multibyte characters, these octal escapes are treated as before.
3694 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3695 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3696 Unicode string.
3698 I have not perpetrated such confusion in PCRE. Up till now, it took just
3699 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3700 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3701 translate to the appropriate multibyte character.
3703 29. Applied some refactoring to reduce the number of warnings from Microsoft
3704 and Borland compilers. This has included removing the fudge introduced
3705 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3706 a warning about an unused variable.
3708 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3709 characters since release 4.0, because Perl (from release 5.004) does not.
3710 [Or at least, is documented not to: some releases seem to be in conflict
3711 with the documentation.] However, when a pattern was studied with
3712 pcre_study() and all its branches started with \s, PCRE still included VT
3713 as a possible starting character. Of course, this did no harm; it just
3714 caused an unnecessary match attempt.
3716 22. Removed a now-redundant internal flag bit that recorded the fact that case
3717 dependency changed within the pattern. This was once needed for "required
3718 byte" processing, but is no longer used. This recovers a now-scarce options
3719 bit. Also moved the least significant internal flag bit to the most-
3720 significant bit of the word, which was not previously used (hangover from
3721 the days when it was an int rather than a uint) to free up another bit for
3722 the future.
3724 23. Added support for CRLF line endings as well as CR and LF. As well as the
3725 default being selectable at build time, it can now be changed at runtime
3726 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3727 specify that it is scanning data with non-default line endings.
3729 24. Changed the definition of CXXLINK to make it agree with the definition of
3730 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3732 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3733 recursions. This makes a big different to stack usage for some patterns.
3735 26. If a subpattern containing a named recursion or subroutine reference such
3736 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3737 the space required for the compiled pattern went wrong and gave too small a
3738 value. Depending on the environment, this could lead to "Failed: internal
3739 error: code overflow at offset 49" or "glibc detected double free or
3740 corruption" errors.
3742 27. Applied patches from Google (a) to support the new newline modes and (b) to
3743 advance over multibyte UTF-8 characters in GlobalReplace.
3745 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3746 difference for some implementation of PCRE in some Windows version.
3748 29. Added some extra testing facilities to pcretest:
3750 \q<number> in a data line sets the "match limit" value
3751 \Q<number> in a data line sets the "match recursion limt" value
3752 -S <number> sets the stack size, where <number> is in megabytes
3754 The -S option isn't available for Windows.
3757 Version 6.6 06-Feb-06
3758 ---------------------
3760 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3761 in pcreposix.h. I have copied the definition from pcre.h.
3763 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3764 because pcre.h is no longer a built file.
3766 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3767 not normally included in the compiled code.
3770 Version 6.5 01-Feb-06
3771 ---------------------
3773 1. When using the partial match feature with pcre_dfa_exec(), it was not
3774 anchoring the second and subsequent partial matches at the new starting
3775 point. This could lead to incorrect results. For example, with the pattern
3776 /1234/, partially matching against "123" and then "a4" gave a match.
3778 2. Changes to pcregrep:
3780 (a) All non-match returns from pcre_exec() were being treated as failures
3781 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3782 error message is output. Some extra information is given for the
3784 probably the only errors that are likely to be caused by users (by
3785 specifying a regex that has nested indefinite repeats, for instance).
3786 If there are more than 20 of these errors, pcregrep is abandoned.
3788 (b) A binary zero was treated as data while matching, but terminated the
3789 output line if it was written out. This has been fixed: binary zeroes
3790 are now no different to any other data bytes.
3792 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3793 used to set a locale for matching. The --locale=xxxx long option has
3794 been added (no short equivalent) to specify a locale explicitly on the
3795 pcregrep command, overriding the environment variables.
3797 (d) When -B was used with -n, some line numbers in the output were one less
3798 than they should have been.
3800 (e) Added the -o (--only-matching) option.
3802 (f) If -A or -C was used with -c (count only), some lines of context were
3803 accidentally printed for the final match.
3805 (g) Added the -H (--with-filename) option.
3807 (h) The combination of options -rh failed to suppress file names for files
3808 that were found from directory arguments.
3810 (i) Added the -D (--devices) and -d (--directories) options.
3812 (j) Added the -F (--fixed-strings) option.
3814 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3816 (l) Added the --colo(u)r option.
3818 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3819 is not present by default.
3821 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3822 items such as (?R) or (?1), when the recursion could match a number of
3823 alternatives. If it matched one of the alternatives, but subsequently,
3824 outside the recursion, there was a failure, the code tried to back up into
3825 the recursion. However, because of the way PCRE is implemented, this is not
3826 possible, and the result was an incorrect result from the match.
3828 In order to prevent this happening, the specification of recursion has
3829 been changed so that all such subpatterns are automatically treated as
3830 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3832 4. I had overlooked the fact that, in some locales, there are characters for
3833 which isalpha() is true but neither isupper() nor islower() are true. In
3834 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3835 and ordfeminine) are like this. This affected the treatment of \w and \W
3836 when they appeared in character classes, but not when they appeared outside
3837 a character class. The bit map for "word" characters is now created
3838 separately from the results of isalnum() instead of just taking it from the
3839 upper, lower, and digit maps. (Plus the underscore character, of course.)
3841 5. The above bug also affected the handling of POSIX character classes such as
3842 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3843 permanent tables. Instead, the bit maps for such a class were previously
3844 created as the appropriate unions of the upper, lower, and digit bitmaps.
3845 Now they are created by subtraction from the [[:word:]] class, which has
3846 its own bitmap.
3848 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3849 It is created by subtracting the vertical space characters (\x09, \x0a,
3850 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3851 subtraction was done in the overall bitmap for a character class, meaning
3852 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3853 be recognized. This bug has been fixed.
3855 7. Patches from the folks at Google:
3857 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3858 real life, but is still worth protecting against".
3860 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3861 regular expressions".
3863 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3864 have it.
3866 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3867 "configure" and the latter not, in order to fix a problem somebody had
3868 with compiling the Arg class on HP-UX.
3870 (e) Improve the error-handling of the C++ wrapper a little bit.
3872 (f) New tests for checking recursion limiting.
3874 8. The pcre_memmove() function, which is used only if the environment does not
3875 have a standard memmove() function (and is therefore rarely compiled),
3876 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3877 returning a result (though PCRE never actually uses the result).
3879 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3880 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3881 returned instead of calling malloc() with an overflowing number that would
3882 most likely cause subsequent chaos.
3884 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3886 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3887 with this option is matched, the nmatch and pmatch options of regexec() are
3888 ignored.
3890 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3891 provided in case anyone wants to the the POSIX interface with UTF-8
3892 strings.
3894 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3895 C++ linking (needed for some HP-UX environments).
3897 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3898 (unused parameter) and in the pcre_printint() function (omitted "default"
3899 switch label when the default is to do nothing).
3901 15. Added some code to make it possible, when PCRE is compiled as a C++
3902 library, to replace subject pointers for pcre_exec() with a smart pointer
3903 class, thus making it possible to process discontinuous strings.
3905 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3906 much the same function. They were added by different people who were trying
3907 to make PCRE easy to compile on non-Unix systems. It has been suggested
3908 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3909 for compiling on Windows systems. I have therefore replaced it with
3910 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3911 defaults to "extern" for C or "extern C" for C++, which works fine on
3912 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3913 SCOPE with something explicit in config.h. In addition:
3915 (a) pcreposix.h still had just "extern" instead of either of these macros;
3916 I have replaced it with PCRE_DATA_SCOPE.
3918 (b) Functions such as _pcre_xclass(), which are internal to the library,
3919 but external in the C sense, all had PCRE_EXPORT in their definitions.
3920 This is apparently wrong for the Windows case, so I have removed it.
3921 (It makes no difference on Unix-like systems.)
3923 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3924 of recursive calls to match(). This is different to MATCH_LIMIT because
3925 that limits the total number of calls to match(), not all of which increase
3926 the depth of recursion. Limiting the recursion depth limits the amount of
3927 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3928 when PCRE is compiled, and changed at run time. A patch from Google adds
3929 this functionality to the C++ interface.
3931 18. Changes to the handling of Unicode character properties:
3933 (a) Updated the table to Unicode 4.1.0.
3935 (b) Recognize characters that are not in the table as "Cn" (undefined).
3937 (c) I revised the way the table is implemented to a much improved format
3938 which includes recognition of ranges. It now supports the ranges that
3939 are defined in UnicodeData.txt, and it also amalgamates other
3940 characters into ranges. This has reduced the number of entries in the
3941 table from around 16,000 to around 3,000, thus reducing its size
3942 considerably. I realized I did not need to use a tree structure after
3943 all - a binary chop search is just as efficient. Having reduced the
3944 number of entries, I extended their size from 6 bytes to 8 bytes to
3945 allow for more data.
3947 (d) Added support for Unicode script names via properties such as \p{Han}.
3949 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3950 matching that character.
3952 20. When matching a repeated Unicode property with a minimum greater than zero,
3953 (for example \pL{2,}), PCRE could look past the end of the subject if it
3954 reached it while seeking the minimum number of characters. This could
3955 happen only if some of the characters were more than one byte long, because
3956 there is a check for at least the minimum number of bytes.
3958 21. Refactored the implementation of \p and \P so as to be more general, to
3959 allow for more different types of property in future. This has changed the
3960 compiled form incompatibly. Anybody with saved compiled patterns that use
3961 \p or \P will have to recompile them.
3963 22. Added "Any" and "L&" to the supported property types.
3965 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3966 but give a compile time error if the value is greater than 0xff.
3968 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3969 accidentally not being installed or uninstalled.
3971 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3972 made were to insert the current release number. This seemed silly, because
3973 it made things harder for people building PCRE on systems that don't run
3974 "configure". I have turned pcre.h into a distributed file, no longer built
3975 by "configure", with the version identification directly included. There is
3976 no longer a pcre.h.in file.
3978 However, this change necessitated a change to the pcre-config script as
3979 well. It is built from pcre-config.in, and one of the substitutions was the
3980 release number. I have updated configure.ac so that ./configure now finds
3981 the release number by grepping pcre.h.
3983 26. Added the ability to run the tests under valgrind.
3986 Version 6.4 05-Sep-05
3987 ---------------------
3989 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3990 "--" to be printed when multiple files were scanned, even when none of the
3991 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3992 consider it to be a bug, and have restored the previous behaviour.
3994 2. A couple of code tidies to get rid of compiler warnings.
3996 3. The pcretest program used to cheat by referring to symbols in the library
3997 whose names begin with _pcre_. These are internal symbols that are not
3998 really supposed to be visible externally, and in some environments it is
3999 possible to suppress them. The cheating is now confined to including
4000 certain files from the library's source, which is a bit cleaner.
4002 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4003 file's purpose clearer.
4005 5. Reorganized pcre_ucp_findchar().
4008 Version 6.3 15-Aug-05
4009 ---------------------
4011 1. The file libpcre.pc.in did not have general read permission in the tarball.
4013 2. There were some problems when building without C++ support:
4015 (a) If C++ support was not built, "make install" and "make test" still
4016 tried to test it.
4018 (b) There were problems when the value of CXX was explicitly set. Some
4019 changes have been made to try to fix these, and ...
4021 (c) --disable-cpp can now be used to explicitly disable C++ support.
4023 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4024 backslash in a target when C++ was disabled. This confuses some
4025 versions of "make", apparently. Using an intermediate variable solves
4026 this. (Same for CPP_LOBJ.)
4028 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4029 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4030 necessary on certain architectures.
4032 4. Added a setting of -export-symbols-regex to the link command to remove
4033 those symbols that are exported in the C sense, but actually are local
4034 within the library, and not documented. Their names all begin with
4035 "_pcre_". This is not a perfect job, because (a) we have to except some
4036 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4037 available (and never for static libraries). I have made a note to try to
4038 find a way round (a) in the future.
4041 Version 6.2 01-Aug-05
4042 ---------------------
4044 1. There was no test for integer overflow of quantifier values. A construction
4045 such as {1111111111111111} would give undefined results. What is worse, if
4046 a minimum quantifier for a parenthesized subpattern overflowed and became
4047 negative, the calculation of the memory size went wrong. This could have
4048 led to memory overwriting.
4050 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4052 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4053 operating environments where this matters.
4055 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4056 PCRE options from within the C++ wrapper.
4058 5. Named capturing subpatterns were not being correctly counted when a pattern
4059 was compiled. This caused two problems: (a) If there were more than 100
4060 such subpatterns, the calculation of the memory needed for the whole
4061 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4062 back references of the form \12, where the number was greater than 9, were
4063 not recognized as back references, even though there were sufficient
4064 previous subpatterns.
4066 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4067 versions of gcc, e.g. 2.95.4.
4070 Version 6.1 21-Jun-05
4071 ---------------------
4073 1. There was one reference to the variable "posix" in pcretest.c that was not
4074 surrounded by "#if !defined NOPOSIX".
4076 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4077 the cross-check on the old pcre_info() function, for the benefit of the
4078 cut-down version of PCRE that is currently imported into Exim.
4080 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4081 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4082 patterns that start (?i) but I don't think that matters. The use of (?i) is
4083 just an example; this all applies to the other options as well.
4085 4. Since libtool seems to echo the compile commands it is issuing, the output
4086 from "make" can be reduced a bit by putting "@" in front of each libtool
4087 compile command.
4089 5. Patch from the folks at Google for configure.in to be a bit more thorough
4090 in checking for a suitable C++ installation before trying to compile the
4091 C++ stuff. This should fix a reported problem when a compiler was present,
4092 but no suitable headers.
4094 6. The man pages all had just "PCRE" as their title. I have changed them to
4095 be the relevant file name. I have also arranged that these names are
4096 retained in the file doc/pcre.txt, which is a concatenation in text format
4097 of all the man pages except the little individual ones for each function.
4099 7. The NON-UNIX-USE file had not been updated for the different set of source
4100 files that come with release 6. I also added a few comments about the C++
4101 wrapper.
4104 Version 6.0 07-Jun-05
4105 ---------------------
4107 1. Some minor internal re-organization to help with my DFA experiments.
4109 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4110 didn't matter for the library itself when fully configured, but did matter
4111 when compiling without UCP support, or within Exim, where the ucp files are
4112 not imported.
4114 3. Refactoring of the library code to split up the various functions into
4115 different source modules. The addition of the new DFA matching code (see
4116 below) to a single monolithic source would have made it really too
4117 unwieldy, quite apart from causing all the code to be include in a
4118 statically linked application, when only some functions are used. This is
4119 relevant even without the DFA addition now that patterns can be compiled in
4120 one application and matched in another.
4122 The downside of splitting up is that there have to be some external
4123 functions and data tables that are used internally in different modules of
4124 the library but which are not part of the API. These have all had their
4125 names changed to start with "_pcre_" so that they are unlikely to clash
4126 with other external names.
4128 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4129 a different (DFA) algorithm. Although it is slower than the original
4130 function, it does have some advantages for certain types of matching
4131 problem.
4133 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4134 including restarting after a partial match.
4136 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4137 defined when compiling for Windows was sent to me. I have put it into the
4138 code, though I have no means of testing or verifying it.
4140 7. Added the pcre_refcount() auxiliary function.
4142 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4143 match before or at the first newline in the subject string. In pcretest,
4144 the /f option on a pattern can be used to set this.
4146 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4147 would behave wrongly. This has been present in PCRE since release 4.0.
4149 10. A number of changes to the pcregrep command:
4151 (a) Refactored how -x works; insert ^(...)$ instead of setting
4152 PCRE_ANCHORED and checking the length, in preparation for adding
4153 something similar for -w.
4155 (b) Added the -w (match as a word) option.
4157 (c) Refactored the way lines are read and buffered so as to have more
4158 than one at a time available.
4160 (d) Implemented a pcregrep test script.
4162 (e) Added the -M (multiline match) option. This allows patterns to match
4163 over several lines of the subject. The buffering ensures that at least
4164 8K, or the rest of the document (whichever is the shorter) is available
4165 for matching (and similarly the previous 8K for lookbehind assertions).
4167 (f) Changed the --help output so that it now says
4169 -w, --word-regex(p)
4171 instead of two lines, one with "regex" and the other with "regexp"
4172 because that confused at least one person since the short forms are the
4173 same. (This required a bit of code, as the output is generated
4174 automatically from a table. It wasn't just a text change.)
4176 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4177 option but starts with a hyphen. Could be a pattern or a path name
4178 starting with a hyphen, for instance.
4180 (h) "-" can be given as a file name to represent stdin.
4182 (i) When file names are being printed, "(standard input)" is used for
4183 the standard input, for compatibility with GNU grep. Previously
4184 "<stdin>" was used.
4186 (j) The option --label=xxx can be used to supply a name to be used for
4187 stdin when file names are being printed. There is no short form.
4189 (k) Re-factored the options decoding logic because we are going to add
4190 two more options that take data. Such options can now be given in four
4191 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4193 (l) Added the -A, -B, and -C options for requesting that lines of context
4194 around matches be printed.
4196 (m) Added the -L option to print the names of files that do not contain
4197 any matching lines, that is, the complement of -l.
4199 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4200 continue to scan other files.
4202 (o) The -s option was incorrectly implemented. For compatibility with other
4203 greps, it now suppresses the error message for a non-existent or non-
4204 accessible file (but not the return code). There is a new option called
4205 -q that suppresses the output of matching lines, which was what -s was
4206 previously doing.
4208 (p) Added --include and --exclude options to specify files for inclusion
4209 and exclusion when recursing.
4211 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4212 Hopefully, it now does.
4214 12. Missing cast in pcre_study().
4216 13. Added an "uninstall" target to the makefile.
4218 14. Replaced "extern" in the function prototypes in Makefile.in with
4219 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4220 world, but is set differently for Windows.
4222 15. Added a second compiling function called pcre_compile2(). The only
4223 difference is that it has an extra argument, which is a pointer to an
4224 integer error code. When there is a compile-time failure, this is set
4225 non-zero, in addition to the error test pointer being set to point to an
4226 error message. The new argument may be NULL if no error number is required
4227 (but then you may as well call pcre_compile(), which is now just a
4228 wrapper). This facility is provided because some applications need a
4229 numeric error indication, but it has also enabled me to tidy up the way
4230 compile-time errors are handled in the POSIX wrapper.
4232 16. Added VPATH=.libs to the makefile; this should help when building with one
4233 prefix path and installing with another. (Or so I'm told by someone who
4234 knows more about this stuff than I do.)
4236 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4237 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4238 match everything, including newlines. This is not POSIX-compatible, but
4239 somebody wanted the feature. From pcretest it can be activated by using
4240 both the P and the s flags.
4242 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4244 19. libpcre.pc was being incorrectly installed as executable.
4246 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4247 it now also looks for '\r' so that it will work unmodified on Windows.
4249 21. Added Google's contributed C++ wrapper to the distribution.
4251 22. Added some untidy missing memory free() calls in pcretest, to keep
4252 Electric Fence happy when testing.
4256 Version 5.0 13-Sep-04
4257 ---------------------
4259 1. Internal change: literal characters are no longer packed up into items
4260 containing multiple characters in a single byte-string. Each character
4261 is now matched using a separate opcode. However, there may be more than one
4262 byte in the character in UTF-8 mode.
4264 2. The pcre_callout_block structure has two new fields: pattern_position and
4265 next_item_length. These contain the offset in the pattern to the next match
4266 item, and its length, respectively.
4268 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4269 insertion of callouts before each pattern item. Added the /C option to
4270 pcretest to make use of this.
4272 4. On the advice of a Windows user, the lines
4274 #if defined(_WIN32) || defined(WIN32)
4275 _setmode( _fileno( stdout ), 0x8000 );
4276 #endif /* defined(_WIN32) || defined(WIN32) */
4278 have been added to the source of pcretest. This apparently does useful
4279 magic in relation to line terminators.
4281 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4282 for the benefit of those environments where the "b" makes a difference.
4284 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4285 to know about it. I have put a hack into configure.in that adds in code
4286 to set GCC=yes if CC=icc. This seems to end up at a point in the
4287 generated configure script that is early enough to affect the setting of
4288 compiler options, which is what is needed, but I have no means of testing
4289 whether it really works. (The user who reported this had patched the
4290 generated configure script, which of course I cannot do.)
4292 LATER: After change 22 below (new libtool files), the configure script
4293 seems to know about icc (and also ecc). Therefore, I have commented out
4294 this hack in configure.in.
4296 7. Added support for pkg-config (2 patches were sent in).
4298 8. Negated POSIX character classes that used a combination of internal tables
4299 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4300 [[:^ascii]]. Typically, they would match almost any characters. The other
4301 POSIX classes were not broken in this way.
4303 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4304 to find the match, as PCRE was deluded into thinking that the match had to
4305 start at the start point or following a newline. The same bug applied to
4306 patterns with negative forward assertions or any backward assertions
4307 preceding ".*" at the start, unless the pattern required a fixed first
4308 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4310 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4311 starting at the last subject character, bytes beyond the end of the subject
4312 string were read.
4314 11. Renamed the variable "class" as "classbits" to make life easier for C++
4315 users. (Previously there was a macro definition, but it apparently wasn't
4316 enough.)
4318 12. Added the new field "tables" to the extra data so that tables can be passed
4319 in at exec time, or the internal tables can be re-selected. This allows
4320 a compiled regex to be saved and re-used at a later time by a different
4321 program that might have everything at different addresses.
4323 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4324 -R library as well as a -L library.
4326 14. The debugging options of pcretest (-d on the command line or D on a
4327 pattern) showed incorrect output for anything following an extended class
4328 that contained multibyte characters and which was followed by a quantifier.
4330 15. Added optional support for general category Unicode character properties
4331 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4332 support. It adds about 90K to the size of the library. The meanings of the
4333 inbuilt class escapes such as \d and \s have NOT been changed.
4335 16. Updated pcredemo.c to include calls to free() to release the memory for the
4336 compiled pattern.
4338 17. The generated file chartables.c was being created in the source directory
4339 instead of in the building directory. This caused the build to fail if the
4340 source directory was different from the building directory, and was
4341 read-only.
4343 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4344 file. No doubt somebody will tell me if they don't make sense... Also added
4345 Dan Mooney's comments about building on OpenVMS.
4347 19. Added support for partial matching via the PCRE_PARTIAL option for
4348 pcre_exec() and the \P data escape in pcretest.
4350 20. Extended pcretest with 3 new pattern features:
4352 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4353 write the compiled pattern to the file whose name is "rest-of-line".
4354 This is a straight binary dump of the data, with the saved pointer to
4355 the character tables forced to be NULL. The study data, if any, is
4356 written too. After writing, pcretest reads a new pattern.
4358 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4359 compiled pattern from the given file. There must not be any
4360 occurrences of "<" in the file name (pretty unlikely); if there are,
4361 pcretest will instead treat the initial "<" as a pattern delimiter.
4362 After reading in the pattern, pcretest goes on to read data lines as
4363 usual.
4365 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4366 and 16-bit fields in a compiled pattern, to simulate a pattern that
4367 was compiled on a host of opposite endianness.
4369 21. The pcre-exec() function can now cope with patterns that were compiled on
4370 hosts of opposite endianness, with this restriction:
4372 As for any compiled expression that is saved and used later, the tables
4373 pointer field cannot be preserved; the extra_data field in the arguments
4374 to pcre_exec() should be used to pass in a tables address if a value
4375 other than the default internal tables were used at compile time.
4377 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4378 now diagnosed as an error. Previously, most of the time, a negative number
4379 would have been treated as zero, but if in addition "ovector" was passed as
4380 NULL, a crash could occur.
4382 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4383 new versions from the libtool 1.5 distribution (the last one is a copy of
4384 a file called libtool.m4). This seems to have fixed the need to patch
4385 "configure" to support Darwin 1.3 (which I used to do). However, I still
4386 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4387 workstation).
4389 24. Changed the PCRE licence to be the more standard "BSD" licence.
4392 Version 4.5 01-Dec-03
4393 ---------------------
4395 1. There has been some re-arrangement of the code for the match() function so
4396 that it can be compiled in a version that does not call itself recursively.
4397 Instead, it keeps those local variables that need separate instances for
4398 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4399 needs to "recurse". Keeping track of where control must go is done by means
4400 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4401 hide most of the details from the main code, and operates only if
4402 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4403 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4404 operating.
4406 To make it easier for callers to provide specially tailored get/free
4407 functions for this usage, two new functions, pcre_stack_malloc, and
4408 pcre_stack_free, are used. They are always called in strict stacking order,
4409 and the size of block requested is always the same.
4411 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4412 PCRE has been compiled to use the stack or the heap for recursion. The
4413 -C option of pcretest uses this to show which version is compiled.
4415 A new data escape \S, is added to pcretest; it causes the amounts of store
4416 obtained and freed by both kinds of malloc/free at match time to be added
4417 to the output.
4419 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4420 what's available on my current Linux desktop machine.
4422 3. When matching a UTF-8 string, the test for a valid string at the start has
4423 been extended. If start_offset is not zero, PCRE now checks that it points
4424 to a byte that is the start of a UTF-8 character. If not, it returns
4425 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4426 this is necessary because there may be backward assertions in the pattern.
4427 When matching the same subject several times, it may save resources to use
4428 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4430 4. The code for checking the validity of UTF-8 strings has been tightened so
4431 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4432 containing "overlong sequences".
4434 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4435 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4436 should have been "&", but it just so happened that all the cases this let
4437 through by mistake were picked up later in the function.
4439 6. I had used a variable called "isblank" - this is a C99 function, causing
4440 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4442 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4443 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4444 so the version is shown; (c) stop "make test" echoing "./RunTest".
4446 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4448 9. The prototype for memmove() for systems that don't have it was using
4449 size_t, but the inclusion of the header that defines size_t was later. I've
4450 moved the #includes for the C headers earlier to avoid this.
4452 10. Added some adjustments to the code to make it easier to compiler on certain
4453 special systems:
4455 (a) Some "const" qualifiers were missing.
4456 (b) Added the macro EXPORT before all exported functions; by default this
4457 is defined to be empty.
4458 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4459 that it reads its output file name as an argument instead of writing
4460 to the standard output and assuming this can be redirected.
4462 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4463 class containing characters with values greater than 255, PCRE compilation
4464 went into a loop.
4466 12. A recursive reference to a subpattern that was within another subpattern
4467 that had a minimum quantifier of zero caused PCRE to crash. For example,
4468 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4469 recursion. If the recursively-called subpattern itself had a zero repeat,
4470 that was OK.
4472 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4473 buffer into which it was copied (for escape processing) was still set at
4474 1024, so long lines caused crashes.
4476 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4477 "internal error: code overflow...". This applied to any character class
4478 that was followed by a possessive quantifier.
4480 15. Modified the Makefile to add libpcre.la as a prerequisite for
4481 libpcreposix.la because I was told this is needed for a parallel build to
4482 work.
4484 16. If a pattern that contained .* following optional items at the start was
4485 studied, the wrong optimizing data was generated, leading to matching
4486 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4487 matching string must start with a or b or c. The correct conclusion for
4488 this pattern is that a match can start with any character.
4491 Version 4.4 13-Aug-03
4492 ---------------------
4494 1. In UTF-8 mode, a character class containing characters with values between
4495 127 and 255 was not handled correctly if the compiled pattern was studied.
4496 In fixing this, I have also improved the studying algorithm for such
4497 classes (slightly).
4499 2. Three internal functions had redundant arguments passed to them. Removal
4500 might give a very teeny performance improvement.
4502 3. Documentation bug: the value of the capture_top field in a callout is *one
4503 more than* the number of the hightest numbered captured substring.
4505 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4506 in incorrectly linking with a previously installed version. They now link
4507 explicitly with libpcre.la.
4509 5. configure.in no longer needs to recognize Cygwin specially.
4511 6. A problem in pcre.in for Windows platforms is fixed.
4513 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4514 pcretest, it used to include the size of the study block as part of its
4515 output. Unfortunately, the structure contains a field that has a different
4516 size on different hardware architectures. This meant that the tests that
4517 showed this size failed. As the block is currently always of a fixed size,
4518 this information isn't actually particularly useful in pcretest output, so
4519 I have just removed it.
4521 8. Three pre-processor statements accidentally did not start in column 1.
4522 Sadly, there are *still* compilers around that complain, even though
4523 standard C has not required this for well over a decade. Sigh.
4525 9. In pcretest, the code for checking callouts passed small integers in the
4526 callout_data field, which is a void * field. However, some picky compilers
4527 complained about the casts involved for this on 64-bit systems. Now
4528 pcretest passes the address of the small integer instead, which should get
4529 rid of the warnings.
4531 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4532 both compile and run time, and gives an error if an invalid UTF-8 sequence
4533 is found. There is a option for disabling this check in cases where the
4534 string is known to be correct and/or the maximum performance is wanted.
4536 11. In response to a bug report, I changed one line in Makefile.in from
4538 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4539 to
4540 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4542 to look similar to other lines, but I have no way of telling whether this
4543 is the right thing to do, as I do not use Windows. No doubt I'll get told
4544 if it's wrong...
4547 Version 4.3 21-May-03
4548 ---------------------
4550 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4551 Makefile.
4553 2. Some refactoring to improve the quality of the code:
4555 (i) The utf8_table... variables are now declared "const".
4557 (ii) The code for \cx, which used the "case flipping" table to upper case
4558 lower case letters, now just substracts 32. This is ASCII-specific,
4559 but the whole concept of \cx is ASCII-specific, so it seems
4560 reasonable.
4562 (iii) PCRE was using its character types table to recognize decimal and
4563 hexadecimal digits in the pattern. This is silly, because it handles
4564 only 0-9, a-f, and A-F, but the character types table is locale-
4565 specific, which means strange things might happen. A private
4566 table is now used for this - though it costs 256 bytes, a table is
4567 much faster than multiple explicit tests. Of course, the standard
4568 character types table is still used for matching digits in subject
4569 strings against \d.
4571 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4572 ending in _t are). So I've renamed it as ESC_tee.
4574 3. The first argument for regexec() in the POSIX wrapper should have been
4575 defined as "const".
4577 4. Changed pcretest to use malloc() for its buffers so that they can be
4578 Electric Fenced for debugging.
4580 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4581 to read one or more bytes before the start of the subject string. Often this
4582 had no effect on PCRE's behaviour, but in some circumstances it could
4583 provoke a segmentation fault.
4585 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4586 to try to read one or more bytes before the start of the subject string.
4588 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4589 UTF-8 support could misbehave in various ways if the subject string
4590 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4591 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4592 back over UTF-8 characters.)
4595 Version 4.2 14-Apr-03
4596 ---------------------
4598 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4600 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4601 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4602 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4603 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4604 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4606 Note: automatic setting of the BUILD variables is not yet working
4607 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4608 compile-time but not at link-time
4609 [LINK]: use for linking executables only
4610 make different versions for Windows and non-Windows
4611 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4612 libraries
4613 [LINK_FOR_BUILD]: new variable
4614 [OBJEXT]: use throughout
4615 [EXEEXT]: use throughout
4616 <winshared>: new target
4617 <wininstall>: new target
4618 <dftables.o>: use native compiler
4619 <dftables>: use native linker
4620 <install>: handle Windows platform correctly
4621 <clean>: ditto
4622 <check>: ditto
4623 copy DLL to top builddir before testing
4625 As part of these changes, -no-undefined was removed again. This was reported
4626 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4627 in any case.
4629 3. Some tidies to get rid of compiler warnings:
4631 . In the match_data structure, match_limit was an unsigned long int, whereas
4632 match_call_count was an int. I've made them both unsigned long ints.
4634 . In pcretest the fact that a const uschar * doesn't automatically cast to
4635 a void * provoked a warning.
4637 . Turning on some more compiler warnings threw up some "shadow" variables
4638 and a few more missing casts.
4640 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4641 option, a class that contained a single character with a value between 128
4642 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4644 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4645 option, a class that contained several characters, but with at least one
4646 whose value was between 128 and 255 caused PCRE to crash.
4649 Version 4.1 12-Mar-03
4650 ---------------------
4652 1. Compiling with gcc -pedantic found a couple of places where casts were
4653 needed, and a string in dftables.c that was longer than standard compilers are
4654 required to support.
4656 2. Compiling with Sun's compiler found a few more places where the code could
4657 be tidied up in order to avoid warnings.
4659 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4660 first of these names is deprecated in the latest Autoconf in favour of the name
4661 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4662 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4663 analogy I have changed it to CFLAGS_FOR_BUILD.
4665 4. Added -no-undefined to the linking command in the Makefile, because this is
4666 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4667 linking step for the pcreposix library.
4669 5. PCRE was failing to diagnose the case of two named groups with the same
4670 name.
4672 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4673 literal character that is needed in the subject for a match, and scans along to
4674 ensure that it is present before embarking on the full matching process. This
4675 saves time in cases of nested unlimited repeats that are never going to match.
4676 Problem: the scan can take a lot of time if the subject is very long (e.g.
4677 megabytes), thus penalizing straightforward matches. It is now done only if the
4678 amount of subject to be scanned is less than 1000 bytes.
4680 7. A lesser problem with the same optimization is that it was recording the
4681 first character of an anchored pattern as "needed", thus provoking a search
4682 right along the subject, even when the first match of the pattern was going to
4683 fail. The "needed" character is now not set for anchored patterns, unless it
4684 follows something in the pattern that is of non-fixed length. Thus, it still
4685 fulfils its original purpose of finding quick non-matches in cases of nested
4686 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4689 Version 4.0 17-Feb-03
4690 ---------------------
4692 1. If a comment in an extended regex that started immediately after a meta-item
4693 extended to the end of string, PCRE compiled incorrect data. This could lead to
4694 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4696 2. Moved to autoconf 2.53 and libtool 1.4.2.
4698 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4699 the special perltest8 script is no longer needed - all the tests can be run
4700 from a single perltest script.
4702 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4703 by \s. It has now been removed in PCRE. This means it isn't recognized as
4704 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4705 class [:space:] *does* include VT, thereby creating a mess.
4707 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4708 space and tab.
4710 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4711 its new features into the main test script, reducing the number of scripts.
4713 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4714 were backward compatible, and made the (?i) apply to the whole pattern, as if
4715 /i were given. Now it behaves more logically, and applies the option setting
4716 only to what follows. PCRE has been changed to follow suit. However, if it
4717 finds options settings right at the start of the pattern, it extracts them into
4718 the global options, as before. Thus, they show up in the info data.
4720 8. Added support for the \Q...\E escape sequence. Characters in between are
4721 treated as literals. This is slightly different from Perl in that $ and @ are
4722 also handled as literals inside the quotes. In Perl, they will cause variable
4723 interpolation. Note the following examples:
4725 Pattern PCRE matches Perl matches
4727 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4728 \Qabc\$xyz\E abc\$xyz abc\$xyz
4729 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4731 For compatibility with Perl, \Q...\E sequences are recognized inside character
4732 classes as well as outside them.
4734 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4735 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4736 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4737 signed/unsigned warnings.
4739 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4740 option for pcretest, so I've replaced it by a simple function that does just
4741 that job.
4743 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4744 "pcregrep -".
4746 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4747 Java package. This provides some syntactic sugar for simple cases of what my
4748 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4749 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4750 item, you can use this simplified notation. Note that only makes sense with
4751 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4752 greediness, whatever the setting of the PCRE_UNGREEDY option.
4754 13. A change of greediness default within a pattern was not taking effect at
4755 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4756 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4757 was abstracted outside.
4759 14. PCRE now supports the \G assertion. It is true when the current matching
4760 position is at the start point of the match. This differs from \A when the
4761 starting offset is non-zero. Used with the /g option of pcretest (or similar
4762 code), it works in the same way as it does for Perl's /g option. If all
4763 alternatives of a regex begin with \G, the expression is anchored to the start
4764 match position, and the "anchored" flag is set in the compiled expression.
4766 15. Some bugs concerning the handling of certain option changes within patterns
4767 have been fixed. These applied to options other than (?ims). For example,
4768 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4769 been the other way round. Some of this was related to change 7 above.
4771 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4772 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4773 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4774 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4776 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4777 mode. Unlike ".", it always matches newline, whatever the setting of
4778 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4779 assertions. Perl allows it, but it doesn't (in general) work because it can't
4780 calculate the length of the lookbehind. At least, that's the case for Perl
4781 5.8.0 - I've been told they are going to document that it doesn't work in
4782 future.
4784 18. Added an error diagnosis for escapes that PCRE does not support: these are
4785 \L, \l, \N, \P, \p, \U, \u, and \X.
4787 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4788 reading past the end of the pattern in cases such as /[abcd/.
4790 20. PCRE was getting more memory than necessary for patterns with classes that
4791 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4793 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4794 compiling PCRE for use with Virtual Pascal.
4796 22. Small fix to the Makefile to make it work properly if the build is done
4797 outside the source tree.
4799 23. Added a new extension: a condition to go with recursion. If a conditional
4800 subpattern starts with (?(R) the "true" branch is used if recursion has
4801 happened, whereas the "false" branch is used only at the top level.
4803 24. When there was a very long string of literal characters (over 255 bytes
4804 without UTF support, over 250 bytes with UTF support), the computation of how
4805 much memory was required could be incorrect, leading to segfaults or other
4806 strange effects.
4808 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4809 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4810 there was a subsequent back reference to those brackets. This meant that, for
4811 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4812 possible to check for precisely this case. All we can do is abandon the
4813 optimization if .* occurs inside capturing brackets when there are any back
4814 references whatsoever. (See below for a better fix that came later.)
4816 26. The handling of the optimization for finding the first character of a
4817 non-anchored pattern, and for finding a character that is required later in the
4818 match were failing in some cases. This didn't break the matching; it just
4819 failed to optimize when it could. The way this is done has been re-implemented.
4821 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4823 28. Added a new feature that provides some of the functionality that Perl
4824 provides with (?{...}). The facility is termed a "callout". The way it is done
4825 in PCRE is for the caller to provide an optional function, by setting
4826 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4827 global variable. By default it is unset, which disables all calling out. To get
4828 the function called, the regex must include (?C) at appropriate points. This
4829 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4830 This provides a means of identifying different callout points. When PCRE
4831 reaches such a point in the regex, if pcre_callout has been set, the external
4832 function is called. It is provided with data in a structure called
4833 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4834 matching continues; if it returns a non-zero value, the match at the current
4835 point fails. However, backtracking will occur if possible. [This was changed
4836 later and other features added - see item 49 below.]
4838 29. pcretest is upgraded to test the callout functionality. It provides a
4839 callout function that displays information. By default, it shows the start of
4840 the match and the current position in the text. There are some new data escapes
4841 to vary what happens:
4843 \C+ in addition, show current contents of captured substrings
4844 \C- do not supply a callout function
4845 \C!n return 1 when callout number n is reached
4846 \C!n!m return 1 when callout number n is reached for the mth time
4848 30. If pcregrep was called with the -l option and just a single file name, it
4849 output "<stdin>" if a match was found, instead of the file name.
4851 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4852 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4853 pcre_exec(). This saves a malloc/free per call. The default value of
4854 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4855 when configuring.
4857 32. The default maximum size of a compiled pattern is 64K. There have been a
4858 few cases of people hitting this limit. The code now uses macros to handle the
4859 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4860 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4861 configuring. Tests 2 and 5 work only with 2-byte links because they output
4862 debugging information about compiled patterns.
4864 33. Internal code re-arrangements:
4866 (a) Moved the debugging function for printing out a compiled regex into
4867 its own source file (printint.c) and used #include to pull it into
4868 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4869 separate copies.
4871 (b) Defined the list of op-code names for debugging as a macro in
4872 internal.h so that it is next to the definition of the opcodes.
4874 (c) Defined a table of op-code lengths for simpler skipping along compiled
4875 code. This is again a macro in internal.h so that it is next to the
4876 definition of the opcodes.
4878 34. Added support for recursive calls to individual subpatterns, along the
4879 lines of Robin Houston's patch (but implemented somewhat differently).
4881 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4882 allow it to read and process whole directories in Win32. This code was
4883 contributed by Lionel Fourquaux; it has not been tested by me.
4885 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4886 used to name a group. Names consist of alphanumerics and underscores, and must
4887 be unique. Back references use the syntax (?P=name) and recursive calls use
4888 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4889 numbers. The function pcre_fullinfo() can be used after compilation to extract
4890 a name/number map. There are three relevant calls:
4892 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4893 PCRE_INFO_NAMECOUNT yields the number of entries
4894 PCRE_INFO_NAMETABLE yields a pointer to the map.
4896 The map is a vector of fixed-size entries. The size of each entry depends on
4897 the length of the longest name used. The first two bytes of each entry are the
4898 group number, most significant byte first. There follows the corresponding
4899 name, zero terminated. The names are in alphabetical order.
4901 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4902 case instead of 255. Making it the same both with and without UTF-8 support
4903 means that the same test output works with both.
4905 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4906 calling malloc() with a zero argument.
4908 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4909 optimization. I've improved things by keeping a bitmap of backreferences with
4910 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4911 fact referenced, the optimization can be applied. It is unlikely that a
4912 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4913 the match to follow \n) will appear inside brackets with a number greater than
4914 31, but if it does, any back reference > 31 suppresses the optimization.
4916 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4917 of disabling numbered capturing parentheses. Any opening parenthesis that is
4918 not followed by ? behaves as if it were followed by ?: but named parentheses
4919 can still be used for capturing (and they will acquire numbers in the usual
4920 way).
4922 41. Redesigned the return codes from the match() function into yes/no/error so
4923 that errors can be passed back from deep inside the nested calls. A malloc
4924 failure while inside a recursive subpattern call now causes the
4925 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4927 42. It is now possible to set a limit on the number of times the match()
4928 function is called in a call to pcre_exec(). This facility makes it possible to
4929 limit the amount of recursion and backtracking, though not in a directly
4930 obvious way, because the match() function is used in a number of different
4931 circumstances. The count starts from zero for each position in the subject
4932 string (for non-anchored patterns). The default limit is, for compatibility, a
4933 large number, namely 10 000 000. You can change this in two ways:
4935 (a) When configuring PCRE before making, you can use --with-match-limit=n
4936 to set a default value for the compiled library.
4938 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4939 a different value is set. See 45 below.
4941 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4943 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4944 of things that can be changed at compile time. The first argument specifies
4945 what is wanted and the second points to where the information is to be placed.
4946 The current list of available information is:
4950 The output is an integer that is set to one if UTF-8 support is available;
4951 otherwise it is set to zero.
4955 The output is an integer that it set to the value of the code that is used for
4956 newline. It is either LF (10) or CR (13).
4960 The output is an integer that contains the number of bytes used for internal
4961 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4965 The output is an integer that contains the threshold above which the POSIX
4966 interface uses malloc() for output vectors. See item 31 above.
4970 The output is an unsigned integer that contains the default limit of the number
4971 of match() calls in a pcre_exec() execution. See 42 above.
4973 44. pcretest has been upgraded by the addition of the -C option. This causes it
4974 to extract all the available output from the new pcre_config() function, and to
4975 output it. The program then exits immediately.
4977 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4978 order to support additional features. One way would have been to define
4979 pcre_exec2() (for example) with extra arguments, but this would not have been
4980 extensible, and would also have required all calls to the original function to
4981 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4982 is used for passing in "extra" data from pcre_study().
4984 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4985 contains the following fields:
4987 flags a bitmap indicating which of the following fields are set
4988 study_data opaque data from pcre_study()
4989 match_limit a way of specifying a limit on match() calls for a specific
4990 call to pcre_exec()
4991 callout_data data for callouts (see 49 below)
4993 The flag bits are also defined in pcre.h, and are
4999 The pcre_study() function now returns one of these new pcre_extra blocks, with
5000 the actual study data pointed to by the study_data field, and the
5001 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5002 before. That is, this change is entirely upwards-compatible and requires no
5003 change to existing code.
5005 If you want to pass in additional data to pcre_exec(), you can either place it
5006 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5007 block.
5009 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5010 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5011 times with different match limits, until it finds the minimum value needed for
5012 pcre_exec() to complete. The value is then output. This can be instructive; for
5013 most simple matches the number is quite small, but for pathological cases it
5014 gets very large very quickly.
5016 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5017 returns the size of the data block pointed to by the study_data field in a
5018 pcre_extra block, that is, the value that was passed as the argument to
5019 pcre_malloc() when PCRE was getting memory in which to place the information
5020 created by pcre_study(). The fourth argument should point to a size_t variable.
5021 pcretest has been extended so that this information is shown after a successful
5022 pcre_study() call when information about the compiled regex is being displayed.
5024 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5025 because what follows is always an absolute path. (Later: it turns out that this
5026 is more than cosmetic for MinGW, because it doesn't like empty path
5027 components.)
5029 49. Some changes have been made to the callout feature (see 28 above):
5031 (i) A callout function now has three choices for what it returns:
5033 0 => success, carry on matching
5034 > 0 => failure at this point, but backtrack if possible
5035 < 0 => serious error, return this value from pcre_exec()
5037 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5038 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5039 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5040 use by callout functions. It will never be used by PCRE itself.
5042 (ii) The pcre_extra structure (see 45 above) has a void * field called
5043 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5044 pcre_callout_block structure has a field of the same name. The contents of
5045 the field passed in the pcre_extra structure are passed to the callout
5046 function in the corresponding field in the callout block. This makes it
5047 easier to use the same callout-containing regex from multiple threads. For
5048 testing, the pcretest program has a new data escape
5050 \C*n pass the number n (may be negative) as callout_data
5052 If the callout function in pcretest receives a non-zero value as
5053 callout_data, it returns that value.
5055 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5056 there were some redundant $(CFLAGS) in commands that are now specified as
5057 $(LINK), which already includes $(CFLAGS).
5059 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5060 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5061 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5062 one-byte characters throughout. Note that case-insensitive matching applies
5063 only to characters whose values are less than 256. PCRE doesn't support the
5064 notion of cases for higher-valued characters.
5066 (i) A character class whose characters are all within 0-255 is handled as
5067 a bit map, and the map is inverted for negative classes. Previously, a
5068 character > 255 always failed to match such a class; however it should
5069 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5071 (ii) A negated character class with a single character < 255 is coded as
5072 "not this character" (OP_NOT). This wasn't working properly when the test
5073 character was multibyte, either singly or repeated.
5075 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5076 mode, for example: \x{100}{2,3}.
5078 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5079 singly or repeated) now correctly test multibyte characters. However,
5080 PCRE doesn't recognize any characters with values greater than 255 as
5081 digits, spaces, or word characters. Such characters always match \D, \S,
5082 and \W, and never match \d, \s, or \w.
5084 (v) Classes may now contain characters and character ranges with values
5085 greater than 255. For example: [ab\x{100}-\x{400}].
5087 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5088 PCRE in UTF-8 mode.
5090 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5091 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5092 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5093 value.)
5095 53. The single man page has become too large. I have therefore split it up into
5096 a number of separate man pages. These also give rise to individual HTML pages;
5097 these are now put in a separate directory, and there is an index.html page that
5098 lists them all. Some hyperlinking between the pages has been installed.
5100 54. Added convenience functions for handling named capturing parentheses.
5102 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5103 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5104 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5105 are faulted.
5107 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5108 calling configure. These values are used when compiling the dftables.c program
5109 which is run to generate the source of the default character tables. They
5110 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5111 you will need to set these values.
5113 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5116 Version 3.9 02-Jan-02
5117 ---------------------
5119 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5121 2. If --disable-static was given, the building process failed when trying to
5122 build pcretest and pcregrep. (For some reason it was using libtool to compile
5123 them, which is not right, as they aren't part of the library.)
5126 Version 3.8 18-Dec-01
5127 ---------------------
5129 1. The experimental UTF-8 code was completely screwed up. It was packing the
5130 bytes in the wrong order. How dumb can you get?
5133 Version 3.7 29-Oct-01
5134 ---------------------
5136 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5137 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5138 this didn't happen under Solaris 8, where I normally test things.
5140 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5141 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5142 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5143 this fix an hour or so after the initial 3.7 release.)
5146 Version 3.6 23-Oct-01
5147 ---------------------
5149 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5150 offsets passed as NULL with zero offset count.
5152 2. The config.guess and config.sub files had not been updated when I moved to
5153 the latest autoconf.
5156 Version 3.5 15-Aug-01
5157 ---------------------
5159 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5160 had been forgotten.
5162 2. By using declared but undefined structures, we can avoid using "void"
5163 definitions in pcre.h while keeping the internal definitions of the structures
5164 private.
5166 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5167 user point of view, this means that both static and shared libraries are built
5168 by default, but this can be individually controlled. More of the work of
5169 handling this static/shared cases is now inside libtool instead of PCRE's make
5170 file.
5172 4. The pcretest utility is now installed along with pcregrep because it is
5173 useful for users (to test regexs) and by doing this, it automatically gets
5174 relinked by libtool. The documentation has been turned into a man page, so
5175 there are now .1, .txt, and .html versions in /doc.
5177 5. Upgrades to pcregrep:
5178 (i) Added long-form option names like gnu grep.
5179 (ii) Added --help to list all options with an explanatory phrase.
5180 (iii) Added -r, --recursive to recurse into sub-directories.
5181 (iv) Added -f, --file to read patterns from a file.
5183 6. pcre_exec() was referring to its "code" argument before testing that
5184 argument for NULL (and giving an error if it was NULL).
5186 7. Upgraded Makefile.in to allow for compiling in a different directory from
5187 the source directory.
5189 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5190 options bits, the pointer it was passed was to an int instead of to an unsigned
5191 long int. This mattered only on 64-bit systems.
5193 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5194 generated) instead of pcre.in, which it its source. Also made the same change
5195 in several of the .c files.
5197 10. A new release of gcc defines printf() as a macro, which broke pcretest
5198 because it had an ifdef in the middle of a string argument for printf(). Fixed
5199 by using separate calls to printf().
5201 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5202 script, to force use of CR or LF instead of \n in the source. On non-Unix
5203 systems, the value can be set in config.h.
5205 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5206 absolute limit. Changed the text of the error message to make this clear, and
5207 likewise updated the man page.
5209 13. The limit of 99 on the number of capturing subpatterns has been removed.
5210 The new limit is 65535, which I hope will not be a "real" limit.
5213 Version 3.4 22-Aug-00
5214 ---------------------
5216 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5218 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5221 Version 3.3 01-Aug-00
5222 ---------------------
5224 1. If an octal character was given, but the value was greater than \377, it
5225 was not getting masked to the least significant bits, as documented. This could
5226 lead to crashes in some systems.
5228 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5229 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5231 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5232 These just pass their arguments on to (pcre_free)(), but they are provided
5233 because some uses of PCRE bind it to non-C systems that can call its functions,
5234 but cannot call free() or pcre_free() directly.
5236 4. Add "make test" as a synonym for "make check". Corrected some comments in
5237 the Makefile.
5239 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5240 Makefile.
5242 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5243 command called pgrep for grepping around the active processes.
5245 7. Added the beginnings of support for UTF-8 character strings.
5247 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5248 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5249 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5250 out for the ar command.)
5253 Version 3.2 12-May-00
5254 ---------------------
5256 This is purely a bug fixing release.
5258 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5259 of ZA. This was just one example of several cases that could provoke this bug,
5260 which was introduced by change 9 of version 2.00. The code for breaking
5261 infinite loops after an iteration that matches an empty string was't working
5262 correctly.
5264 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5265 when matched against abbab (for example). After matching an empty string, it
5266 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5267 caused it to match further down the string than it should.
5269 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5270 was there because it doesn't seem to be needed, and it causes trouble on some
5271 systems, as it is not a Standard C header. It has been removed.
5273 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5274 were reported on the Macintosh. The changes were from
5276 while ((c = *(++ptr)) != 0 && c != '\n');
5277 to
5278 while ((c = *(++ptr)) != 0 && c != '\n') ;
5280 Totally extraordinary, but if that's what it takes...
5282 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5283 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5284 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5285 assumes the way PCRE uses memmove() (always moving upwards).
5287 6. PCRE is being used in one environment where strchr() is not available. There
5288 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5289 faster code anyway.
5292 Version 3.1 09-Feb-00
5293 ---------------------
5295 The only change in this release is the fixing of some bugs in Makefile.in for
5296 the "install" target:
5298 (1) It was failing to install pcreposix.h.
5300 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5303 Version 3.0 01-Feb-00
5304 ---------------------
5306 1. Add support for the /+ modifier to perltest (to output $` like it does in
5307 pcretest).
5309 2. Add support for the /g modifier to perltest.
5311 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5312 matches null strings.
5314 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5315 pattern. Perl treats empty patterns specially - it reuses the most recent
5316 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5317 effect.
5319 5. The POSIX interface was broken in that it was just handing over the POSIX
5320 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5321 required a bigger vector, with some working space on the end. This means that
5322 the POSIX wrapper now has to get and free some memory, and copy the results.
5324 6. Added some simple autoconf support, placing the test data and the
5325 documentation in separate directories, re-organizing some of the
5326 information files, and making it build pcre-config (a GNU standard). Also added
5327 libtool support for building PCRE as a shared library, which is now the
5328 default.
5330 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5331 09 are not valid octal constants. Single digits will be used for minor values
5332 less than 10.
5334 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5335 existing programs that set these in the POSIX interface can use PCRE without
5336 modification.
5338 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5339 return all that pcre_info() returns, plus additional data. The pcre_info()
5340 function is retained for compatibility, but is considered to be obsolete.
5342 10. Added experimental recursion feature (?R) to handle one common case that
5343 Perl 5.6 will be able to do with (?p{...}).
5345 11. Added support for POSIX character classes like [:alpha:], which Perl is
5346 adopting.
5349 Version 2.08 31-Aug-99
5350 ----------------------
5352 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5353 trying to match at the startoffset position, but instead was moving forward to
5354 the next newline as if a previous match had failed.
5356 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5357 and could get into a loop if a null string was matched other than at the start
5358 of the subject.
5360 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5361 be distinguished at compile time, and for completeness also added PCRE_DATE.
5363 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5364 in GnuWin32 environments.
5367 Version 2.07 29-Jul-99
5368 ----------------------
5370 1. The documentation is now supplied in plain text form and HTML as well as in
5371 the form of man page sources.
5373 2. C++ compilers don't like assigning (void *) values to other pointer types.
5374 In particular this affects malloc(). Although there is no problem in Standard
5375 C, I've put in casts to keep C++ compilers happy.
5377 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5378 should be (const char *).
5380 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5381 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5382 However, I haven't made this a standard facility. The documentation doesn't
5383 mention it, and the Makefile doesn't support it.
5385 5. The Makefile now contains an "install" target, with editable destinations at
5386 the top of the file. The pcretest program is not installed.
5388 6. pgrep -V now gives the PCRE version number and date.
5390 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5391 causing the entire string to be ignored, instead of just the last character.
5393 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5394 non-matching string, it can take a very, very long time, even for strings of
5395 quite modest length, because of the nested recursion. PCRE now does better in
5396 some of these cases. It does this by remembering the last required literal
5397 character in the pattern, and pre-searching the subject to ensure it is present
5398 before running the real match. In other words, it applies a heuristic to detect
5399 some types of certain failure quickly, and in the above example, if presented
5400 with a string that has no trailing " it gives "no match" very quickly.
5402 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5403 other alternatives are tried instead.
5406 Version 2.06 09-Jun-99
5407 ----------------------
5409 1. Change pcretest's output for amount of store used to show just the code
5410 space, because the remainder (the data block) varies in size between 32-bit and
5411 64-bit systems.
5413 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5414 start matching at. This allows lookbehinds to work when searching for multiple
5415 occurrences in a string.
5417 3. Added additional options to pcretest for testing multiple occurrences:
5419 /+ outputs the rest of the string that follows a match
5420 /g loops for multiple occurrences, using the new startoffset argument
5421 /G loops for multiple occurrences by passing an incremented pointer
5423 4. PCRE wasn't doing the "first character" optimization for patterns starting
5424 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5425 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5426 the letter 'x'. On long subject strings, this gives a significant speed-up.
5429 Version 2.05 21-Apr-99
5430 ----------------------
5432 1. Changed the type of magic_number from int to long int so that it works
5433 properly on 16-bit systems.
5435 2. Fixed a bug which caused patterns starting with .* not to work correctly
5436 when the subject string contained newline characters. PCRE was assuming
5437 anchoring for such patterns in all cases, which is not correct because .* will
5438 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5439 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5440 must be retried after every newline in the subject.
5443 Version 2.04 18-Feb-99
5444 ----------------------
5446 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5447 computation of the store needed to hold the pattern was incorrect (too large).
5448 If such patterns were nested a few deep, this could multiply and become a real
5449 problem.
5451 2. Added /M option to pcretest to show the memory requirement of a specific
5452 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5454 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5455 compiled in such a way that the backtracking after subsequent failure was
5456 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5457 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5460 Version 2.03 02-Feb-99
5461 ----------------------
5463 1. Fixed typo and small mistake in man page.
5465 2. Added 4th condition (GPL supersedes if conflict) and created separate
5466 LICENCE file containing the conditions.
5468 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5469 Perl, that is the internal \ allows the delimiter to be included in the
5470 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5471 the final delimiter, add \ to the end of the pattern (to test the error).
5473 4. Added the convenience functions for extracting substrings after a successful
5474 match. Updated pcretest to make it able to test these functions.
5477 Version 2.02 14-Jan-99
5478 ----------------------
5480 1. Initialized the working variables associated with each extraction so that
5481 their saving and restoring doesn't refer to uninitialized store.
5483 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5484 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5485 fix the problem.
5487 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5488 calls, and wasn't printing the correct value for compiling calls. Increased the
5489 default value of LOOPREPEAT, and the number of significant figures in the
5490 times.
5492 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5494 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5495 a building problem on Windows NT with a FAT file system.
5498 Version 2.01 21-Oct-98
5499 ----------------------
5501 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5502 to character tables built by pcre_maketables() in the current locale. If NULL
5503 is passed, the default tables are used.
5506 Version 2.00 24-Sep-98
5507 ----------------------
5509 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5510 it any more.
5512 2. Allow quantification of (?>) groups, and make it work correctly.
5514 3. The first character computation wasn't working for (?>) groups.
5516 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5517 end of the subject) and add 5.005's \z, which really does match only at the
5518 very end of the subject.
5520 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5522 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5523 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5524 localized options. All options to pcre_study() were also removed.
5526 7. Add other new features from 5.005:
5528 $(?<= positive lookbehind
5529 $(?<! negative lookbehind
5530 (?imsx-imsx) added the unsetting capability
5531 such a setting is global if at outer level; local otherwise
5532 (?imsx-imsx:) non-capturing groups with option setting
5533 (?(cond)re|re) conditional pattern matching
5535 A backreference to itself in a repeated group matches the previous
5536 captured string.
5538 8. General tidying up of studying (both automatic and via "study")
5539 consequential on the addition of new assertions.
5541 9. As in 5.005, unlimited repeated groups that could match an empty substring
5542 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5543 runtime if any iteration does actually match an empty substring.
5545 10. Include the RunTest script in the distribution.
5547 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5548 discrepancies, some of which were old and were also with respect to 5.004. They
5549 have now been fixed.
5552 Version 1.09 28-Apr-98
5553 ----------------------
5555 1. A negated single character class followed by a quantifier with a minimum
5556 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5557 program crashes, or just wrong answers. This did not apply to negated classes
5558 containing more than one character, or to minima other than one.
5561 Version 1.08 27-Mar-98
5562 ----------------------
5564 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5566 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5567 latter must appear before anything that relies on it in the pattern.
5570 Version 1.07 16-Feb-98
5571 ----------------------
5573 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5574 repeat of a potentially empty string).
5577 Version 1.06 23-Jan-98
5578 ----------------------
5580 1. Added Markus Oberhumer's little patches for C++.
5582 2. Literal strings longer than 255 characters were broken.
5585 Version 1.05 23-Dec-97
5586 ----------------------
5588 1. Negated character classes containing more than one character were failing if
5589 PCRE_CASELESS was set at run time.
5592 Version 1.04 19-Dec-97
5593 ----------------------
5595 1. Corrected the man page, where some "const" qualifiers had been omitted.
5597 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5598 input syntax.
5600 3. Fixed memory leak which occurred when a regex with back references was
5601 matched with an offsets vector that wasn't big enough. The temporary memory
5602 that is used in this case wasn't being freed if the match failed.
5604 4. Tidied pcretest to ensure it frees memory that it gets.
5606 5. Temporary memory was being obtained in the case where the passed offsets
5607 vector was exactly big enough.
5609 6. Corrected definition of offsetof() from change 5 below.
5611 7. I had screwed up change 6 below and broken the rules for the use of
5612 setjmp(). Now fixed.
5615 Version 1.03 18-Dec-97
5616 ----------------------
5618 1. A erroneous regex with a missing opening parenthesis was correctly
5619 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5620 on some systems.
5622 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5623 it was reported that one broken compiler failed on the former because "code" is
5624 also an independent variable.
5626 3. The erroneous regex a[]b caused an array overrun reference.
5628 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5629 fail on data ending with that character. (It was going on too far, and checking
5630 the next character, typically a binary zero.) This was specific to the
5631 optimized code for single-character negative classes.
5633 5. Added a contributed patch from the TIN world which does the following:
5635 + Add an undef for memmove, in case the the system defines a macro for it.
5637 + Add a definition of offsetof(), in case there isn't one. (I don't know
5638 the reason behind this - offsetof() is part of the ANSI standard - but
5639 it does no harm).
5641 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5642 most of the places where whitespace preceded '#'. I have given up and
5643 allowed the remaining 2 cases to be at the margin.
5645 + Rename some variables in pcre to eliminate shadowing. This seems very
5646 pedantic, but does no harm, of course.
5648 6. Moved the call to setjmp() into its own function, to get rid of warnings
5649 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5651 7. Constructs such as \d{8,} were compiling into the equivalent of
5652 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5653 outcome, but in this particular case used more store than had been allocated,
5654 which caused the bug to be discovered because it threw up an internal error.
5656 8. The debugging code in both pcre and pcretest for outputting the compiled
5657 form of a regex was going wrong in the case of back references followed by
5658 curly-bracketed repeats.
5661 Version 1.02 12-Dec-97
5662 ----------------------
5664 1. Typos in pcre.3 and comments in the source fixed.
5666 2. Applied a contributed patch to get rid of places where it used to remove
5667 'const' from variables, and fixed some signed/unsigned and uninitialized
5668 variable warnings.
5670 3. Added the "runtest" target to Makefile.
5672 4. Set default compiler flag to -O2 rather than just -O.
5675 Version 1.01 19-Nov-97
5676 ----------------------
5678 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5679 like /([ab]*)*/, that is, for classes with more than one character in them.
5681 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5682 as /((?>a*))*/ (a PCRE_EXTRA facility).
5685 Version 1.00 18-Nov-97
5686 ----------------------
5688 1. Added compile-time macros to support systems such as SunOS4 which don't have
5689 memmove() or strerror() but have other things that can be used instead.
5691 2. Arranged that "make clean" removes the executables.
5694 Version 0.99 27-Oct-97
5695 ----------------------
5697 1. Fixed bug in code for optimizing classes with only one character. It was
5698 initializing a 32-byte map regardless, which could cause it to run off the end
5699 of the memory it had got.
5701 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5704 Version 0.98 22-Oct-97
5705 ----------------------
5707 1. Fixed bug in code for handling temporary memory usage when there are more
5708 back references than supplied space in the ovector. This could cause segfaults.
5711 Version 0.97 21-Oct-97
5712 ----------------------
5714 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5716 2. Optimized negated single characters not to use a bit map.
5718 3. Brought error texts together as macro definitions; clarified some of them;
5719 fixed one that was wrong - it said "range out of order" when it meant "invalid
5720 escape sequence".
5722 4. Changed some char * arguments to const char *.
5724 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5726 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5727 pcretest.
5730 Version 0.96 16-Oct-97
5731 ----------------------
5733 1. Added a simple "pgrep" utility to the distribution.
5735 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5736 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5737 where "ddd" means "one or more decimal digits".
5739 3. Fixed serious bug. If a pattern had a back reference, but the call to
5740 pcre_exec() didn't supply a large enough ovector to record the related
5741 identifying subpattern, the match always failed. PCRE now remembers the number
5742 of the largest back reference, and gets some temporary memory in which to save
5743 the offsets during matching if necessary, in order to ensure that
5744 backreferences always work.
5746 4. Increased the compatibility with Perl in a number of ways:
5748 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5749 to request this handling. The option can be set at compile or exec time.
5751 (b) $ matches before a terminating newline by default; an option
5752 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5753 mode). The option can be set at compile or exec time.
5755 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5756 the same as Perl's. If the decimal number it represents is less than 10
5757 or there aren't that many previous left capturing parentheses, an octal
5758 escape is read. Inside a character class, it's always an octal escape,
5759 even if it is a single digit.
5761 (d) An escaped but undefined alphabetic character is taken as a literal,
5762 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5763 escapes.
5765 (e) {0} is now permitted. (The previous item is removed from the compiled
5766 pattern).
5768 5. Changed all the names of code files so that the basic parts are no longer
5769 than 10 characters, and abolished the teeny "globals.c" file.
5771 6. Changed the handling of character classes; they are now done with a 32-byte
5772 bit map always.
5774 7. Added the -d and /D options to pcretest to make it possible to look at the
5775 internals of compilation without having to recompile pcre.
5778 Version 0.95 23-Sep-97
5779 ----------------------
5781 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5782 \x20 at the start of a run of normal characters. These were being treated as
5783 real characters, instead of the source characters being re-checked.
5786 Version 0.94 18-Sep-97
5787 ----------------------
5789 1. The functions are now thread-safe, with the caveat that the global variables
5790 containing pointers to malloc() and free() or alternative functions are the
5791 same for all threads.
5793 2. Get pcre_study() to generate a bitmap of initial characters for non-
5794 anchored patterns when this is possible, and use it if passed to pcre_exec().
5797 Version 0.93 15-Sep-97
5798 ----------------------
5800 1. /(b)|(:+)/ was computing an incorrect first character.
5802 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5803 but not actually doing anything yet.
5805 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5806 as Perl does (e.g. [-az] or [az-]).
5808 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5809 all possible positions.
5811 5. Split up into different modules to avoid including unneeded functions in a
5812 compiled binary. However, compile and exec are still in one module. The "study"
5813 function is split off.
5815 6. The character tables are now in a separate module whose source is generated
5816 by an auxiliary program - but can then be edited by hand if required. There are
5817 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5818 toupper() in the code.
5820 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5821 make them global. Abolish the function for setting them, as the caller can now
5822 set them directly.
5825 Version 0.92 11-Sep-97
5826 ----------------------
5828 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5829 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5831 2. Caseless matching was not working in character classes if the characters in
5832 the pattern were in upper case.
5834 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5836 4. Make PCRE_ANCHORED public and accept as a compile option.
5838 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5839 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5840 pass them.
5842 6. Give an error if bad option bits passed at compile or run time.
5844 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5845 pcretest to cause it to pass that flag.
5847 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5848 options, and the first character, if set.
5850 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5853 Version 0.91 10-Sep-97
5854 ----------------------
5856 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5857 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5859 2. PCRE was looping on encountering an indefinitely repeated back reference to
5860 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5861 Perl does - treats the match as successful.
5863 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5