/[pcre]/code/trunk/ChangeLog
ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1543 - (show annotations)
Tue Apr 7 15:52:11 2015 UTC (4 years, 6 months ago) by ph10
File size: 276332 byte(s)
Error occurred while calculating annotation data.
Fix pcregrep loop when \K is used in a lookbehind assertion.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-2015
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20 4. An attempt to do global matching in pcretest with a zero-length ovector
21 caused a crash.
22
23 5. Fixed a memory leak during matching that could occur for a subpattern
24 subroutine call (recursive or otherwise) if the number of captured groups
25 that had to be saved was greater than ten.
26
27 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
28 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30
31 7. A UTF pattern containing a "not" match of a non-ASCII character and a
32 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33
34 8. When a pattern is compiled, it remembers the highest back reference so that
35 when matching, if the ovector is too small, extra memory can be obtained to
36 use instead. A conditional subpattern whose condition is a check on a
37 capture having happened, such as, for example in the pattern
38 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39 setting the highest backreference number. This mattered only if pcre_exec()
40 was called with an ovector that was too small to hold the capture, and there
41 was no other kind of back reference (a situation which is probably quite
42 rare). The effect of the bug was that the condition was always treated as
43 FALSE when the capture could not be consulted, leading to a incorrect
44 behaviour by pcre2_match(). This bug has been fixed.
45
46 9. A reference to a duplicated named group (either a back reference or a test
47 for being set in a conditional) that occurred in a part of the pattern where
48 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49 to be incorrectly calculated, leading to overwriting.
50
51 10. A mutually recursive set of back references such as (\2)(\1) caused a
52 segfault at study time (while trying to find the minimum matching length).
53 The infinite loop is now broken (with the minimum length unset, that is,
54 zero).
55
56 11. If an assertion that was used as a condition was quantified with a minimum
57 of zero, matching went wrong. In particular, if the whole group had
58 unlimited repetition and could match an empty string, a segfault was
59 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60 assertions to be quantified, but not if they are being used as conditions,
61 so the above pattern is faulted by Perl. PCRE has now been changed so that
62 it also rejects such patterns.
63
64 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65 failed to allow the zero-repeat case if pcre2_exec() was called with an
66 ovector too small to capture the group.
67
68 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69 Red Hat Product Security:
70
71 (a) A crash if /K and /F were both set with the option to save the compiled
72 pattern.
73
74 (b) Another crash if the option to print captured substrings in a callout
75 was combined with setting a null ovector, for example \O\C+ as a subject
76 string.
77
78 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79 forward reference repeated a large (but limited) number of times within a
80 repeated outer group that has a zero minimum quantifier, caused incorrect
81 code to be compiled, leading to the error "internal error:
82 previously-checked referenced subpattern not found" when an incorrect
83 memory address was read. This bug was reported as "heap overflow",
84 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
85 CVE-2015-2325.
86
87 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
88 call within a group that also contained a recursive back reference caused
89 incorrect code to be compiled. This bug was reported as "heap overflow",
90 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
91 number CVE-2015-2326.
92
93 24. Computing the size of the JIT read-only data in advance has been a source
94 of various issues, and new ones are still appear unfortunately. To fix
95 existing and future issues, size computation is eliminated from the code,
96 and replaced by on-demand memory allocation.
97
98 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
99 adjacent to the end of the range, and the range contained characters with
100 more than one other case, caused incorrect behaviour when compiled in UTF
101 mode. In that example, the range a-j was left out of the class.
102
103 26. Fix JIT compilation of conditional blocks, which assertion
104 is converted to (*FAIL). E.g: /(?(?!))/.
105
106 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
107 discovered by the LLVM fuzzer.
108
109 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
110 when this assertion was used as a condition, for example (?(?!)a|b). In
111 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
112 error about an unsupported item.
113
114 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
115 possessification code could take exponential time to complete. A recursion
116 depth limit of 1000 has been imposed to limit the resources used by this
117 optimization.
118
119 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
120 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
121 because \S ensures they are all in the class. The code for doing this was
122 interacting badly with the code for computing the amount of space needed to
123 compile the pattern, leading to a buffer overflow. This bug was discovered
124 by the LLVM fuzzer.
125
126 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
127 other kinds of group caused stack overflow at compile time. This bug was
128 discovered by the LLVM fuzzer.
129
130 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
131 between a subroutine call and its quantifier was incorrectly compiled,
132 leading to buffer overflow or other errors. This bug was discovered by the
133 LLVM fuzzer.
134
135 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
136 assertion after (?(. The code was failing to check the character after
137 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
138 was discovered by the LLVM fuzzer.
139
140 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
141 a fixed maximum following a group that contains a subroutine reference was
142 incorrectly compiled and could trigger buffer overflow. This bug was
143 discovered by the LLVM fuzzer.
144
145 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
146 caused a stack overflow instead of the diagnosis of a non-fixed length
147 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
148
149 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
150 (e.g. /(?<=\Ka)/) could make pcregrep loop.
151
152
153 Version 8.36 26-September-2014
154 ------------------------------
155
156 1. Got rid of some compiler warnings in the C++ modules that were shown up by
157 -Wmissing-field-initializers and -Wunused-parameter.
158
159 2. The tests for quantifiers being too big (greater than 65535) were being
160 applied after reading the number, and stupidly assuming that integer
161 overflow would give a negative number. The tests are now applied as the
162 numbers are read.
163
164 3. Tidy code in pcre_exec.c where two branches that used to be different are
165 now the same.
166
167 4. The JIT compiler did not generate match limit checks for certain
168 bracketed expressions with quantifiers. This may lead to exponential
169 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
170 issue should be resolved now.
171
172 5. Fixed an issue, which occures when nested alternatives are optimized
173 with table jumps.
174
175 6. Inserted two casts and changed some ints to size_t in the light of some
176 reported 64-bit compiler warnings (Bugzilla 1477).
177
178 7. Fixed a bug concerned with zero-minimum possessive groups that could match
179 an empty string, which sometimes were behaving incorrectly in the
180 interpreter (though correctly in the JIT matcher). This pcretest input is
181 an example:
182
183 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
184 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
185
186 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
187 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
188 for an empty string was breaking the inner loop and carrying on at a lower
189 level, when possessive repeated groups should always return to a higher
190 level as they have no backtrack points in them. The empty string test now
191 occurs at the outer level.
192
193 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
194 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
195
196 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
197 doesn't).
198
199 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
200 to hold a pattern to be incorrectly computed (too small) when there were
201 named back references to duplicated names. This could cause "internal
202 error: code overflow" or "double free or corruption" or other memory
203 handling errors.
204
205 11. When named subpatterns had the same prefixes, back references could be
206 confused. For example, in this pattern:
207
208 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
209
210 the reference to 'Name' was incorrectly treated as a reference to a
211 duplicate name.
212
213 12. A pattern such as /^s?c/mi8 where the optional character has more than
214 one "other case" was incorrectly compiled such that it would only try to
215 match starting at "c".
216
217 13. When a pattern starting with \s was studied, VT was not included in the
218 list of possible starting characters; this should have been part of the
219 8.34/18 patch.
220
221 14. If a character class started [\Qx]... where x is any character, the class
222 was incorrectly terminated at the ].
223
224 15. If a pattern that started with a caseless match for a character with more
225 than one "other case" was studied, PCRE did not set up the starting code
226 unit bit map for the list of possible characters. Now it does. This is an
227 optimization improvement, not a bug fix.
228
229 16. The Unicode data tables have been updated to Unicode 7.0.0.
230
231 17. Fixed a number of memory leaks in pcregrep.
232
233 18. Avoid a compiler warning (from some compilers) for a function call with
234 a cast that removes "const" from an lvalue by using an intermediate
235 variable (to which the compiler does not object).
236
237 19. Incorrect code was compiled if a group that contained an internal recursive
238 back reference was optional (had quantifier with a minimum of zero). This
239 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
240 caused segmentation faults because of stack overflows at compile time.
241
242 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
243 group that is quantified with an indefinite repeat, caused a compile-time
244 loop which used up all the system stack and provoked a segmentation fault.
245 This was not the same bug as 19 above.
246
247 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
248 Patch by Mike Frysinger.
249
250
251 Version 8.35 04-April-2014
252 --------------------------
253
254 1. A new flag is set, when property checks are present in an XCLASS.
255 When this flag is not set, PCRE can perform certain optimizations
256 such as studying these XCLASS-es.
257
258 2. The auto-possessification of character sets were improved: a normal
259 and an extended character set can be compared now. Furthermore
260 the JIT compiler optimizes more character set checks.
261
262 3. Got rid of some compiler warnings for potentially uninitialized variables
263 that show up only when compiled with -O2.
264
265 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
266 of a match later then the end of the match. The pcretest program was not
267 handling the case sensibly - it was outputting from the start to the next
268 binary zero. It now reports this situation in a message, and outputs the
269 text from the end to the start.
270
271 5. Fast forward search is improved in JIT. Instead of the first three
272 characters, any three characters with fixed position can be searched.
273 Search order: first, last, middle.
274
275 6. Improve character range checks in JIT. Characters are read by an inprecise
276 function now, which returns with an unknown value if the character code is
277 above a certain threshold (e.g: 256). The only limitation is that the value
278 must be bigger than the threshold as well. This function is useful when
279 the characters above the threshold are handled in the same way.
280
281 7. The macros whose names start with RAWUCHAR are placeholders for a future
282 mode in which only the bottom 21 bits of 32-bit data items are used. To
283 make this more memorable for those maintaining the code, the names have
284 been changed to start with UCHAR21, and an extensive comment has been added
285 to their definition.
286
287 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
288 to the export list in Makefile.am (they were accidentally omitted from the
289 8.34 tarball).
290
291 9. The informational output from pcretest used the phrase "starting byte set"
292 which is inappropriate for the 16-bit and 32-bit libraries. As the output
293 for "first char" and "need char" really means "non-UTF-char", I've changed
294 "byte" to "char", and slightly reworded the output. The documentation about
295 these values has also been (I hope) clarified.
296
297 10. Another JIT related optimization: use table jumps for selecting the correct
298 backtracking path, when more than four alternatives are present inside a
299 bracket.
300
301 11. Empty match is not possible, when the minimum length is greater than zero,
302 and there is no \K in the pattern. JIT should avoid empty match checks in
303 such cases.
304
305 12. In a caseless character class with UCP support, when a character with more
306 than one alternative case was not the first character of a range, not all
307 the alternative cases were added to the class. For example, s and \x{17f}
308 are both alternative cases for S: the class [RST] was handled correctly,
309 but [R-T] was not.
310
311 13. The configure.ac file always checked for pthread support when JIT was
312 enabled. This is not used in Windows, so I have put this test inside a
313 check for the presence of windows.h (which was already tested for).
314
315 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
316 The algorithm provides a way to skip certain starting offsets, and usually
317 faster than linear prefix searches.
318
319 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
320 as for 'fr_FR' and 'french'. For some reason, however, it then used the
321 Windows-specific input and output files, which have 'french' screwed in.
322 So this could never have worked. One of the problems with locales is that
323 they aren't always the same. I have now updated RunTest so that it checks
324 the output of the locale test (test 3) against three different output
325 files, and it allows the test to pass if any one of them matches. With luck
326 this should make the test pass on some versions of Solaris where it was
327 failing. Because of the uncertainty, the script did not used to stop if
328 test 3 failed; it now does. If further versions of a French locale ever
329 come to light, they can now easily be added.
330
331 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
332 there was a message during ./configure, but it did not stop. This now
333 provokes an error. The invalid example in README has been corrected.
334 If a value less than the minimum is given, the minimum value has always
335 been used, but now a warning is given.
336
337 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
338 was a bug in the test system, which is now fixed. Also, the list of various
339 configurations that are tested for each release did not have one with both
340 16/32 bits and --enable-bar-anycrlf. It now does.
341
342 18. pcretest was missing "-C bsr" for displaying the \R default setting.
343
344 19. Little endian PowerPC systems are supported now by the JIT compiler.
345
346 20. The fast forward newline mechanism could enter to an infinite loop on
347 certain invalid UTF-8 input. Although we don't support these cases
348 this issue can be fixed by a performance optimization.
349
350 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
351 not take account if existing stack usage. There is now a new global
352 variable called pcre_stack_guard that can be set to point to an external
353 function to check stack availability. It is called at the start of
354 processing every parenthesized group.
355
356 22. A typo in the code meant that in ungreedy mode the max/min qualifier
357 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
358 not match "ab".
359
360 23. When UTF was disabled, the JIT program reported some incorrect compile
361 errors. These messages are silenced now.
362
363 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
364 compiler.
365
366 25. Change all the temporary files used in RunGrepTest to be different to those
367 used by RunTest so that the tests can be run simultaneously, for example by
368 "make -j check".
369
370
371 Version 8.34 15-December-2013
372 -----------------------------
373
374 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
375 executable memory. Patch inspired by Carsten Klein.
376
377 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
378 this macro is never tested and has no effect, because the work to support
379 coverage involves only compiling and linking options and special targets in
380 the Makefile. The comment in config.h implied that defining the macro would
381 enable coverage support, which is totally false. There was also support for
382 setting this macro in the CMake files (my fault, I just copied it from
383 configure). SUPPORT_GCOV has now been removed.
384
385 3. Make a small performance improvement in strlen16() and strlen32() in
386 pcretest.
387
388 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
389 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
390 these cases). There was also one in pcretest.c.
391
392 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
393
394 6. In UTF mode, the code for checking whether a group could match an empty
395 string (which is used for indefinitely repeated groups to allow for
396 breaking an infinite loop) was broken when the group contained a repeated
397 negated single-character class with a character that occupied more than one
398 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
399 UTF-8 mode). The effect was undefined: the group might or might not be
400 deemed as matching an empty string, or the program might have crashed.
401
402 7. The code for checking whether a group could match an empty string was not
403 recognizing that \h, \H, \v, \V, and \R must match a character.
404
405 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
406 an empty string. If it can, pcretest shows this in its information output.
407
408 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
409 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
410 matched by pcre_exec() without using JIT:
411
412 (a) If the rest of the pattern did not match after a maximal run of
413 grapheme clusters, the code for backing up to try with fewer of them
414 did not always back up over a full grapheme when characters that do not
415 have the modifier quality were involved, e.g. Hangul syllables.
416
417 (b) If the match point in a subject started with modifier character, and
418 there was no match, the code could incorrectly back up beyond the match
419 point, and potentially beyond the first character in the subject,
420 leading to a segfault or an incorrect match result.
421
422 10. A conditional group with an assertion condition could lead to PCRE
423 recording an incorrect first data item for a match if no other first data
424 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
425 first data item, and therefore matched "ca" after "c" instead of at the
426 start.
427
428 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
429 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
430
431 12. The source of pcregrep now includes z/OS-specific code so that it can be
432 compiled for z/OS as part of the special z/OS distribution.
433
434 13. Added the -T and -TM options to pcretest.
435
436 14. The code in pcre_compile.c for creating the table of named capturing groups
437 has been refactored. Instead of creating the table dynamically during the
438 actual compiling pass, the information is remembered during the pre-compile
439 pass (on the stack unless there are more than 20 named groups, in which
440 case malloc() is used) and the whole table is created before the actual
441 compile happens. This has simplified the code (it is now nearly 150 lines
442 shorter) and prepared the way for better handling of references to groups
443 with duplicate names.
444
445 15. A back reference to a named subpattern when there is more than one of the
446 same name now checks them in the order in which they appear in the pattern.
447 The first one that is set is used for the reference. Previously only the
448 first one was inspected. This change makes PCRE more compatible with Perl.
449
450 16. Unicode character properties were updated from Unicode 6.3.0.
451
452 17. The compile-time code for auto-possessification has been refactored, based
453 on a patch by Zoltan Herczeg. It now happens after instead of during
454 compilation. The code is cleaner, and more cases are handled. The option
455 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
456 options in pcretest are provided to set it. It can also be set by
457 (*NO_AUTO_POSSESS) at the start of a pattern.
458
459 18. The character VT has been added to the default ("C" locale) set of
460 characters that match \s and are generally treated as white space,
461 following this same change in Perl 5.18. There is now no difference between
462 "Perl space" and "POSIX space". Whether VT is treated as white space in
463 other locales depends on the locale.
464
465 19. The code for checking named groups as conditions, either for being set or
466 for being recursed, has been refactored (this is related to 14 and 15
467 above). Processing unduplicated named groups should now be as fast at
468 numerical groups, and processing duplicated groups should be faster than
469 before.
470
471 20. Two patches to the CMake build system, by Alexander Barkov:
472
473 (1) Replace the "source" command by "." in CMakeLists.txt because
474 "source" is a bash-ism.
475
476 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
477 without these the CMake build does not work on Solaris.
478
479 21. Perl has changed its handling of \8 and \9. If there is no previously
480 encountered capturing group of those numbers, they are treated as the
481 literal characters 8 and 9 instead of a binary zero followed by the
482 literals. PCRE now does the same.
483
484 22. Following Perl, added \o{} to specify codepoints in octal, making it
485 possible to specify values greater than 0777 and also making them
486 unambiguous.
487
488 23. Perl now gives an error for missing closing braces after \x{... instead of
489 treating the string as literal. PCRE now does the same.
490
491 24. RunTest used to grumble if an inappropriate test was selected explicitly,
492 but just skip it when running all tests. This make it awkward to run ranges
493 of tests when one of them was inappropriate. Now it just skips any
494 inappropriate tests, as it always did when running all tests.
495
496 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
497 character types such as \d or \w, too many callouts were inserted, and the
498 data that they returned was rubbish.
499
500 26. In UCP mode, \s was not matching two of the characters that Perl matches,
501 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
502 were matched by \h. The code has now been refactored so that the lists of
503 the horizontal and vertical whitespace characters used for \h and \v (which
504 are defined only in one place) are now also used for \s.
505
506 27. Add JIT support for the 64 bit TileGX architecture.
507 Patch by Jiong Wang (Tilera Corporation).
508
509 28. Possessive quantifiers for classes (both explicit and automatically
510 generated) now use special opcodes instead of wrapping in ONCE brackets.
511
512 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
513 (because it's meaningless), this was not happening when PCRE_CASELESS was
514 set. Not wrong, but inefficient.
515
516 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
517 properties for \w, \d, etc) is present in a test regex. Otherwise if the
518 test contains no characters greater than 255, Perl doesn't realise it
519 should be using Unicode semantics.
520
521 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
522 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
523 does in Unicode mode.
524
525 32. Added the "forbid" facility to pcretest so that putting tests into the
526 wrong test files can sometimes be quickly detected.
527
528 33. There is now a limit (default 250) on the depth of nesting of parentheses.
529 This limit is imposed to control the amount of system stack used at compile
530 time. It can be changed at build time by --with-parens-nest-limit=xxx or
531 the equivalent in CMake.
532
533 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
534 errors. Perl warns for these when in warning mode, but PCRE has no facility
535 for giving warnings.
536
537 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
538 However, this was not working for (?!) because it is optimized to (*FAIL),
539 for which PCRE does not allow quantifiers. The optimization is now disabled
540 when a quantifier follows (?!). I can't see any use for this, but it makes
541 things uniform.
542
543 36. Perl no longer allows group names to start with digits, so I have made this
544 change also in PCRE. It simplifies the code a bit.
545
546 37. In extended mode, Perl ignores spaces before a + that indicates a
547 possessive quantifier. PCRE allowed a space before the quantifier, but not
548 before the possessive +. It now does.
549
550 38. The use of \K (reset reported match start) within a repeated possessive
551 group such as (a\Kb)*+ was not working.
552
553 40. Document that the same character tables must be used at compile time and
554 run time, and that the facility to pass tables to pcre_exec() and
555 pcre_dfa_exec() is for use only with saved/restored patterns.
556
557 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
558 features for Builds with MSVC:
559
560 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
561 builds). The .rc files can be used to set FileDescription and many other
562 attributes.
563
564 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
565 This allows higher-level build scripts which want .pdb files to avoid
566 hard-coding the exact files needed."
567
568 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
569 mean "start of word" and "end of word", respectively, as a transition aid.
570
571 43. A minimizing repeat of a class containing codepoints greater than 255 in
572 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
573 compiled to use the heap for recursion.
574
575 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
576 is configured.
577
578
579 Version 8.33 28-May-2013
580 ------------------------
581
582 1. Added 'U' to some constants that are compared to unsigned integers, to
583 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
584 variables that are added to signed variables, to ensure the result is
585 signed and can be negated.
586
587 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
588 CMake config files.
589
590 3. Revise the creation of config.h.generic so that all boolean macros are
591 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
592 overriding via -D on the command line possible.
593
594 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
595 to unsigned int is reported to make a quite noticeable speed difference in
596 a specific Windows environment. Testing on Linux did also appear to show
597 some benefit (and it is clearly not harmful). Also fixed the definition of
598 Xop which should be unsigned.
599
600 5. Related to (4), changing the definition of the intermediate variable cc
601 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
602 improvements.
603
604 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
605 unnecessary spaces.
606
607 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
608 and later.
609
610 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
611
612 9. Optimizing fast_forward_start_bits in JIT.
613
614 10. Adding support for callouts in JIT, and fixing some issues revealed
615 during this work. Namely:
616
617 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
618
619 (b) Minimum length was not checked before the matching is started.
620
621 11. The value of capture_last that is passed to callouts was incorrect in some
622 cases when there was a capture on one path that was subsequently abandoned
623 after a backtrack. Also, the capture_last value is now reset after a
624 recursion, since all captures are also reset in this case.
625
626 12. The interpreter no longer returns the "too many substrings" error in the
627 case when an overflowing capture is in a branch that is subsequently
628 abandoned after a backtrack.
629
630 13. In the pathological case when an offset vector of size 2 is used, pcretest
631 now prints out the matched string after a yield of 0 or 1.
632
633 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
634 Only supported by the JIT compiler at the moment.
635
636 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
637
638 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
639 offset of the starting point of the matching process, provided the offsets
640 vector is large enough.
641
642 17. The \A escape now records a lookbehind value of 1, though its execution
643 does not actually inspect the previous character. This is to ensure that,
644 in partial multi-segment matching, at least one character from the old
645 segment is retained when a new segment is processed. Otherwise, if there
646 are no lookbehinds in the pattern, \A might match incorrectly at the start
647 of a new segment.
648
649 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
650
651 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
652 gives some modest performance improvement in 8-bit mode.
653
654 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
655 be expressed in certain programming languages using Universal Character
656 Names.
657
658 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
659 which points out that "non characters" are not "characters that may not
660 appear in Unicode strings" but rather "characters that are reserved for
661 internal use and have only local meaning".
662
663 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
664 there was a conditional group that depended on an assertion, if the
665 assertion was false, the callout that immediately followed the alternation
666 in the condition was skipped when pcre_exec() was used for matching.
667
668 23. Allow an explicit callout to be inserted before an assertion that is the
669 condition for a conditional group, for compatibility with automatic
670 callouts, which always insert a callout at this point.
671
672 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
673 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
674
675 25. (*PRUNE) is now supported by the JIT compiler.
676
677 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
678
679 27. Fix the case where there are two or more SKIPs with arguments that may be
680 ignored.
681
682 28. (*SKIP) is now supported by the JIT compiler.
683
684 29. (*THEN) is now supported by the JIT compiler.
685
686 30. Update RunTest with additional test selector options.
687
688 31. The way PCRE handles backtracking verbs has been changed in two ways.
689
690 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
691 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
692 backtracking. In some cases this makes it more Perl-compatible, but Perl's
693 rather obscure rules do not always do the same thing.
694
695 (2) Previously, backtracking verbs were confined within assertions. This is
696 no longer the case for positive assertions, except for (*ACCEPT). Again,
697 this sometimes improves Perl compatibility, and sometimes does not.
698
699 32. A number of tests that were in test 2 because Perl did things differently
700 have been moved to test 1, because either Perl or PCRE has changed, and
701 these tests are now compatible.
702
703 32. Backtracking control verbs are now handled in the same way in JIT and
704 interpreter.
705
706 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
707 contained a forward subroutine reference caused a compile error.
708
709 34. Auto-detect and optimize limited repetitions in JIT.
710
711 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
712 blocking (*UTF) etc.
713
714 36. In the interpreter, maximizing pattern repetitions for characters and
715 character types now use tail recursion, which reduces stack usage.
716
717 37. The value of the max lookbehind was not correctly preserved if a compiled
718 and saved regex was reloaded on a host of different endianness.
719
720 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
721 of the compiled pattern block, expand the flags field from 16 to 32 bits
722 because it was almost full.
723
724 39. Try madvise first before posix_madvise.
725
726 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
727 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
728 original change locked out all matches of empty strings. This has been
729 changed so that one match of an empty string per line is recognized.
730 Subsequent searches on the same line (for colouring or for --only-matching,
731 for example) do not recognize empty strings.
732
733 41. Applied a user patch to fix a number of spelling mistakes in comments.
734
735 42. Data lines longer than 65536 caused pcretest to crash.
736
737 43. Clarified the data type for length and startoffset arguments for pcre_exec
738 and pcre_dfa_exec in the function-specific man pages, where they were
739 explicitly stated to be in bytes, never having been updated. I also added
740 some clarification to the pcreapi man page.
741
742 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
743 a segmentation fault.
744
745
746 Version 8.32 30-November-2012
747 -----------------------------
748
749 1. Improved JIT compiler optimizations for first character search and single
750 character iterators.
751
752 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
753 Patch by Daniel Richard G.
754
755 3. Single character iterator optimizations in the JIT compiler.
756
757 4. Improved JIT compiler optimizations for character ranges.
758
759 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
760 Reported by Giuseppe D'Angelo.
761
762 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
763 of a line, was being set incorrectly in cases where .* appeared inside
764 atomic brackets at the start of a pattern, or where there was a subsequent
765 *PRUNE or *SKIP.
766
767 7. Improved instruction cache flush for POWER/PowerPC.
768 Patch by Daniel Richard G.
769
770 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
771 grep:
772
773 (a) There is now no limit to the number of patterns to be matched.
774
775 (b) An error is given if a pattern is too long.
776
777 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
778 are now supported.
779
780 (d) --exclude-from and --include-from (multiple use) have been added.
781
782 (e) Exclusions and inclusions now apply to all files and directories, not
783 just to those obtained from scanning a directory recursively.
784
785 (f) Multiple uses of -f and --file-list are now supported.
786
787 (g) In a Windows environment, the default for -d has been changed from
788 "read" (the GNU grep default) to "skip", because otherwise the presence
789 of a directory in the file list provokes an error.
790
791 (h) The documentation has been revised and clarified in places.
792
793 9. Improve the matching speed of capturing brackets.
794
795 10. Changed the meaning of \X so that it now matches a Unicode extended
796 grapheme cluster.
797
798 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
799 out POSIX threads when JIT support is configured.
800
801 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
802
803 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
804 structure could go wrong in environments where size_t is not the same size
805 as int.
806
807 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
808 set.
809
810 15. The EBCDIC support had decayed; later updates to the code had included
811 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
812 general tidy up of EBCDIC-related issues, and the documentation was also
813 not quite right. There is now a test that can be run on ASCII systems to
814 check some of the EBCDIC-related things (but is it not a full test).
815
816 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
817 in a small tidy to the code.
818
819 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
820
821 18. If the --only-matching (-o) option in pcregrep is specified multiple
822 times, each one causes appropriate output. For example, -o1 -o2 outputs the
823 substrings matched by the 1st and 2nd capturing parentheses. A separating
824 string can be specified by --om-separator (default empty).
825
826 19. Improving the first n character searches.
827
828 20. Turn case lists for horizontal and vertical white space into macros so that
829 they are defined only once.
830
831 21. This set of changes together give more compatible Unicode case-folding
832 behaviour for characters that have more than one other case when UCP
833 support is available.
834
835 (a) The Unicode property table now has offsets into a new table of sets of
836 three or more characters that are case-equivalent. The MultiStage2.py
837 script that generates these tables (the pcre_ucd.c file) now scans
838 CaseFolding.txt instead of UnicodeData.txt for character case
839 information.
840
841 (b) The code for adding characters or ranges of characters to a character
842 class has been abstracted into a generalized function that also handles
843 case-independence. In UTF-mode with UCP support, this uses the new data
844 to handle characters with more than one other case.
845
846 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
847 whose other case is greater than 256 are now correctly matched
848 caselessly. Previously, the high codepoint matched the low one, but not
849 vice versa.
850
851 (d) The processing of \h, \H, \v, and \ in character classes now makes use
852 of the new class addition function, using character lists defined as
853 macros alongside the case definitions of 20 above.
854
855 (e) Caseless back references now work with characters that have more than
856 one other case.
857
858 (f) General caseless matching of characters with more than one other case
859 is supported.
860
861 22. Unicode character properties were updated from Unicode 6.2.0
862
863 23. Improved CMake support under Windows. Patch by Daniel Richard G.
864
865 24. Add support for 32-bit character strings, and UTF-32
866
867 25. Major JIT compiler update (code refactoring and bugfixing).
868 Experimental Sparc 32 support is added.
869
870 26. Applied a modified version of Daniel Richard G's patch to create
871 pcre.h.generic and config.h.generic by "make" instead of in the
872 PrepareRelease script.
873
874 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
875 pcre_compile.c when checking for a zero character.
876
877 28. Introducing a native interface for JIT. Through this interface, the compiled
878 machine code can be directly executed. The purpose of this interface is to
879 provide fast pattern matching, so several sanity checks are not performed.
880 However, feature tests are still performed. The new interface provides
881 1.4x speedup compared to the old one.
882
883 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
884 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
885 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
886
887 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
888 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
889 the "old" RFC 2279). Instead, it ended up passing a negative length to
890 pcre_exec().
891
892 31. Add support for GCC's visibility feature to hide internal functions.
893
894 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
895 "unknown -C option" after outputting 0 or 1.
896
897 33. There is now support for generating a code coverage report for the test
898 suite in environments where gcc is the compiler and lcov is installed. This
899 is mainly for the benefit of the developers.
900
901 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
902 unaddressable using valgrind annotations, allowing valgrind to detect
903 invalid memory accesses. This is mainly for the benefit of the developers.
904
905 25. (*UTF) can now be used to start a pattern in any of the three libraries.
906
907 26. Give configure error if --enable-cpp but no C++ compiler found.
908
909
910 Version 8.31 06-July-2012
911 -------------------------
912
913 1. Fixing a wrong JIT test case and some compiler warnings.
914
915 2. Removed a bashism from the RunTest script.
916
917 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
918 to unsigned type, result still unsigned" that was given by an MS compiler
919 on encountering the code "-sizeof(xxx)".
920
921 4. Partial matching support is added to the JIT compiler.
922
923 5. Fixed several bugs concerned with partial matching of items that consist
924 of more than one character:
925
926 (a) /^(..)\1/ did not partially match "aba" because checking references was
927 done on an "all or nothing" basis. This also applied to repeated
928 references.
929
930 (b) \R did not give a hard partial match if \r was found at the end of the
931 subject.
932
933 (c) \X did not give a hard partial match after matching one or more
934 characters at the end of the subject.
935
936 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
937 a partial match for the string "\r".
938
939 (e) When newline was set to CRLF, the metacharacter "." did not recognize
940 a partial match for a CR character at the end of the subject string.
941
942 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
943 running pcretest, the text "(JIT)" added to the output whenever JIT is
944 actually used to run the match.
945
946 7. Individual JIT compile options can be set in pcretest by following -s+[+]
947 or /S+[+] with a digit between 1 and 7.
948
949 8. OP_NOT now supports any UTF character not just single-byte ones.
950
951 9. (*MARK) control verb is now supported by the JIT compiler.
952
953 10. The command "./RunTest list" lists the available tests without actually
954 running any of them. (Because I keep forgetting what they all are.)
955
956 11. Add PCRE_INFO_MAXLOOKBEHIND.
957
958 12. Applied a (slightly modified) user-supplied patch that improves performance
959 when the heap is used for recursion (compiled with --disable-stack-for-
960 recursion). Instead of malloc and free for each heap frame each time a
961 logical recursion happens, frames are retained on a chain and re-used where
962 possible. This sometimes gives as much as 30% improvement.
963
964 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
965 call.
966
967 14. As documented, (*COMMIT) is now confined to within a positive assertion.
968
969 15. It is now possible to link pcretest with libedit as an alternative to
970 libreadline.
971
972 16. (*COMMIT) control verb is now supported by the JIT compiler.
973
974 17. The Unicode data tables have been updated to Unicode 6.1.0.
975
976 18. Added --file-list option to pcregrep.
977
978 19. Added binary file support to pcregrep, including the -a, --binary-files,
979 -I, and --text options.
980
981 20. The madvise function is renamed for posix_madvise for QNX compatibility
982 reasons. Fixed by Giuseppe D'Angelo.
983
984 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
985
986 22. Changed the option for creating symbolic links for 16-bit man pages from
987 -s to -sf so that re-installing does not cause issues.
988
989 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
990
991 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
992 matches in certain environments (the workspace was not being correctly
993 retained). Also added to pcre_dfa_exec() a simple plausibility check on
994 some of the workspace data at the beginning of a restart.
995
996 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
997 was not doing so when it should - probably a typo introduced by SVN 528
998 (change 8.10/14).
999
1000 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1001 \w+ when the character tables indicated that \x{c4} was a word character.
1002 There were several related cases, all because the tests for doing a table
1003 lookup were testing for characters less than 127 instead of 255.
1004
1005 27. If a pattern contains capturing parentheses that are not used in a match,
1006 their slots in the ovector are set to -1. For those that are higher than
1007 any matched groups, this happens at the end of processing. In the case when
1008 there were back references that the ovector was too small to contain
1009 (causing temporary malloc'd memory to be used during matching), and the
1010 highest capturing number was not used, memory off the end of the ovector
1011 was incorrectly being set to -1. (It was using the size of the temporary
1012 memory instead of the true size.)
1013
1014 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1015 ovector size, it uses memory at the end of the block that it has got.
1016
1017 29. Check for an overlong MARK name and give an error at compile time. The
1018 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1019
1020 30. JIT compiler update.
1021
1022 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1023 Rill for the patch.
1024
1025 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1026
1027 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1028
1029 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1030 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1031 was enabled.
1032
1033 35. Improve JIT code generation for greedy plus quantifier.
1034
1035 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1036 1 to "aa" instead of to an empty string. The bug affected repeated groups
1037 that could potentially match an empty string.
1038
1039 37. Optimizing single character iterators in JIT.
1040
1041 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1042 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1043 codepoints that are too big for the mode are faulted, and in a UTF mode,
1044 disallowed codepoints are also faulted.
1045
1046 39. If PCRE was compiled with UTF support, in three places in the DFA
1047 matcher there was code that should only have been obeyed in UTF mode, but
1048 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1049 processing when bytes with values greater than 127 were present. In 16-bit
1050 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1051 both cases the values are those that cannot be the first data item in a UTF
1052 character. The three items that might have provoked this were recursions,
1053 possessively repeated groups, and atomic groups.
1054
1055 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1056 and pcregrep, because some OS require shared objects to be explicitly
1057 passed to ld, causing the link step to fail if they are not.
1058
1059 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1060 mode, patterns that started with \h* or \R* might be incorrectly matched.
1061
1062
1063 Version 8.30 04-February-2012
1064 -----------------------------
1065
1066 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1067 name is defined in ctype.h.
1068
1069 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1070 only in quite long subpatterns.
1071
1072 3. Removed the function pcre_info(), which has been obsolete and deprecated
1073 since it was replaced by pcre_fullinfo() in February 2000.
1074
1075 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1076 match a (*MARK), and the match failed at the start of the subject, a
1077 reference to memory before the start of the subject could occur. This bug
1078 was introduced by fix 17 of release 8.21.
1079
1080 5. A reference to an unset group with zero minimum repetition was giving
1081 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1082 /(another)?(\1?)test/ matched against "hello world test". This bug was
1083 introduced in release 8.13.
1084
1085 6. Add support for 16-bit character strings (a large amount of work involving
1086 many changes and refactorings).
1087
1088 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1089 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1090 from a file.
1091
1092 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1093 rounding is not applied in this particular case).
1094
1095 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1096 if they appear, or are escaped, in patterns.
1097
1098 10. Get rid of a number of -Wunused-but-set-variable warnings.
1099
1100 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1101 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1102 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1103 also returns the mark "x". This bug applied to capturing parentheses,
1104 non-capturing parentheses, and atomic parentheses. It also applied to some
1105 assertions.
1106
1107 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1108 information out of configure.ac instead of relying on pcre.h.generic, which
1109 is not stored in the repository.
1110
1111 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1112 -lreadline.
1113
1114 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1115
1116 15. Applied Graycode's patch to put the top-level frame on the stack rather
1117 than the heap when not using the stack for recursion. This gives a
1118 performance improvement in many cases when recursion is not deep.
1119
1120 16. Experimental code added to "pcretest -C" to output the stack frame size.
1121
1122
1123 Version 8.21 12-Dec-2011
1124 ------------------------
1125
1126 1. Updating the JIT compiler.
1127
1128 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1129 are added as well.
1130
1131 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1132 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1133 calling _pcre_jit_exec. Some extra comments are added.
1134
1135 4. (*MARK) settings inside atomic groups that do not contain any capturing
1136 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1137 was introduced by change 18 for 8.20.
1138
1139 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1140 ECMA-262 standard.
1141
1142 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1143 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1144 This bug was probably introduced by change 9 of 8.13.
1145
1146 7. While fixing 6 above, I noticed that a number of other items were being
1147 incorrectly rejected as "not fixed length". This arose partly because newer
1148 opcodes had not been added to the fixed-length checking code. I have (a)
1149 corrected the bug and added tests for these items, and (b) arranged for an
1150 error to occur if an unknown opcode is encountered while checking for fixed
1151 length instead of just assuming "not fixed length". The items that were
1152 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1153 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1154 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1155
1156 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1157 being incorrectly compiled and would have given unpredicatble results.
1158
1159 9. A possessively repeated subpattern with minimum repeat count greater than
1160 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1161 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1162 the first (A) could occur when it should not.
1163
1164 10. Add a cast and remove a redundant test from the code.
1165
1166 11. JIT should use pcre_malloc/pcre_free for allocation.
1167
1168 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1169 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1170 is anything other than /usr, -L is still shown).
1171
1172 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1173
1174 14. Perl does not support \N without a following name in a [] class; PCRE now
1175 also gives an error.
1176
1177 15. If a forward reference was repeated with an upper limit of around 2000,
1178 it caused the error "internal error: overran compiling workspace". The
1179 maximum number of forward references (including repeats) was limited by the
1180 internal workspace, and dependent on the LINK_SIZE. The code has been
1181 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1182 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1183 of around 200,000 forward references. While doing this, I also speeded up
1184 the filling in of repeated forward references.
1185
1186 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1187 incorrectly expecting the subject to contain another "a" after the start.
1188
1189 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1190 in the match, the SKIP should be ignored. This was not happening; instead
1191 the SKIP was being treated as NOMATCH. For patterns such as
1192 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1193 tested.
1194
1195 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1196 now much more compatible with Perl, in particular in cases where the result
1197 is a non-match for a non-anchored pattern. For example, if
1198 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1199 "m", where previously it did not return a name. A side effect of this
1200 change is that for partial matches, the last encountered mark name is
1201 returned, as for non matches. A number of tests that were previously not
1202 Perl-compatible have been moved into the Perl-compatible test files. The
1203 refactoring has had the pleasing side effect of removing one argument from
1204 the match() function, thus reducing its stack requirements.
1205
1206 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1207 subsequent uses of /S (without +) incorrectly behaved like /S+.
1208
1209 21. Retrieve executable code size support for the JIT compiler and fixing
1210 some warnings.
1211
1212 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1213 not work when the shorter character appeared right at the end of the
1214 subject string.
1215
1216 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1217 systems.
1218
1219 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1220 output it when the /M option is used in pcretest.
1221
1222 25. The CheckMan script was not being included in the distribution. Also, added
1223 an explicit "perl" to run Perl scripts from the PrepareRelease script
1224 because this is reportedly needed in Windows.
1225
1226 26. If study data was being save in a file and studying had not found a set of
1227 "starts with" bytes for the pattern, the data written to the file (though
1228 never used) was taken from uninitialized memory and so caused valgrind to
1229 complain.
1230
1231 27. Updated RunTest.bat as provided by Sheri Pierce.
1232
1233 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1234
1235 29. Computation of memory usage for the table of capturing group names was
1236 giving an unnecessarily large value.
1237
1238
1239 Version 8.20 21-Oct-2011
1240 ------------------------
1241
1242 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1243 a POSIX class. After further experiments with Perl, which convinced me that
1244 Perl has bugs and confusions, a closing square bracket is no longer allowed
1245 in a POSIX name. This bug also affected patterns with classes that started
1246 with full stops.
1247
1248 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1249 captured substring, but while checking the failing first alternative,
1250 substring 1 is temporarily captured. If the output vector supplied to
1251 pcre_exec() was not big enough for this capture, the yield of the function
1252 was still zero ("insufficient space for captured substrings"). This cannot
1253 be totally fixed without adding another stack variable, which seems a lot
1254 of expense for a edge case. However, I have improved the situation in cases
1255 such as /(a)(b)x|abc/ matched against "abc", where the return code
1256 indicates that fewer than the maximum number of slots in the ovector have
1257 been set.
1258
1259 3. Related to (2) above: when there are more back references in a pattern than
1260 slots in the output vector, pcre_exec() uses temporary memory during
1261 matching, and copies in the captures as far as possible afterwards. It was
1262 using the entire output vector, but this conflicts with the specification
1263 that only 2/3 is used for passing back captured substrings. Now it uses
1264 only the first 2/3, for compatibility. This is, of course, another edge
1265 case.
1266
1267 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1268 main code base, and can be used by building with --enable-jit. When this is
1269 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1270 runtime --no-jit option is given.
1271
1272 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1273 ovector, the return from the function was zero, implying that there were
1274 other matches that did not fit. The correct "exactly full" value is now
1275 returned.
1276
1277 6. If a subpattern that was called recursively or as a subroutine contained
1278 (*PRUNE) or any other control that caused it to give a non-standard return,
1279 invalid errors such as "Error -26 (nested recursion at the same subject
1280 position)" or even infinite loops could occur.
1281
1282 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1283 computing the minimum length on reaching *ACCEPT, and so ended up with the
1284 wrong value of 1 rather than 0. Further investigation indicates that
1285 computing a minimum subject length in the presence of *ACCEPT is difficult
1286 (think back references, subroutine calls), and so I have changed the code
1287 so that no minimum is registered for a pattern that contains *ACCEPT.
1288
1289 8. If (*THEN) was present in the first (true) branch of a conditional group,
1290 it was not handled as intended. [But see 16 below.]
1291
1292 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1293 Sheri Pierce.
1294
1295 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1296 the first byte in a match must be "a".
1297
1298 11. Change 17 for 8.13 increased the recursion depth for patterns like
1299 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1300 pattern contains any instances of (*THEN). If it does not, the old
1301 optimizations are restored. It would be nice to do this on a per-group
1302 basis, but at the moment that is not feasible.
1303
1304 12. In some environments, the output of pcretest -C is CRLF terminated. This
1305 broke RunTest's code that checks for the link size. A single white space
1306 character after the value is now allowed for.
1307
1308 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1309 For "fr", it uses the Windows-specific input and output files.
1310
1311 14. If (*THEN) appeared in a group that was called recursively or as a
1312 subroutine, it did not work as intended. [But see next item.]
1313
1314 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1315 pattern fragments (but not containing any | characters). If A and B are
1316 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1317 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1318 D. In other words, Perl considers parentheses that do not contain any |
1319 characters to be part of a surrounding alternative, whereas PCRE was
1320 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1321 differently. PCRE now behaves in the same way as Perl, except in the case
1322 of subroutine/recursion calls such as (?1) which have in any case always
1323 been different (but PCRE had them first :-).
1324
1325 16. Related to 15 above: Perl does not treat the | in a conditional group as
1326 creating alternatives. Such a group is treated in the same way as an
1327 ordinary group without any | characters when processing (*THEN). PCRE has
1328 been changed to match Perl's behaviour.
1329
1330 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1331 RunGrepTest script failed.
1332
1333 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1334 inevitable for groups that contain captures, but it can lead to a lot of
1335 stack use in large patterns. The old behaviour has been restored for atomic
1336 groups that do not contain any capturing parentheses.
1337
1338 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1339 suppress the check for a minimum subject length at run time. (If it was
1340 given to pcre_exec() or pcre_dfa_exec() it did work.)
1341
1342 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1343 fail to work when decoding hex characters in data strings in EBCDIC
1344 environments.
1345
1346 21. It appears that in at least one Mac OS environment, the isxdigit() function
1347 is implemented as a macro that evaluates to its argument more than once,
1348 contravening the C 90 Standard (I haven't checked a later standard). There
1349 was an instance in pcretest which caused it to go wrong when processing
1350 \x{...} escapes in subject strings. The has been rewritten to avoid using
1351 things like p++ in the argument of isxdigit().
1352
1353
1354 Version 8.13 16-Aug-2011
1355 ------------------------
1356
1357 1. The Unicode data tables have been updated to Unicode 6.0.0.
1358
1359 2. Two minor typos in pcre_internal.h have been fixed.
1360
1361 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1362 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1363 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1364
1365 4. There were a number of related bugs in the code for matching backrefences
1366 caselessly in UTF-8 mode when codes for the characters concerned were
1367 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1368 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1369 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1370 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1371 2-byte code at the end of the subject (it thought there wasn't enough data
1372 left).
1373
1374 5. Comprehensive information about what went wrong is now returned by
1375 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1376 as the output vector has at least 2 elements. The offset of the start of
1377 the failing character and a reason code are placed in the vector.
1378
1379 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1380 now returned is for the first byte of the failing character, instead of the
1381 last byte inspected. This is an incompatible change, but I hope it is small
1382 enough not to be a problem. It makes the returned offset consistent with
1383 pcre_exec() and pcre_dfa_exec().
1384
1385 7. pcretest now gives a text phrase as well as the error number when
1386 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1387 failure, the offset and reason code are output.
1388
1389 8. When \R was used with a maximizing quantifier it failed to skip backwards
1390 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1391 back over a single character (\n). This seems wrong (because it treated the
1392 two characters as a single entity when going forwards), conflicts with the
1393 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1394 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1395 has been changed.
1396
1397 9. Some internal refactoring has changed the processing so that the handling
1398 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1399 time (the PCRE_DOTALL option was changed this way some time ago: version
1400 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1401 which was always a bit of a fudge. It also means that there is one less
1402 argument for the match() function, which reduces its stack requirements
1403 slightly. This change also fixes an incompatibility with Perl: the pattern
1404 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1405
1406 10. More internal refactoring has drastically reduced the number of recursive
1407 calls to match() for possessively repeated groups such as (abc)++ when
1408 using pcre_exec().
1409
1410 11. While implementing 10, a number of bugs in the handling of groups were
1411 discovered and fixed:
1412
1413 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1414 (a|)*(?1) gave a compile-time internal error.
1415 ((a|)+)+ did not notice that the outer group could match an empty string.
1416 (^a|^)+ was not marked as anchored.
1417 (.*a|.*)+ was not marked as matching at start or after a newline.
1418
1419 12. Yet more internal refactoring has removed another argument from the match()
1420 function. Special calls to this function are now indicated by setting a
1421 value in a variable in the "match data" data block.
1422
1423 13. Be more explicit in pcre_study() instead of relying on "default" for
1424 opcodes that mean there is no starting character; this means that when new
1425 ones are added and accidentally left out of pcre_study(), testing should
1426 pick them up.
1427
1428 14. The -s option of pcretest has been documented for ages as being an old
1429 synonym of -m (show memory usage). I have changed it to mean "force study
1430 for every regex", that is, assume /S for every regex. This is similar to -i
1431 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1432 using it. It makes it easier to run collections of tests with and without
1433 study enabled, and thereby test pcre_study() more easily. All the standard
1434 tests are now run with and without -s (but some patterns can be marked as
1435 "never study" - see 20 below).
1436
1437 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1438 restoration of the capturing data to the outer values was not happening
1439 correctly.
1440
1441 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1442 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1443 pattern had matched an empty string, and so incorrectly returned a no
1444 match.
1445
1446 17. There was optimizing code for the last branch of non-capturing parentheses,
1447 and also for the obeyed branch of a conditional subexpression, which used
1448 tail recursion to cut down on stack usage. Unfortunately, now that there is
1449 the possibility of (*THEN) occurring in these branches, tail recursion is
1450 no longer possible because the return has to be checked for (*THEN). These
1451 two optimizations have therefore been removed. [But see 8.20/11 above.]
1452
1453 18. If a pattern containing \R was studied, it was assumed that \R always
1454 matched two bytes, thus causing the minimum subject length to be
1455 incorrectly computed because \R can also match just one byte.
1456
1457 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1458 was incorrectly computed.
1459
1460 20. If /S is present twice on a test pattern in pcretest input, it now
1461 *disables* studying, thereby overriding the use of -s on the command line
1462 (see 14 above). This is necessary for one or two tests to keep the output
1463 identical in both cases.
1464
1465 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1466 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1467
1468 22. When an atomic group that contained a capturing parenthesis was
1469 successfully matched, but the branch in which it appeared failed, the
1470 capturing was not being forgotten if a higher numbered group was later
1471 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1472 group 1 to "a", when in fact it should be unset. This applied to multi-
1473 branched capturing and non-capturing groups, repeated or not, and also to
1474 positive assertions (capturing in negative assertions does not happen
1475 in PCRE) and also to nested atomic groups.
1476
1477 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1478 subject after a captured substring, to make it easier to tell which of a
1479 number of identical substrings has been captured.
1480
1481 24. The way atomic groups are processed by pcre_exec() has been changed so that
1482 if they are repeated, backtracking one repetition now resets captured
1483 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1484 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1485 "aaa". Previously, it would have been "a". As part of this code
1486 refactoring, the way recursive calls are handled has also been changed.
1487
1488 25. If an assertion condition captured any substrings, they were not passed
1489 back unless some other capturing happened later. For example, if
1490 (?(?=(a))a) was matched against "a", no capturing was returned.
1491
1492 26. When studying a pattern that contained subroutine calls or assertions,
1493 the code for finding the minimum length of a possible match was handling
1494 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1495 group 1 called group 2 while simultaneously a separate group 2 called group
1496 1). A stack overflow occurred in this case. I have fixed this by limiting
1497 the recursion depth to 10.
1498
1499 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1500 Fortmann. This supports explicit test numbers on the command line, and has
1501 argument validation and error reporting.
1502
1503 28. An instance of \X with an unlimited repeat could fail if at any point the
1504 first character it looked at was a mark character.
1505
1506 29. Some minor code refactoring concerning Unicode properties and scripts
1507 should reduce the stack requirement of match() slightly.
1508
1509 30. Added the '=' option to pcretest to check the setting of unused capturing
1510 slots at the end of the pattern, which are documented as being -1, but are
1511 not included in the return count.
1512
1513 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1514 compiled something random. Now it gives a compile-time error (as does
1515 Perl).
1516
1517 32. A *MARK encountered during the processing of a positive assertion is now
1518 recorded and passed back (compatible with Perl).
1519
1520 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1521 had alternative anchored branches, the search for a second match in a line
1522 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1523 matched the line "0102" twice. The same bug affected patterns that started
1524 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1525 twice.
1526
1527 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1528 does, and because of capturing effects, quantifying parenthesized
1529 assertions may at times be useful. Quantifiers are now allowed for
1530 parenthesized assertions.
1531
1532 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1533
1534 36. \g was being checked for fancy things in a character class, when it should
1535 just be a literal "g".
1536
1537 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1538 appearance of a nested POSIX class supersedes an apparent external class.
1539 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1540 unescaped square brackets may also appear as part of class names. For
1541 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1542 more like Perl. (But see 8.20/1 above.)
1543
1544 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1545 was because it thought it was \N{name}, which is not supported).
1546
1547 39. Add minix to OS list not supporting the -S option in pcretest.
1548
1549 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1550 cannot analyze patterns in sufficient detail to catch mutual recursions
1551 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1552 subgroup is called recursively as a subpattern for a second time at the
1553 same position in the subject string. In previous releases this might have
1554 been caught by the recursion limit, or it might have run out of stack.
1555
1556 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1557 happen only once. PCRE was, however incorrectly giving a compile time error
1558 "recursive call could loop indefinitely" because it cannot analyze the
1559 pattern in sufficient detail. The compile time test no longer happens when
1560 PCRE is compiling a conditional subpattern, but actual runaway loops are
1561 now caught at runtime (see 40 above).
1562
1563 42. It seems that Perl allows any characters other than a closing parenthesis
1564 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1565 has been changed to be the same.
1566
1567 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1568 as not to get warnings when autogen.sh is called. Also changed
1569 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1570
1571 44. To help people who use pcregrep to scan files containing exceedingly long
1572 lines, the following changes have been made:
1573
1574 (a) The default value of the buffer size parameter has been increased from
1575 8K to 20K. (The actual buffer used is three times this size.)
1576
1577 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1578 PCRE is built.
1579
1580 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1581 to be set at run time.
1582
1583 (d) Numerical values in pcregrep options can be followed by K or M, for
1584 example --buffer-size=50K.
1585
1586 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1587 given and the return code is set to 2.
1588
1589 45. Add a pointer to the latest mark to the callout data block.
1590
1591 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1592 partial match of an empty string instead of no match. This was specific to
1593 the use of ".".
1594
1595 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1596 complete match instead of a partial match. This bug was dependent on both
1597 the PCRE_UTF8 and PCRE_DOTALL options being set.
1598
1599 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1600 starting byte set, because \b was not being ignored.
1601
1602
1603 Version 8.12 15-Jan-2011
1604 ------------------------
1605
1606 1. Fixed some typos in the markup of the man pages, and wrote a script that
1607 checks for such things as part of the documentation building process.
1608
1609 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1610 --match-limit and --recursion-limit options (added for 8.11). In
1611 particular, this made one of the standard tests fail. (The integer value
1612 went into the wrong half of a long int.)
1613
1614 3. If the --colour option was given to pcregrep with -v (invert match), it
1615 did strange things, either producing crazy output, or crashing. It should,
1616 of course, ignore a request for colour when reporting lines that do not
1617 match.
1618
1619 4. Another pcregrep bug caused similar problems if --colour was specified with
1620 -M (multiline) and the pattern match finished with a line ending.
1621
1622 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1623 matched in multiline mode, the following line was shown as part of the
1624 match. This seems wrong, so I have changed it.
1625
1626 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1627 the check for further matches in the same line (so they could be coloured)
1628 to overrun the end of the current line. If another match was found, it was
1629 incorrectly shown (and then shown again when found in the next line).
1630
1631 7. If pcregrep was compiled under Windows, there was a reference to the
1632 function pcregrep_exit() before it was defined. I am assuming this was
1633 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1634 reported by a user. I've moved the definition above the reference.
1635
1636
1637 Version 8.11 10-Dec-2010
1638 ------------------------
1639
1640 1. (*THEN) was not working properly if there were untried alternatives prior
1641 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1642 backtracked to try for "b" instead of moving to the next alternative branch
1643 at the same level (in this case, to look for "c"). The Perl documentation
1644 is clear that when (*THEN) is backtracked onto, it goes to the "next
1645 alternative in the innermost enclosing group".
1646
1647 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1648 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1649 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1650 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1651 (*THEN).
1652
1653 3. If \s appeared in a character class, it removed the VT character from
1654 the class, even if it had been included by some previous item, for example
1655 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1656 of \s, but is part of the POSIX "space" class.)
1657
1658 4. A partial match never returns an empty string (because you can always
1659 match an empty string at the end of the subject); however the checking for
1660 an empty string was starting at the "start of match" point. This has been
1661 changed to the "earliest inspected character" point, because the returned
1662 data for a partial match starts at this character. This means that, for
1663 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1664 (previously it gave "no match").
1665
1666 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1667 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1668 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1669 has an implication that the given string is incomplete (because a partial
1670 match is preferred over a full match). For this reason, these items now
1671 give a partial match in this situation. [Aside: previously, the one case
1672 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1673 match rather than a full match, which was wrong by the old rules, but is
1674 now correct.]
1675
1676 6. There was a bug in the handling of #-introduced comments, recognized when
1677 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1678 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1679 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1680 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1681 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1682 places in pcre_compile().
1683
1684 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1685 comments when looking ahead for named forward references to subpatterns,
1686 the only newline sequence it recognized was NL. It now handles newlines
1687 according to the set newline convention.
1688
1689 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1690 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1691 cater for a lack of strerror(). These oversights have been fixed.
1692
1693 9. Added --match-limit and --recursion-limit to pcregrep.
1694
1695 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1696
1697 11. When the -o option was used, pcregrep was setting a return code of 1, even
1698 when matches were found, and --line-buffered was not being honoured.
1699
1700 12. Added an optional parentheses number to the -o and --only-matching options
1701 of pcregrep.
1702
1703 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1704 can match an empty string. The code to do it in pcretest and pcredemo
1705 needed fixing:
1706
1707 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1708 only one byte after an empty string match just before CRLF (this case
1709 just got forgotten; "any" and "anycrlf" were OK).
1710
1711 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1712 mode when an empty string match preceded an ASCII character followed by
1713 a non-ASCII character. (The code for advancing by one character rather
1714 than one byte was nonsense.)
1715
1716 (c) The pcredemo.c sample program did not have any code at all to handle
1717 the cases when CRLF is a valid newline sequence.
1718
1719 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1720 as a starting offset was within the subject string. There is now a new
1721 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1722 negative or greater than the length of the string. In order to test this,
1723 pcretest is extended to allow the setting of negative starting offsets.
1724
1725 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1726 starting offset points to the beginning of a UTF-8 character was
1727 unnecessarily clumsy. I tidied it up.
1728
1729 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1730 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1731
1732 17. Nobody had reported that the --include_dir option, which was added in
1733 release 7.7 should have been called --include-dir (hyphen, not underscore)
1734 for compatibility with GNU grep. I have changed it to --include-dir, but
1735 left --include_dir as an undocumented synonym, and the same for
1736 --exclude-dir, though that is not available in GNU grep, at least as of
1737 release 2.5.4.
1738
1739 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1740 characters from a string of bytes) have been redefined so as not to use
1741 loops, in order to improve performance in some environments. At the same
1742 time, I abstracted some of the common code into auxiliary macros to save
1743 repetition (this should not affect the compiled code).
1744
1745 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1746 compile-time error is now given if \c is not followed by an ASCII
1747 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1748 different, and any byte value is allowed.)
1749
1750 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1751 START_OPTIMIZE option, which is now allowed at compile time - but just
1752 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1753 to pcregrep and other applications that have no direct access to PCRE
1754 options. The new /Y option in pcretest sets this option when calling
1755 pcre_compile().
1756
1757 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1758 back references. Groups containing recursive back references were forced to
1759 be atomic by that change, but in the case of named groups, the amount of
1760 memory required was incorrectly computed, leading to "Failed: internal
1761 error: code overflow". This has been fixed.
1762
1763 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1764 pcretest.c, to avoid build problems in some Borland environments.
1765
1766
1767 Version 8.10 25-Jun-2010
1768 ------------------------
1769
1770 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1771 THEN.
1772
1773 2. (*ACCEPT) was not working when inside an atomic group.
1774
1775 3. Inside a character class, \B is treated as a literal by default, but
1776 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1777 causes the error). The code is unchanged, but I tidied the documentation.
1778
1779 4. Inside a character class, PCRE always treated \R and \X as literals,
1780 whereas Perl faults them if its -w option is set. I have changed PCRE so
1781 that it faults them when PCRE_EXTRA is set.
1782
1783 5. Added support for \N, which always matches any character other than
1784 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1785
1786 6. When compiling pcregrep with newer versions of gcc which may have
1787 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1788 declared with attribute warn_unused_result" were given. Just casting the
1789 result to (void) does not stop the warnings; a more elaborate fudge is
1790 needed. I've used a macro to implement this.
1791
1792 7. Minor change to pcretest.c to avoid a compiler warning.
1793
1794 8. Added four artifical Unicode properties to help with an option to make
1795 \s etc use properties (see next item). The new properties are: Xan
1796 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1797
1798 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1799 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1800 this option. Modified pcretest to add /W to test this facility. Added
1801 REG_UCP to make it available via the POSIX interface.
1802
1803 10. Added --line-buffered to pcregrep.
1804
1805 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1806 studied, and the match started with a letter with a code point greater than
1807 127 whose first byte was different to the first byte of the other case of
1808 the letter, the other case of this starting letter was not recognized
1809 (#976).
1810
1811 12. If a pattern that was studied started with a repeated Unicode property
1812 test, for example, \p{Nd}+, there was the theoretical possibility of
1813 setting up an incorrect bitmap of starting bytes, but fortunately it could
1814 not have actually happened in practice until change 8 above was made (it
1815 added property types that matched character-matching opcodes).
1816
1817 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1818 possible starting bytes for non-anchored patterns.
1819
1820 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1821 \R, and also a number of cases that involve Unicode properties, both
1822 explicit and implicit when PCRE_UCP is set.
1823
1824 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1825 input, it could crash or give wrong results if characters with values
1826 greater than 0xc0 were present in the subject string. (Detail: it assumed
1827 UTF-8 input when processing these items.)
1828
1829 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1830 size_t is 64-bit (#991).
1831
1832 17. Added a check for running out of memory when PCRE is compiled with
1833 --disable-stack-for-recursion (#990).
1834
1835 18. If the last data line in a file for pcretest does not have a newline on
1836 the end, a newline was missing in the output.
1837
1838 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1839 less than 128) in its various bitmaps. However, there is a facility for
1840 generating tables according to the current locale when PCRE is compiled. It
1841 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1842 characters, are recognized by isspace() and therefore were getting set in
1843 these tables, and indeed these tables seem to approximate to ISO 8859. This
1844 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1845 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1846 which of course cannot start UTF-8 characters. I have changed the code so
1847 that only real ASCII characters (less than 128) and the correct starting
1848 bytes for UTF-8 encodings are set for characters greater than 127 when in
1849 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1850 altogether.)
1851
1852 20. Added the /T option to pcretest so as to be able to run tests with non-
1853 standard character tables, thus making it possible to include the tests
1854 used for 19 above in the standard set of tests.
1855
1856 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1857 reference to a subpattern the other side of a comment that contains an
1858 opening parenthesis caused either an internal compiling error, or a
1859 reference to the wrong subpattern.
1860
1861
1862 Version 8.02 19-Mar-2010
1863 ------------------------
1864
1865 1. The Unicode data tables have been updated to Unicode 5.2.0.
1866
1867 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1868 configured.
1869
1870 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1871 original author of that file, following a query about its status.
1872
1873 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1874 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1875
1876 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1877 quantifier applied to a forward-referencing subroutine call, could compile
1878 incorrect code or give the error "internal error: previously-checked
1879 referenced subpattern not found".
1880
1881 6. Both MS Visual Studio and Symbian OS have problems with initializing
1882 variables to point to external functions. For these systems, therefore,
1883 pcre_malloc etc. are now initialized to local functions that call the
1884 relevant global functions.
1885
1886 7. There were two entries missing in the vectors called coptable and poptable
1887 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1888 I've fixed the data, and added a kludgy way of testing at compile time that
1889 the lengths are correct (equal to the number of opcodes).
1890
1891 8. Following on from 7, I added a similar kludge to check the length of the
1892 eint vector in pcreposix.c.
1893
1894 9. Error texts for pcre_compile() are held as one long string to avoid too
1895 much relocation at load time. To find a text, the string is searched,
1896 counting zeros. There was no check for running off the end of the string,
1897 which could happen if a new error number was added without updating the
1898 string.
1899
1900 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1901
1902 11. \K was not working if it appeared in an atomic group or in a group that
1903 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1904 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1905 the assertion is positive, but not if it is negative.
1906
1907 12. Change 11 fortuitously reduced the size of the stack frame used in the
1908 "match()" function of pcre_exec.c by one pointer. Forthcoming
1909 implementation of support for (*MARK) will need an extra pointer on the
1910 stack; I have reserved it now, so that the stack frame size does not
1911 decrease.
1912
1913 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1914 item in branch that calls a recursion is a subroutine call - as in the
1915 second branch in the above example - was incorrectly given the compile-
1916 time error "recursive call could loop indefinitely" because pcre_compile()
1917 was not correctly checking the subroutine for matching a non-empty string.
1918
1919 14. The checks for overrunning compiling workspace could trigger after an
1920 overrun had occurred. This is a "should never occur" error, but it can be
1921 triggered by pathological patterns such as hundreds of nested parentheses.
1922 The checks now trigger 100 bytes before the end of the workspace.
1923
1924 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1925
1926
1927 Version 8.01 19-Jan-2010
1928 ------------------------
1929
1930 1. If a pattern contained a conditional subpattern with only one branch (in
1931 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1932 computed the wrong minimum data length (which is of course zero for such
1933 subpatterns). This could cause incorrect "no match" results.
1934
1935 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1936 the pattern is reset in the first branch, pcre_compile() failed with
1937 "internal error: code overflow at offset...". This happened only when
1938 the reset was to the original external option setting. (An optimization
1939 abstracts leading options settings into an external setting, which was the
1940 cause of this.)
1941
1942 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1943 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1944 assertion pattern did not match (meaning that the assertion was true), it
1945 was incorrectly treated as false if the SKIP had been reached during the
1946 matching. This also applied to assertions used as conditions.
1947
1948 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1949 assertion subpattern, including such a pattern used as a condition,
1950 unpredictable results occurred, instead of the error return
1951 PCRE_ERROR_DFA_UITEM.
1952
1953 5. The C++ GlobalReplace function was not working like Perl for the special
1954 situation when an empty string is matched. It now does the fancy magic
1955 stuff that is necessary.
1956
1957 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1958 removed. (These were left over from very, very early versions of PCRE.)
1959
1960 7. Some cosmetic changes to the code to make life easier when compiling it
1961 as part of something else:
1962
1963 (a) Change DEBUG to PCRE_DEBUG.
1964
1965 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1966 called "current" as "current_branch", to prevent a collision with the
1967 Linux macro when compiled as a kernel module.
1968
1969 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1970 prevent a collision with the Linux macro when compiled as a kernel
1971 module.
1972
1973 8. In pcre_compile() there are some checks for integer overflows that used to
1974 cast potentially large values to (double). This has been changed to that
1975 when building, a check for int64_t is made, and if it is found, it is used
1976 instead, thus avoiding the use of floating point arithmetic. (There is no
1977 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1978 double.
1979
1980 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1981 2005 (difference between two addresses compared to an unsigned value).
1982
1983 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1984 custom one, because of the following reported problem in Windows:
1985
1986 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1987 under Win32.
1988 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1989 therefore missing the function definition.
1990 - The compiler thus generates a "C" signature for the test function.
1991 - The linker fails to find the "C" function.
1992 - PCRE fails to configure if asked to do so against libbz2.
1993
1994 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1995 messages were output:
1996
1997 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1998 rerunning libtoolize, to keep the correct libtool macros in-tree.
1999 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2000
2001 I have done both of these things.
2002
2003 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2004 most of the time, it *can* run out if it is given a pattern that contains a
2005 runaway infinite recursion. I updated the discussion in the pcrestack man
2006 page.
2007
2008 13. Now that we have gone to the x.xx style of version numbers, the minor
2009 version may start with zero. Using 08 or 09 is a bad idea because users
2010 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2011 interpreted as invalid octal numbers. I've updated the previous comment in
2012 configure.ac, and also added a check that gives an error if 08 or 09 are
2013 used.
2014
2015 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2016 causing partial matching to fail when the end of the subject matched \W
2017 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2018
2019 15. There were some discrepancies between the declarations in pcre_internal.h
2020 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2021 their definitions. The declarations used "const uschar *" and the
2022 definitions used USPTR. Even though USPTR is normally defined as "const
2023 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2024 reported that: "This difference in casting confuses some C++ compilers, for
2025 example, SunCC recognizes above declarations as different functions and
2026 generates broken code for hbpcre." I have changed the declarations to use
2027 USPTR.
2028
2029 16. GNU libtool is named differently on some systems. The autogen.sh script now
2030 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2031 (FreeBSD).
2032
2033 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2034 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2035 comment: "Figure out how to create a longlong from a string: strtoll and
2036 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2037 instance, but it only takes 2 args instead of 3!"
2038
2039 18. A subtle bug concerned with back references has been fixed by a change of
2040 specification, with a corresponding code fix. A pattern such as
2041 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2042 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2043 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2044 same bug. Such groups have to be quantified to be useful, or contained
2045 inside another quantified group. (If there's no repetition, the reference
2046 can never match.) The problem arises because, having left the group and
2047 moved on to the rest of the pattern, a later failure that backtracks into
2048 the group uses the captured value from the final iteration of the group
2049 rather than the correct earlier one. I have fixed this in PCRE by forcing
2050 any group that contains a reference to itself to be an atomic group; that
2051 is, there cannot be any backtracking into it once it has completed. This is
2052 similar to recursive and subroutine calls.
2053
2054
2055 Version 8.00 19-Oct-09
2056 ----------------------
2057
2058 1. The table for translating pcre_compile() error codes into POSIX error codes
2059 was out-of-date, and there was no check on the pcre_compile() error code
2060 being within the table. This could lead to an OK return being given in
2061 error.
2062
2063 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2064 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2065 in a Windows environment.
2066
2067 3. The pcregrep --count option prints the count for each file even when it is
2068 zero, as does GNU grep. However, pcregrep was also printing all files when
2069 --files-with-matches was added. Now, when both options are given, it prints
2070 counts only for those files that have at least one match. (GNU grep just
2071 prints the file name in this circumstance, but including the count seems
2072 more useful - otherwise, why use --count?) Also ensured that the
2073 combination -clh just lists non-zero counts, with no names.
2074
2075 4. The long form of the pcregrep -F option was incorrectly implemented as
2076 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2077 but it seems right to fix it, and I didn't think it was worth preserving
2078 the old behaviour.
2079
2080 5. The command line items --regex=pattern and --regexp=pattern were not
2081 recognized by pcregrep, which required --regex pattern or --regexp pattern
2082 (with a space rather than an '='). The man page documented the '=' forms,
2083 which are compatible with GNU grep; these now work.
2084
2085 6. No libpcreposix.pc file was created for pkg-config; there was just
2086 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2087
2088 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2089 when UCP support is not needed, by modifying the Python script that
2090 generates it from Unicode data files. This should not matter if the module
2091 is correctly used as a library, but I received one complaint about 50K of
2092 unwanted data. My guess is that the person linked everything into his
2093 program rather than using a library. Anyway, it does no harm.
2094
2095 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2096 was a minimum greater than 1 for a wide character in a possessive
2097 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2098 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2099 character. Chaos in the form of incorrect output or a compiling loop could
2100 result.
2101
2102 9. The restrictions on what a pattern can contain when partial matching is
2103 requested for pcre_exec() have been removed. All patterns can now be
2104 partially matched by this function. In addition, if there are at least two
2105 slots in the offset vector, the offset of the earliest inspected character
2106 for the match and the offset of the end of the subject are set in them when
2107 PCRE_ERROR_PARTIAL is returned.
2108
2109 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2110 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2111 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2112 and may be more useful for multi-segment matching.
2113
2114 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2115 used to be given if ever the end of the subject was reached; now it is
2116 given only if matching could not proceed because another character was
2117 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2118 string "Z", which now yields "no match" instead of "partial match". In the
2119 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2120 final character ended with (*FAIL).
2121
2122 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2123 if the pattern had a "must contain" character that was already found in the
2124 earlier partial match, unless partial matching was again requested. For
2125 example, with the pattern /dog.(body)?/, the "must contain" character is
2126 "g". If the first part-match was for the string "dog", restarting with
2127 "sbody" failed. This bug has been fixed.
2128
2129 13. The string returned by pcre_dfa_exec() after a partial match has been
2130 changed so that it starts at the first inspected character rather than the
2131 first character of the match. This makes a difference only if the pattern
2132 starts with a lookbehind assertion or \b or \B (\K is not supported by
2133 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2134 matching functions compatible, and I think it's the right thing to do.
2135
2136 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2137 so that the demonstration program is easily available in environments where
2138 PCRE has not been installed from source.
2139
2140 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2141 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2142 library.
2143
2144 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2145 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2146 is not the first non-POSIX option to be added. Clearly some people find
2147 these options useful.
2148
2149 17. If a caller to the POSIX matching function regexec() passes a non-zero
2150 value for nmatch with a NULL value for pmatch, the value of
2151 nmatch is forced to zero.
2152
2153 18. RunGrepTest did not have a test for the availability of the -u option of
2154 the diff command, as RunTest does. It now checks in the same way as
2155 RunTest, and also checks for the -b option.
2156
2157 19. If an odd number of negated classes containing just a single character
2158 interposed, within parentheses, between a forward reference to a named
2159 subpattern and the definition of the subpattern, compilation crashed with
2160 an internal error, complaining that it could not find the referenced
2161 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2162 [The bug was that it was starting one character too far in when skipping
2163 over the character class, thus treating the ] as data rather than
2164 terminating the class. This meant it could skip too much.]
2165
2166 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2167 /g option in pcretest when the pattern contains \K, which makes it possible
2168 to have an empty string match not at the start, even when the pattern is
2169 anchored. Updated pcretest and pcredemo to use this option.
2170
2171 21. If the maximum number of capturing subpatterns in a recursion was greater
2172 than the maximum at the outer level, the higher number was returned, but
2173 with unset values at the outer level. The correct (outer level) value is
2174 now given.
2175
2176 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2177 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2178 make it do so. The string so far is captured, making this feature
2179 compatible with Perl.
2180
2181 23. The tests have been re-organized, adding tests 11 and 12, to make it
2182 possible to check the Perl 5.10 features against Perl 5.10.
2183
2184 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2185 pattern matches a fixed length string. PCRE did not allow this; now it
2186 does. Neither allows recursion.
2187
2188 25. I finally figured out how to implement a request to provide the minimum
2189 length of subject string that was needed in order to match a given pattern.
2190 (It was back references and recursion that I had previously got hung up
2191 on.) This code has now been added to pcre_study(); it finds a lower bound
2192 to the length of subject needed. It is not necessarily the greatest lower
2193 bound, but using it to avoid searching strings that are too short does give
2194 some useful speed-ups. The value is available to calling programs via
2195 pcre_fullinfo().
2196
2197 26. While implementing 25, I discovered to my embarrassment that pcretest had
2198 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2199 study optimizations had never been tested with that matching function.
2200 Oops. What is worse, even when it was passed study data, there was a bug in
2201 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2202 were also very few tests of studied patterns with pcre_dfa_exec().
2203
2204 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2205 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2206 on the other side of the coin, they are no longer allowed to have different
2207 names, because these cannot be distinguished in PCRE, and this has caused
2208 confusion. (This is a difference from Perl.)
2209
2210 28. When duplicate subpattern names are present (necessarily with different
2211 numbers, as required by 27 above), and a test is made by name in a
2212 conditional pattern, either for a subpattern having been matched, or for
2213 recursion in such a pattern, all the associated numbered subpatterns are
2214 tested, and the overall condition is true if the condition is true for any
2215 one of them. This is the way Perl works, and is also more like the way
2216 testing by number works.
2217
2218
2219 Version 7.9 11-Apr-09
2220 ---------------------
2221
2222 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2223 (pcretest), all targets were linked against these libraries. This included
2224 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2225 libraries. This caused unwanted dependencies to be created. This problem
2226 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2227 pcretest is linked with readline.
2228
2229 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2230 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2231 moved outside it again, because FALSE and TRUE are already defined in AIX,
2232 but BOOL is not.
2233
2234 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2235 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2236
2237 4. The pcregrep documentation said spaces were inserted as well as colons (or
2238 hyphens) following file names and line numbers when outputting matching
2239 lines. This is not true; no spaces are inserted. I have also clarified the
2240 wording for the --colour (or --color) option.
2241
2242 5. In pcregrep, when --colour was used with -o, the list of matching strings
2243 was not coloured; this is different to GNU grep, so I have changed it to be
2244 the same.
2245
2246 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2247 each matching line was coloured. Now it goes on to look for further matches
2248 of any of the test patterns, which is the same behaviour as GNU grep.
2249
2250 7. A pattern that could match an empty string could cause pcregrep to loop; it
2251 doesn't make sense to accept an empty string match in pcregrep, so I have
2252 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2253 seems to be how GNU grep behaves. [But see later change 40 for release
2254 8.33.]
2255
2256 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2257 start or after a newline", because the conditional assertion was not being
2258 correctly handled. The rule now is that both the assertion and what follows
2259 in the first alternative must satisfy the test.
2260
2261 9. If auto-callout was enabled in a pattern with a conditional group whose
2262 condition was an assertion, PCRE could crash during matching, both with
2263 pcre_exec() and pcre_dfa_exec().
2264
2265 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2266 used for matching.
2267
2268 11. Unicode property support in character classes was not working for
2269 characters (bytes) greater than 127 when not in UTF-8 mode.
2270
2271 12. Added the -M command line option to pcretest.
2272
2273 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2274
2275 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2276
2277 16. Added comments and documentation about mis-use of no_arg in the C++
2278 wrapper.
2279
2280 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2281 from Martin Jerabek that uses macro names for all relevant character and
2282 string constants.
2283
2284 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2285 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2286 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2287 these, but not everybody uses configure.
2288
2289 19. A conditional group that had only one branch was not being correctly
2290 recognized as an item that could match an empty string. This meant that an
2291 enclosing group might also not be so recognized, causing infinite looping
2292 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2293 with the subject "ab", where knowledge that the repeated group can match
2294 nothing is needed in order to break the loop.
2295
2296 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2297 exec(), but without supplying a callout function, matching went wrong.
2298
2299 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2300 leak if the size of the offset vector was greater than 30. When the vector
2301 is smaller, the saved offsets during recursion go onto a local stack
2302 vector, but for larger vectors malloc() is used. It was failing to free
2303 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2304 error, in fact).
2305
2306 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2307 heapframe that is used only when UTF-8 support is enabled. This caused no
2308 problem, but was untidy.
2309
2310 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2311 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2312 included within another project.
2313
2314 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2315 slightly modified by me:
2316
2317 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2318 not building pcregrep.
2319
2320 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2321 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2322
2323 25. Forward references, both numeric and by name, in patterns that made use of
2324 duplicate group numbers, could behave incorrectly or give incorrect errors,
2325 because when scanning forward to find the reference group, PCRE was not
2326 taking into account the duplicate group numbers. A pattern such as
2327 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2328
2329 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2330 the feature of a custom pointer more persuasive (as requested by a user).
2331
2332 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2333 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2334 pre-defined.
2335
2336 28. Added support for (*UTF8) at the start of a pattern.
2337
2338 29. Arrange for flags added by the "release type" setting in CMake to be shown
2339 in the configuration summary.
2340
2341
2342 Version 7.8 05-Sep-08
2343 ---------------------
2344
2345 1. Replaced UCP searching code with optimized version as implemented for Ad
2346 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2347 stage table and inline lookup instead of a function, giving speed ups of 2
2348 to 5 times on some simple patterns that I tested. Permission was given to
2349 distribute the MultiStage2.py script that generates the tables (it's not in
2350 the tarball, but is in the Subversion repository).
2351
2352 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2353 scripts.
2354
2355 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2356 a group with a zero qualifier. The result of the study could be incorrect,
2357 or the function might crash, depending on the pattern.
2358
2359 4. Caseless matching was not working for non-ASCII characters in back
2360 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2361 It now works when Unicode Property Support is available.
2362
2363 5. In pcretest, an escape such as \x{de} in the data was always generating
2364 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2365 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2366 truncation.
2367
2368 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2369
2370 7. Added two (int) casts to pcregrep when printing the difference of two
2371 pointers, in case they are 64-bit values.
2372
2373 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2374 test 2 if it fails.
2375
2376 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2377 and a #define of that name to empty if it is not externally set. This is to
2378 allow users of MSVC to set it if necessary.
2379
2380 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2381 the convenience functions in the pcre_get.c source file.
2382
2383 11. An option change at the start of a pattern that had top-level alternatives
2384 could cause overwriting and/or a crash. This command provoked a crash in
2385 some environments:
2386
2387 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2388
2389 This potential security problem was recorded as CVE-2008-2371.
2390
2391 12. For a pattern where the match had to start at the beginning or immediately
2392 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2393 pcre_dfa_exec() could read past the end of the passed subject if there was
2394 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2395 pcretest so that it places the subject at the end of its malloc-ed buffer.
2396
2397 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2398 exec() might read past the end of the data buffer in UTF-8 mode.
2399
2400 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2401 the data contained the byte 0x85 as part of a UTF-8 character within its
2402 first line. This applied both to normal and DFA matching.
2403
2404 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2405 /^[^d]*?$/8 failed to match "abc".
2406
2407 16. Added a missing copyright notice to pcrecpp_internal.h.
2408
2409 17. Make it more clear in the documentation that values returned from
2410 pcre_exec() in ovector are byte offsets, not character counts.
2411
2412 18. Tidied a few places to stop certain compilers from issuing warnings.
2413
2414 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2415 supplied by Stefan Weber. I made a further small update for 7.8 because
2416 there is a change of source arrangements: the pcre_searchfuncs.c module is
2417 replaced by pcre_ucd.c.
2418
2419
2420 Version 7.7 07-May-08
2421 ---------------------
2422
2423 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2424 a string to a long long, pretend we don't even have a long long." This is
2425 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2426
2427 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2428 pre-7.6 versions, which defined a global no_arg variable instead of putting
2429 it in the RE class. (See also #8 below.)
2430
2431 3. Remove a line of dead code, identified by coverity and reported by Nuno
2432 Lopes.
2433
2434 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2435
2436 (1) The include/exclude patterns were being applied to the whole pathnames
2437 of files, instead of just to the final components.
2438
2439 (2) If there was more than one level of directory, the subdirectories were
2440 skipped unless they satisfied the include/exclude conditions. This is
2441 inconsistent with GNU grep (and could even be seen as contrary to the
2442 pcregrep specification - which I improved to make it absolutely clear).
2443 The action now is always to scan all levels of directory, and just
2444 apply the include/exclude patterns to regular files.
2445
2446 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2447 --exclude_dir in the tests to avoid scanning .svn directories.
2448
2449 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2450 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2451 doesn't support NULs in patterns.
2452
2453 7. Added some missing "const"s to declarations of static tables in
2454 pcre_compile.c and pcre_dfa_exec.c.
2455
2456 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2457 caused by fix #2 above. (Subsequently also a second patch to fix the
2458 first patch. And a third patch - this was a messy problem.)
2459
2460 9. Applied Craig's patch to remove the use of push_back().
2461
2462 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2463 matching function regexec().
2464
2465 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2466 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2467 references. PCRE supports relative numbers with this syntax (I don't think
2468 Oniguruma does).
2469
2470 12. Previously, a group with a zero repeat such as (...){0} was completely
2471 omitted from the compiled regex. However, this means that if the group
2472 was called as a subroutine from elsewhere in the pattern, things went wrong
2473 (an internal error was given). Such groups are now left in the compiled
2474 pattern, with a new opcode that causes them to be skipped at execution
2475 time.
2476
2477 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2478 to the way PCRE behaves:
2479
2480 (a) A lone ] character is dis-allowed (Perl treats it as data).
2481
2482 (b) A back reference to an unmatched subpattern matches an empty string
2483 (Perl fails the current match path).
2484
2485 (c) A data ] in a character class must be notated as \] because if the
2486 first data character in a class is ], it defines an empty class. (In
2487 Perl it is not possible to have an empty class.) The empty class []
2488 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2489 The negative empty class [^] matches any one character, independently
2490 of the DOTALL setting.
2491
2492 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2493 non-existent subpattern following a character class starting with ']' and
2494 containing () gave an internal compiling error instead of "reference to
2495 non-existent subpattern". Fortunately, when the pattern did exist, the
2496 compiled code was correct. (When scanning forwards to check for the
2497 existence of the subpattern, it was treating the data ']' as terminating
2498 the class, so got the count wrong. When actually compiling, the reference
2499 was subsequently set up correctly.)
2500
2501 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2502 it was being rejected as not supported by pcre_dfa_exec(), even though
2503 other assertions are supported. I have made pcre_dfa_exec() support
2504 (*FAIL).
2505
2506 16. The implementation of 13c above involved the invention of a new opcode,
2507 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2508 cannot be changed at match time, I realized I could make a small
2509 improvement to matching performance by compiling OP_ALLANY instead of
2510 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2511 on the OP_ANY path.
2512
2513 17. Compiling pcretest on Windows with readline support failed without the
2514 following two fixes: (1) Make the unistd.h include conditional on
2515 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2516
2517 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2518 ncurses library to be included for pcretest when ReadLine support is
2519 requested, but also to allow for it to be overridden. This patch came from
2520 Daniel Bergström.
2521
2522 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2523 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2524 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2525 spotting this.
2526
2527
2528 Version 7.6 28-Jan-08
2529 ---------------------
2530
2531 1. A character class containing a very large number of characters with
2532 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2533 overflow.
2534
2535 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2536 HAVE_LONG_LONG is not defined.
2537
2538 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2539 bring them up to date and include new features. This patch includes:
2540
2541 - Fixed PH's badly added libz and libbz2 support.
2542 - Fixed a problem with static linking.
2543 - Added pcredemo. [But later removed - see 7 below.]
2544 - Fixed dftables problem and added an option.
2545 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2546 HAVE_LONG_LONG.
2547 - Added readline support for pcretest.
2548 - Added an listing of the option settings after cmake has run.
2549
2550 4. A user submitted a patch to Makefile that makes it easy to create
2551 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2552 Makefile.am that cause it to include this special target, without
2553 affecting anything else. Note that the same mingw target plus all
2554 the other distribution libraries and programs are now supported
2555 when configuring with CMake (see 6 below) instead of with
2556 Configure/Make.
2557
2558 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2559 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2560 exported in the Windows port". It has not yet been confirmed that the patch
2561 solves the problem, but it does no harm.
2562
2563 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2564 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2565 with CMake, and also correct the comment about stack recursion.
2566
2567 7. Remove the automatic building of pcredemo from the ./configure system and
2568 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2569 of a program that users should build themselves after PCRE is installed, so
2570 building it automatically is not really right. What is more, it gave
2571 trouble in some build environments.
2572
2573 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2574
2575
2576 Version 7.5 10-Jan-08
2577 ---------------------
2578
2579 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2580 values in parens when parsing an RE using the C++ wrapper."
2581
2582 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2583 Characters greater than 255 were excluded from the class instead of being
2584 included.
2585
2586 3. The same bug as (2) above applied to negated POSIX classes such as
2587 [:^space:].
2588
2589 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2590 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2591 I have changed it.
2592
2593 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2594 first named subpattern) and a construct such as (?&a) would reference the
2595 first named subpattern whose name started with "a" (in other words, the
2596 length check was missing). Both these problems are fixed. "Subpattern name
2597 expected" is now given for (?&) (a zero-length name), and this patch also
2598 makes it give the same error for \k'' (previously it complained that that
2599 was a reference to a non-existent subpattern).
2600
2601 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2602 this is right because (?- can be followed by option settings as well as by
2603 digits. I have, however, made the messages clearer.
2604
2605 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2606 than the number used in the conditional) now cause a compile-time error.
2607 This is actually not compatible with Perl, which accepts such patterns, but
2608 treats the conditional as always being FALSE (as PCRE used to), but it
2609 seems to me that giving a diagnostic is better.
2610
2611 8. Change "alphameric" to the more common word "alphanumeric" in comments
2612 and messages.
2613
2614 9. Fix two occurrences of "backslash" in comments that should have been
2615 "backspace".
2616
2617 10. Remove two redundant lines of code that can never be obeyed (their function
2618 was moved elsewhere).
2619
2620 11. The program that makes PCRE's Unicode character property table had a bug
2621 which caused it to generate incorrect table entries for sequences of
2622 characters that have the same character type, but are in different scripts.
2623 It amalgamated them into a single range, with the script of the first of
2624 them. In other words, some characters were in the wrong script. There were
2625 thirteen such cases, affecting characters in the following ranges:
2626
2627 U+002b0 - U+002c1
2628 U+0060c - U+0060d
2629 U+0061e - U+00612
2630 U+0064b - U+0065e
2631 U+0074d - U+0076d
2632 U+01800 - U+01805
2633 U+01d00 - U+01d77
2634 U+01d9b - U+01dbf
2635 U+0200b - U+0200f
2636 U+030fc - U+030fe
2637 U+03260 - U+0327f
2638 U+0fb46 - U+0fbb1
2639 U+10450 - U+1049d
2640
2641 12. The -o option (show only the matching part of a line) for pcregrep was not
2642 compatible with GNU grep in that, if there was more than one match in a
2643 line, it showed only the first of them. It now behaves in the same way as
2644 GNU grep.
2645
2646 13. If the -o and -v options were combined for pcregrep, it printed a blank
2647 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2648 does the same. The return code can be used to tell if there were any
2649 non-matching lines.
2650
2651 14. Added --file-offsets and --line-offsets to pcregrep.
2652
2653 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2654 infinitely looping recursion. The bug was that positive lookaheads were not
2655 being skipped when checking for a possible empty match (negative lookaheads
2656 and both kinds of lookbehind were skipped).
2657
2658 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2659 inclusion of <windows.h> to before rather than after the definition of
2660 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2661
2662 17. Specifying a possessive quantifier with a specific limit for a Unicode
2663 character property caused pcre_compile() to compile bad code, which led at
2664 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2665 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2666 caused the error; without that there was no problem.
2667
2668 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2669
2670 19. Added --enable-pcretest-libreadline.
2671
2672 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2673 RE::GlobalReplace(). As a result, the number of replacements returned was
2674 double what it should be. I removed one of the increments, but Craig sent a
2675 later patch that removed the other one (the right fix) and added unit tests
2676 that check the return values (which was not done before).
2677
2678 21. Several CMake things:
2679
2680 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2681 the names libpcre and libpcreposix, not just pcre and pcreposix.
2682
2683 (2) The above change means that pcretest and pcregrep are now correctly
2684 linked with the newly-built libraries, not previously installed ones.
2685
2686 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2687
2688 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2689 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2690 UTF-8 newline character). The key issue is that the pattern starts .*;
2691 this means that the match must be either at the beginning, or after a
2692 newline. The bug was in the code for advancing after a failed match and
2693 checking that the new position followed a newline. It was not taking
2694 account of UTF-8 characters correctly.
2695
2696 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2697 character classes. PCRE was not treating the sequence [:...:] as a
2698 character class unless the ... were all letters. Perl, however, seems to
2699 allow any characters between [: and :], though of course it rejects as
2700 unknown any "names" that contain non-letters, because all the known class
2701 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2702 for example, whereas PCRE did not - it did not recognize a POSIX character
2703 class. This seemed a bit dangerous, so the code has been changed to be
2704 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2705 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2706 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2707 Perl does, and where it didn't before.
2708
2709 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2710 Windows environments %n is disabled by default.
2711
2712
2713 Version 7.4 21-Sep-07
2714 ---------------------
2715
2716 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2717 means that a class such as [\s] counted as "explicit reference to CR or
2718 LF". That isn't really right - the whole point of the change was to try to
2719 help when there was an actual mention of one of the two characters. So now
2720 the change happens only if \r or \n (or a literal CR or LF) character is
2721 encountered.
2722
2723 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2724 of both had grown to the point where there were only 3 bits left.
2725 Fortunately, there was spare space in the data structure, and so I have
2726 moved the internal flags into a new 16-bit field to free up more option
2727 bits.
2728
2729 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2730 but did not set the internal JCHANGED flag - either of these is enough to
2731 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2732 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2733 start sets both bits.
2734
2735 4. Added options (at build time, compile time, exec time) to change \R from
2736 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2737
2738 5. doc/pcresyntax.html was missing from the distribution.
2739
2740 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2741 compatibility, even though it is no longer used.
2742
2743 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2744 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2745 windows.h file is present (where different names are used). [This was
2746 reversed later after testing - see 16 below.]
2747
2748 8. Changed all #include <config.h> to #include "config.h". There were also
2749 some further <pcre.h> cases that I changed to "pcre.h".
2750
2751 9. When pcregrep was used with the --colour option, it missed the line ending
2752 sequence off the lines that it output.
2753
2754 10. It was pointed out to me that arrays of string pointers cause lots of
2755 relocations when a shared library is dynamically loaded. A technique of
2756 using a single long string with a table of offsets can drastically reduce
2757 these. I have refactored PCRE in four places to do this. The result is
2758 dramatic:
2759
2760 Originally: 290
2761 After changing UCP table: 187
2762 After changing error message table: 43
2763 After changing table of "verbs" 36
2764 After changing table of Posix names 22
2765
2766 Thanks to the folks working on Gregex for glib for this insight.
2767
2768 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2769 unicode-properties was also set.
2770
2771 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2772
2773 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2774 checked only for CRLF.
2775
2776 14. Added casts to pcretest.c to avoid compiler warnings.
2777
2778 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2779
2780 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2781 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2782 entirely. This removes changes made in 7 above.
2783
2784 17. The CMake files have been updated, and there is now more information about
2785 building with CMake in the NON-UNIX-USE document.
2786
2787
2788 Version 7.3 28-Aug-07
2789 ---------------------
2790
2791 1. In the rejigging of the build system that eventually resulted in 7.1, the
2792 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2793 brackets there is not right, since it causes compilers to look for an
2794 installed pcre.h, not the version that is in the source that is being
2795 compiled (which of course may be different). I have changed it back to:
2796
2797 #include "pcre.h"
2798
2799 I have a vague recollection that the change was concerned with compiling in
2800 different directories, but in the new build system, that is taken care of
2801 by the VPATH setting the Makefile.
2802
2803 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2804 when the subject happened to end in the byte 0x85 (e.g. if the last
2805 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2806 characters but of course it shouldn't be taken as a newline when it is part
2807 of another character. The bug was that, for an unlimited repeat of . in
2808 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2809 characters when looking for a newline.
2810
2811 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2812
2813 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2814 in debug output.
2815
2816 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2817 long printing in the pcrecpp unittest when running under MinGW.
2818
2819 6. ESC_K was left out of the EBCDIC table.
2820
2821 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2822 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2823 limit also applies to "virtual nesting" when a pattern is recursive, and in
2824 this case 1000 isn't so big. I have been able to remove this limit at the
2825 expense of backing off one optimization in certain circumstances. Normally,
2826 when pcre_exec() would call its internal match() function recursively and
2827 immediately return the result unconditionally, it uses a "tail recursion"
2828 feature to save stack. However, when a subpattern that can match an empty
2829 string has an unlimited repetition quantifier, it no longer makes this
2830 optimization. That gives it a stack frame in which to save the data for
2831 checking that an empty string has been matched. Previously this was taken
2832 from the 1000-entry workspace that had been reserved. So now there is no
2833 explicit limit, but more stack is used.
2834
2835 8. Applied Daniel's patches to solve problems with the import/export magic
2836 syntax that is required for Windows, and which was going wrong for the
2837 pcreposix and pcrecpp parts of the library. These were overlooked when this
2838 problem was solved for the main library.
2839
2840 9. There were some crude static tests to avoid integer overflow when computing
2841 the size of patterns that contain repeated groups with explicit upper
2842 limits. As the maximum quantifier is 65535, the maximum group length was
2843 set at 30,000 so that the product of these two numbers did not overflow a
2844 32-bit integer. However, it turns out that people want to use groups that
2845 are longer than 30,000 bytes (though not repeat them that many times).
2846 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2847 made it possible to implement the integer overflow checks in a much more
2848 dynamic way, which I have now done. The artificial limitation on group
2849 length has been removed - we now have only the limit on the total length of
2850 the compiled pattern, which depends on the LINK_SIZE setting.
2851
2852 10. Fixed a bug in the documentation for get/copy named substring when
2853 duplicate names are permitted. If none of the named substrings are set, the
2854 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2855 empty string.
2856
2857 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2858 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2859 because the ] is interpreted as the first data character and the
2860 terminating ] is not found. PCRE has been made compatible with Perl in this
2861 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2862 cause memory overwriting.
2863
2864 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2865 string has been matched (to stop an infinite loop). It was not recognizing
2866 a conditional subpattern that could match an empty string if that
2867 subpattern was within another subpattern. For example, it looped when
2868 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2869 condition was not nested. This bug has been fixed.
2870
2871 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2872 past the start of the subject in the presence of bytes with the top bit
2873 set, for example "\x8aBCD".
2874
2875 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2876 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2877
2878 14. Optimized (?!) to (*FAIL).
2879
2880 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2881 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2882 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2883 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2884 does: it's just the validity check that is more restrictive.
2885
2886 16. Inserted checks for integer overflows during escape sequence (backslash)
2887 processing, and also fixed erroneous offset values for syntax errors during
2888 backslash processing.
2889
2890 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2891 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2892
2893 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2894 caused an overrun.
2895
2896 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2897 something other than just ASCII characters) inside a group that had an
2898 unlimited repeat caused a loop at compile time (while checking to see
2899 whether the group could match an empty string).
2900
2901 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2902 [\P{Any}] did so. (Error in the code for printing property names.)
2903
2904 21. An orphan \E inside a character class could cause a crash.
2905
2906 22. A repeated capturing bracket such as (A)? could cause a wild memory
2907 reference during compilation.
2908
2909 23. There are several functions in pcre_compile() that scan along a compiled
2910 expression for various reasons (e.g. to see if it's fixed length for look
2911 behind). There were bugs in these functions when a repeated \p or \P was
2912 present in the pattern. These operators have additional parameters compared
2913 with \d, etc, and these were not being taken into account when moving along
2914 the compiled data. Specifically:
2915
2916 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2917 length.
2918
2919 (b) An item such as \pL+ within a repeated group could cause crashes or
2920 loops.
2921
2922 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2923 "reference to non-existent subpattern" error.
2924
2925 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2926
2927 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2928 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2929
2930 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2931
2932 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2933 character were causing crashes (broken optimization).
2934
2935 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2936 \p or \P) caused a compile-time loop.
2937
2938 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2939 break. For example, the unstudied pattern [\r\n]A does not match the string
2940 "\r\nA" because change 7.0/46 below moves the current point on by two
2941 characters after failing to match at the start. However, the pattern \nA
2942 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2943 the same is true. There doesn't seem any very clean way out of this, but
2944 what I have chosen to do makes the common cases work: PCRE now takes note
2945 of whether there can be an explicit match for \r or \n anywhere in the
2946 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2947 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2948 pattern has explicit CR or LF references.
2949
2950 29. Added (*CR) etc for changing newline setting at start of pattern.
2951
2952
2953 Version 7.2 19-Jun-07
2954 ---------------------
2955
2956 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2957 which is apparently normally available under Windows.
2958
2959 2. Re-jig the pcregrep tests with different newline settings in an attempt
2960 to make them independent of the local environment's newline setting.
2961
2962 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2963
2964 4. Some of the "internals" tests were previously cut out when the link size
2965 was not 2, because the output contained actual offsets. The recent new
2966 "Z" feature of pcretest means that these can be cut out, making the tests
2967 usable with all link sizes.
2968
2969 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2970 stack recursion. This gives a massive performance boost under BSD, but just
2971 a small improvement under Linux. However, it saves one field in the frame
2972 in all cases.
2973
2974 6. Added more features from the forthcoming Perl 5.10:
2975
2976 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2977 recursion call. It refers to the nth most recently opened parentheses.
2978
2979 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2980 to be opened parentheses.
2981
2982 (c) Conditions that refer to capturing parentheses can be specified
2983 relatively, for example, (?(-2)... or (?(+3)...
2984
2985 (d) \K resets the start of the current match so that everything before
2986 is not part of it.
2987
2988 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2989
2990 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2991 reference syntax.
2992
2993 (g) (?| introduces a group in which the numbering of parentheses in each
2994 alternative starts with the same number.
2995
2996 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2997
2998 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2999 PCRE_INFO_JCHANGED.
3000
3001 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3002 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3003 for detecting groups that can match an empty string.
3004
3005 9. A pattern with a very large number of alternatives (more than several
3006 hundred) was running out of internal workspace during the pre-compile
3007 phase, where pcre_compile() figures out how much memory will be needed. A
3008 bit of new cunning has reduced the workspace needed for groups with
3009 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3010 workspace instead of running out of the 4096 that are available.
3011
3012 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3013
3014 11. Applied patch from Google to remove an optimization that didn't quite work.
3015 The report of the bug said:
3016
3017 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3018 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3019 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3020
3021 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3022 it matched the wrong number of bytes.
3023
3024
3025 Version 7.1 24-Apr-07
3026 ---------------------
3027
3028 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3029 that is more "standard", making use of automake and other Autotools. There
3030 is some re-arrangement of the files and adjustment of comments consequent
3031 on this.
3032
3033 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3034 for recursive directory scanning broke on some systems because the files
3035 are not scanned in any specific order and on different systems the order
3036 was different. A call to "sort" has been inserted into RunGrepTest for the
3037 approprate test as a short-term fix. In the longer term there may be an
3038 alternative.
3039
3040 3. I had an email from Eric Raymond about problems translating some of PCRE's
3041 man pages to HTML (despite the fact that I distribute HTML pages, some
3042 people do their own conversions for various reasons). The problems
3043 concerned the use of low-level troff macros .br and .in. I have therefore
3044 removed all such uses from the man pages (some were redundant, some could
3045 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3046 HTML has been updated to handle .nf/.fi and to complain if it encounters
3047 .br or .in.
3048
3049 4. Updated comments in configure.ac that get placed in config.h.in and also
3050 arranged for config.h to be included in the distribution, with the name
3051 config.h.generic, for the benefit of those who have to compile without
3052 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3053
3054 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3055 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3056 makevp.bat for the current PCRE, using the additional files
3057 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3058
3059 6. A Windows user reported a minor discrepancy with test 2, which turned out
3060 to be caused by a trailing space on an input line that had got lost in his
3061 copy. The trailing space was an accident, so I've just removed it.
3062
3063 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3064 that is needed.
3065
3066 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3067 as "const" (a) because they are and (b) because it helps the PHP
3068 maintainers who have recently made a script to detect big data structures
3069 in the php code that should be moved to the .rodata section. I remembered
3070 to update Builducptable as well, so it won't revert if ucptable.h is ever
3071 re-created.
3072
3073 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3074 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3075 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3076 support is not required. This saves 1.5-2K of code, which is important in
3077 some applications.
3078
3079 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3080 so as not to refer to the tables, even though these functions will never be
3081 called when UTF-8 support is disabled. Otherwise there are problems with a
3082 shared library.
3083
3084 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3085
3086 (a) It was defining its arguments as char * instead of void *.
3087
3088 (b) It was assuming that all moves were upwards in memory; this was true
3089 a long time ago when I wrote it, but is no longer the case.
3090
3091 The emulated memove() is provided for those environments that have neither
3092 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3093 is clearly not the case, as these two bugs were recently reported.
3094
3095 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3096 and Detrail to create the HTML documentation, the .txt form of the man
3097 pages, and it removes trailing spaces from listed files. It also creates
3098 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3099 case, it wraps all the #defines with #ifndefs. This script should be run
3100 before "make dist".
3101
3102 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3103 with Unicode property support.
3104
3105 (a) For a maximizing quantifier, if the two different cases of the
3106 character were of different lengths in their UTF-8 codings (there are
3107 some cases like this - I found 11), and the matching function had to
3108 back up over a mixture of the two cases, it incorrectly assumed they
3109 were both the same length.
3110
3111 (b) When PCRE was configured to use the heap rather than the stack for
3112 recursion during matching, it was not correctly preserving the data for
3113 the other case of a UTF-8 character when checking ahead for a match
3114 while processing a minimizing repeat. If the check also involved
3115 matching a wide character, but failed, corruption could cause an
3116 erroneous result when trying to check for a repeat of the original
3117 character.
3118
3119 13. Some tidying changes to the testing mechanism:
3120
3121 (a) The RunTest script now detects the internal link size and whether there
3122 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3123 values substituted by "configure". (The RunGrepTest script already did
3124 this for UTF-8.) The configure.ac script no longer substitutes the
3125 relevant variables.
3126
3127 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3128 with length and offset values. This means that the output is different
3129 for different internal link sizes. Test 2 is skipped for link sizes
3130 other than 2 because of this, bypassing the problem. Unfortunately,
3131 there was also a test in test 3 (the locale tests) that used /B and
3132 failed for link sizes other than 2. Rather than cut the whole test out,
3133 I have added a new /Z option to pcretest that replaces the length and
3134 offset values with spaces. This is now used to make test 3 independent
3135 of link size. (Test 2 will be tidied up later.)
3136
3137 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3138 segmentation fault instead of returning the appropriate error message.
3139
3140 15. In multiline mode when the newline sequence was set to "any", the pattern
3141 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3142 This doesn't seem right; it now treats the CRLF combination as the line
3143 ending, and so does not match in that case. It's only a pattern such as ^$
3144 that would hit this one: something like ^ABC$ would have failed after \r
3145 and then tried again after \r\n.
3146
3147 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3148 in an attempt to make files that differ only in their line terminators
3149 compare equal. This works on Linux.
3150
3151 17. Under certain error circumstances pcregrep might try to free random memory
3152 as it exited. This is now fixed, thanks to valgrind.
3153
3154 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3155 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3156 was because its rules for how to advance for /g after matching an empty
3157 string at the end of a line did not allow for this case. They now check for
3158 it specially.
3159
3160 20. pcretest is supposed to handle patterns and data of any length, by
3161 extending its buffers when necessary. It was getting this wrong when the
3162 buffer for a data line had to be extended.
3163
3164 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3165 CRLF as a newline sequence.
3166
3167 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3168 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3169 I have nevertheless tidied it up.
3170
3171 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3172
3173 24. Added a man page for pcre-config.
3174
3175
3176 Version 7.0 19-Dec-06
3177 ---------------------
3178
3179 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3180 moving to gcc 4.1.1.
3181
3182 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3183 sys/time.h, which is documented as needed for this function. It doesn't
3184 seem to matter on Linux, but it showed up on some releases of OS X.
3185
3186 3. It seems that there are systems where bytes whose values are greater than
3187 127 match isprint() in the "C" locale. The "C" locale should be the
3188 default when a C program starts up. In most systems, only ASCII printing
3189 characters match isprint(). This difference caused the output from pcretest
3190 to vary, making some of the tests fail. I have changed pcretest so that:
3191
3192 (a) When it is outputting text in the compiled version of a pattern, bytes
3193 other than 32-126 are always shown as hex escapes.
3194
3195 (b) When it is outputting text that is a matched part of a subject string,
3196 it does the same, unless a different locale has been set for the match
3197 (using the /L modifier). In this case, it uses isprint() to decide.
3198
3199 4. Fixed a major bug that caused incorrect computation of the amount of memory
3200 required for a compiled pattern when options that changed within the
3201 pattern affected the logic of the preliminary scan that determines the
3202 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3203 that the computed length was too small. The symptoms of this bug were
3204 either the PCRE error "internal error: code overflow" from pcre_compile(),
3205 or a glibc crash with a message such as "pcretest: free(): invalid next
3206 size (fast)". Examples of patterns that provoked this bug (shown in
3207 pcretest format) are:
3208
3209 /(?-x: )/x
3210 /(?x)(?-x: \s*#\s*)/
3211 /((?i)[\x{c0}])/8
3212 /(?i:[\x{c0}])/8
3213
3214 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3215 is now done differently.
3216
3217 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3218 wrapper classes; (b) implement a new function in the C++ scanner that is
3219 more efficient than the old way of doing things because it avoids levels of
3220 recursion in the regex matching; (c) add a paragraph to the documentation
3221 for the FullMatch() function.
3222
3223 6. The escape sequence \n was being treated as whatever was defined as
3224 "newline". Not only was this contrary to the documentation, which states
3225 that \n is character 10 (hex 0A), but it also went horribly wrong when
3226 "newline" was defined as CRLF. This has been fixed.
3227
3228 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3229 was being set to -1 for the "end of line" case (supposedly a value that no
3230 character can have). Though this value is never used (the check for end of
3231 line is "zero bytes in current character"), it caused compiler complaints.
3232 I've changed it to 0xffffffff.
3233
3234 8. In pcre_version.c, the version string was being built by a sequence of
3235 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3236 string (as it is for production releases) called a macro with an empty
3237 argument. The C standard says the result of this is undefined. The gcc
3238 compiler treats it as an empty string (which was what was wanted) but it is
3239 reported that Visual C gives an error. The source has been hacked around to
3240 avoid this problem.
3241
3242 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3243 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3244 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3245 of them did).
3246
3247 10. Originally, pcretest opened its input and output without "b"; then I was
3248 told that "b" was needed in some environments, so it was added for release
3249 5.0 to both the input and output. (It makes no difference on Unix-like
3250 systems.) Later I was told that it is wrong for the input on Windows. I've
3251 now abstracted the modes into two macros, to make it easier to fiddle with
3252 them, and removed "b" from the input mode under Windows.
3253
3254 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3255
3256 12. Added -help and --help to pcretest as an official way of being reminded
3257 of the options.
3258
3259 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3260 and pcrecpp.cc because they annoy compilers at high warning levels.
3261
3262 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3263
3264 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3265 scripts are not C programs :-) and which was not noticed because it works
3266 on Linux.
3267
3268 16. pcretest is supposed to handle any length of pattern and data line (as one
3269 line or as a continued sequence of lines) by extending its input buffer if
3270 necessary. This feature was broken for very long pattern lines, leading to
3271 a string of junk being passed to pcre_compile() if the pattern was longer
3272 than about 50K.
3273
3274 17. I have done a major re-factoring of the way pcre_compile() computes the
3275 amount of memory needed for a compiled pattern. Previously, there was code
3276 that made a preliminary scan of the pattern in order to do this. That was
3277 OK when PCRE was new, but as the facilities have expanded, it has become
3278 harder and harder to keep it in step with the real compile phase, and there
3279 have been a number of bugs (see for example, 4 above). I have now found a
3280 cunning way of running the real compile function in a "fake" mode that
3281 enables it to compute how much memory it would need, while actually only
3282 ever using a few hundred bytes of working memory and without too many
3283 tests of the mode. This should make future maintenance and development
3284 easier. A side effect of this work is that the limit of 200 on the nesting
3285 depth of parentheses has been removed (though this was never a serious
3286 limitation, I suspect). However, there is a downside: pcre_compile() now
3287 runs more slowly than before (30% or more, depending on the pattern). I
3288 hope this isn't a big issue. There is no effect on runtime performance.
3289
3290 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3291 newline (only possible for the last line of a file) and it was a
3292 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3293
3294 19. Added additional timing features to pcretest. (1) The -tm option now times
3295 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3296 separate command line item, by a number that specifies the number of
3297 repeats to use when timing. The default is 50000; this gives better
3298 precision, but takes uncomfortably long for very large patterns.
3299
3300 20. Extended pcre_study() to be more clever in cases where a branch of a
3301 subpattern has no definite first character. For example, (a*|b*)[cd] would
3302 previously give no result from pcre_study(). Now it recognizes that the
3303 first character must be a, b, c, or d.
3304
3305 21. There was an incorrect error "recursive call could loop indefinitely" if
3306 a subpattern (or the entire pattern) that was being tested for matching an
3307 empty string contained only one non-empty item after a nested subpattern.
3308 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3309 incorrectly, because the \d was being skipped in the check.
3310
3311 22. The pcretest program now has a new pattern option /B and a command line
3312 option -b, which is equivalent to adding /B to every pattern. This causes
3313 it to show the compiled bytecode, without the additional information that
3314 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3315 is the same as /B/I).
3316
3317 23. A new optimization is now able automatically to treat some sequences such
3318 as a*b as a*+b. More specifically, if something simple (such as a character
3319 or a simple class like \d) has an unlimited quantifier, and is followed by
3320 something that cannot possibly match the quantified thing, the quantifier
3321 is automatically "possessified".
3322
3323 24. A recursive reference to a subpattern whose number was greater than 39
3324 went wrong under certain circumstances in UTF-8 mode. This bug could also
3325 have affected the operation of pcre_study().
3326
3327 25. Realized that a little bit of performance could be had by replacing
3328 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3329
3330 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3331
3332 27. Possessive quantifiers such as a++ were previously implemented by turning
3333 them into atomic groups such as ($>a+). Now they have their own opcodes,
3334 which improves performance. This includes the automatically created ones
3335 from 23 above.
3336
3337 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3338 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3339 the first matched character to be a colon. This applied both to named and
3340 numbered groups.
3341
3342 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3343
3344 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3345 building PCRE on Windows easier, so I have included it in the distribution.
3346
3347 31. There is now a check in pcretest against a ridiculously large number being
3348 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3349 loop, the loop is abandoned.
3350
3351 32. Forward references to subpatterns in conditions such as (?(2)...) where
3352 subpattern 2 is defined later cause pcre_compile() to search forwards in
3353 the pattern for the relevant set of parentheses. This search went wrong
3354 when there were unescaped parentheses in a character class, parentheses
3355 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3356
3357 33. "Subroutine" calls and backreferences were previously restricted to
3358 referencing subpatterns earlier in the regex. This restriction has now
3359 been removed.
3360
3361 34. Added a number of extra features that are going to be in Perl 5.10. On the
3362 whole, these are just syntactic alternatives for features that PCRE had
3363 previously implemented using the Python syntax or my own invention. The
3364 other formats are all retained for compatibility.
3365
3366 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3367 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3368 also .NET compatible.
3369
3370 (b) A recursion or subroutine call to a named group can now be defined as
3371 (?&name) as well as (?P>name).
3372
3373 (c) A backreference to a named group can now be defined as \k<name> or
3374 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3375 5.10, are also .NET compatible.
3376
3377 (d) A conditional reference to a named group can now use the syntax
3378 (?(<name>) or (?('name') as well as (?(name).
3379
3380 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3381 groups (named and numbered) that are never evaluated inline, but can be
3382 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3383 is always false. There may be only one alternative in such a group.
3384
3385 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3386 as the simple (?(R). The condition is true only if the most recent
3387 recursion is that of the given number or name. It does not search out
3388 through the entire recursion stack.
3389
3390 (g) The escape \gN or \g{N} has been added, where N is a positive or
3391 negative number, specifying an absolute or relative reference.
3392
3393 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3394 some "unreachable code" warnings.
3395
3396 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3397 things, this adds five new scripts.
3398
3399 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3400 There were also incompatibilities regarding the handling of \Q..\E inside
3401 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3402 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3403
3404 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3405 matches an empty string, and forcibly breaks the loop. There were bugs in
3406 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3407 against aaaa the result was just "a" rather than "aaaa", for example. Two
3408 separate and independent bugs (that affected different cases) have been
3409 fixed.
3410
3411 39. Refactored the code to abolish the use of different opcodes for small
3412 capturing bracket numbers. This is a tidy that I avoided doing when I
3413 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3414 The new approach is not only tidier, it makes it possible to reduce the
3415 memory needed to fix the previous bug (38).
3416
3417 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3418 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3419 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3420 mode.
3421
3422 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3423 report.
3424
3425 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3426 copy construction and assignment in the C++ wrapper.
3427
3428 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3429 couple of bugs that could have given wrong results in the "--newline=crlf"
3430 case.
3431
3432 44. Added a number of casts and did some reorganization of signed/unsigned int
3433 variables following suggestions from Dair Grant. Also renamed the variable
3434 "this" as "item" because it is a C++ keyword.
3435
3436 45. Arranged for dftables to add
3437
3438 #include "pcre_internal.h"
3439
3440 to pcre_chartables.c because without it, gcc 4.x may remove the array
3441 definition from the final binary if PCRE is built into a static library and
3442 dead code stripping is activated.
3443
3444 46. For an unanchored pattern, if a match attempt fails at the start of a
3445 newline sequence, and the newline setting is CRLF or ANY, and the next two
3446 characters are CRLF, advance by two characters instead of one.
3447
3448
3449 Version 6.7 04-Jul-06
3450 ---------------------
3451
3452 1. In order to handle tests when input lines are enormously long, pcretest has
3453 been re-factored so that it automatically extends its buffers when
3454 necessary. The code is crude, but this _is_ just a test program. The
3455 default size has been increased from 32K to 50K.
3456
3457 2. The code in pcre_study() was using the value of the re argument before
3458 testing it for NULL. (Of course, in any sensible call of the function, it
3459 won't be NULL.)
3460
3461 3. The memmove() emulation function in pcre_internal.h, which is used on
3462 systems that lack both memmove() and bcopy() - that is, hardly ever -
3463 was missing a "static" storage class specifier.
3464
3465 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3466 containing an extended class (one that cannot be represented by a bitmap
3467 because it contains high-valued characters or Unicode property items, e.g.
3468 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3469 pattern, but PCRE should not loop if you do not (it no longer does).
3470 [Detail: two cases were found: (a) a repeated subpattern containing an
3471 extended class; (b) a recursive reference to a subpattern that followed a
3472 previous extended class. It wasn't skipping over the extended class
3473 correctly when UTF-8 mode was not set.]
3474
3475 5. A negated single-character class was not being recognized as fixed-length
3476 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3477 compile error "lookbehind assertion is not fixed length".
3478
3479 6. The RunPerlTest auxiliary script was showing an unexpected difference
3480 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3481 write a Perl script that can interpret lines of an input file either as
3482 byte characters or as UTF-8, which is what "perltest" was being required to
3483 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3484 can't do is switch easily at run time between having the "use utf8;" pragma
3485 or not. In the end, I fudged it by using the RunPerlTest script to insert
3486 "use utf8;" explicitly for the UTF-8 tests.
3487
3488 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3489 the end of the subject string, contrary to the documentation and to what
3490 Perl does. This was true of both matching functions. Now it matches only at
3491 the start of the subject and immediately after *internal* newlines.
3492
3493 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3494 a pointer to an int instead of a pointer to an unsigned long int. This
3495 caused problems on 64-bit systems.
3496
3497 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3498 instance of the 'standard' template library not being so standard".
3499
3500 10. There was no check on the number of named subpatterns nor the maximum
3501 length of a subpattern name. The product of these values is used to compute
3502 the size of the memory block for a compiled pattern. By supplying a very
3503 long subpattern name and a large number of named subpatterns, the size
3504 computation could be caused to overflow. This is now prevented by limiting
3505 the length of names to 32 characters, and the number of named subpatterns
3506 to 10,000.
3507
3508 11. Subpatterns that are repeated with specific counts have to be replicated in
3509 the compiled pattern. The size of memory for this was computed from the
3510 length of the subpattern and the repeat count. The latter is limited to
3511 65535, but there was no limit on the former, meaning that integer overflow
3512 could in principle occur. The compiled length of a repeated subpattern is
3513 now limited to 30,000 bytes in order to prevent this.
3514
3515 12. Added the optional facility to have named substrings with the same name.
3516
3517 13. Added the ability to use a named substring as a condition, using the
3518 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3519 are numbers (not recommended). Forward references are permitted.
3520
3521 14. Added forward references in named backreferences (if you see what I mean).
3522
3523 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3524 pattern could run off the end of the subject. For example, the pattern
3525 "(?s)(.{1,5})"8 did this with the subject "ab".
3526
3527 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3528 PCRE_CASELESS was set when matching characters that were quantified with ?
3529 or *.
3530
3531 17. A character class other than a single negated character that had a minimum
3532 but no maximum quantifier - for example [ab]{6,} - was not handled
3533 correctly by pce_dfa_exec(). It would match only one character.
3534
3535 18. A valid (though odd) pattern that looked like a POSIX character
3536 class but used an invalid character after [ (for example [[,abc,]]) caused
3537 pcre_compile() to give the error "Failed: internal error: code overflow" or
3538 in some cases to crash with a glibc free() error. This could even happen if
3539 the pattern terminated after [[ but there just happened to be a sequence of
3540 letters, a binary zero, and a closing ] in the memory that followed.
3541
3542 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3543 over the years. Originally (before any Unicode support), just the bottom 8
3544 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3545 output from "man perlunicode" includes this:
3546
3547 The regular expression compiler produces polymorphic opcodes. That
3548 is, the pattern adapts to the data and automatically switches to
3549 the Unicode character scheme when presented with Unicode data--or
3550 instead uses a traditional byte scheme when presented with byte
3551 data.
3552
3553 Sadly, a wide octal escape does not cause a switch, and in a string with
3554 no other multibyte characters, these octal escapes are treated as before.
3555 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3556 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3557 Unicode string.
3558
3559 I have not perpetrated such confusion in PCRE. Up till now, it took just
3560 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3561 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3562 translate to the appropriate multibyte character.
3563
3564 29. Applied some refactoring to reduce the number of warnings from Microsoft
3565 and Borland compilers. This has included removing the fudge introduced
3566 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3567 a warning about an unused variable.
3568
3569 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3570 characters since release 4.0, because Perl (from release 5.004) does not.
3571 [Or at least, is documented not to: some releases seem to be in conflict
3572 with the documentation.] However, when a pattern was studied with
3573 pcre_study() and all its branches started with \s, PCRE still included VT
3574 as a possible starting character. Of course, this did no harm; it just
3575 caused an unnecessary match attempt.
3576
3577 22. Removed a now-redundant internal flag bit that recorded the fact that case
3578 dependency changed within the pattern. This was once needed for "required
3579 byte" processing, but is no longer used. This recovers a now-scarce options
3580 bit. Also moved the least significant internal flag bit to the most-
3581 significant bit of the word, which was not previously used (hangover from
3582 the days when it was an int rather than a uint) to free up another bit for
3583 the future.
3584
3585 23. Added support for CRLF line endings as well as CR and LF. As well as the
3586 default being selectable at build time, it can now be changed at runtime
3587 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3588 specify that it is scanning data with non-default line endings.
3589
3590 24. Changed the definition of CXXLINK to make it agree with the definition of
3591 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3592
3593 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3594 recursions. This makes a big different to stack usage for some patterns.
3595
3596 26. If a subpattern containing a named recursion or subroutine reference such
3597 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3598 the space required for the compiled pattern went wrong and gave too small a
3599 value. Depending on the environment, this could lead to "Failed: internal
3600 error: code overflow at offset 49" or "glibc detected double free or
3601 corruption" errors.
3602
3603 27. Applied patches from Google (a) to support the new newline modes and (b) to
3604 advance over multibyte UTF-8 characters in GlobalReplace.
3605
3606 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3607 difference for some implementation of PCRE in some Windows version.
3608
3609 29. Added some extra testing facilities to pcretest:
3610
3611 \q<number> in a data line sets the "match limit" value
3612 \Q<number> in a data line sets the "match recursion limt" value
3613 -S <number> sets the stack size, where <number> is in megabytes
3614
3615 The -S option isn't available for Windows.
3616
3617
3618 Version 6.6 06-Feb-06
3619 ---------------------
3620
3621 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3622 in pcreposix.h. I have copied the definition from pcre.h.
3623
3624 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3625 because pcre.h is no longer a built file.
3626
3627 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3628 not normally included in the compiled code.
3629
3630
3631 Version 6.5 01-Feb-06
3632 ---------------------
3633
3634 1. When using the partial match feature with pcre_dfa_exec(), it was not
3635 anchoring the second and subsequent partial matches at the new starting
3636 point. This could lead to incorrect results. For example, with the pattern
3637 /1234/, partially matching against "123" and then "a4" gave a match.
3638
3639 2. Changes to pcregrep:
3640
3641 (a) All non-match returns from pcre_exec() were being treated as failures
3642 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3643 error message is output. Some extra information is given for the
3644 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3645 probably the only errors that are likely to be caused by users (by
3646 specifying a regex that has nested indefinite repeats, for instance).
3647 If there are more than 20 of these errors, pcregrep is abandoned.
3648
3649 (b) A binary zero was treated as data while matching, but terminated the
3650 output line if it was written out. This has been fixed: binary zeroes
3651 are now no different to any other data bytes.
3652
3653 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3654 used to set a locale for matching. The --locale=xxxx long option has
3655 been added (no short equivalent) to specify a locale explicitly on the
3656 pcregrep command, overriding the environment variables.
3657
3658 (d) When -B was used with -n, some line numbers in the output were one less
3659 than they should have been.
3660
3661 (e) Added the -o (--only-matching) option.
3662
3663 (f) If -A or -C was used with -c (count only), some lines of context were
3664 accidentally printed for the final match.
3665
3666 (g) Added the -H (--with-filename) option.
3667
3668 (h) The combination of options -rh failed to suppress file names for files
3669 that were found from directory arguments.
3670
3671 (i) Added the -D (--devices) and -d (--directories) options.
3672
3673 (j) Added the -F (--fixed-strings) option.
3674
3675 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3676
3677 (l) Added the --colo(u)r option.
3678
3679 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3680 is not present by default.
3681
3682 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3683 items such as (?R) or (?1), when the recursion could match a number of
3684 alternatives. If it matched one of the alternatives, but subsequently,
3685 outside the recursion, there was a failure, the code tried to back up into
3686 the recursion. However, because of the way PCRE is implemented, this is not
3687 possible, and the result was an incorrect result from the match.
3688
3689 In order to prevent this happening, the specification of recursion has
3690 been changed so that all such subpatterns are automatically treated as
3691 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3692
3693 4. I had overlooked the fact that, in some locales, there are characters for
3694 which isalpha() is true but neither isupper() nor islower() are true. In
3695 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3696 and ordfeminine) are like this. This affected the treatment of \w and \W
3697 when they appeared in character classes, but not when they appeared outside
3698 a character class. The bit map for "word" characters is now created
3699 separately from the results of isalnum() instead of just taking it from the
3700 upper, lower, and digit maps. (Plus the underscore character, of course.)
3701
3702 5. The above bug also affected the handling of POSIX character classes such as
3703 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3704 permanent tables. Instead, the bit maps for such a class were previously
3705 created as the appropriate unions of the upper, lower, and digit bitmaps.
3706 Now they are created by subtraction from the [[:word:]] class, which has
3707 its own bitmap.
3708
3709 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3710 It is created by subtracting the vertical space characters (\x09, \x0a,
3711 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3712 subtraction was done in the overall bitmap for a character class, meaning
3713 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3714 be recognized. This bug has been fixed.
3715
3716 7. Patches from the folks at Google:
3717
3718 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3719 real life, but is still worth protecting against".
3720
3721 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3722 regular expressions".
3723
3724 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3725 have it.
3726
3727 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3728 "configure" and the latter not, in order to fix a problem somebody had
3729 with compiling the Arg class on HP-UX.
3730
3731 (e) Improve the error-handling of the C++ wrapper a little bit.
3732
3733 (f) New tests for checking recursion limiting.
3734
3735 8. The pcre_memmove() function, which is used only if the environment does not
3736 have a standard memmove() function (and is therefore rarely compiled),
3737 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3738 returning a result (though PCRE never actually uses the result).
3739
3740 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3741 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3742 returned instead of calling malloc() with an overflowing number that would
3743 most likely cause subsequent chaos.
3744
3745 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3746
3747 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3748 with this option is matched, the nmatch and pmatch options of regexec() are
3749 ignored.
3750
3751 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3752 provided in case anyone wants to the the POSIX interface with UTF-8
3753 strings.
3754
3755 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3756 C++ linking (needed for some HP-UX environments).
3757
3758 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3759 (unused parameter) and in the pcre_printint() function (omitted "default"
3760 switch label when the default is to do nothing).
3761
3762 15. Added some code to make it possible, when PCRE is compiled as a C++
3763 library, to replace subject pointers for pcre_exec() with a smart pointer
3764 class, thus making it possible to process discontinuous strings.
3765
3766 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3767 much the same function. They were added by different people who were trying
3768 to make PCRE easy to compile on non-Unix systems. It has been suggested
3769 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3770 for compiling on Windows systems. I have therefore replaced it with
3771 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3772 defaults to "extern" for C or "extern C" for C++, which works fine on
3773 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3774 SCOPE with something explicit in config.h. In addition:
3775
3776 (a) pcreposix.h still had just "extern" instead of either of these macros;
3777 I have replaced it with PCRE_DATA_SCOPE.
3778
3779 (b) Functions such as _pcre_xclass(), which are internal to the library,
3780 but external in the C sense, all had PCRE_EXPORT in their definitions.
3781 This is apparently wrong for the Windows case, so I have removed it.
3782 (It makes no difference on Unix-like systems.)
3783
3784 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3785 of recursive calls to match(). This is different to MATCH_LIMIT because
3786 that limits the total number of calls to match(), not all of which increase
3787 the depth of recursion. Limiting the recursion depth limits the amount of
3788 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3789 when PCRE is compiled, and changed at run time. A patch from Google adds
3790 this functionality to the C++ interface.
3791
3792 18. Changes to the handling of Unicode character properties:
3793
3794 (a) Updated the table to Unicode 4.1.0.
3795
3796 (b) Recognize characters that are not in the table as "Cn" (undefined).
3797
3798 (c) I revised the way the table is implemented to a much improved format
3799 which includes recognition of ranges. It now supports the ranges that
3800 are defined in UnicodeData.txt, and it also amalgamates other
3801 characters into ranges. This has reduced the number of entries in the
3802 table from around 16,000 to around 3,000, thus reducing its size
3803 considerably. I realized I did not need to use a tree structure after
3804 all - a binary chop search is just as efficient. Having reduced the
3805 number of entries, I extended their size from 6 bytes to 8 bytes to
3806 allow for more data.
3807
3808 (d) Added support for Unicode script names via properties such as \p{Han}.
3809
3810 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3811 matching that character.
3812
3813 20. When matching a repeated Unicode property with a minimum greater than zero,
3814 (for example \pL{2,}), PCRE could look past the end of the subject if it
3815 reached it while seeking the minimum number of characters. This could
3816 happen only if some of the characters were more than one byte long, because
3817 there is a check for at least the minimum number of bytes.
3818
3819 21. Refactored the implementation of \p and \P so as to be more general, to
3820 allow for more different types of property in future. This has changed the
3821 compiled form incompatibly. Anybody with saved compiled patterns that use
3822 \p or \P will have to recompile them.
3823
3824 22. Added "Any" and "L&" to the supported property types.
3825
3826 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3827 but give a compile time error if the value is greater than 0xff.
3828
3829 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3830 accidentally not being installed or uninstalled.
3831
3832 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3833 made were to insert the current release number. This seemed silly, because
3834 it made things harder for people building PCRE on systems that don't run
3835 "configure". I have turned pcre.h into a distributed file, no longer built
3836 by "configure", with the version identification directly included. There is
3837 no longer a pcre.h.in file.
3838
3839 However, this change necessitated a change to the pcre-config script as
3840 well. It is built from pcre-config.in, and one of the substitutions was the
3841 release number. I have updated configure.ac so that ./configure now finds
3842 the release number by grepping pcre.h.
3843
3844 26. Added the ability to run the tests under valgrind.
3845
3846
3847 Version 6.4 05-Sep-05
3848 ---------------------
3849
3850 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3851 "--" to be printed when multiple files were scanned, even when none of the
3852 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3853 consider it to be a bug, and have restored the previous behaviour.
3854
3855 2. A couple of code tidies to get rid of compiler warnings.
3856
3857 3. The pcretest program used to cheat by referring to symbols in the library
3858 whose names begin with _pcre_. These are internal symbols that are not
3859 really supposed to be visible externally, and in some environments it is
3860 possible to suppress them. The cheating is now confined to including
3861 certain files from the library's source, which is a bit cleaner.
3862
3863 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3864 file's purpose clearer.
3865
3866 5. Reorganized pcre_ucp_findchar().
3867
3868
3869 Version 6.3 15-Aug-05
3870 ---------------------
3871
3872 1. The file libpcre.pc.in did not have general read permission in the tarball.
3873
3874 2. There were some problems when building without C++ support:
3875
3876 (a) If C++ support was not built, "make install" and "make test" still
3877 tried to test it.
3878
3879 (b) There were problems when the value of CXX was explicitly set. Some
3880 changes have been made to try to fix these, and ...
3881
3882 (c) --disable-cpp can now be used to explicitly disable C++ support.
3883
3884 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3885 backslash in a target when C++ was disabled. This confuses some
3886 versions of "make", apparently. Using an intermediate variable solves
3887 this. (Same for CPP_LOBJ.)
3888
3889 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3890 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3891 necessary on certain architectures.
3892
3893 4. Added a setting of -export-symbols-regex to the link command to remove
3894 those symbols that are exported in the C sense, but actually are local
3895 within the library, and not documented. Their names all begin with
3896 "_pcre_". This is not a perfect job, because (a) we have to except some
3897 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3898 available (and never for static libraries). I have made a note to try to
3899 find a way round (a) in the future.
3900
3901
3902 Version 6.2 01-Aug-05
3903 ---------------------
3904
3905 1. There was no test for integer overflow of quantifier values. A construction
3906 such as {1111111111111111} would give undefined results. What is worse, if
3907 a minimum quantifier for a parenthesized subpattern overflowed and became
3908 negative, the calculation of the memory size went wrong. This could have
3909 led to memory overwriting.
3910
3911 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3912
3913 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3914 operating environments where this matters.
3915
3916 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3917 PCRE options from within the C++ wrapper.
3918
3919 5. Named capturing subpatterns were not being correctly counted when a pattern
3920 was compiled. This caused two problems: (a) If there were more than 100
3921 such subpatterns, the calculation of the memory needed for the whole
3922 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3923 back references of the form \12, where the number was greater than 9, were
3924 not recognized as back references, even though there were sufficient
3925 previous subpatterns.
3926
3927 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3928 versions of gcc, e.g. 2.95.4.
3929
3930
3931 Version 6.1 21-Jun-05
3932 ---------------------
3933
3934 1. There was one reference to the variable "posix" in pcretest.c that was not
3935 surrounded by "#if !defined NOPOSIX".
3936
3937 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3938 the cross-check on the old pcre_info() function, for the benefit of the
3939 cut-down version of PCRE that is currently imported into Exim.
3940
3941 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3942 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3943 patterns that start (?i) but I don't think that matters. The use of (?i) is
3944 just an example; this all applies to the other options as well.
3945
3946 4. Since libtool seems to echo the compile commands it is issuing, the output
3947 from "make" can be reduced a bit by putting "@" in front of each libtool
3948 compile command.
3949
3950 5. Patch from the folks at Google for configure.in to be a bit more thorough
3951 in checking for a suitable C++ installation before trying to compile the
3952 C++ stuff. This should fix a reported problem when a compiler was present,
3953 but no suitable headers.
3954
3955 6. The man pages all had just "PCRE" as their title. I have changed them to
3956 be the relevant file name. I have also arranged that these names are
3957 retained in the file doc/pcre.txt, which is a concatenation in text format
3958 of all the man pages except the little individual ones for each function.
3959
3960 7. The NON-UNIX-USE file had not been updated for the different set of source
3961 files that come with release 6. I also added a few comments about the C++
3962 wrapper.
3963
3964
3965 Version 6.0 07-Jun-05
3966 ---------------------
3967
3968 1. Some minor internal re-organization to help with my DFA experiments.
3969
3970 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3971 didn't matter for the library itself when fully configured, but did matter
3972 when compiling without UCP support, or within Exim, where the ucp files are
3973 not imported.
3974
3975 3. Refactoring of the library code to split up the various functions into
3976 different source modules. The addition of the new DFA matching code (see
3977 below) to a single monolithic source would have made it really too
3978 unwieldy, quite apart from causing all the code to be include in a
3979 statically linked application, when only some functions are used. This is
3980 relevant even without the DFA addition now that patterns can be compiled in
3981 one application and matched in another.
3982
3983 The downside of splitting up is that there have to be some external
3984 functions and data tables that are used internally in different modules of
3985 the library but which are not part of the API. These have all had their
3986 names changed to start with "_pcre_" so that they are unlikely to clash
3987 with other external names.
3988
3989 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3990 a different (DFA) algorithm. Although it is slower than the original
3991 function, it does have some advantages for certain types of matching
3992 problem.
3993
3994 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3995 including restarting after a partial match.
3996
3997 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3998 defined when compiling for Windows was sent to me. I have put it into the
3999 code, though I have no means of testing or verifying it.
4000
4001 7. Added the pcre_refcount() auxiliary function.
4002
4003 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4004 match before or at the first newline in the subject string. In pcretest,
4005 the /f option on a pattern can be used to set this.
4006
4007 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4008 would behave wrongly. This has been present in PCRE since release 4.0.
4009
4010 10. A number of changes to the pcregrep command:
4011
4012 (a) Refactored how -x works; insert ^(...)$ instead of setting
4013 PCRE_ANCHORED and checking the length, in preparation for adding
4014 something similar for -w.
4015
4016 (b) Added the -w (match as a word) option.
4017
4018 (c) Refactored the way lines are read and buffered so as to have more
4019 than one at a time available.
4020
4021 (d) Implemented a pcregrep test script.
4022
4023 (e) Added the -M (multiline match) option. This allows patterns to match
4024 over several lines of the subject. The buffering ensures that at least
4025 8K, or the rest of the document (whichever is the shorter) is available
4026 for matching (and similarly the previous 8K for lookbehind assertions).
4027
4028 (f) Changed the --help output so that it now says
4029
4030 -w, --word-regex(p)
4031
4032 instead of two lines, one with "regex" and the other with "regexp"
4033 because that confused at least one person since the short forms are the
4034 same. (This required a bit of code, as the output is generated
4035 automatically from a table. It wasn't just a text change.)
4036
4037 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4038 option but starts with a hyphen. Could be a pattern or a path name
4039 starting with a hyphen, for instance.
4040
4041 (h) "-" can be given as a file name to represent stdin.
4042
4043 (i) When file names are being printed, "(standard input)" is used for
4044 the standard input, for compatibility with GNU grep. Previously
4045 "<stdin>" was used.
4046
4047 (j) The option --label=xxx can be used to supply a name to be used for
4048 stdin when file names are being printed. There is no short form.
4049
4050 (k) Re-factored the options decoding logic because we are going to add
4051 two more options that take data. Such options can now be given in four
4052 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4053
4054 (l) Added the -A, -B, and -C options for requesting that lines of context
4055 around matches be printed.
4056
4057 (m) Added the -L option to print the names of files that do not contain
4058 any matching lines, that is, the complement of -l.
4059
4060 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4061 continue to scan other files.
4062
4063 (o) The -s option was incorrectly implemented. For compatibility with other
4064 greps, it now suppresses the error message for a non-existent or non-
4065 accessible file (but not the return code). There is a new option called
4066 -q that suppresses the output of matching lines, which was what -s was
4067 previously doing.
4068
4069 (p) Added --include and --exclude options to specify files for inclusion
4070 and exclusion when recursing.
4071
4072 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4073 Hopefully, it now does.
4074
4075 12. Missing cast in pcre_study().
4076
4077 13. Added an "uninstall" target to the makefile.
4078
4079 14. Replaced "extern" in the function prototypes in Makefile.in with
4080 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4081 world, but is set differently for Windows.
4082
4083 15. Added a second compiling function called pcre_compile2(). The only
4084 difference is that it has an extra argument, which is a pointer to an
4085 integer error code. When there is a compile-time failure, this is set
4086 non-zero, in addition to the error test pointer being set to point to an
4087 error message. The new argument may be NULL if no error number is required
4088 (but then you may as well call pcre_compile(), which is now just a
4089 wrapper). This facility is provided because some applications need a
4090 numeric error indication, but it has also enabled me to tidy up the way
4091 compile-time errors are handled in the POSIX wrapper.
4092
4093 16. Added VPATH=.libs to the makefile; this should help when building with one
4094 prefix path and installing with another. (Or so I'm told by someone who
4095 knows more about this stuff than I do.)
4096
4097 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4098 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4099 match everything, including newlines. This is not POSIX-compatible, but
4100 somebody wanted the feature. From pcretest it can be activated by using
4101 both the P and the s flags.
4102
4103 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4104
4105 19. libpcre.pc was being incorrectly installed as executable.
4106
4107 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4108 it now also looks for '\r' so that it will work unmodified on Windows.
4109
4110 21. Added Google's contributed C++ wrapper to the distribution.
4111
4112 22. Added some untidy missing memory free() calls in pcretest, to keep
4113 Electric Fence happy when testing.
4114
4115
4116
4117 Version 5.0 13-Sep-04
4118 ---------------------
4119
4120 1. Internal change: literal characters are no longer packed up into items
4121 containing multiple characters in a single byte-string. Each character
4122 is now matched using a separate opcode. However, there may be more than one
4123 byte in the character in UTF-8 mode.
4124
4125 2. The pcre_callout_block structure has two new fields: pattern_position and
4126 next_item_length. These contain the offset in the pattern to the next match
4127 item, and its length, respectively.
4128
4129 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4130 insertion of callouts before each pattern item. Added the /C option to
4131 pcretest to make use of this.
4132
4133 4. On the advice of a Windows user, the lines
4134
4135 #if defined(_WIN32) || defined(WIN32)
4136 _setmode( _fileno( stdout ), 0x8000 );
4137 #endif /* defined(_WIN32) || defined(WIN32) */
4138
4139 have been added to the source of pcretest. This apparently does useful
4140 magic in relation to line terminators.
4141
4142 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4143 for the benefit of those environments where the "b" makes a difference.
4144
4145 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4146 to know about it. I have put a hack into configure.in that adds in code
4147 to set GCC=yes if CC=icc. This seems to end up at a point in the
4148 generated configure script that is early enough to affect the setting of
4149 compiler options, which is what is needed, but I have no means of testing
4150 whether it really works. (The user who reported this had patched the
4151 generated configure script, which of course I cannot do.)
4152
4153 LATER: After change 22 below (new libtool files), the configure script
4154 seems to know about icc (and also ecc). Therefore, I have commented out
4155 this hack in configure.in.
4156
4157 7. Added support for pkg-config (2 patches were sent in).
4158
4159 8. Negated POSIX character classes that used a combination of internal tables
4160 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4161 [[:^ascii]]. Typically, they would match almost any characters. The other
4162 POSIX classes were not broken in this way.
4163
4164 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4165 to find the match, as PCRE was deluded into thinking that the match had to
4166 start at the start point or following a newline. The same bug applied to
4167 patterns with negative forward assertions or any backward assertions
4168 preceding ".*" at the start, unless the pattern required a fixed first
4169 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4170
4171 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4172 starting at the last subject character, bytes beyond the end of the subject
4173 string were read.
4174
4175 11. Renamed the variable "class" as "classbits" to make life easier for C++
4176 users. (Previously there was a macro definition, but it apparently wasn't
4177 enough.)
4178
4179 12. Added the new field "tables" to the extra data so that tables can be passed
4180 in at exec time, or the internal tables can be re-selected. This allows
4181 a compiled regex to be saved and re-used at a later time by a different
4182 program that might have everything at different addresses.
4183
4184 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4185 -R library as well as a -L library.
4186
4187 14. The debugging options of pcretest (-d on the command line or D on a
4188 pattern) showed incorrect output for anything following an extended class
4189 that contained multibyte characters and which was followed by a quantifier.
4190
4191 15. Added optional support for general category Unicode character properties
4192 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4193 support. It adds about 90K to the size of the library. The meanings of the
4194 inbuilt class escapes such as \d and \s have NOT been changed.
4195
4196 16. Updated pcredemo.c to include calls to free() to release the memory for the
4197 compiled pattern.
4198
4199 17. The generated file chartables.c was being created in the source directory
4200 instead of in the building directory. This caused the build to fail if the
4201 source directory was different from the building directory, and was
4202 read-only.
4203
4204 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4205 file. No doubt somebody will tell me if they don't make sense... Also added
4206 Dan Mooney's comments about building on OpenVMS.
4207
4208 19. Added support for partial matching via the PCRE_PARTIAL option for
4209 pcre_exec() and the \P data escape in pcretest.
4210
4211 20. Extended pcretest with 3 new pattern features:
4212
4213 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4214 write the compiled pattern to the file whose name is "rest-of-line".
4215 This is a straight binary dump of the data, with the saved pointer to
4216 the character tables forced to be NULL. The study data, if any, is
4217 written too. After writing, pcretest reads a new pattern.
4218
4219 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4220 compiled pattern from the given file. There must not be any
4221 occurrences of "<" in the file name (pretty unlikely); if there are,
4222 pcretest will instead treat the initial "<" as a pattern delimiter.
4223 After reading in the pattern, pcretest goes on to read data lines as
4224 usual.
4225
4226 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4227 and 16-bit fields in a compiled pattern, to simulate a pattern that
4228 was compiled on a host of opposite endianness.
4229
4230 21. The pcre-exec() function can now cope with patterns that were compiled on
4231 hosts of opposite endianness, with this restriction:
4232
4233 As for any compiled expression that is saved and used later, the tables
4234 pointer field cannot be preserved; the extra_data field in the arguments
4235 to pcre_exec() should be used to pass in a tables address if a value
4236 other than the default internal tables were used at compile time.
4237
4238 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4239 now diagnosed as an error. Previously, most of the time, a negative number
4240 would have been treated as zero, but if in addition "ovector" was passed as
4241 NULL, a crash could occur.
4242
4243 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4244 new versions from the libtool 1.5 distribution (the last one is a copy of
4245 a file called libtool.m4). This seems to have fixed the need to patch
4246 "configure" to support Darwin 1.3 (which I used to do). However, I still
4247 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4248 workstation).
4249
4250 24. Changed the PCRE licence to be the more standard "BSD" licence.
4251
4252
4253 Version 4.5 01-Dec-03
4254 ---------------------
4255
4256 1. There has been some re-arrangement of the code for the match() function so
4257 that it can be compiled in a version that does not call itself recursively.
4258 Instead, it keeps those local variables that need separate instances for
4259 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4260 needs to "recurse". Keeping track of where control must go is done by means
4261 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4262 hide most of the details from the main code, and operates only if
4263 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4264 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4265 operating.
4266
4267 To make it easier for callers to provide specially tailored get/free
4268 functions for this usage, two new functions, pcre_stack_malloc, and
4269 pcre_stack_free, are used. They are always called in strict stacking order,
4270 and the size of block requested is always the same.
4271
4272 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4273 PCRE has been compiled to use the stack or the heap for recursion. The
4274 -C option of pcretest uses this to show which version is compiled.
4275
4276 A new data escape \S, is added to pcretest; it causes the amounts of store
4277 obtained and freed by both kinds of malloc/free at match time to be added
4278 to the output.
4279
4280 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4281 what's available on my current Linux desktop machine.
4282
4283 3. When matching a UTF-8 string, the test for a valid string at the start has
4284 been extended. If start_offset is not zero, PCRE now checks that it points
4285 to a byte that is the start of a UTF-8 character. If not, it returns
4286 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4287 this is necessary because there may be backward assertions in the pattern.
4288 When matching the same subject several times, it may save resources to use
4289 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4290
4291 4. The code for checking the validity of UTF-8 strings has been tightened so
4292 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4293 containing "overlong sequences".
4294
4295 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4296 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4297 should have been "&", but it just so happened that all the cases this let
4298 through by mistake were picked up later in the function.
4299
4300 6. I had used a variable called "isblank" - this is a C99 function, causing
4301 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4302
4303 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4304 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4305 so the version is shown; (c) stop "make test" echoing "./RunTest".
4306
4307 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4308
4309 9. The prototype for memmove() for systems that don't have it was using
4310 size_t, but the inclusion of the header that defines size_t was later. I've
4311 moved the #includes for the C headers earlier to avoid this.
4312
4313 10. Added some adjustments to the code to make it easier to compiler on certain
4314 special systems:
4315
4316 (a) Some "const" qualifiers were missing.
4317 (b) Added the macro EXPORT before all exported functions; by default this
4318 is defined to be empty.
4319 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4320 that it reads its output file name as an argument instead of writing
4321 to the standard output and assuming this can be redirected.
4322
4323 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4324 class containing characters with values greater than 255, PCRE compilation
4325 went into a loop.
4326
4327 12. A recursive reference to a subpattern that was within another subpattern
4328 that had a minimum quantifier of zero caused PCRE to crash. For example,
4329 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4330 recursion. If the recursively-called subpattern itself had a zero repeat,
4331 that was OK.
4332
4333 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4334 buffer into which it was copied (for escape processing) was still set at
4335 1024, so long lines caused crashes.
4336
4337 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4338 "internal error: code overflow...". This applied to any character class
4339 that was followed by a possessive quantifier.
4340
4341 15. Modified the Makefile to add libpcre.la as a prerequisite for
4342 libpcreposix.la because I was told this is needed for a parallel build to
4343 work.
4344
4345 16. If a pattern that contained .* following optional items at the start was
4346 studied, the wrong optimizing data was generated, leading to matching
4347 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4348 matching string must start with a or b or c. The correct conclusion for
4349 this pattern is that a match can start with any character.
4350
4351
4352 Version 4.4 13-Aug-03
4353 ---------------------
4354
4355 1. In UTF-8 mode, a character class containing characters with values between
4356 127 and 255 was not handled correctly if the compiled pattern was studied.
4357 In fixing this, I have also improved the studying algorithm for such
4358 classes (slightly).
4359
4360 2. Three internal functions had redundant arguments passed to them. Removal
4361 might give a very teeny performance improvement.
4362
4363 3. Documentation bug: the value of the capture_top field in a callout is *one
4364 more than* the number of the hightest numbered captured substring.
4365
4366 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4367 in incorrectly linking with a previously installed version. They now link
4368 explicitly with libpcre.la.
4369
4370 5. configure.in no longer needs to recognize Cygwin specially.
4371
4372 6. A problem in pcre.in for Windows platforms is fixed.
4373
4374 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4375 pcretest, it used to include the size of the study block as part of its
4376 output. Unfortunately, the structure contains a field that has a different
4377 size on different hardware architectures. This meant that the tests that
4378 showed this size failed. As the block is currently always of a fixed size,
4379 this information isn't actually particularly useful in pcretest output, so
4380 I have just removed it.
4381
4382 8. Three pre-processor statements accidentally did not start in column 1.
4383 Sadly, there are *still* compilers around that complain, even though
4384 standard C has not required this for well over a decade. Sigh.
4385
4386 9. In pcretest, the code for checking callouts passed small integers in the
4387 callout_data field, which is a void * field. However, some picky compilers
4388 complained about the casts involved for this on 64-bit systems. Now
4389 pcretest passes the address of the small integer instead, which should get
4390 rid of the warnings.
4391
4392 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4393 both compile and run time, and gives an error if an invalid UTF-8 sequence
4394 is found. There is a option for disabling this check in cases where the
4395 string is known to be correct and/or the maximum performance is wanted.
4396
4397 11. In response to a bug report, I changed one line in Makefile.in from
4398
4399 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4400 to
4401 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4402
4403 to look similar to other lines, but I have no way of telling whether this
4404 is the right thing to do, as I do not use Windows. No doubt I'll get told
4405 if it's wrong...
4406
4407
4408 Version 4.3 21-May-03
4409 ---------------------
4410
4411 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4412 Makefile.
4413
4414 2. Some refactoring to improve the quality of the code:
4415
4416 (i) The utf8_table... variables are now declared "const".
4417
4418 (ii) The code for \cx, which used the "case flipping" table to upper case
4419 lower case letters, now just substracts 32. This is ASCII-specific,
4420 but the whole concept of \cx is ASCII-specific, so it seems
4421 reasonable.
4422
4423 (iii) PCRE was using its character types table to recognize decimal and
4424 hexadecimal digits in the pattern. This is silly, because it handles
4425 only 0-9, a-f, and A-F, but the character types table is locale-
4426 specific, which means strange things might happen. A private
4427 table is now used for this - though it costs 256 bytes, a table is
4428 much faster than multiple explicit tests. Of course, the standard
4429 character types table is still used for matching digits in subject
4430 strings against \d.
4431
4432 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4433 ending in _t are). So I've renamed it as ESC_tee.
4434
4435 3. The first argument for regexec() in the POSIX wrapper should have been
4436 defined as "const".
4437
4438 4. Changed pcretest to use malloc() for its buffers so that they can be
4439 Electric Fenced for debugging.
4440
4441 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4442 to read one or more bytes before the start of the subject string. Often this
4443 had no effect on PCRE's behaviour, but in some circumstances it could
4444 provoke a segmentation fault.
4445
4446 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4447 to try to read one or more bytes before the start of the subject string.
4448
4449 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4450 UTF-8 support could misbehave in various ways if the subject string
4451 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4452 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4453 back over UTF-8 characters.)
4454
4455
4456 Version 4.2 14-Apr-03
4457 ---------------------
4458
4459 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4460
4461 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4462 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4463 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4464 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4465 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4466 and BUILD_EXEEXT
4467 Note: automatic setting of the BUILD variables is not yet working
4468 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4469 compile-time but not at link-time
4470 [LINK]: use for linking executables only
4471 make different versions for Windows and non-Windows
4472 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4473 libraries
4474 [LINK_FOR_BUILD]: new variable
4475 [OBJEXT]: use throughout
4476 [EXEEXT]: use throughout
4477 <winshared>: new target
4478 <wininstall>: new target
4479 <dftables.o>: use native compiler
4480 <dftables>: use native linker
4481 <install>: handle Windows platform correctly
4482 <clean>: ditto
4483 <check>: ditto
4484 copy DLL to top builddir before testing
4485
4486 As part of these changes, -no-undefined was removed again. This was reported
4487 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4488 in any case.
4489
4490 3. Some tidies to get rid of compiler warnings:
4491
4492 . In the match_data structure, match_limit was an unsigned long int, whereas
4493 match_call_count was an int. I've made them both unsigned long ints.
4494
4495 . In pcretest the fact that a const uschar * doesn't automatically cast to
4496 a void * provoked a warning.
4497
4498 . Turning on some more compiler warnings threw up some "shadow" variables
4499 and a few more missing casts.
4500
4501 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4502 option, a class that contained a single character with a value between 128
4503 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4504
4505 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4506 option, a class that contained several characters, but with at least one
4507 whose value was between 128 and 255 caused PCRE to crash.
4508
4509
4510 Version 4.1 12-Mar-03
4511 ---------------------
4512
4513 1. Compiling with gcc -pedantic found a couple of places where casts were
4514 needed, and a string in dftables.c that was longer than standard compilers are
4515 required to support.
4516
4517 2. Compiling with Sun's compiler found a few more places where the code could
4518 be tidied up in order to avoid warnings.
4519
4520 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4521 first of these names is deprecated in the latest Autoconf in favour of the name
4522 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4523 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4524 analogy I have changed it to CFLAGS_FOR_BUILD.
4525
4526 4. Added -no-undefined to the linking command in the Makefile, because this is
4527 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4528 linking step for the pcreposix library.
4529
4530 5. PCRE was failing to diagnose the case of two named groups with the same
4531 name.
4532
4533 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4534 literal character that is needed in the subject for a match, and scans along to
4535 ensure that it is present before embarking on the full matching process. This
4536 saves time in cases of nested unlimited repeats that are never going to match.
4537 Problem: the scan can take a lot of time if the subject is very long (e.g.
4538 megabytes), thus penalizing straightforward matches. It is now done only if the
4539 amount of subject to be scanned is less than 1000 bytes.
4540
4541 7. A lesser problem with the same optimization is that it was recording the
4542 first character of an anchored pattern as "needed", thus provoking a search
4543 right along the subject, even when the first match of the pattern was going to
4544 fail. The "needed" character is now not set for anchored patterns, unless it
4545 follows something in the pattern that is of non-fixed length. Thus, it still
4546 fulfils its original purpose of finding quick non-matches in cases of nested
4547 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4548
4549
4550 Version 4.0 17-Feb-03
4551 ---------------------
4552
4553 1. If a comment in an extended regex that started immediately after a meta-item
4554 extended to the end of string, PCRE compiled incorrect data. This could lead to
4555 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4556
4557 2. Moved to autoconf 2.53 and libtool 1.4.2.
4558
4559 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4560 the special perltest8 script is no longer needed - all the tests can be run
4561 from a single perltest script.
4562
4563 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4564 by \s. It has now been removed in PCRE. This means it isn't recognized as
4565 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4566 class [:space:] *does* include VT, thereby creating a mess.
4567
4568 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4569 space and tab.
4570
4571 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4572 its new features into the main test script, reducing the number of scripts.
4573
4574 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4575 were backward compatible, and made the (?i) apply to the whole pattern, as if
4576 /i were given. Now it behaves more logically, and applies the option setting
4577 only to what follows. PCRE has been changed to follow suit. However, if it
4578 finds options settings right at the start of the pattern, it extracts them into
4579 the global options, as before. Thus, they show up in the info data.
4580
4581 8. Added support for the \Q...\E escape sequence. Characters in between are
4582 treated as literals. This is slightly different from Perl in that $ and @ are
4583 also handled as literals inside the quotes. In Perl, they will cause variable
4584 interpolation. Note the following examples:
4585
4586 Pattern PCRE matches Perl matches
4587
4588 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4589 \Qabc\$xyz\E abc\$xyz abc\$xyz
4590 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4591
4592 For compatibility with Perl, \Q...\E sequences are recognized inside character
4593 classes as well as outside them.
4594
4595 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4596 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4597 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4598 signed/unsigned warnings.
4599
4600 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4601 option for pcretest, so I've replaced it by a simple function that does just
4602 that job.
4603
4604 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4605 "pcregrep -".
4606
4607 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4608 Java package. This provides some syntactic sugar for simple cases of what my
4609 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4610 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4611 item, you can use this simplified notation. Note that only makes sense with
4612 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4613 greediness, whatever the setting of the PCRE_UNGREEDY option.
4614
4615 13. A change of greediness default within a pattern was not taking effect at
4616 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4617 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4618 was abstracted outside.
4619
4620 14. PCRE now supports the \G assertion. It is true when the current matching
4621 position is at the start point of the match. This differs from \A when the
4622 starting offset is non-zero. Used with the /g option of pcretest (or similar
4623 code), it works in the same way as it does for Perl's /g option. If all
4624 alternatives of a regex begin with \G, the expression is anchored to the start
4625 match position, and the "anchored" flag is set in the compiled expression.
4626
4627 15. Some bugs concerning the handling of certain option changes within patterns
4628 have been fixed. These applied to options other than (?ims). For example,
4629 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4630 been the other way round. Some of this was related to change 7 above.
4631
4632 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4633 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4634 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4635 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4636
4637 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4638 mode. Unlike ".", it always matches newline, whatever the setting of
4639 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4640 assertions. Perl allows it, but it doesn't (in general) work because it can't
4641 calculate the length of the lookbehind. At least, that's the case for Perl
4642 5.8.0 - I've been told they are going to document that it doesn't work in
4643 future.
4644
4645 18. Added an error diagnosis for escapes that PCRE does not support: these are
4646 \L, \l, \N, \P, \p, \U, \u, and \X.
4647
4648 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4649 reading past the end of the pattern in cases such as /[abcd/.
4650
4651 20. PCRE was getting more memory than necessary for patterns with classes that
4652 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4653
4654 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4655 compiling PCRE for use with Virtual Pascal.
4656
4657 22. Small fix to the Makefile to make it work properly if the build is done
4658 outside the source tree.
4659
4660 23. Added a new extension: a condition to go with recursion. If a conditional
4661 subpattern starts with (?(R) the "true" branch is used if recursion has
4662 happened, whereas the "false" branch is used only at the top level.
4663
4664 24. When there was a very long string of literal characters (over 255 bytes
4665 without UTF support, over 250 bytes with UTF support), the computation of how
4666 much memory was required could be incorrect, leading to segfaults or other
4667 strange effects.
4668
4669 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4670 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4671 there was a subsequent back reference to those brackets. This meant that, for
4672 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4673 possible to check for precisely this case. All we can do is abandon the
4674 optimization if .* occurs inside capturing brackets when there are any back
4675 references whatsoever. (See below for a better fix that came later.)
4676
4677 26. The handling of the optimization for finding the first character of a
4678 non-anchored pattern, and for finding a character that is required later in the
4679 match were failing in some cases. This didn't break the matching; it just
4680 failed to optimize when it could. The way this is done has been re-implemented.
4681
4682 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4683
4684 28. Added a new feature that provides some of the functionality that Perl
4685 provides with (?{...}). The facility is termed a "callout". The way it is done
4686 in PCRE is for the caller to provide an optional function, by setting
4687 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4688 global variable. By default it is unset, which disables all calling out. To get
4689 the function called, the regex must include (?C) at appropriate points. This
4690 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4691 This provides a means of identifying different callout points. When PCRE
4692 reaches such a point in the regex, if pcre_callout has been set, the external
4693 function is called. It is provided with data in a structure called
4694 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4695 matching continues; if it returns a non-zero value, the match at the current
4696 point fails. However, backtracking will occur if possible. [This was changed
4697 later and other features added - see item 49 below.]
4698
4699 29. pcretest is upgraded to test the callout functionality. It provides a
4700 callout function that displays information. By default, it shows the start of
4701 the match and the current position in the text. There are some new data escapes
4702 to vary what happens:
4703
4704 \C+ in addition, show current contents of captured substrings
4705 \C- do not supply a callout function
4706 \C!n return 1 when callout number n is reached
4707 \C!n!m return 1 when callout number n is reached for the mth time
4708
4709 30. If pcregrep was called with the -l option and just a single file name, it
4710 output "<stdin>" if a match was found, instead of the file name.
4711
4712 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4713 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4714 pcre_exec(). This saves a malloc/free per call. The default value of
4715 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4716 when configuring.
4717
4718 32. The default maximum size of a compiled pattern is 64K. There have been a
4719 few cases of people hitting this limit. The code now uses macros to handle the
4720 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4721 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4722 configuring. Tests 2 and 5 work only with 2-byte links because they output
4723 debugging information about compiled patterns.
4724
4725 33. Internal code re-arrangements:
4726
4727 (a) Moved the debugging function for printing out a compiled regex into
4728 its own source file (printint.c) and used #include to pull it into
4729 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4730 separate copies.
4731
4732 (b) Defined the list of op-code names for debugging as a macro in
4733 internal.h so that it is next to the definition of the opcodes.
4734
4735 (c) Defined a table of op-code lengths for simpler skipping along compiled
4736 code. This is again a macro in internal.h so that it is next to the
4737 definition of the opcodes.
4738
4739 34. Added support for recursive calls to individual subpatterns, along the
4740 lines of Robin Houston's patch (but implemented somewhat differently).
4741
4742 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4743 allow it to read and process whole directories in Win32. This code was
4744 contributed by Lionel Fourquaux; it has not been tested by me.
4745
4746 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4747 used to name a group. Names consist of alphanumerics and underscores, and must
4748 be unique. Back references use the syntax (?P=name) and recursive calls use
4749 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4750 numbers. The function pcre_fullinfo() can be used after compilation to extract
4751 a name/number map. There are three relevant calls:
4752
4753 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4754 PCRE_INFO_NAMECOUNT yields the number of entries
4755 PCRE_INFO_NAMETABLE yields a pointer to the map.
4756
4757 The map is a vector of fixed-size entries. The size of each entry depends on
4758 the length of the longest name used. The first two bytes of each entry are the
4759 group number, most significant byte first. There follows the corresponding
4760 name, zero terminated. The names are in alphabetical order.
4761
4762 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4763 case instead of 255. Making it the same both with and without UTF-8 support
4764 means that the same test output works with both.
4765
4766 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4767 calling malloc() with a zero argument.
4768
4769 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4770 optimization. I've improved things by keeping a bitmap of backreferences with
4771 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4772 fact referenced, the optimization can be applied. It is unlikely that a
4773 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4774 the match to follow \n) will appear inside brackets with a number greater than
4775 31, but if it does, any back reference > 31 suppresses the optimization.
4776
4777 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4778 of disabling numbered capturing parentheses. Any opening parenthesis that is
4779 not followed by ? behaves as if it were followed by ?: but named parentheses
4780 can still be used for capturing (and they will acquire numbers in the usual
4781 way).
4782
4783 41. Redesigned the return codes from the match() function into yes/no/error so
4784 that errors can be passed back from deep inside the nested calls. A malloc
4785 failure while inside a recursive subpattern call now causes the
4786 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4787
4788 42. It is now possible to set a limit on the number of times the match()
4789 function is called in a call to pcre_exec(). This facility makes it possible to
4790 limit the amount of recursion and backtracking, though not in a directly
4791 obvious way, because the match() function is used in a number of different
4792 circumstances. The count starts from zero for each position in the subject
4793 string (for non-anchored patterns). The default limit is, for compatibility, a
4794 large number, namely 10 000 000. You can change this in two ways:
4795
4796 (a) When configuring PCRE before making, you can use --with-match-limit=n
4797 to set a default value for the compiled library.
4798
4799 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4800 a different value is set. See 45 below.
4801
4802 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4803
4804 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4805 of things that can be changed at compile time. The first argument specifies
4806 what is wanted and the second points to where the information is to be placed.
4807 The current list of available information is:
4808
4809 PCRE_CONFIG_UTF8
4810
4811 The output is an integer that is set to one if UTF-8 support is available;
4812 otherwise it is set to zero.
4813
4814 PCRE_CONFIG_NEWLINE
4815
4816 The output is an integer that it set to the value of the code that is used for
4817 newline. It is either LF (10) or CR (13).
4818
4819 PCRE_CONFIG_LINK_SIZE
4820
4821 The output is an integer that contains the number of bytes used for internal
4822 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4823
4824 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4825
4826 The output is an integer that contains the threshold above which the POSIX
4827 interface uses malloc() for output vectors. See item 31 above.
4828
4829 PCRE_CONFIG_MATCH_LIMIT
4830
4831 The output is an unsigned integer that contains the default limit of the number
4832 of match() calls in a pcre_exec() execution. See 42 above.
4833
4834 44. pcretest has been upgraded by the addition of the -C option. This causes it
4835 to extract all the available output from the new pcre_config() function, and to
4836 output it. The program then exits immediately.
4837
4838 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4839 order to support additional features. One way would have been to define
4840 pcre_exec2() (for example) with extra arguments, but this would not have been
4841 extensible, and would also have required all calls to the original function to
4842 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4843 is used for passing in "extra" data from pcre_study().
4844
4845 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4846 contains the following fields:
4847
4848 flags a bitmap indicating which of the following fields are set
4849 study_data opaque data from pcre_study()
4850 match_limit a way of specifying a limit on match() calls for a specific
4851 call to pcre_exec()
4852 callout_data data for callouts (see 49 below)
4853
4854 The flag bits are also defined in pcre.h, and are
4855
4856 PCRE_EXTRA_STUDY_DATA
4857 PCRE_EXTRA_MATCH_LIMIT
4858 PCRE_EXTRA_CALLOUT_DATA
4859
4860 The pcre_study() function now returns one of these new pcre_extra blocks, with
4861 the actual study data pointed to by the study_data field, and the
4862 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4863 before. That is, this change is entirely upwards-compatible and requires no
4864 change to existing code.
4865
4866 If you want to pass in additional data to pcre_exec(), you can either place it
4867 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4868 block.
4869
4870 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4871 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4872 times with different match limits, until it finds the minimum value needed for
4873 pcre_exec() to complete. The value is then output. This can be instructive; for
4874 most simple matches the number is quite small, but for pathological cases it
4875 gets very large very quickly.
4876
4877 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4878 returns the size of the data block pointed to by the study_data field in a
4879 pcre_extra block, that is, the value that was passed as the argument to
4880 pcre_malloc() when PCRE was getting memory in which to place the information
4881 created by pcre_study(). The fourth argument should point to a size_t variable.
4882 pcretest has been extended so that this information is shown after a successful
4883 pcre_study() call when information about the compiled regex is being displayed.
4884
4885 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4886 because what follows is always an absolute path. (Later: it turns out that this
4887 is more than cosmetic for MinGW, because it doesn't like empty path
4888 components.)
4889
4890 49. Some changes have been made to the callout feature (see 28 above):
4891
4892 (i) A callout function now has three choices for what it returns:
4893
4894 0 => success, carry on matching
4895 > 0 => failure at this point, but backtrack if possible
4896 < 0 => serious error, return this value from pcre_exec()
4897
4898 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4899 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4900 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4901 use by callout functions. It will never be used by PCRE itself.
4902
4903 (ii) The pcre_extra structure (see 45 above) has a void * field called
4904 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4905 pcre_callout_block structure has a field of the same name. The contents of
4906 the field passed in the pcre_extra structure are passed to the callout
4907 function in the corresponding field in the callout block. This makes it
4908 easier to use the same callout-containing regex from multiple threads. For
4909 testing, the pcretest program has a new data escape
4910
4911 \C*n pass the number n (may be negative) as callout_data
4912
4913 If the callout function in pcretest receives a non-zero value as
4914 callout_data, it returns that value.
4915
4916 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4917 there were some redundant $(CFLAGS) in commands that are now specified as
4918 $(LINK), which already includes $(CFLAGS).
4919
4920 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4921 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4922 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4923 one-byte characters throughout. Note that case-insensitive matching applies
4924 only to characters whose values are less than 256. PCRE doesn't support the
4925 notion of cases for higher-valued characters.
4926
4927 (i) A character class whose characters are all within 0-255 is handled as
4928 a bit map, and the map is inverted for negative classes. Previously, a
4929 character > 255 always failed to match such a class; however it should
4930 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4931
4932 (ii) A negated character class with a single character < 255 is coded as
4933 "not this character" (OP_NOT). This wasn't working properly when the test
4934 character was multibyte, either singly or repeated.
4935
4936 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4937 mode, for example: \x{100}{2,3}.
4938
4939 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4940 singly or repeated) now correctly test multibyte characters. However,
4941 PCRE doesn't recognize any characters with values greater than 255 as
4942 digits, spaces, or word characters. Such characters always match \D, \S,
4943 and \W, and never match \d, \s, or \w.
4944
4945 (v) Classes may now contain characters and character ranges with values
4946 greater than 255. For example: [ab\x{100}-\x{400}].
4947
4948 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4949 PCRE in UTF-8 mode.
4950
4951 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4952 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4953 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4954 value.)
4955
4956 53. The single man page has become too large. I have therefore split it up into
4957 a number of separate man pages. These also give rise to individual HTML pages;
4958 these are now put in a separate directory, and there is an index.html page that
4959 lists them all. Some hyperlinking between the pages has been installed.
4960
4961 54. Added convenience functions for handling named capturing parentheses.
4962
4963 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4964 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4965 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4966 are faulted.
4967
4968 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4969 calling configure. These values are used when compiling the dftables.c program
4970 which is run to generate the source of the default character tables. They
4971 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4972 you will need to set these values.
4973
4974 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4975
4976
4977 Version 3.9 02-Jan-02
4978 ---------------------
4979
4980 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4981
4982 2. If --disable-static was given, the building process failed when trying to
4983 build pcretest and pcregrep. (For some reason it was using libtool to compile
4984 them, which is not right, as they aren't part of the library.)
4985
4986
4987 Version 3.8 18-Dec-01
4988 ---------------------
4989
4990 1. The experimental UTF-8 code was completely screwed up. It was packing the
4991 bytes in the wrong order. How dumb can you get?
4992
4993
4994 Version 3.7 29-Oct-01
4995 ---------------------
4996
4997 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4998 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4999 this didn't happen under Solaris 8, where I normally test things.
5000
5001 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5002 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5003 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5004 this fix an hour or so after the initial 3.7 release.)
5005
5006
5007 Version 3.6 23-Oct-01
5008 ---------------------
5009
5010 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5011 offsets passed as NULL with zero offset count.
5012
5013 2. The config.guess and config.sub files had not been updated when I moved to
5014 the latest autoconf.
5015
5016
5017 Version 3.5 15-Aug-01
5018 ---------------------
5019
5020 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5021 had been forgotten.
5022
5023 2. By using declared but undefined structures, we can avoid using "void"
5024 definitions in pcre.h while keeping the internal definitions of the structures
5025 private.
5026
5027 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5028 user point of view, this means that both static and shared libraries are built
5029 by default, but this can be individually controlled. More of the work of
5030 handling this static/shared cases is now inside libtool instead of PCRE's make
5031 file.
5032
5033 4. The pcretest utility is now installed along with pcregrep because it is
5034 useful for users (to test regexs) and by doing this, it automatically gets
5035 relinked by libtool. The documentation has been turned into a man page, so
5036 there are now .1, .txt, and .html versions in /doc.
5037
5038 5. Upgrades to pcregrep:
5039 (i) Added long-form option names like gnu grep.
5040 (ii) Added --help to list all options with an explanatory phrase.
5041 (iii) Added -r, --recursive to recurse into sub-directories.
5042 (iv) Added -f, --file to read patterns from a file.
5043
5044 6. pcre_exec() was referring to its "code" argument before testing that
5045 argument for NULL (and giving an error if it was NULL).
5046
5047 7. Upgraded Makefile.in to allow for compiling in a different directory from
5048 the source directory.
5049
5050 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5051 options bits, the pointer it was passed was to an int instead of to an unsigned
5052 long int. This mattered only on 64-bit systems.
5053
5054 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5055 generated) instead of pcre.in, which it its source. Also made the same change
5056 in several of the .c files.
5057
5058 10. A new release of gcc defines printf() as a macro, which broke pcretest
5059 because it had an ifdef in the middle of a string argument for printf(). Fixed
5060 by using separate calls to printf().
5061
5062 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5063 script, to force use of CR or LF instead of \n in the source. On non-Unix
5064 systems, the value can be set in config.h.
5065
5066 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5067 absolute limit. Changed the text of the error message to make this clear, and
5068 likewise updated the man page.
5069
5070 13. The limit of 99 on the number of capturing subpatterns has been removed.
5071 The new limit is 65535, which I hope will not be a "real" limit.
5072
5073
5074 Version 3.4 22-Aug-00
5075 ---------------------
5076
5077 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5078
5079 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5080
5081
5082 Version 3.3 01-Aug-00
5083 ---------------------
5084
5085 1. If an octal character was given, but the value was greater than \377, it
5086 was not getting masked to the least significant bits, as documented. This could
5087 lead to crashes in some systems.
5088
5089 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5090 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5091
5092 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5093 These just pass their arguments on to (pcre_free)(), but they are provided
5094 because some uses of PCRE bind it to non-C systems that can call its functions,
5095 but cannot call free() or pcre_free() directly.
5096
5097 4. Add "make test" as a synonym for "make check". Corrected some comments in
5098 the Makefile.
5099
5100 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5101 Makefile.
5102
5103 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5104 command called pgrep for grepping around the active processes.
5105
5106 7. Added the beginnings of support for UTF-8 character strings.
5107
5108 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5109 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5110 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5111 out for the ar command.)
5112
5113
5114 Version 3.2 12-May-00
5115 ---------------------
5116
5117 This is purely a bug fixing release.
5118
5119 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5120 of ZA. This was just one example of several cases that could provoke this bug,
5121 which was introduced by change 9 of version 2.00. The code for breaking
5122 infinite loops after an iteration that matches an empty string was't working
5123 correctly.
5124
5125 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5126 when matched against abbab (for example). After matching an empty string, it
5127 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5128 caused it to match further down the string than it should.
5129
5130 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5131 was there because it doesn't seem to be needed, and it causes trouble on some
5132 systems, as it is not a Standard C header. It has been removed.
5133
5134 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5135 were reported on the Macintosh. The changes were from
5136
5137 while ((c = *(++ptr)) != 0 && c != '\n');
5138 to
5139 while ((c = *(++ptr)) != 0 && c != '\n') ;
5140
5141 Totally extraordinary, but if that's what it takes...
5142
5143 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5144 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5145 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5146 assumes the way PCRE uses memmove() (always moving upwards).
5147
5148 6. PCRE is being used in one environment where strchr() is not available. There
5149 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5150 faster code anyway.
5151
5152
5153 Version 3.1 09-Feb-00
5154 ---------------------
5155
5156 The only change in this release is the fixing of some bugs in Makefile.in for
5157 the "install" target:
5158
5159 (1) It was failing to install pcreposix.h.
5160
5161 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5162
5163
5164 Version 3.0 01-Feb-00
5165 ---------------------
5166
5167 1. Add support for the /+ modifier to perltest (to output $` like it does in
5168 pcretest).
5169
5170 2. Add support for the /g modifier to perltest.
5171
5172 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5173 matches null strings.
5174
5175 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5176 pattern. Perl treats empty patterns specially - it reuses the most recent
5177 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5178 effect.
5179
5180 5. The POSIX interface was broken in that it was just handing over the POSIX
5181 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5182 required a bigger vector, with some working space on the end. This means that
5183 the POSIX wrapper now has to get and free some memory, and copy the results.
5184
5185 6. Added some simple autoconf support, placing the test data and the
5186 documentation in separate directories, re-organizing some of the
5187 information files, and making it build pcre-config (a GNU standard). Also added
5188 libtool support for building PCRE as a shared library, which is now the
5189 default.
5190
5191 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5192 09 are not valid octal constants. Single digits will be used for minor values
5193 less than 10.
5194
5195 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5196 existing programs that set these in the POSIX interface can use PCRE without
5197 modification.
5198
5199 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5200 return all that pcre_info() returns, plus additional data. The pcre_info()
5201 function is retained for compatibility, but is considered to be obsolete.
5202
5203 10. Added experimental recursion feature (?R) to handle one common case that
5204 Perl 5.6 will be able to do with (?p{...}).
5205
5206 11. Added support for POSIX character classes like [:alpha:], which Perl is
5207 adopting.
5208
5209
5210 Version 2.08 31-Aug-99
5211 ----------------------
5212
5213 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5214 trying to match at the startoffset position, but instead was moving forward to
5215 the next newline as if a previous match had failed.
5216
5217 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5218 and could get into a loop if a null string was matched other than at the start
5219 of the subject.
5220
5221 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5222 be distinguished at compile time, and for completeness also added PCRE_DATE.
5223
5224 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5225 in GnuWin32 environments.
5226
5227
5228 Version 2.07 29-Jul-99
5229 ----------------------
5230
5231 1. The documentation is now supplied in plain text form and HTML as well as in
5232 the form of man page sources.
5233
5234 2. C++ compilers don't like assigning (void *) values to other pointer types.
5235 In particular this affects malloc(). Although there is no problem in Standard
5236 C, I've put in casts to keep C++ compilers happy.
5237
5238 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5239 should be (const char *).
5240
5241 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5242 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5243 However, I haven't made this a standard facility. The documentation doesn't
5244 mention it, and the Makefile doesn't support it.
5245
5246 5. The Makefile now contains an "install" target, with editable destinations at
5247 the top of the file. The pcretest program is not installed.
5248
5249 6. pgrep -V now gives the PCRE version number and date.
5250
5251 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5252 causing the entire string to be ignored, instead of just the last character.
5253
5254 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5255 non-matching string, it can take a very, very long time, even for strings of
5256 quite modest length, because of the nested recursion. PCRE now does better in
5257 some of these cases. It does this by remembering the last required literal
5258 character in the pattern, and pre-searching the subject to ensure it is present
5259 before running the real match. In other words, it applies a heuristic to detect
5260 some types of certain failure quickly, and in the above example, if presented
5261 with a string that has no trailing " it gives "no match" very quickly.
5262
5263 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5264 other alternatives are tried instead.
5265
5266
5267 Version 2.06 09-Jun-99
5268 ----------------------
5269
5270 1. Change pcretest's output for amount of store used to show just the code
5271 space, because the remainder (the data block) varies in size between 32-bit and
5272 64-bit systems.
5273
5274 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5275 start matching at. This allows lookbehinds to work when searching for multiple
5276 occurrences in a string.
5277
5278 3. Added additional options to pcretest for testing multiple occurrences:
5279
5280 /+ outputs the rest of the string that follows a match
5281 /g loops for multiple occurrences, using the new startoffset argument
5282 /G loops for multiple occurrences by passing an incremented pointer
5283
5284 4. PCRE wasn't doing the "first character" optimization for patterns starting
5285 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5286 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5287 the letter 'x'. On long subject strings, this gives a significant speed-up.
5288
5289
5290 Version 2.05 21-Apr-99
5291 ----------------------
5292
5293 1. Changed the type of magic_number from int to long int so that it works
5294 properly on 16-bit systems.
5295
5296 2. Fixed a bug which caused patterns starting with .* not to work correctly
5297 when the subject string contained newline characters. PCRE was assuming
5298 anchoring for such patterns in all cases, which is not correct because .* will
5299 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5300 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5301 must be retried after every newline in the subject.
5302
5303
5304 Version 2.04 18-Feb-99
5305 ----------------------
5306
5307 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5308 computation of the store needed to hold the pattern was incorrect (too large).
5309 If such patterns were nested a few deep, this could multiply and become a real
5310 problem.
5311
5312 2. Added /M option to pcretest to show the memory requirement of a specific
5313 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5314
5315 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5316 compiled in such a way that the backtracking after subsequent failure was
5317 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5318 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5319
5320
5321 Version 2.03 02-Feb-99
5322 ----------------------
5323
5324 1. Fixed typo and small mistake in man page.
5325
5326 2. Added 4th condition (GPL supersedes if conflict) and created separate
5327 LICENCE file containing the conditions.
5328
5329 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5330 Perl, that is the internal \ allows the delimiter to be included in the
5331 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5332 the final delimiter, add \ to the end of the pattern (to test the error).
5333
5334 4. Added the convenience functions for extracting substrings after a successful
5335 match. Updated pcretest to make it able to test these functions.
5336
5337
5338 Version 2.02 14-Jan-99
5339 ----------------------
5340
5341 1. Initialized the working variables associated with each extraction so that
5342 their saving and restoring doesn't refer to uninitialized store.
5343
5344 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5345 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5346 fix the problem.
5347
5348 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5349 calls, and wasn't printing the correct value for compiling calls. Increased the
5350 default value of LOOPREPEAT, and the number of significant figures in the
5351 times.
5352
5353 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5354
5355 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5356 a building problem on Windows NT with a FAT file system.
5357
5358
5359 Version 2.01 21-Oct-98
5360 ----------------------
5361
5362 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5363 to character tables built by pcre_maketables() in the current locale. If NULL
5364 is passed, the default tables are used.
5365
5366
5367 Version 2.00 24-Sep-98
5368 ----------------------
5369
5370 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5371 it any more.
5372
5373 2. Allow quantification of (?>) groups, and make it work correctly.
5374
5375 3. The first character computation wasn't working for (?>) groups.
5376
5377 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5378 end of the subject) and add 5.005's \z, which really does match only at the
5379 very end of the subject.
5380
5381 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5382
5383 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5384 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5385 localized options. All options to pcre_study() were also removed.
5386
5387 7. Add other new features from 5.005:
5388
5389 $(?<= positive lookbehind
5390 $(?<! negative lookbehind
5391 (?imsx-imsx) added the unsetting capability
5392 such a setting is global if at outer level; local otherwise
5393 (?imsx-imsx:) non-capturing groups with option setting
5394 (?(cond)re|re) conditional pattern matching
5395
5396 A backreference to itself in a repeated group matches the previous
5397 captured string.
5398
5399 8. General tidying up of studying (both automatic and via "study")
5400 consequential on the addition of new assertions.
5401
5402 9. As in 5.005, unlimited repeated groups that could match an empty substring
5403 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5404 runtime if any iteration does actually match an empty substring.
5405
5406 10. Include the RunTest script in the distribution.
5407
5408 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5409 discrepancies, some of which were old and were also with respect to 5.004. They
5410 have now been fixed.
5411
5412
5413 Version 1.09 28-Apr-98
5414 ----------------------
5415
5416 1. A negated single character class followed by a quantifier with a minimum
5417 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5418 program crashes, or just wrong answers. This did not apply to negated classes
5419 containing more than one character, or to minima other than one.
5420
5421
5422 Version 1.08 27-Mar-98
5423 ----------------------
5424
5425 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5426
5427 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5428 latter must appear before anything that relies on it in the pattern.
5429
5430
5431 Version 1.07 16-Feb-98
5432 ----------------------
5433
5434 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5435 repeat of a potentially empty string).
5436
5437
5438 Version 1.06 23-Jan-98
5439 ----------------------
5440
5441 1. Added Markus Oberhumer's little patches for C++.
5442
5443 2. Literal strings longer than 255 characters were broken.
5444
5445
5446 Version 1.05 23-Dec-97
5447 ----------------------
5448
5449 1. Negated character classes containing more than one character were failing if
5450 PCRE_CASELESS was set at run time.
5451
5452
5453 Version 1.04 19-Dec-97
5454 ----------------------
5455
5456 1. Corrected the man page, where some "const" qualifiers had been omitted.
5457
5458 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5459 input syntax.
5460
5461 3. Fixed memory leak which occurred when a regex with back references was
5462 matched with an offsets vector that wasn't big enough. The temporary memory
5463 that is used in this case wasn't being freed if the match failed.
5464
5465 4. Tidied pcretest to ensure it frees memory that it gets.
5466
5467 5. Temporary memory was being obtained in the case where the passed offsets
5468 vector was exactly big enough.
5469
5470 6. Corrected definition of offsetof() from change 5 below.
5471
5472 7. I had screwed up change 6 below and broken the rules for the use of
5473 setjmp(). Now fixed.
5474
5475
5476 Version 1.03 18-Dec-97
5477 ----------------------
5478
5479 1. A erroneous regex with a missing opening parenthesis was correctly
5480 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5481 on some systems.
5482
5483 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5484 it was reported that one broken compiler failed on the former because "code" is
5485 also an independent variable.
5486
5487 3. The erroneous regex a[]b caused an array overrun reference.
5488
5489 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5490 fail on data ending with that character. (It was going on too far, and checking
5491 the next character, typically a binary zero.) This was specific to the
5492 optimized code for single-character negative classes.
5493
5494 5. Added a contributed patch from the TIN world which does the following:
5495
5496 + Add an undef for memmove, in case the the system defines a macro for it.
5497
5498 + Add a definition of offsetof(), in case there isn't one. (I don't know
5499 the reason behind this - offsetof() is part of the ANSI standard - but
5500 it does no harm).
5501
5502 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5503 most of the places where whitespace preceded '#'. I have given up and
5504 allowed the remaining 2 cases to be at the margin.
5505
5506 + Rename some variables in pcre to eliminate shadowing. This seems very
5507 pedantic, but does no harm, of course.
5508
5509 6. Moved the call to setjmp() into its own function, to get rid of warnings
5510 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5511
5512 7. Constructs such as \d{8,} were compiling into the equivalent of
5513 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5514 outcome, but in this particular case used more store than had been allocated,
5515 which caused the bug to be discovered because it threw up an internal error.
5516
5517 8. The debugging code in both pcre and pcretest for outputting the compiled
5518 form of a regex was going wrong in the case of back references followed by
5519 curly-bracketed repeats.
5520
5521
5522 Version 1.02 12-Dec-97
5523 ----------------------
5524
5525 1. Typos in pcre.3 and comments in the source fixed.
5526
5527 2. Applied a contributed patch to get rid of places where it used to remove
5528 'const' from variables, and fixed some signed/unsigned and uninitialized
5529 variable warnings.
5530
5531 3. Added the "runtest" target to Makefile.
5532
5533 4. Set default compiler flag to -O2 rather than just -O.
5534
5535
5536 Version 1.01 19-Nov-97
5537 ----------------------
5538
5539 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5540 like /([ab]*)*/, that is, for classes with more than one character in them.
5541
5542 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5543 as /((?>a*))*/ (a PCRE_EXTRA facility).
5544
5545
5546 Version 1.00 18-Nov-97
5547 ----------------------
5548
5549 1. Added compile-time macros to support systems such as SunOS4 which don't have
5550 memmove() or strerror() but have other things that can be used instead.
5551
5552 2. Arranged that "make clean" removes the executables.
5553
5554
5555 Version 0.99 27-Oct-97
5556 ----------------------
5557
5558 1. Fixed bug in code for optimizing classes with only one character. It was
5559 initializing a 32-byte map regardless, which could cause it to run off the end
5560 of the memory it had got.
5561
5562 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5563
5564
5565 Version 0.98 22-Oct-97
5566 ----------------------
5567
5568 1. Fixed bug in code for handling temporary memory usage when there are more
5569 back references than supplied space in the ovector. This could cause segfaults.
5570
5571
5572 Version 0.97 21-Oct-97
5573 ----------------------
5574
5575 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5576
5577 2. Optimized negated single characters not to use a bit map.
5578
5579 3. Brought error texts together as macro definitions; clarified some of them;
5580 fixed one that was wrong - it said "range out of order" when it meant "invalid
5581 escape sequence".
5582
5583 4. Changed some char * arguments to const char *.
5584
5585 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5586
5587 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5588 pcretest.
5589
5590
5591 Version 0.96 16-Oct-97
5592 ----------------------
5593
5594 1. Added a simple "pgrep" utility to the distribution.
5595
5596 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5597 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5598 where "ddd" means "one or more decimal digits".
5599
5600 3. Fixed serious bug. If a pattern had a back reference, but the call to
5601 pcre_exec() didn't supply a large enough ovector to record the related
5602 identifying subpattern, the match always failed. PCRE now remembers the number
5603 of the largest back reference, and gets some temporary memory in which to save
5604 the offsets during matching if necessary, in order to ensure that
5605 backreferences always work.
5606
5607 4. Increased the compatibility with Perl in a number of ways:
5608
5609 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5610 to request this handling. The option can be set at compile or exec time.
5611
5612 (b) $ matches before a terminating newline by default; an option
5613 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5614 mode). The option can be set at compile or exec time.
5615
5616 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5617 the same as Perl's. If the decimal number it represents is less than 10
5618 or there aren't that many previous left capturing parentheses, an octal
5619 escape is read. Inside a character class, it's always an octal escape,
5620 even if it is a single digit.
5621
5622 (d) An escaped but undefined alphabetic character is taken as a literal,
5623 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5624 escapes.
5625
5626 (e) {0} is now permitted. (The previous item is removed from the compiled
5627 pattern).
5628
5629 5. Changed all the names of code files so that the basic parts are no longer
5630 than 10 characters, and abolished the teeny "globals.c" file.
5631
5632 6. Changed the handling of character classes; they are now done with a 32-byte
5633 bit map always.
5634
5635 7. Added the -d and /D options to pcretest to make it possible to look at the
5636 internals of compilation without having to recompile pcre.
5637
5638
5639 Version 0.95 23-Sep-97
5640 ----------------------
5641
5642 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5643 \x20 at the start of a run of normal characters. These were being treated as
5644 real characters, instead of the source characters being re-checked.
5645
5646
5647 Version 0.94 18-Sep-97
5648 ----------------------
5649
5650 1. The functions are now thread-safe, with the caveat that the global variables
5651 containing pointers to malloc() and free() or alternative functions are the
5652 same for all threads.
5653
5654 2. Get pcre_study() to generate a bitmap of initial characters for non-
5655 anchored patterns when this is possible, and use it if passed to pcre_exec().
5656
5657
5658 Version 0.93 15-Sep-97
5659 ----------------------
5660
5661 1. /(b)|(:+)/ was computing an incorrect first character.
5662
5663 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5664 but not actually doing anything yet.
5665
5666 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5667 as Perl does (e.g. [-az] or [az-]).
5668
5669 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5670 all possible positions.
5671
5672 5. Split up into different modules to avoid including unneeded functions in a
5673 compiled binary. However, compile and exec are still in one module. The "study"
5674 function is split off.
5675
5676 6. The character tables are now in a separate module whose source is generated
5677 by an auxiliary program - but can then be edited by hand if required. There are
5678 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5679 toupper() in the code.
5680
5681 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5682 make them global. Abolish the function for setting them, as the caller can now
5683 set them directly.
5684
5685
5686 Version 0.92 11-Sep-97
5687 ----------------------
5688
5689 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5690 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5691
5692 2. Caseless matching was not working in character classes if the characters in
5693 the pattern were in upper case.
5694
5695 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5696
5697 4. Make PCRE_ANCHORED public and accept as a compile option.
5698
5699 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5700 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5701 pass them.
5702
5703 6. Give an error if bad option bits passed at compile or run time.
5704
5705 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5706 pcretest to cause it to pass that flag.
5707
5708 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5709 options, and the first character, if set.
5710
5711 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5712
5713
5714 Version 0.91 10-Sep-97
5715 ----------------------
5716
5717 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5718 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5719
5720 2. PCRE was looping on encountering an indefinitely repeated back reference to
5721 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5722 Perl does - treats the match as successful.
5723
5724 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5