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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1555 - (show annotations)
Sun May 3 17:58:17 2015 UTC (4 years, 5 months ago) by ph10
File size: 277652 byte(s)
Fix forward reference offset bug.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.38 xx-xxx-xxxx
5 ------------------------
6
7 1. If a group that contained a recursive back reference also contained a
8 forward reference subroutine call followed by a non-forward-reference
9 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
10 compile correct code, leading to undefined behaviour or an internally
11 detected error. This bug was discovered by the LLVM fuzzer.
12
13
14 Version 8.37 28-April-2015
15 --------------------------
16
17 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
18 for those parentheses to be closed with whatever has been captured so far.
19 However, it was failing to mark any other groups between the hightest
20 capture so far and the currrent group as "unset". Thus, the ovector for
21 those groups contained whatever was previously there. An example is the
22 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
23
24 2. If an assertion condition was quantified with a minimum of zero (an odd
25 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
26
27 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
28 unrecognized modifier, a crash could occur.
29
30 4. An attempt to do global matching in pcretest with a zero-length ovector
31 caused a crash.
32
33 5. Fixed a memory leak during matching that could occur for a subpattern
34 subroutine call (recursive or otherwise) if the number of captured groups
35 that had to be saved was greater than ten.
36
37 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
38 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
39 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
40
41 7. A UTF pattern containing a "not" match of a non-ASCII character and a
42 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
43
44 8. When a pattern is compiled, it remembers the highest back reference so that
45 when matching, if the ovector is too small, extra memory can be obtained to
46 use instead. A conditional subpattern whose condition is a check on a
47 capture having happened, such as, for example in the pattern
48 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
49 setting the highest backreference number. This mattered only if pcre_exec()
50 was called with an ovector that was too small to hold the capture, and there
51 was no other kind of back reference (a situation which is probably quite
52 rare). The effect of the bug was that the condition was always treated as
53 FALSE when the capture could not be consulted, leading to a incorrect
54 behaviour by pcre_exec(). This bug has been fixed.
55
56 9. A reference to a duplicated named group (either a back reference or a test
57 for being set in a conditional) that occurred in a part of the pattern where
58 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
59 to be incorrectly calculated, leading to overwriting.
60
61 10. A mutually recursive set of back references such as (\2)(\1) caused a
62 segfault at study time (while trying to find the minimum matching length).
63 The infinite loop is now broken (with the minimum length unset, that is,
64 zero).
65
66 11. If an assertion that was used as a condition was quantified with a minimum
67 of zero, matching went wrong. In particular, if the whole group had
68 unlimited repetition and could match an empty string, a segfault was
69 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
70 assertions to be quantified, but not if they are being used as conditions,
71 so the above pattern is faulted by Perl. PCRE has now been changed so that
72 it also rejects such patterns.
73
74 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
75 failed to allow the zero-repeat case if pcre2_exec() was called with an
76 ovector too small to capture the group.
77
78 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
79 Red Hat Product Security:
80
81 (a) A crash if /K and /F were both set with the option to save the compiled
82 pattern.
83
84 (b) Another crash if the option to print captured substrings in a callout
85 was combined with setting a null ovector, for example \O\C+ as a subject
86 string.
87
88 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
89 forward reference repeated a large (but limited) number of times within a
90 repeated outer group that has a zero minimum quantifier, caused incorrect
91 code to be compiled, leading to the error "internal error:
92 previously-checked referenced subpattern not found" when an incorrect
93 memory address was read. This bug was reported as "heap overflow",
94 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
95 CVE-2015-2325.
96
97 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
98 call within a group that also contained a recursive back reference caused
99 incorrect code to be compiled. This bug was reported as "heap overflow",
100 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
101 number CVE-2015-2326.
102
103 24. Computing the size of the JIT read-only data in advance has been a source
104 of various issues, and new ones are still appear unfortunately. To fix
105 existing and future issues, size computation is eliminated from the code,
106 and replaced by on-demand memory allocation.
107
108 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
109 adjacent to the end of the range, and the range contained characters with
110 more than one other case, caused incorrect behaviour when compiled in UTF
111 mode. In that example, the range a-j was left out of the class.
112
113 26. Fix JIT compilation of conditional blocks, which assertion
114 is converted to (*FAIL). E.g: /(?(?!))/.
115
116 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
117 discovered by the LLVM fuzzer.
118
119 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
120 when this assertion was used as a condition, for example (?(?!)a|b). In
121 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
122 error about an unsupported item.
123
124 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
125 possessification code could take exponential time to complete. A recursion
126 depth limit of 1000 has been imposed to limit the resources used by this
127 optimization.
128
129 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
130 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
131 because \S ensures they are all in the class. The code for doing this was
132 interacting badly with the code for computing the amount of space needed to
133 compile the pattern, leading to a buffer overflow. This bug was discovered
134 by the LLVM fuzzer.
135
136 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
137 other kinds of group caused stack overflow at compile time. This bug was
138 discovered by the LLVM fuzzer.
139
140 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
141 between a subroutine call and its quantifier was incorrectly compiled,
142 leading to buffer overflow or other errors. This bug was discovered by the
143 LLVM fuzzer.
144
145 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
146 assertion after (?(. The code was failing to check the character after
147 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
148 was discovered by the LLVM fuzzer.
149
150 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
151 a fixed maximum following a group that contains a subroutine reference was
152 incorrectly compiled and could trigger buffer overflow. This bug was
153 discovered by the LLVM fuzzer.
154
155 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
156 caused a stack overflow instead of the diagnosis of a non-fixed length
157 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
158
159 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
160 (e.g. /(?<=\Ka)/) could make pcregrep loop.
161
162 37. There was a similar problem to 36 in pcretest for global matches.
163
164 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
165 and a subsequent item in the pattern caused a non-match, backtracking over
166 the repeated \X did not stop, but carried on past the start of the subject,
167 causing reference to random memory and/or a segfault. There were also some
168 other cases where backtracking after \C could crash. This set of bugs was
169 discovered by the LLVM fuzzer.
170
171 39. The function for finding the minimum length of a matching string could take
172 a very long time if mutual recursion was present many times in a pattern,
173 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
174 method has been implemented. This infelicity was discovered by the LLVM
175 fuzzer.
176
177 40. Static linking against the PCRE library using the pkg-config module was
178 failing on missing pthread symbols.
179
180
181 Version 8.36 26-September-2014
182 ------------------------------
183
184 1. Got rid of some compiler warnings in the C++ modules that were shown up by
185 -Wmissing-field-initializers and -Wunused-parameter.
186
187 2. The tests for quantifiers being too big (greater than 65535) were being
188 applied after reading the number, and stupidly assuming that integer
189 overflow would give a negative number. The tests are now applied as the
190 numbers are read.
191
192 3. Tidy code in pcre_exec.c where two branches that used to be different are
193 now the same.
194
195 4. The JIT compiler did not generate match limit checks for certain
196 bracketed expressions with quantifiers. This may lead to exponential
197 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
198 issue should be resolved now.
199
200 5. Fixed an issue, which occures when nested alternatives are optimized
201 with table jumps.
202
203 6. Inserted two casts and changed some ints to size_t in the light of some
204 reported 64-bit compiler warnings (Bugzilla 1477).
205
206 7. Fixed a bug concerned with zero-minimum possessive groups that could match
207 an empty string, which sometimes were behaving incorrectly in the
208 interpreter (though correctly in the JIT matcher). This pcretest input is
209 an example:
210
211 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
212 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
213
214 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
215 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
216 for an empty string was breaking the inner loop and carrying on at a lower
217 level, when possessive repeated groups should always return to a higher
218 level as they have no backtrack points in them. The empty string test now
219 occurs at the outer level.
220
221 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
222 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
223
224 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
225 doesn't).
226
227 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
228 to hold a pattern to be incorrectly computed (too small) when there were
229 named back references to duplicated names. This could cause "internal
230 error: code overflow" or "double free or corruption" or other memory
231 handling errors.
232
233 11. When named subpatterns had the same prefixes, back references could be
234 confused. For example, in this pattern:
235
236 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
237
238 the reference to 'Name' was incorrectly treated as a reference to a
239 duplicate name.
240
241 12. A pattern such as /^s?c/mi8 where the optional character has more than
242 one "other case" was incorrectly compiled such that it would only try to
243 match starting at "c".
244
245 13. When a pattern starting with \s was studied, VT was not included in the
246 list of possible starting characters; this should have been part of the
247 8.34/18 patch.
248
249 14. If a character class started [\Qx]... where x is any character, the class
250 was incorrectly terminated at the ].
251
252 15. If a pattern that started with a caseless match for a character with more
253 than one "other case" was studied, PCRE did not set up the starting code
254 unit bit map for the list of possible characters. Now it does. This is an
255 optimization improvement, not a bug fix.
256
257 16. The Unicode data tables have been updated to Unicode 7.0.0.
258
259 17. Fixed a number of memory leaks in pcregrep.
260
261 18. Avoid a compiler warning (from some compilers) for a function call with
262 a cast that removes "const" from an lvalue by using an intermediate
263 variable (to which the compiler does not object).
264
265 19. Incorrect code was compiled if a group that contained an internal recursive
266 back reference was optional (had quantifier with a minimum of zero). This
267 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
268 caused segmentation faults because of stack overflows at compile time.
269
270 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
271 group that is quantified with an indefinite repeat, caused a compile-time
272 loop which used up all the system stack and provoked a segmentation fault.
273 This was not the same bug as 19 above.
274
275 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
276 Patch by Mike Frysinger.
277
278
279 Version 8.35 04-April-2014
280 --------------------------
281
282 1. A new flag is set, when property checks are present in an XCLASS.
283 When this flag is not set, PCRE can perform certain optimizations
284 such as studying these XCLASS-es.
285
286 2. The auto-possessification of character sets were improved: a normal
287 and an extended character set can be compared now. Furthermore
288 the JIT compiler optimizes more character set checks.
289
290 3. Got rid of some compiler warnings for potentially uninitialized variables
291 that show up only when compiled with -O2.
292
293 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
294 of a match later then the end of the match. The pcretest program was not
295 handling the case sensibly - it was outputting from the start to the next
296 binary zero. It now reports this situation in a message, and outputs the
297 text from the end to the start.
298
299 5. Fast forward search is improved in JIT. Instead of the first three
300 characters, any three characters with fixed position can be searched.
301 Search order: first, last, middle.
302
303 6. Improve character range checks in JIT. Characters are read by an inprecise
304 function now, which returns with an unknown value if the character code is
305 above a certain threshold (e.g: 256). The only limitation is that the value
306 must be bigger than the threshold as well. This function is useful when
307 the characters above the threshold are handled in the same way.
308
309 7. The macros whose names start with RAWUCHAR are placeholders for a future
310 mode in which only the bottom 21 bits of 32-bit data items are used. To
311 make this more memorable for those maintaining the code, the names have
312 been changed to start with UCHAR21, and an extensive comment has been added
313 to their definition.
314
315 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
316 to the export list in Makefile.am (they were accidentally omitted from the
317 8.34 tarball).
318
319 9. The informational output from pcretest used the phrase "starting byte set"
320 which is inappropriate for the 16-bit and 32-bit libraries. As the output
321 for "first char" and "need char" really means "non-UTF-char", I've changed
322 "byte" to "char", and slightly reworded the output. The documentation about
323 these values has also been (I hope) clarified.
324
325 10. Another JIT related optimization: use table jumps for selecting the correct
326 backtracking path, when more than four alternatives are present inside a
327 bracket.
328
329 11. Empty match is not possible, when the minimum length is greater than zero,
330 and there is no \K in the pattern. JIT should avoid empty match checks in
331 such cases.
332
333 12. In a caseless character class with UCP support, when a character with more
334 than one alternative case was not the first character of a range, not all
335 the alternative cases were added to the class. For example, s and \x{17f}
336 are both alternative cases for S: the class [RST] was handled correctly,
337 but [R-T] was not.
338
339 13. The configure.ac file always checked for pthread support when JIT was
340 enabled. This is not used in Windows, so I have put this test inside a
341 check for the presence of windows.h (which was already tested for).
342
343 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
344 The algorithm provides a way to skip certain starting offsets, and usually
345 faster than linear prefix searches.
346
347 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
348 as for 'fr_FR' and 'french'. For some reason, however, it then used the
349 Windows-specific input and output files, which have 'french' screwed in.
350 So this could never have worked. One of the problems with locales is that
351 they aren't always the same. I have now updated RunTest so that it checks
352 the output of the locale test (test 3) against three different output
353 files, and it allows the test to pass if any one of them matches. With luck
354 this should make the test pass on some versions of Solaris where it was
355 failing. Because of the uncertainty, the script did not used to stop if
356 test 3 failed; it now does. If further versions of a French locale ever
357 come to light, they can now easily be added.
358
359 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
360 there was a message during ./configure, but it did not stop. This now
361 provokes an error. The invalid example in README has been corrected.
362 If a value less than the minimum is given, the minimum value has always
363 been used, but now a warning is given.
364
365 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
366 was a bug in the test system, which is now fixed. Also, the list of various
367 configurations that are tested for each release did not have one with both
368 16/32 bits and --enable-bar-anycrlf. It now does.
369
370 18. pcretest was missing "-C bsr" for displaying the \R default setting.
371
372 19. Little endian PowerPC systems are supported now by the JIT compiler.
373
374 20. The fast forward newline mechanism could enter to an infinite loop on
375 certain invalid UTF-8 input. Although we don't support these cases
376 this issue can be fixed by a performance optimization.
377
378 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
379 not take account if existing stack usage. There is now a new global
380 variable called pcre_stack_guard that can be set to point to an external
381 function to check stack availability. It is called at the start of
382 processing every parenthesized group.
383
384 22. A typo in the code meant that in ungreedy mode the max/min qualifier
385 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
386 not match "ab".
387
388 23. When UTF was disabled, the JIT program reported some incorrect compile
389 errors. These messages are silenced now.
390
391 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
392 compiler.
393
394 25. Change all the temporary files used in RunGrepTest to be different to those
395 used by RunTest so that the tests can be run simultaneously, for example by
396 "make -j check".
397
398
399 Version 8.34 15-December-2013
400 -----------------------------
401
402 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
403 executable memory. Patch inspired by Carsten Klein.
404
405 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
406 this macro is never tested and has no effect, because the work to support
407 coverage involves only compiling and linking options and special targets in
408 the Makefile. The comment in config.h implied that defining the macro would
409 enable coverage support, which is totally false. There was also support for
410 setting this macro in the CMake files (my fault, I just copied it from
411 configure). SUPPORT_GCOV has now been removed.
412
413 3. Make a small performance improvement in strlen16() and strlen32() in
414 pcretest.
415
416 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
417 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
418 these cases). There was also one in pcretest.c.
419
420 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
421
422 6. In UTF mode, the code for checking whether a group could match an empty
423 string (which is used for indefinitely repeated groups to allow for
424 breaking an infinite loop) was broken when the group contained a repeated
425 negated single-character class with a character that occupied more than one
426 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
427 UTF-8 mode). The effect was undefined: the group might or might not be
428 deemed as matching an empty string, or the program might have crashed.
429
430 7. The code for checking whether a group could match an empty string was not
431 recognizing that \h, \H, \v, \V, and \R must match a character.
432
433 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
434 an empty string. If it can, pcretest shows this in its information output.
435
436 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
437 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
438 matched by pcre_exec() without using JIT:
439
440 (a) If the rest of the pattern did not match after a maximal run of
441 grapheme clusters, the code for backing up to try with fewer of them
442 did not always back up over a full grapheme when characters that do not
443 have the modifier quality were involved, e.g. Hangul syllables.
444
445 (b) If the match point in a subject started with modifier character, and
446 there was no match, the code could incorrectly back up beyond the match
447 point, and potentially beyond the first character in the subject,
448 leading to a segfault or an incorrect match result.
449
450 10. A conditional group with an assertion condition could lead to PCRE
451 recording an incorrect first data item for a match if no other first data
452 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
453 first data item, and therefore matched "ca" after "c" instead of at the
454 start.
455
456 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
457 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
458
459 12. The source of pcregrep now includes z/OS-specific code so that it can be
460 compiled for z/OS as part of the special z/OS distribution.
461
462 13. Added the -T and -TM options to pcretest.
463
464 14. The code in pcre_compile.c for creating the table of named capturing groups
465 has been refactored. Instead of creating the table dynamically during the
466 actual compiling pass, the information is remembered during the pre-compile
467 pass (on the stack unless there are more than 20 named groups, in which
468 case malloc() is used) and the whole table is created before the actual
469 compile happens. This has simplified the code (it is now nearly 150 lines
470 shorter) and prepared the way for better handling of references to groups
471 with duplicate names.
472
473 15. A back reference to a named subpattern when there is more than one of the
474 same name now checks them in the order in which they appear in the pattern.
475 The first one that is set is used for the reference. Previously only the
476 first one was inspected. This change makes PCRE more compatible with Perl.
477
478 16. Unicode character properties were updated from Unicode 6.3.0.
479
480 17. The compile-time code for auto-possessification has been refactored, based
481 on a patch by Zoltan Herczeg. It now happens after instead of during
482 compilation. The code is cleaner, and more cases are handled. The option
483 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
484 options in pcretest are provided to set it. It can also be set by
485 (*NO_AUTO_POSSESS) at the start of a pattern.
486
487 18. The character VT has been added to the default ("C" locale) set of
488 characters that match \s and are generally treated as white space,
489 following this same change in Perl 5.18. There is now no difference between
490 "Perl space" and "POSIX space". Whether VT is treated as white space in
491 other locales depends on the locale.
492
493 19. The code for checking named groups as conditions, either for being set or
494 for being recursed, has been refactored (this is related to 14 and 15
495 above). Processing unduplicated named groups should now be as fast at
496 numerical groups, and processing duplicated groups should be faster than
497 before.
498
499 20. Two patches to the CMake build system, by Alexander Barkov:
500
501 (1) Replace the "source" command by "." in CMakeLists.txt because
502 "source" is a bash-ism.
503
504 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
505 without these the CMake build does not work on Solaris.
506
507 21. Perl has changed its handling of \8 and \9. If there is no previously
508 encountered capturing group of those numbers, they are treated as the
509 literal characters 8 and 9 instead of a binary zero followed by the
510 literals. PCRE now does the same.
511
512 22. Following Perl, added \o{} to specify codepoints in octal, making it
513 possible to specify values greater than 0777 and also making them
514 unambiguous.
515
516 23. Perl now gives an error for missing closing braces after \x{... instead of
517 treating the string as literal. PCRE now does the same.
518
519 24. RunTest used to grumble if an inappropriate test was selected explicitly,
520 but just skip it when running all tests. This make it awkward to run ranges
521 of tests when one of them was inappropriate. Now it just skips any
522 inappropriate tests, as it always did when running all tests.
523
524 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
525 character types such as \d or \w, too many callouts were inserted, and the
526 data that they returned was rubbish.
527
528 26. In UCP mode, \s was not matching two of the characters that Perl matches,
529 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
530 were matched by \h. The code has now been refactored so that the lists of
531 the horizontal and vertical whitespace characters used for \h and \v (which
532 are defined only in one place) are now also used for \s.
533
534 27. Add JIT support for the 64 bit TileGX architecture.
535 Patch by Jiong Wang (Tilera Corporation).
536
537 28. Possessive quantifiers for classes (both explicit and automatically
538 generated) now use special opcodes instead of wrapping in ONCE brackets.
539
540 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
541 (because it's meaningless), this was not happening when PCRE_CASELESS was
542 set. Not wrong, but inefficient.
543
544 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
545 properties for \w, \d, etc) is present in a test regex. Otherwise if the
546 test contains no characters greater than 255, Perl doesn't realise it
547 should be using Unicode semantics.
548
549 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
550 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
551 does in Unicode mode.
552
553 32. Added the "forbid" facility to pcretest so that putting tests into the
554 wrong test files can sometimes be quickly detected.
555
556 33. There is now a limit (default 250) on the depth of nesting of parentheses.
557 This limit is imposed to control the amount of system stack used at compile
558 time. It can be changed at build time by --with-parens-nest-limit=xxx or
559 the equivalent in CMake.
560
561 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
562 errors. Perl warns for these when in warning mode, but PCRE has no facility
563 for giving warnings.
564
565 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
566 However, this was not working for (?!) because it is optimized to (*FAIL),
567 for which PCRE does not allow quantifiers. The optimization is now disabled
568 when a quantifier follows (?!). I can't see any use for this, but it makes
569 things uniform.
570
571 36. Perl no longer allows group names to start with digits, so I have made this
572 change also in PCRE. It simplifies the code a bit.
573
574 37. In extended mode, Perl ignores spaces before a + that indicates a
575 possessive quantifier. PCRE allowed a space before the quantifier, but not
576 before the possessive +. It now does.
577
578 38. The use of \K (reset reported match start) within a repeated possessive
579 group such as (a\Kb)*+ was not working.
580
581 40. Document that the same character tables must be used at compile time and
582 run time, and that the facility to pass tables to pcre_exec() and
583 pcre_dfa_exec() is for use only with saved/restored patterns.
584
585 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
586 features for Builds with MSVC:
587
588 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
589 builds). The .rc files can be used to set FileDescription and many other
590 attributes.
591
592 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
593 This allows higher-level build scripts which want .pdb files to avoid
594 hard-coding the exact files needed."
595
596 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
597 mean "start of word" and "end of word", respectively, as a transition aid.
598
599 43. A minimizing repeat of a class containing codepoints greater than 255 in
600 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
601 compiled to use the heap for recursion.
602
603 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
604 is configured.
605
606
607 Version 8.33 28-May-2013
608 ------------------------
609
610 1. Added 'U' to some constants that are compared to unsigned integers, to
611 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
612 variables that are added to signed variables, to ensure the result is
613 signed and can be negated.
614
615 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
616 CMake config files.
617
618 3. Revise the creation of config.h.generic so that all boolean macros are
619 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
620 overriding via -D on the command line possible.
621
622 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
623 to unsigned int is reported to make a quite noticeable speed difference in
624 a specific Windows environment. Testing on Linux did also appear to show
625 some benefit (and it is clearly not harmful). Also fixed the definition of
626 Xop which should be unsigned.
627
628 5. Related to (4), changing the definition of the intermediate variable cc
629 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
630 improvements.
631
632 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
633 unnecessary spaces.
634
635 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
636 and later.
637
638 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
639
640 9. Optimizing fast_forward_start_bits in JIT.
641
642 10. Adding support for callouts in JIT, and fixing some issues revealed
643 during this work. Namely:
644
645 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
646
647 (b) Minimum length was not checked before the matching is started.
648
649 11. The value of capture_last that is passed to callouts was incorrect in some
650 cases when there was a capture on one path that was subsequently abandoned
651 after a backtrack. Also, the capture_last value is now reset after a
652 recursion, since all captures are also reset in this case.
653
654 12. The interpreter no longer returns the "too many substrings" error in the
655 case when an overflowing capture is in a branch that is subsequently
656 abandoned after a backtrack.
657
658 13. In the pathological case when an offset vector of size 2 is used, pcretest
659 now prints out the matched string after a yield of 0 or 1.
660
661 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
662 Only supported by the JIT compiler at the moment.
663
664 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
665
666 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
667 offset of the starting point of the matching process, provided the offsets
668 vector is large enough.
669
670 17. The \A escape now records a lookbehind value of 1, though its execution
671 does not actually inspect the previous character. This is to ensure that,
672 in partial multi-segment matching, at least one character from the old
673 segment is retained when a new segment is processed. Otherwise, if there
674 are no lookbehinds in the pattern, \A might match incorrectly at the start
675 of a new segment.
676
677 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
678
679 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
680 gives some modest performance improvement in 8-bit mode.
681
682 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
683 be expressed in certain programming languages using Universal Character
684 Names.
685
686 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
687 which points out that "non characters" are not "characters that may not
688 appear in Unicode strings" but rather "characters that are reserved for
689 internal use and have only local meaning".
690
691 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
692 there was a conditional group that depended on an assertion, if the
693 assertion was false, the callout that immediately followed the alternation
694 in the condition was skipped when pcre_exec() was used for matching.
695
696 23. Allow an explicit callout to be inserted before an assertion that is the
697 condition for a conditional group, for compatibility with automatic
698 callouts, which always insert a callout at this point.
699
700 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
701 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
702
703 25. (*PRUNE) is now supported by the JIT compiler.
704
705 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
706
707 27. Fix the case where there are two or more SKIPs with arguments that may be
708 ignored.
709
710 28. (*SKIP) is now supported by the JIT compiler.
711
712 29. (*THEN) is now supported by the JIT compiler.
713
714 30. Update RunTest with additional test selector options.
715
716 31. The way PCRE handles backtracking verbs has been changed in two ways.
717
718 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
719 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
720 backtracking. In some cases this makes it more Perl-compatible, but Perl's
721 rather obscure rules do not always do the same thing.
722
723 (2) Previously, backtracking verbs were confined within assertions. This is
724 no longer the case for positive assertions, except for (*ACCEPT). Again,
725 this sometimes improves Perl compatibility, and sometimes does not.
726
727 32. A number of tests that were in test 2 because Perl did things differently
728 have been moved to test 1, because either Perl or PCRE has changed, and
729 these tests are now compatible.
730
731 32. Backtracking control verbs are now handled in the same way in JIT and
732 interpreter.
733
734 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
735 contained a forward subroutine reference caused a compile error.
736
737 34. Auto-detect and optimize limited repetitions in JIT.
738
739 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
740 blocking (*UTF) etc.
741
742 36. In the interpreter, maximizing pattern repetitions for characters and
743 character types now use tail recursion, which reduces stack usage.
744
745 37. The value of the max lookbehind was not correctly preserved if a compiled
746 and saved regex was reloaded on a host of different endianness.
747
748 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
749 of the compiled pattern block, expand the flags field from 16 to 32 bits
750 because it was almost full.
751
752 39. Try madvise first before posix_madvise.
753
754 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
755 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
756 original change locked out all matches of empty strings. This has been
757 changed so that one match of an empty string per line is recognized.
758 Subsequent searches on the same line (for colouring or for --only-matching,
759 for example) do not recognize empty strings.
760
761 41. Applied a user patch to fix a number of spelling mistakes in comments.
762
763 42. Data lines longer than 65536 caused pcretest to crash.
764
765 43. Clarified the data type for length and startoffset arguments for pcre_exec
766 and pcre_dfa_exec in the function-specific man pages, where they were
767 explicitly stated to be in bytes, never having been updated. I also added
768 some clarification to the pcreapi man page.
769
770 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
771 a segmentation fault.
772
773
774 Version 8.32 30-November-2012
775 -----------------------------
776
777 1. Improved JIT compiler optimizations for first character search and single
778 character iterators.
779
780 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
781 Patch by Daniel Richard G.
782
783 3. Single character iterator optimizations in the JIT compiler.
784
785 4. Improved JIT compiler optimizations for character ranges.
786
787 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
788 Reported by Giuseppe D'Angelo.
789
790 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
791 of a line, was being set incorrectly in cases where .* appeared inside
792 atomic brackets at the start of a pattern, or where there was a subsequent
793 *PRUNE or *SKIP.
794
795 7. Improved instruction cache flush for POWER/PowerPC.
796 Patch by Daniel Richard G.
797
798 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
799 grep:
800
801 (a) There is now no limit to the number of patterns to be matched.
802
803 (b) An error is given if a pattern is too long.
804
805 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
806 are now supported.
807
808 (d) --exclude-from and --include-from (multiple use) have been added.
809
810 (e) Exclusions and inclusions now apply to all files and directories, not
811 just to those obtained from scanning a directory recursively.
812
813 (f) Multiple uses of -f and --file-list are now supported.
814
815 (g) In a Windows environment, the default for -d has been changed from
816 "read" (the GNU grep default) to "skip", because otherwise the presence
817 of a directory in the file list provokes an error.
818
819 (h) The documentation has been revised and clarified in places.
820
821 9. Improve the matching speed of capturing brackets.
822
823 10. Changed the meaning of \X so that it now matches a Unicode extended
824 grapheme cluster.
825
826 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
827 out POSIX threads when JIT support is configured.
828
829 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
830
831 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
832 structure could go wrong in environments where size_t is not the same size
833 as int.
834
835 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
836 set.
837
838 15. The EBCDIC support had decayed; later updates to the code had included
839 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
840 general tidy up of EBCDIC-related issues, and the documentation was also
841 not quite right. There is now a test that can be run on ASCII systems to
842 check some of the EBCDIC-related things (but is it not a full test).
843
844 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
845 in a small tidy to the code.
846
847 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
848
849 18. If the --only-matching (-o) option in pcregrep is specified multiple
850 times, each one causes appropriate output. For example, -o1 -o2 outputs the
851 substrings matched by the 1st and 2nd capturing parentheses. A separating
852 string can be specified by --om-separator (default empty).
853
854 19. Improving the first n character searches.
855
856 20. Turn case lists for horizontal and vertical white space into macros so that
857 they are defined only once.
858
859 21. This set of changes together give more compatible Unicode case-folding
860 behaviour for characters that have more than one other case when UCP
861 support is available.
862
863 (a) The Unicode property table now has offsets into a new table of sets of
864 three or more characters that are case-equivalent. The MultiStage2.py
865 script that generates these tables (the pcre_ucd.c file) now scans
866 CaseFolding.txt instead of UnicodeData.txt for character case
867 information.
868
869 (b) The code for adding characters or ranges of characters to a character
870 class has been abstracted into a generalized function that also handles
871 case-independence. In UTF-mode with UCP support, this uses the new data
872 to handle characters with more than one other case.
873
874 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
875 whose other case is greater than 256 are now correctly matched
876 caselessly. Previously, the high codepoint matched the low one, but not
877 vice versa.
878
879 (d) The processing of \h, \H, \v, and \ in character classes now makes use
880 of the new class addition function, using character lists defined as
881 macros alongside the case definitions of 20 above.
882
883 (e) Caseless back references now work with characters that have more than
884 one other case.
885
886 (f) General caseless matching of characters with more than one other case
887 is supported.
888
889 22. Unicode character properties were updated from Unicode 6.2.0
890
891 23. Improved CMake support under Windows. Patch by Daniel Richard G.
892
893 24. Add support for 32-bit character strings, and UTF-32
894
895 25. Major JIT compiler update (code refactoring and bugfixing).
896 Experimental Sparc 32 support is added.
897
898 26. Applied a modified version of Daniel Richard G's patch to create
899 pcre.h.generic and config.h.generic by "make" instead of in the
900 PrepareRelease script.
901
902 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
903 pcre_compile.c when checking for a zero character.
904
905 28. Introducing a native interface for JIT. Through this interface, the compiled
906 machine code can be directly executed. The purpose of this interface is to
907 provide fast pattern matching, so several sanity checks are not performed.
908 However, feature tests are still performed. The new interface provides
909 1.4x speedup compared to the old one.
910
911 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
912 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
913 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
914
915 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
916 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
917 the "old" RFC 2279). Instead, it ended up passing a negative length to
918 pcre_exec().
919
920 31. Add support for GCC's visibility feature to hide internal functions.
921
922 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
923 "unknown -C option" after outputting 0 or 1.
924
925 33. There is now support for generating a code coverage report for the test
926 suite in environments where gcc is the compiler and lcov is installed. This
927 is mainly for the benefit of the developers.
928
929 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
930 unaddressable using valgrind annotations, allowing valgrind to detect
931 invalid memory accesses. This is mainly for the benefit of the developers.
932
933 25. (*UTF) can now be used to start a pattern in any of the three libraries.
934
935 26. Give configure error if --enable-cpp but no C++ compiler found.
936
937
938 Version 8.31 06-July-2012
939 -------------------------
940
941 1. Fixing a wrong JIT test case and some compiler warnings.
942
943 2. Removed a bashism from the RunTest script.
944
945 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
946 to unsigned type, result still unsigned" that was given by an MS compiler
947 on encountering the code "-sizeof(xxx)".
948
949 4. Partial matching support is added to the JIT compiler.
950
951 5. Fixed several bugs concerned with partial matching of items that consist
952 of more than one character:
953
954 (a) /^(..)\1/ did not partially match "aba" because checking references was
955 done on an "all or nothing" basis. This also applied to repeated
956 references.
957
958 (b) \R did not give a hard partial match if \r was found at the end of the
959 subject.
960
961 (c) \X did not give a hard partial match after matching one or more
962 characters at the end of the subject.
963
964 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
965 a partial match for the string "\r".
966
967 (e) When newline was set to CRLF, the metacharacter "." did not recognize
968 a partial match for a CR character at the end of the subject string.
969
970 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
971 running pcretest, the text "(JIT)" added to the output whenever JIT is
972 actually used to run the match.
973
974 7. Individual JIT compile options can be set in pcretest by following -s+[+]
975 or /S+[+] with a digit between 1 and 7.
976
977 8. OP_NOT now supports any UTF character not just single-byte ones.
978
979 9. (*MARK) control verb is now supported by the JIT compiler.
980
981 10. The command "./RunTest list" lists the available tests without actually
982 running any of them. (Because I keep forgetting what they all are.)
983
984 11. Add PCRE_INFO_MAXLOOKBEHIND.
985
986 12. Applied a (slightly modified) user-supplied patch that improves performance
987 when the heap is used for recursion (compiled with --disable-stack-for-
988 recursion). Instead of malloc and free for each heap frame each time a
989 logical recursion happens, frames are retained on a chain and re-used where
990 possible. This sometimes gives as much as 30% improvement.
991
992 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
993 call.
994
995 14. As documented, (*COMMIT) is now confined to within a positive assertion.
996
997 15. It is now possible to link pcretest with libedit as an alternative to
998 libreadline.
999
1000 16. (*COMMIT) control verb is now supported by the JIT compiler.
1001
1002 17. The Unicode data tables have been updated to Unicode 6.1.0.
1003
1004 18. Added --file-list option to pcregrep.
1005
1006 19. Added binary file support to pcregrep, including the -a, --binary-files,
1007 -I, and --text options.
1008
1009 20. The madvise function is renamed for posix_madvise for QNX compatibility
1010 reasons. Fixed by Giuseppe D'Angelo.
1011
1012 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1013
1014 22. Changed the option for creating symbolic links for 16-bit man pages from
1015 -s to -sf so that re-installing does not cause issues.
1016
1017 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1018
1019 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1020 matches in certain environments (the workspace was not being correctly
1021 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1022 some of the workspace data at the beginning of a restart.
1023
1024 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1025 was not doing so when it should - probably a typo introduced by SVN 528
1026 (change 8.10/14).
1027
1028 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1029 \w+ when the character tables indicated that \x{c4} was a word character.
1030 There were several related cases, all because the tests for doing a table
1031 lookup were testing for characters less than 127 instead of 255.
1032
1033 27. If a pattern contains capturing parentheses that are not used in a match,
1034 their slots in the ovector are set to -1. For those that are higher than
1035 any matched groups, this happens at the end of processing. In the case when
1036 there were back references that the ovector was too small to contain
1037 (causing temporary malloc'd memory to be used during matching), and the
1038 highest capturing number was not used, memory off the end of the ovector
1039 was incorrectly being set to -1. (It was using the size of the temporary
1040 memory instead of the true size.)
1041
1042 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1043 ovector size, it uses memory at the end of the block that it has got.
1044
1045 29. Check for an overlong MARK name and give an error at compile time. The
1046 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1047
1048 30. JIT compiler update.
1049
1050 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1051 Rill for the patch.
1052
1053 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1054
1055 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1056
1057 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1058 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1059 was enabled.
1060
1061 35. Improve JIT code generation for greedy plus quantifier.
1062
1063 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1064 1 to "aa" instead of to an empty string. The bug affected repeated groups
1065 that could potentially match an empty string.
1066
1067 37. Optimizing single character iterators in JIT.
1068
1069 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1070 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1071 codepoints that are too big for the mode are faulted, and in a UTF mode,
1072 disallowed codepoints are also faulted.
1073
1074 39. If PCRE was compiled with UTF support, in three places in the DFA
1075 matcher there was code that should only have been obeyed in UTF mode, but
1076 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1077 processing when bytes with values greater than 127 were present. In 16-bit
1078 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1079 both cases the values are those that cannot be the first data item in a UTF
1080 character. The three items that might have provoked this were recursions,
1081 possessively repeated groups, and atomic groups.
1082
1083 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1084 and pcregrep, because some OS require shared objects to be explicitly
1085 passed to ld, causing the link step to fail if they are not.
1086
1087 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1088 mode, patterns that started with \h* or \R* might be incorrectly matched.
1089
1090
1091 Version 8.30 04-February-2012
1092 -----------------------------
1093
1094 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1095 name is defined in ctype.h.
1096
1097 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1098 only in quite long subpatterns.
1099
1100 3. Removed the function pcre_info(), which has been obsolete and deprecated
1101 since it was replaced by pcre_fullinfo() in February 2000.
1102
1103 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1104 match a (*MARK), and the match failed at the start of the subject, a
1105 reference to memory before the start of the subject could occur. This bug
1106 was introduced by fix 17 of release 8.21.
1107
1108 5. A reference to an unset group with zero minimum repetition was giving
1109 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1110 /(another)?(\1?)test/ matched against "hello world test". This bug was
1111 introduced in release 8.13.
1112
1113 6. Add support for 16-bit character strings (a large amount of work involving
1114 many changes and refactorings).
1115
1116 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1117 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1118 from a file.
1119
1120 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1121 rounding is not applied in this particular case).
1122
1123 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1124 if they appear, or are escaped, in patterns.
1125
1126 10. Get rid of a number of -Wunused-but-set-variable warnings.
1127
1128 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1129 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1130 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1131 also returns the mark "x". This bug applied to capturing parentheses,
1132 non-capturing parentheses, and atomic parentheses. It also applied to some
1133 assertions.
1134
1135 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1136 information out of configure.ac instead of relying on pcre.h.generic, which
1137 is not stored in the repository.
1138
1139 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1140 -lreadline.
1141
1142 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1143
1144 15. Applied Graycode's patch to put the top-level frame on the stack rather
1145 than the heap when not using the stack for recursion. This gives a
1146 performance improvement in many cases when recursion is not deep.
1147
1148 16. Experimental code added to "pcretest -C" to output the stack frame size.
1149
1150
1151 Version 8.21 12-Dec-2011
1152 ------------------------
1153
1154 1. Updating the JIT compiler.
1155
1156 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1157 are added as well.
1158
1159 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1160 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1161 calling _pcre_jit_exec. Some extra comments are added.
1162
1163 4. (*MARK) settings inside atomic groups that do not contain any capturing
1164 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1165 was introduced by change 18 for 8.20.
1166
1167 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1168 ECMA-262 standard.
1169
1170 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1171 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1172 This bug was probably introduced by change 9 of 8.13.
1173
1174 7. While fixing 6 above, I noticed that a number of other items were being
1175 incorrectly rejected as "not fixed length". This arose partly because newer
1176 opcodes had not been added to the fixed-length checking code. I have (a)
1177 corrected the bug and added tests for these items, and (b) arranged for an
1178 error to occur if an unknown opcode is encountered while checking for fixed
1179 length instead of just assuming "not fixed length". The items that were
1180 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1181 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1182 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1183
1184 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1185 being incorrectly compiled and would have given unpredicatble results.
1186
1187 9. A possessively repeated subpattern with minimum repeat count greater than
1188 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1189 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1190 the first (A) could occur when it should not.
1191
1192 10. Add a cast and remove a redundant test from the code.
1193
1194 11. JIT should use pcre_malloc/pcre_free for allocation.
1195
1196 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1197 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1198 is anything other than /usr, -L is still shown).
1199
1200 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1201
1202 14. Perl does not support \N without a following name in a [] class; PCRE now
1203 also gives an error.
1204
1205 15. If a forward reference was repeated with an upper limit of around 2000,
1206 it caused the error "internal error: overran compiling workspace". The
1207 maximum number of forward references (including repeats) was limited by the
1208 internal workspace, and dependent on the LINK_SIZE. The code has been
1209 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1210 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1211 of around 200,000 forward references. While doing this, I also speeded up
1212 the filling in of repeated forward references.
1213
1214 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1215 incorrectly expecting the subject to contain another "a" after the start.
1216
1217 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1218 in the match, the SKIP should be ignored. This was not happening; instead
1219 the SKIP was being treated as NOMATCH. For patterns such as
1220 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1221 tested.
1222
1223 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1224 now much more compatible with Perl, in particular in cases where the result
1225 is a non-match for a non-anchored pattern. For example, if
1226 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1227 "m", where previously it did not return a name. A side effect of this
1228 change is that for partial matches, the last encountered mark name is
1229 returned, as for non matches. A number of tests that were previously not
1230 Perl-compatible have been moved into the Perl-compatible test files. The
1231 refactoring has had the pleasing side effect of removing one argument from
1232 the match() function, thus reducing its stack requirements.
1233
1234 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1235 subsequent uses of /S (without +) incorrectly behaved like /S+.
1236
1237 21. Retrieve executable code size support for the JIT compiler and fixing
1238 some warnings.
1239
1240 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1241 not work when the shorter character appeared right at the end of the
1242 subject string.
1243
1244 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1245 systems.
1246
1247 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1248 output it when the /M option is used in pcretest.
1249
1250 25. The CheckMan script was not being included in the distribution. Also, added
1251 an explicit "perl" to run Perl scripts from the PrepareRelease script
1252 because this is reportedly needed in Windows.
1253
1254 26. If study data was being save in a file and studying had not found a set of
1255 "starts with" bytes for the pattern, the data written to the file (though
1256 never used) was taken from uninitialized memory and so caused valgrind to
1257 complain.
1258
1259 27. Updated RunTest.bat as provided by Sheri Pierce.
1260
1261 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1262
1263 29. Computation of memory usage for the table of capturing group names was
1264 giving an unnecessarily large value.
1265
1266
1267 Version 8.20 21-Oct-2011
1268 ------------------------
1269
1270 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1271 a POSIX class. After further experiments with Perl, which convinced me that
1272 Perl has bugs and confusions, a closing square bracket is no longer allowed
1273 in a POSIX name. This bug also affected patterns with classes that started
1274 with full stops.
1275
1276 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1277 captured substring, but while checking the failing first alternative,
1278 substring 1 is temporarily captured. If the output vector supplied to
1279 pcre_exec() was not big enough for this capture, the yield of the function
1280 was still zero ("insufficient space for captured substrings"). This cannot
1281 be totally fixed without adding another stack variable, which seems a lot
1282 of expense for a edge case. However, I have improved the situation in cases
1283 such as /(a)(b)x|abc/ matched against "abc", where the return code
1284 indicates that fewer than the maximum number of slots in the ovector have
1285 been set.
1286
1287 3. Related to (2) above: when there are more back references in a pattern than
1288 slots in the output vector, pcre_exec() uses temporary memory during
1289 matching, and copies in the captures as far as possible afterwards. It was
1290 using the entire output vector, but this conflicts with the specification
1291 that only 2/3 is used for passing back captured substrings. Now it uses
1292 only the first 2/3, for compatibility. This is, of course, another edge
1293 case.
1294
1295 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1296 main code base, and can be used by building with --enable-jit. When this is
1297 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1298 runtime --no-jit option is given.
1299
1300 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1301 ovector, the return from the function was zero, implying that there were
1302 other matches that did not fit. The correct "exactly full" value is now
1303 returned.
1304
1305 6. If a subpattern that was called recursively or as a subroutine contained
1306 (*PRUNE) or any other control that caused it to give a non-standard return,
1307 invalid errors such as "Error -26 (nested recursion at the same subject
1308 position)" or even infinite loops could occur.
1309
1310 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1311 computing the minimum length on reaching *ACCEPT, and so ended up with the
1312 wrong value of 1 rather than 0. Further investigation indicates that
1313 computing a minimum subject length in the presence of *ACCEPT is difficult
1314 (think back references, subroutine calls), and so I have changed the code
1315 so that no minimum is registered for a pattern that contains *ACCEPT.
1316
1317 8. If (*THEN) was present in the first (true) branch of a conditional group,
1318 it was not handled as intended. [But see 16 below.]
1319
1320 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1321 Sheri Pierce.
1322
1323 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1324 the first byte in a match must be "a".
1325
1326 11. Change 17 for 8.13 increased the recursion depth for patterns like
1327 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1328 pattern contains any instances of (*THEN). If it does not, the old
1329 optimizations are restored. It would be nice to do this on a per-group
1330 basis, but at the moment that is not feasible.
1331
1332 12. In some environments, the output of pcretest -C is CRLF terminated. This
1333 broke RunTest's code that checks for the link size. A single white space
1334 character after the value is now allowed for.
1335
1336 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1337 For "fr", it uses the Windows-specific input and output files.
1338
1339 14. If (*THEN) appeared in a group that was called recursively or as a
1340 subroutine, it did not work as intended. [But see next item.]
1341
1342 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1343 pattern fragments (but not containing any | characters). If A and B are
1344 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1345 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1346 D. In other words, Perl considers parentheses that do not contain any |
1347 characters to be part of a surrounding alternative, whereas PCRE was
1348 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1349 differently. PCRE now behaves in the same way as Perl, except in the case
1350 of subroutine/recursion calls such as (?1) which have in any case always
1351 been different (but PCRE had them first :-).
1352
1353 16. Related to 15 above: Perl does not treat the | in a conditional group as
1354 creating alternatives. Such a group is treated in the same way as an
1355 ordinary group without any | characters when processing (*THEN). PCRE has
1356 been changed to match Perl's behaviour.
1357
1358 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1359 RunGrepTest script failed.
1360
1361 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1362 inevitable for groups that contain captures, but it can lead to a lot of
1363 stack use in large patterns. The old behaviour has been restored for atomic
1364 groups that do not contain any capturing parentheses.
1365
1366 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1367 suppress the check for a minimum subject length at run time. (If it was
1368 given to pcre_exec() or pcre_dfa_exec() it did work.)
1369
1370 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1371 fail to work when decoding hex characters in data strings in EBCDIC
1372 environments.
1373
1374 21. It appears that in at least one Mac OS environment, the isxdigit() function
1375 is implemented as a macro that evaluates to its argument more than once,
1376 contravening the C 90 Standard (I haven't checked a later standard). There
1377 was an instance in pcretest which caused it to go wrong when processing
1378 \x{...} escapes in subject strings. The has been rewritten to avoid using
1379 things like p++ in the argument of isxdigit().
1380
1381
1382 Version 8.13 16-Aug-2011
1383 ------------------------
1384
1385 1. The Unicode data tables have been updated to Unicode 6.0.0.
1386
1387 2. Two minor typos in pcre_internal.h have been fixed.
1388
1389 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1390 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1391 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1392
1393 4. There were a number of related bugs in the code for matching backrefences
1394 caselessly in UTF-8 mode when codes for the characters concerned were
1395 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1396 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1397 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1398 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1399 2-byte code at the end of the subject (it thought there wasn't enough data
1400 left).
1401
1402 5. Comprehensive information about what went wrong is now returned by
1403 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1404 as the output vector has at least 2 elements. The offset of the start of
1405 the failing character and a reason code are placed in the vector.
1406
1407 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1408 now returned is for the first byte of the failing character, instead of the
1409 last byte inspected. This is an incompatible change, but I hope it is small
1410 enough not to be a problem. It makes the returned offset consistent with
1411 pcre_exec() and pcre_dfa_exec().
1412
1413 7. pcretest now gives a text phrase as well as the error number when
1414 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1415 failure, the offset and reason code are output.
1416
1417 8. When \R was used with a maximizing quantifier it failed to skip backwards
1418 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1419 back over a single character (\n). This seems wrong (because it treated the
1420 two characters as a single entity when going forwards), conflicts with the
1421 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1422 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1423 has been changed.
1424
1425 9. Some internal refactoring has changed the processing so that the handling
1426 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1427 time (the PCRE_DOTALL option was changed this way some time ago: version
1428 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1429 which was always a bit of a fudge. It also means that there is one less
1430 argument for the match() function, which reduces its stack requirements
1431 slightly. This change also fixes an incompatibility with Perl: the pattern
1432 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1433
1434 10. More internal refactoring has drastically reduced the number of recursive
1435 calls to match() for possessively repeated groups such as (abc)++ when
1436 using pcre_exec().
1437
1438 11. While implementing 10, a number of bugs in the handling of groups were
1439 discovered and fixed:
1440
1441 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1442 (a|)*(?1) gave a compile-time internal error.
1443 ((a|)+)+ did not notice that the outer group could match an empty string.
1444 (^a|^)+ was not marked as anchored.
1445 (.*a|.*)+ was not marked as matching at start or after a newline.
1446
1447 12. Yet more internal refactoring has removed another argument from the match()
1448 function. Special calls to this function are now indicated by setting a
1449 value in a variable in the "match data" data block.
1450
1451 13. Be more explicit in pcre_study() instead of relying on "default" for
1452 opcodes that mean there is no starting character; this means that when new
1453 ones are added and accidentally left out of pcre_study(), testing should
1454 pick them up.
1455
1456 14. The -s option of pcretest has been documented for ages as being an old
1457 synonym of -m (show memory usage). I have changed it to mean "force study
1458 for every regex", that is, assume /S for every regex. This is similar to -i
1459 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1460 using it. It makes it easier to run collections of tests with and without
1461 study enabled, and thereby test pcre_study() more easily. All the standard
1462 tests are now run with and without -s (but some patterns can be marked as
1463 "never study" - see 20 below).
1464
1465 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1466 restoration of the capturing data to the outer values was not happening
1467 correctly.
1468
1469 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1470 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1471 pattern had matched an empty string, and so incorrectly returned a no
1472 match.
1473
1474 17. There was optimizing code for the last branch of non-capturing parentheses,
1475 and also for the obeyed branch of a conditional subexpression, which used
1476 tail recursion to cut down on stack usage. Unfortunately, now that there is
1477 the possibility of (*THEN) occurring in these branches, tail recursion is
1478 no longer possible because the return has to be checked for (*THEN). These
1479 two optimizations have therefore been removed. [But see 8.20/11 above.]
1480
1481 18. If a pattern containing \R was studied, it was assumed that \R always
1482 matched two bytes, thus causing the minimum subject length to be
1483 incorrectly computed because \R can also match just one byte.
1484
1485 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1486 was incorrectly computed.
1487
1488 20. If /S is present twice on a test pattern in pcretest input, it now
1489 *disables* studying, thereby overriding the use of -s on the command line
1490 (see 14 above). This is necessary for one or two tests to keep the output
1491 identical in both cases.
1492
1493 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1494 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1495
1496 22. When an atomic group that contained a capturing parenthesis was
1497 successfully matched, but the branch in which it appeared failed, the
1498 capturing was not being forgotten if a higher numbered group was later
1499 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1500 group 1 to "a", when in fact it should be unset. This applied to multi-
1501 branched capturing and non-capturing groups, repeated or not, and also to
1502 positive assertions (capturing in negative assertions does not happen
1503 in PCRE) and also to nested atomic groups.
1504
1505 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1506 subject after a captured substring, to make it easier to tell which of a
1507 number of identical substrings has been captured.
1508
1509 24. The way atomic groups are processed by pcre_exec() has been changed so that
1510 if they are repeated, backtracking one repetition now resets captured
1511 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1512 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1513 "aaa". Previously, it would have been "a". As part of this code
1514 refactoring, the way recursive calls are handled has also been changed.
1515
1516 25. If an assertion condition captured any substrings, they were not passed
1517 back unless some other capturing happened later. For example, if
1518 (?(?=(a))a) was matched against "a", no capturing was returned.
1519
1520 26. When studying a pattern that contained subroutine calls or assertions,
1521 the code for finding the minimum length of a possible match was handling
1522 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1523 group 1 called group 2 while simultaneously a separate group 2 called group
1524 1). A stack overflow occurred in this case. I have fixed this by limiting
1525 the recursion depth to 10.
1526
1527 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1528 Fortmann. This supports explicit test numbers on the command line, and has
1529 argument validation and error reporting.
1530
1531 28. An instance of \X with an unlimited repeat could fail if at any point the
1532 first character it looked at was a mark character.
1533
1534 29. Some minor code refactoring concerning Unicode properties and scripts
1535 should reduce the stack requirement of match() slightly.
1536
1537 30. Added the '=' option to pcretest to check the setting of unused capturing
1538 slots at the end of the pattern, which are documented as being -1, but are
1539 not included in the return count.
1540
1541 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1542 compiled something random. Now it gives a compile-time error (as does
1543 Perl).
1544
1545 32. A *MARK encountered during the processing of a positive assertion is now
1546 recorded and passed back (compatible with Perl).
1547
1548 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1549 had alternative anchored branches, the search for a second match in a line
1550 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1551 matched the line "0102" twice. The same bug affected patterns that started
1552 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1553 twice.
1554
1555 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1556 does, and because of capturing effects, quantifying parenthesized
1557 assertions may at times be useful. Quantifiers are now allowed for
1558 parenthesized assertions.
1559
1560 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1561
1562 36. \g was being checked for fancy things in a character class, when it should
1563 just be a literal "g".
1564
1565 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1566 appearance of a nested POSIX class supersedes an apparent external class.
1567 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1568 unescaped square brackets may also appear as part of class names. For
1569 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1570 more like Perl. (But see 8.20/1 above.)
1571
1572 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1573 was because it thought it was \N{name}, which is not supported).
1574
1575 39. Add minix to OS list not supporting the -S option in pcretest.
1576
1577 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1578 cannot analyze patterns in sufficient detail to catch mutual recursions
1579 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1580 subgroup is called recursively as a subpattern for a second time at the
1581 same position in the subject string. In previous releases this might have
1582 been caught by the recursion limit, or it might have run out of stack.
1583
1584 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1585 happen only once. PCRE was, however incorrectly giving a compile time error
1586 "recursive call could loop indefinitely" because it cannot analyze the
1587 pattern in sufficient detail. The compile time test no longer happens when
1588 PCRE is compiling a conditional subpattern, but actual runaway loops are
1589 now caught at runtime (see 40 above).
1590
1591 42. It seems that Perl allows any characters other than a closing parenthesis
1592 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1593 has been changed to be the same.
1594
1595 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1596 as not to get warnings when autogen.sh is called. Also changed
1597 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1598
1599 44. To help people who use pcregrep to scan files containing exceedingly long
1600 lines, the following changes have been made:
1601
1602 (a) The default value of the buffer size parameter has been increased from
1603 8K to 20K. (The actual buffer used is three times this size.)
1604
1605 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1606 PCRE is built.
1607
1608 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1609 to be set at run time.
1610
1611 (d) Numerical values in pcregrep options can be followed by K or M, for
1612 example --buffer-size=50K.
1613
1614 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1615 given and the return code is set to 2.
1616
1617 45. Add a pointer to the latest mark to the callout data block.
1618
1619 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1620 partial match of an empty string instead of no match. This was specific to
1621 the use of ".".
1622
1623 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1624 complete match instead of a partial match. This bug was dependent on both
1625 the PCRE_UTF8 and PCRE_DOTALL options being set.
1626
1627 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1628 starting byte set, because \b was not being ignored.
1629
1630
1631 Version 8.12 15-Jan-2011
1632 ------------------------
1633
1634 1. Fixed some typos in the markup of the man pages, and wrote a script that
1635 checks for such things as part of the documentation building process.
1636
1637 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1638 --match-limit and --recursion-limit options (added for 8.11). In
1639 particular, this made one of the standard tests fail. (The integer value
1640 went into the wrong half of a long int.)
1641
1642 3. If the --colour option was given to pcregrep with -v (invert match), it
1643 did strange things, either producing crazy output, or crashing. It should,
1644 of course, ignore a request for colour when reporting lines that do not
1645 match.
1646
1647 4. Another pcregrep bug caused similar problems if --colour was specified with
1648 -M (multiline) and the pattern match finished with a line ending.
1649
1650 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1651 matched in multiline mode, the following line was shown as part of the
1652 match. This seems wrong, so I have changed it.
1653
1654 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1655 the check for further matches in the same line (so they could be coloured)
1656 to overrun the end of the current line. If another match was found, it was
1657 incorrectly shown (and then shown again when found in the next line).
1658
1659 7. If pcregrep was compiled under Windows, there was a reference to the
1660 function pcregrep_exit() before it was defined. I am assuming this was
1661 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1662 reported by a user. I've moved the definition above the reference.
1663
1664
1665 Version 8.11 10-Dec-2010
1666 ------------------------
1667
1668 1. (*THEN) was not working properly if there were untried alternatives prior
1669 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1670 backtracked to try for "b" instead of moving to the next alternative branch
1671 at the same level (in this case, to look for "c"). The Perl documentation
1672 is clear that when (*THEN) is backtracked onto, it goes to the "next
1673 alternative in the innermost enclosing group".
1674
1675 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1676 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1677 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1678 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1679 (*THEN).
1680
1681 3. If \s appeared in a character class, it removed the VT character from
1682 the class, even if it had been included by some previous item, for example
1683 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1684 of \s, but is part of the POSIX "space" class.)
1685
1686 4. A partial match never returns an empty string (because you can always
1687 match an empty string at the end of the subject); however the checking for
1688 an empty string was starting at the "start of match" point. This has been
1689 changed to the "earliest inspected character" point, because the returned
1690 data for a partial match starts at this character. This means that, for
1691 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1692 (previously it gave "no match").
1693
1694 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1695 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1696 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1697 has an implication that the given string is incomplete (because a partial
1698 match is preferred over a full match). For this reason, these items now
1699 give a partial match in this situation. [Aside: previously, the one case
1700 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1701 match rather than a full match, which was wrong by the old rules, but is
1702 now correct.]
1703
1704 6. There was a bug in the handling of #-introduced comments, recognized when
1705 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1706 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1707 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1708 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1709 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1710 places in pcre_compile().
1711
1712 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1713 comments when looking ahead for named forward references to subpatterns,
1714 the only newline sequence it recognized was NL. It now handles newlines
1715 according to the set newline convention.
1716
1717 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1718 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1719 cater for a lack of strerror(). These oversights have been fixed.
1720
1721 9. Added --match-limit and --recursion-limit to pcregrep.
1722
1723 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1724
1725 11. When the -o option was used, pcregrep was setting a return code of 1, even
1726 when matches were found, and --line-buffered was not being honoured.
1727
1728 12. Added an optional parentheses number to the -o and --only-matching options
1729 of pcregrep.
1730
1731 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1732 can match an empty string. The code to do it in pcretest and pcredemo
1733 needed fixing:
1734
1735 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1736 only one byte after an empty string match just before CRLF (this case
1737 just got forgotten; "any" and "anycrlf" were OK).
1738
1739 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1740 mode when an empty string match preceded an ASCII character followed by
1741 a non-ASCII character. (The code for advancing by one character rather
1742 than one byte was nonsense.)
1743
1744 (c) The pcredemo.c sample program did not have any code at all to handle
1745 the cases when CRLF is a valid newline sequence.
1746
1747 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1748 as a starting offset was within the subject string. There is now a new
1749 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1750 negative or greater than the length of the string. In order to test this,
1751 pcretest is extended to allow the setting of negative starting offsets.
1752
1753 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1754 starting offset points to the beginning of a UTF-8 character was
1755 unnecessarily clumsy. I tidied it up.
1756
1757 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1758 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1759
1760 17. Nobody had reported that the --include_dir option, which was added in
1761 release 7.7 should have been called --include-dir (hyphen, not underscore)
1762 for compatibility with GNU grep. I have changed it to --include-dir, but
1763 left --include_dir as an undocumented synonym, and the same for
1764 --exclude-dir, though that is not available in GNU grep, at least as of
1765 release 2.5.4.
1766
1767 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1768 characters from a string of bytes) have been redefined so as not to use
1769 loops, in order to improve performance in some environments. At the same
1770 time, I abstracted some of the common code into auxiliary macros to save
1771 repetition (this should not affect the compiled code).
1772
1773 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1774 compile-time error is now given if \c is not followed by an ASCII
1775 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1776 different, and any byte value is allowed.)
1777
1778 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1779 START_OPTIMIZE option, which is now allowed at compile time - but just
1780 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1781 to pcregrep and other applications that have no direct access to PCRE
1782 options. The new /Y option in pcretest sets this option when calling
1783 pcre_compile().
1784
1785 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1786 back references. Groups containing recursive back references were forced to
1787 be atomic by that change, but in the case of named groups, the amount of
1788 memory required was incorrectly computed, leading to "Failed: internal
1789 error: code overflow". This has been fixed.
1790
1791 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1792 pcretest.c, to avoid build problems in some Borland environments.
1793
1794
1795 Version 8.10 25-Jun-2010
1796 ------------------------
1797
1798 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1799 THEN.
1800
1801 2. (*ACCEPT) was not working when inside an atomic group.
1802
1803 3. Inside a character class, \B is treated as a literal by default, but
1804 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1805 causes the error). The code is unchanged, but I tidied the documentation.
1806
1807 4. Inside a character class, PCRE always treated \R and \X as literals,
1808 whereas Perl faults them if its -w option is set. I have changed PCRE so
1809 that it faults them when PCRE_EXTRA is set.
1810
1811 5. Added support for \N, which always matches any character other than
1812 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1813
1814 6. When compiling pcregrep with newer versions of gcc which may have
1815 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1816 declared with attribute warn_unused_result" were given. Just casting the
1817 result to (void) does not stop the warnings; a more elaborate fudge is
1818 needed. I've used a macro to implement this.
1819
1820 7. Minor change to pcretest.c to avoid a compiler warning.
1821
1822 8. Added four artifical Unicode properties to help with an option to make
1823 \s etc use properties (see next item). The new properties are: Xan
1824 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1825
1826 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1827 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1828 this option. Modified pcretest to add /W to test this facility. Added
1829 REG_UCP to make it available via the POSIX interface.
1830
1831 10. Added --line-buffered to pcregrep.
1832
1833 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1834 studied, and the match started with a letter with a code point greater than
1835 127 whose first byte was different to the first byte of the other case of
1836 the letter, the other case of this starting letter was not recognized
1837 (#976).
1838
1839 12. If a pattern that was studied started with a repeated Unicode property
1840 test, for example, \p{Nd}+, there was the theoretical possibility of
1841 setting up an incorrect bitmap of starting bytes, but fortunately it could
1842 not have actually happened in practice until change 8 above was made (it
1843 added property types that matched character-matching opcodes).
1844
1845 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1846 possible starting bytes for non-anchored patterns.
1847
1848 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1849 \R, and also a number of cases that involve Unicode properties, both
1850 explicit and implicit when PCRE_UCP is set.
1851
1852 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1853 input, it could crash or give wrong results if characters with values
1854 greater than 0xc0 were present in the subject string. (Detail: it assumed
1855 UTF-8 input when processing these items.)
1856
1857 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1858 size_t is 64-bit (#991).
1859
1860 17. Added a check for running out of memory when PCRE is compiled with
1861 --disable-stack-for-recursion (#990).
1862
1863 18. If the last data line in a file for pcretest does not have a newline on
1864 the end, a newline was missing in the output.
1865
1866 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1867 less than 128) in its various bitmaps. However, there is a facility for
1868 generating tables according to the current locale when PCRE is compiled. It
1869 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1870 characters, are recognized by isspace() and therefore were getting set in
1871 these tables, and indeed these tables seem to approximate to ISO 8859. This
1872 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1873 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1874 which of course cannot start UTF-8 characters. I have changed the code so
1875 that only real ASCII characters (less than 128) and the correct starting
1876 bytes for UTF-8 encodings are set for characters greater than 127 when in
1877 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1878 altogether.)
1879
1880 20. Added the /T option to pcretest so as to be able to run tests with non-
1881 standard character tables, thus making it possible to include the tests
1882 used for 19 above in the standard set of tests.
1883
1884 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1885 reference to a subpattern the other side of a comment that contains an
1886 opening parenthesis caused either an internal compiling error, or a
1887 reference to the wrong subpattern.
1888
1889
1890 Version 8.02 19-Mar-2010
1891 ------------------------
1892
1893 1. The Unicode data tables have been updated to Unicode 5.2.0.
1894
1895 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1896 configured.
1897
1898 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1899 original author of that file, following a query about its status.
1900
1901 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1902 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1903
1904 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1905 quantifier applied to a forward-referencing subroutine call, could compile
1906 incorrect code or give the error "internal error: previously-checked
1907 referenced subpattern not found".
1908
1909 6. Both MS Visual Studio and Symbian OS have problems with initializing
1910 variables to point to external functions. For these systems, therefore,
1911 pcre_malloc etc. are now initialized to local functions that call the
1912 relevant global functions.
1913
1914 7. There were two entries missing in the vectors called coptable and poptable
1915 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1916 I've fixed the data, and added a kludgy way of testing at compile time that
1917 the lengths are correct (equal to the number of opcodes).
1918
1919 8. Following on from 7, I added a similar kludge to check the length of the
1920 eint vector in pcreposix.c.
1921
1922 9. Error texts for pcre_compile() are held as one long string to avoid too
1923 much relocation at load time. To find a text, the string is searched,
1924 counting zeros. There was no check for running off the end of the string,
1925 which could happen if a new error number was added without updating the
1926 string.
1927
1928 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1929
1930 11. \K was not working if it appeared in an atomic group or in a group that
1931 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1932 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1933 the assertion is positive, but not if it is negative.
1934
1935 12. Change 11 fortuitously reduced the size of the stack frame used in the
1936 "match()" function of pcre_exec.c by one pointer. Forthcoming
1937 implementation of support for (*MARK) will need an extra pointer on the
1938 stack; I have reserved it now, so that the stack frame size does not
1939 decrease.
1940
1941 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1942 item in branch that calls a recursion is a subroutine call - as in the
1943 second branch in the above example - was incorrectly given the compile-
1944 time error "recursive call could loop indefinitely" because pcre_compile()
1945 was not correctly checking the subroutine for matching a non-empty string.
1946
1947 14. The checks for overrunning compiling workspace could trigger after an
1948 overrun had occurred. This is a "should never occur" error, but it can be
1949 triggered by pathological patterns such as hundreds of nested parentheses.
1950 The checks now trigger 100 bytes before the end of the workspace.
1951
1952 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1953
1954
1955 Version 8.01 19-Jan-2010
1956 ------------------------
1957
1958 1. If a pattern contained a conditional subpattern with only one branch (in
1959 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1960 computed the wrong minimum data length (which is of course zero for such
1961 subpatterns). This could cause incorrect "no match" results.
1962
1963 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1964 the pattern is reset in the first branch, pcre_compile() failed with
1965 "internal error: code overflow at offset...". This happened only when
1966 the reset was to the original external option setting. (An optimization
1967 abstracts leading options settings into an external setting, which was the
1968 cause of this.)
1969
1970 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1971 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1972 assertion pattern did not match (meaning that the assertion was true), it
1973 was incorrectly treated as false if the SKIP had been reached during the
1974 matching. This also applied to assertions used as conditions.
1975
1976 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1977 assertion subpattern, including such a pattern used as a condition,
1978 unpredictable results occurred, instead of the error return
1979 PCRE_ERROR_DFA_UITEM.
1980
1981 5. The C++ GlobalReplace function was not working like Perl for the special
1982 situation when an empty string is matched. It now does the fancy magic
1983 stuff that is necessary.
1984
1985 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1986 removed. (These were left over from very, very early versions of PCRE.)
1987
1988 7. Some cosmetic changes to the code to make life easier when compiling it
1989 as part of something else:
1990
1991 (a) Change DEBUG to PCRE_DEBUG.
1992
1993 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1994 called "current" as "current_branch", to prevent a collision with the
1995 Linux macro when compiled as a kernel module.
1996
1997 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1998 prevent a collision with the Linux macro when compiled as a kernel
1999 module.
2000
2001 8. In pcre_compile() there are some checks for integer overflows that used to
2002 cast potentially large values to (double). This has been changed to that
2003 when building, a check for int64_t is made, and if it is found, it is used
2004 instead, thus avoiding the use of floating point arithmetic. (There is no
2005 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2006 double.
2007
2008 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2009 2005 (difference between two addresses compared to an unsigned value).
2010
2011 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2012 custom one, because of the following reported problem in Windows:
2013
2014 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2015 under Win32.
2016 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2017 therefore missing the function definition.
2018 - The compiler thus generates a "C" signature for the test function.
2019 - The linker fails to find the "C" function.
2020 - PCRE fails to configure if asked to do so against libbz2.
2021
2022 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2023 messages were output:
2024
2025 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2026 rerunning libtoolize, to keep the correct libtool macros in-tree.
2027 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2028
2029 I have done both of these things.
2030
2031 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2032 most of the time, it *can* run out if it is given a pattern that contains a
2033 runaway infinite recursion. I updated the discussion in the pcrestack man
2034 page.
2035
2036 13. Now that we have gone to the x.xx style of version numbers, the minor
2037 version may start with zero. Using 08 or 09 is a bad idea because users
2038 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2039 interpreted as invalid octal numbers. I've updated the previous comment in
2040 configure.ac, and also added a check that gives an error if 08 or 09 are
2041 used.
2042
2043 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2044 causing partial matching to fail when the end of the subject matched \W
2045 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2046
2047 15. There were some discrepancies between the declarations in pcre_internal.h
2048 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2049 their definitions. The declarations used "const uschar *" and the
2050 definitions used USPTR. Even though USPTR is normally defined as "const
2051 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2052 reported that: "This difference in casting confuses some C++ compilers, for
2053 example, SunCC recognizes above declarations as different functions and
2054 generates broken code for hbpcre." I have changed the declarations to use
2055 USPTR.
2056
2057 16. GNU libtool is named differently on some systems. The autogen.sh script now
2058 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2059 (FreeBSD).
2060
2061 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2062 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2063 comment: "Figure out how to create a longlong from a string: strtoll and
2064 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2065 instance, but it only takes 2 args instead of 3!"
2066
2067 18. A subtle bug concerned with back references has been fixed by a change of
2068 specification, with a corresponding code fix. A pattern such as
2069 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2070 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2071 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2072 same bug. Such groups have to be quantified to be useful, or contained
2073 inside another quantified group. (If there's no repetition, the reference
2074 can never match.) The problem arises because, having left the group and
2075 moved on to the rest of the pattern, a later failure that backtracks into
2076 the group uses the captured value from the final iteration of the group
2077 rather than the correct earlier one. I have fixed this in PCRE by forcing
2078 any group that contains a reference to itself to be an atomic group; that
2079 is, there cannot be any backtracking into it once it has completed. This is
2080 similar to recursive and subroutine calls.
2081
2082
2083 Version 8.00 19-Oct-09
2084 ----------------------
2085
2086 1. The table for translating pcre_compile() error codes into POSIX error codes
2087 was out-of-date, and there was no check on the pcre_compile() error code
2088 being within the table. This could lead to an OK return being given in
2089 error.
2090
2091 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2092 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2093 in a Windows environment.
2094
2095 3. The pcregrep --count option prints the count for each file even when it is
2096 zero, as does GNU grep. However, pcregrep was also printing all files when
2097 --files-with-matches was added. Now, when both options are given, it prints
2098 counts only for those files that have at least one match. (GNU grep just
2099 prints the file name in this circumstance, but including the count seems
2100 more useful - otherwise, why use --count?) Also ensured that the
2101 combination -clh just lists non-zero counts, with no names.
2102
2103 4. The long form of the pcregrep -F option was incorrectly implemented as
2104 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2105 but it seems right to fix it, and I didn't think it was worth preserving
2106 the old behaviour.
2107
2108 5. The command line items --regex=pattern and --regexp=pattern were not
2109 recognized by pcregrep, which required --regex pattern or --regexp pattern
2110 (with a space rather than an '='). The man page documented the '=' forms,
2111 which are compatible with GNU grep; these now work.
2112
2113 6. No libpcreposix.pc file was created for pkg-config; there was just
2114 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2115
2116 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2117 when UCP support is not needed, by modifying the Python script that
2118 generates it from Unicode data files. This should not matter if the module
2119 is correctly used as a library, but I received one complaint about 50K of
2120 unwanted data. My guess is that the person linked everything into his
2121 program rather than using a library. Anyway, it does no harm.
2122
2123 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2124 was a minimum greater than 1 for a wide character in a possessive
2125 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2126 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2127 character. Chaos in the form of incorrect output or a compiling loop could
2128 result.
2129
2130 9. The restrictions on what a pattern can contain when partial matching is
2131 requested for pcre_exec() have been removed. All patterns can now be
2132 partially matched by this function. In addition, if there are at least two
2133 slots in the offset vector, the offset of the earliest inspected character
2134 for the match and the offset of the end of the subject are set in them when
2135 PCRE_ERROR_PARTIAL is returned.
2136
2137 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2138 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2139 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2140 and may be more useful for multi-segment matching.
2141
2142 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2143 used to be given if ever the end of the subject was reached; now it is
2144 given only if matching could not proceed because another character was
2145 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2146 string "Z", which now yields "no match" instead of "partial match". In the
2147 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2148 final character ended with (*FAIL).
2149
2150 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2151 if the pattern had a "must contain" character that was already found in the
2152 earlier partial match, unless partial matching was again requested. For
2153 example, with the pattern /dog.(body)?/, the "must contain" character is
2154 "g". If the first part-match was for the string "dog", restarting with
2155 "sbody" failed. This bug has been fixed.
2156
2157 13. The string returned by pcre_dfa_exec() after a partial match has been
2158 changed so that it starts at the first inspected character rather than the
2159 first character of the match. This makes a difference only if the pattern
2160 starts with a lookbehind assertion or \b or \B (\K is not supported by
2161 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2162 matching functions compatible, and I think it's the right thing to do.
2163
2164 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2165 so that the demonstration program is easily available in environments where
2166 PCRE has not been installed from source.
2167
2168 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2169 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2170 library.
2171
2172 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2173 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2174 is not the first non-POSIX option to be added. Clearly some people find
2175 these options useful.
2176
2177 17. If a caller to the POSIX matching function regexec() passes a non-zero
2178 value for nmatch with a NULL value for pmatch, the value of
2179 nmatch is forced to zero.
2180
2181 18. RunGrepTest did not have a test for the availability of the -u option of
2182 the diff command, as RunTest does. It now checks in the same way as
2183 RunTest, and also checks for the -b option.
2184
2185 19. If an odd number of negated classes containing just a single character
2186 interposed, within parentheses, between a forward reference to a named
2187 subpattern and the definition of the subpattern, compilation crashed with
2188 an internal error, complaining that it could not find the referenced
2189 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2190 [The bug was that it was starting one character too far in when skipping
2191 over the character class, thus treating the ] as data rather than
2192 terminating the class. This meant it could skip too much.]
2193
2194 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2195 /g option in pcretest when the pattern contains \K, which makes it possible
2196 to have an empty string match not at the start, even when the pattern is
2197 anchored. Updated pcretest and pcredemo to use this option.
2198
2199 21. If the maximum number of capturing subpatterns in a recursion was greater
2200 than the maximum at the outer level, the higher number was returned, but
2201 with unset values at the outer level. The correct (outer level) value is
2202 now given.
2203
2204 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2205 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2206 make it do so. The string so far is captured, making this feature
2207 compatible with Perl.
2208
2209 23. The tests have been re-organized, adding tests 11 and 12, to make it
2210 possible to check the Perl 5.10 features against Perl 5.10.
2211
2212 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2213 pattern matches a fixed length string. PCRE did not allow this; now it
2214 does. Neither allows recursion.
2215
2216 25. I finally figured out how to implement a request to provide the minimum
2217 length of subject string that was needed in order to match a given pattern.
2218 (It was back references and recursion that I had previously got hung up
2219 on.) This code has now been added to pcre_study(); it finds a lower bound
2220 to the length of subject needed. It is not necessarily the greatest lower
2221 bound, but using it to avoid searching strings that are too short does give
2222 some useful speed-ups. The value is available to calling programs via
2223 pcre_fullinfo().
2224
2225 26. While implementing 25, I discovered to my embarrassment that pcretest had
2226 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2227 study optimizations had never been tested with that matching function.
2228 Oops. What is worse, even when it was passed study data, there was a bug in
2229 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2230 were also very few tests of studied patterns with pcre_dfa_exec().
2231
2232 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2233 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2234 on the other side of the coin, they are no longer allowed to have different
2235 names, because these cannot be distinguished in PCRE, and this has caused
2236 confusion. (This is a difference from Perl.)
2237
2238 28. When duplicate subpattern names are present (necessarily with different
2239 numbers, as required by 27 above), and a test is made by name in a
2240 conditional pattern, either for a subpattern having been matched, or for
2241 recursion in such a pattern, all the associated numbered subpatterns are
2242 tested, and the overall condition is true if the condition is true for any
2243 one of them. This is the way Perl works, and is also more like the way
2244 testing by number works.
2245
2246
2247 Version 7.9 11-Apr-09
2248 ---------------------
2249
2250 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2251 (pcretest), all targets were linked against these libraries. This included
2252 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2253 libraries. This caused unwanted dependencies to be created. This problem
2254 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2255 pcretest is linked with readline.
2256
2257 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2258 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2259 moved outside it again, because FALSE and TRUE are already defined in AIX,
2260 but BOOL is not.
2261
2262 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2263 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2264
2265 4. The pcregrep documentation said spaces were inserted as well as colons (or
2266 hyphens) following file names and line numbers when outputting matching
2267 lines. This is not true; no spaces are inserted. I have also clarified the
2268 wording for the --colour (or --color) option.
2269
2270 5. In pcregrep, when --colour was used with -o, the list of matching strings
2271 was not coloured; this is different to GNU grep, so I have changed it to be
2272 the same.
2273
2274 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2275 each matching line was coloured. Now it goes on to look for further matches
2276 of any of the test patterns, which is the same behaviour as GNU grep.
2277
2278 7. A pattern that could match an empty string could cause pcregrep to loop; it
2279 doesn't make sense to accept an empty string match in pcregrep, so I have
2280 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2281 seems to be how GNU grep behaves. [But see later change 40 for release
2282 8.33.]
2283
2284 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2285 start or after a newline", because the conditional assertion was not being
2286 correctly handled. The rule now is that both the assertion and what follows
2287 in the first alternative must satisfy the test.
2288
2289 9. If auto-callout was enabled in a pattern with a conditional group whose
2290 condition was an assertion, PCRE could crash during matching, both with
2291 pcre_exec() and pcre_dfa_exec().
2292
2293 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2294 used for matching.
2295
2296 11. Unicode property support in character classes was not working for
2297 characters (bytes) greater than 127 when not in UTF-8 mode.
2298
2299 12. Added the -M command line option to pcretest.
2300
2301 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2302
2303 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2304
2305 16. Added comments and documentation about mis-use of no_arg in the C++
2306 wrapper.
2307
2308 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2309 from Martin Jerabek that uses macro names for all relevant character and
2310 string constants.
2311
2312 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2313 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2314 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2315 these, but not everybody uses configure.
2316
2317 19. A conditional group that had only one branch was not being correctly
2318 recognized as an item that could match an empty string. This meant that an
2319 enclosing group might also not be so recognized, causing infinite looping
2320 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2321 with the subject "ab", where knowledge that the repeated group can match
2322 nothing is needed in order to break the loop.
2323
2324 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2325 exec(), but without supplying a callout function, matching went wrong.
2326
2327 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2328 leak if the size of the offset vector was greater than 30. When the vector
2329 is smaller, the saved offsets during recursion go onto a local stack
2330 vector, but for larger vectors malloc() is used. It was failing to free
2331 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2332 error, in fact).
2333
2334 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2335 heapframe that is used only when UTF-8 support is enabled. This caused no
2336 problem, but was untidy.
2337
2338 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2339 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2340 included within another project.
2341
2342 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2343 slightly modified by me:
2344
2345 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2346 not building pcregrep.
2347
2348 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2349 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2350
2351 25. Forward references, both numeric and by name, in patterns that made use of
2352 duplicate group numbers, could behave incorrectly or give incorrect errors,
2353 because when scanning forward to find the reference group, PCRE was not
2354 taking into account the duplicate group numbers. A pattern such as
2355 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2356
2357 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2358 the feature of a custom pointer more persuasive (as requested by a user).
2359
2360 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2361 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2362 pre-defined.
2363
2364 28. Added support for (*UTF8) at the start of a pattern.
2365
2366 29. Arrange for flags added by the "release type" setting in CMake to be shown
2367 in the configuration summary.
2368
2369
2370 Version 7.8 05-Sep-08
2371 ---------------------
2372
2373 1. Replaced UCP searching code with optimized version as implemented for Ad
2374 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2375 stage table and inline lookup instead of a function, giving speed ups of 2
2376 to 5 times on some simple patterns that I tested. Permission was given to
2377 distribute the MultiStage2.py script that generates the tables (it's not in
2378 the tarball, but is in the Subversion repository).
2379
2380 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2381 scripts.
2382
2383 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2384 a group with a zero qualifier. The result of the study could be incorrect,
2385 or the function might crash, depending on the pattern.
2386
2387 4. Caseless matching was not working for non-ASCII characters in back
2388 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2389 It now works when Unicode Property Support is available.
2390
2391 5. In pcretest, an escape such as \x{de} in the data was always generating
2392 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2393 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2394 truncation.
2395
2396 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2397
2398 7. Added two (int) casts to pcregrep when printing the difference of two
2399 pointers, in case they are 64-bit values.
2400
2401 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2402 test 2 if it fails.
2403
2404 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2405 and a #define of that name to empty if it is not externally set. This is to
2406 allow users of MSVC to set it if necessary.
2407
2408 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2409 the convenience functions in the pcre_get.c source file.
2410
2411 11. An option change at the start of a pattern that had top-level alternatives
2412 could cause overwriting and/or a crash. This command provoked a crash in
2413 some environments:
2414
2415 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2416
2417 This potential security problem was recorded as CVE-2008-2371.
2418
2419 12. For a pattern where the match had to start at the beginning or immediately
2420 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2421 pcre_dfa_exec() could read past the end of the passed subject if there was
2422 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2423 pcretest so that it places the subject at the end of its malloc-ed buffer.
2424
2425 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2426 exec() might read past the end of the data buffer in UTF-8 mode.
2427
2428 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2429 the data contained the byte 0x85 as part of a UTF-8 character within its
2430 first line. This applied both to normal and DFA matching.
2431
2432 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2433 /^[^d]*?$/8 failed to match "abc".
2434
2435 16. Added a missing copyright notice to pcrecpp_internal.h.
2436
2437 17. Make it more clear in the documentation that values returned from
2438 pcre_exec() in ovector are byte offsets, not character counts.
2439
2440 18. Tidied a few places to stop certain compilers from issuing warnings.
2441
2442 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2443 supplied by Stefan Weber. I made a further small update for 7.8 because
2444 there is a change of source arrangements: the pcre_searchfuncs.c module is
2445 replaced by pcre_ucd.c.
2446
2447
2448 Version 7.7 07-May-08
2449 ---------------------
2450
2451 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2452 a string to a long long, pretend we don't even have a long long." This is
2453 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2454
2455 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2456 pre-7.6 versions, which defined a global no_arg variable instead of putting
2457 it in the RE class. (See also #8 below.)
2458
2459 3. Remove a line of dead code, identified by coverity and reported by Nuno
2460 Lopes.
2461
2462 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2463
2464 (1) The include/exclude patterns were being applied to the whole pathnames
2465 of files, instead of just to the final components.
2466
2467 (2) If there was more than one level of directory, the subdirectories were
2468 skipped unless they satisfied the include/exclude conditions. This is
2469 inconsistent with GNU grep (and could even be seen as contrary to the
2470 pcregrep specification - which I improved to make it absolutely clear).
2471 The action now is always to scan all levels of directory, and just
2472 apply the include/exclude patterns to regular files.
2473
2474 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2475 --exclude_dir in the tests to avoid scanning .svn directories.
2476
2477 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2478 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2479 doesn't support NULs in patterns.
2480
2481 7. Added some missing "const"s to declarations of static tables in
2482 pcre_compile.c and pcre_dfa_exec.c.
2483
2484 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2485 caused by fix #2 above. (Subsequently also a second patch to fix the
2486 first patch. And a third patch - this was a messy problem.)
2487
2488 9. Applied Craig's patch to remove the use of push_back().
2489
2490 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2491 matching function regexec().
2492
2493 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2494 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2495 references. PCRE supports relative numbers with this syntax (I don't think
2496 Oniguruma does).
2497
2498 12. Previously, a group with a zero repeat such as (...){0} was completely
2499 omitted from the compiled regex. However, this means that if the group
2500 was called as a subroutine from elsewhere in the pattern, things went wrong
2501 (an internal error was given). Such groups are now left in the compiled
2502 pattern, with a new opcode that causes them to be skipped at execution
2503 time.
2504
2505 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2506 to the way PCRE behaves:
2507
2508 (a) A lone ] character is dis-allowed (Perl treats it as data).
2509
2510 (b) A back reference to an unmatched subpattern matches an empty string
2511 (Perl fails the current match path).
2512
2513 (c) A data ] in a character class must be notated as \] because if the
2514 first data character in a class is ], it defines an empty class. (In
2515 Perl it is not possible to have an empty class.) The empty class []
2516 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2517 The negative empty class [^] matches any one character, independently
2518 of the DOTALL setting.
2519
2520 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2521 non-existent subpattern following a character class starting with ']' and
2522 containing () gave an internal compiling error instead of "reference to
2523 non-existent subpattern". Fortunately, when the pattern did exist, the
2524 compiled code was correct. (When scanning forwards to check for the
2525 existence of the subpattern, it was treating the data ']' as terminating
2526 the class, so got the count wrong. When actually compiling, the reference
2527 was subsequently set up correctly.)
2528
2529 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2530 it was being rejected as not supported by pcre_dfa_exec(), even though
2531 other assertions are supported. I have made pcre_dfa_exec() support
2532 (*FAIL).
2533
2534 16. The implementation of 13c above involved the invention of a new opcode,
2535 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2536 cannot be changed at match time, I realized I could make a small
2537 improvement to matching performance by compiling OP_ALLANY instead of
2538 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2539 on the OP_ANY path.
2540
2541 17. Compiling pcretest on Windows with readline support failed without the
2542 following two fixes: (1) Make the unistd.h include conditional on
2543 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2544
2545 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2546 ncurses library to be included for pcretest when ReadLine support is
2547 requested, but also to allow for it to be overridden. This patch came from
2548 Daniel Bergström.
2549
2550 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2551 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2552 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2553 spotting this.
2554
2555
2556 Version 7.6 28-Jan-08
2557 ---------------------
2558
2559 1. A character class containing a very large number of characters with
2560 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2561 overflow.
2562
2563 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2564 HAVE_LONG_LONG is not defined.
2565
2566 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2567 bring them up to date and include new features. This patch includes:
2568
2569 - Fixed PH's badly added libz and libbz2 support.
2570 - Fixed a problem with static linking.
2571 - Added pcredemo. [But later removed - see 7 below.]
2572 - Fixed dftables problem and added an option.
2573 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2574 HAVE_LONG_LONG.
2575 - Added readline support for pcretest.
2576 - Added an listing of the option settings after cmake has run.
2577
2578 4. A user submitted a patch to Makefile that makes it easy to create
2579 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2580 Makefile.am that cause it to include this special target, without
2581 affecting anything else. Note that the same mingw target plus all
2582 the other distribution libraries and programs are now supported
2583 when configuring with CMake (see 6 below) instead of with
2584 Configure/Make.
2585
2586 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2587 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2588 exported in the Windows port". It has not yet been confirmed that the patch
2589 solves the problem, but it does no harm.
2590
2591 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2592 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2593 with CMake, and also correct the comment about stack recursion.
2594
2595 7. Remove the automatic building of pcredemo from the ./configure system and
2596 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2597 of a program that users should build themselves after PCRE is installed, so
2598 building it automatically is not really right. What is more, it gave
2599 trouble in some build environments.
2600
2601 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2602
2603
2604 Version 7.5 10-Jan-08
2605 ---------------------
2606
2607 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2608 values in parens when parsing an RE using the C++ wrapper."
2609
2610 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2611 Characters greater than 255 were excluded from the class instead of being
2612 included.
2613
2614 3. The same bug as (2) above applied to negated POSIX classes such as
2615 [:^space:].
2616
2617 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2618 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2619 I have changed it.
2620
2621 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2622 first named subpattern) and a construct such as (?&a) would reference the
2623 first named subpattern whose name started with "a" (in other words, the
2624 length check was missing). Both these problems are fixed. "Subpattern name
2625 expected" is now given for (?&) (a zero-length name), and this patch also
2626 makes it give the same error for \k'' (previously it complained that that
2627 was a reference to a non-existent subpattern).
2628
2629 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2630 this is right because (?- can be followed by option settings as well as by
2631 digits. I have, however, made the messages clearer.
2632
2633 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2634 than the number used in the conditional) now cause a compile-time error.
2635 This is actually not compatible with Perl, which accepts such patterns, but
2636 treats the conditional as always being FALSE (as PCRE used to), but it
2637 seems to me that giving a diagnostic is better.
2638
2639 8. Change "alphameric" to the more common word "alphanumeric" in comments
2640 and messages.
2641
2642 9. Fix two occurrences of "backslash" in comments that should have been
2643 "backspace".
2644
2645 10. Remove two redundant lines of code that can never be obeyed (their function
2646 was moved elsewhere).
2647
2648 11. The program that makes PCRE's Unicode character property table had a bug
2649 which caused it to generate incorrect table entries for sequences of
2650 characters that have the same character type, but are in different scripts.
2651 It amalgamated them into a single range, with the script of the first of
2652 them. In other words, some characters were in the wrong script. There were
2653 thirteen such cases, affecting characters in the following ranges:
2654
2655 U+002b0 - U+002c1
2656 U+0060c - U+0060d
2657 U+0061e - U+00612
2658 U+0064b - U+0065e
2659 U+0074d - U+0076d
2660 U+01800 - U+01805
2661 U+01d00 - U+01d77
2662 U+01d9b - U+01dbf
2663 U+0200b - U+0200f
2664 U+030fc - U+030fe
2665 U+03260 - U+0327f
2666 U+0fb46 - U+0fbb1
2667 U+10450 - U+1049d
2668
2669 12. The -o option (show only the matching part of a line) for pcregrep was not
2670 compatible with GNU grep in that, if there was more than one match in a
2671 line, it showed only the first of them. It now behaves in the same way as
2672 GNU grep.
2673
2674 13. If the -o and -v options were combined for pcregrep, it printed a blank
2675 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2676 does the same. The return code can be used to tell if there were any
2677 non-matching lines.
2678
2679 14. Added --file-offsets and --line-offsets to pcregrep.
2680
2681 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2682 infinitely looping recursion. The bug was that positive lookaheads were not
2683 being skipped when checking for a possible empty match (negative lookaheads
2684 and both kinds of lookbehind were skipped).
2685
2686 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2687 inclusion of <windows.h> to before rather than after the definition of
2688 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2689
2690 17. Specifying a possessive quantifier with a specific limit for a Unicode
2691 character property caused pcre_compile() to compile bad code, which led at
2692 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2693 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2694 caused the error; without that there was no problem.
2695
2696 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2697
2698 19. Added --enable-pcretest-libreadline.
2699
2700 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2701 RE::GlobalReplace(). As a result, the number of replacements returned was
2702 double what it should be. I removed one of the increments, but Craig sent a
2703 later patch that removed the other one (the right fix) and added unit tests
2704 that check the return values (which was not done before).
2705
2706 21. Several CMake things:
2707
2708 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2709 the names libpcre and libpcreposix, not just pcre and pcreposix.
2710
2711 (2) The above change means that pcretest and pcregrep are now correctly
2712 linked with the newly-built libraries, not previously installed ones.
2713
2714 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2715
2716 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2717 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2718 UTF-8 newline character). The key issue is that the pattern starts .*;
2719 this means that the match must be either at the beginning, or after a
2720 newline. The bug was in the code for advancing after a failed match and
2721 checking that the new position followed a newline. It was not taking
2722 account of UTF-8 characters correctly.
2723
2724 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2725 character classes. PCRE was not treating the sequence [:...:] as a
2726 character class unless the ... were all letters. Perl, however, seems to
2727 allow any characters between [: and :], though of course it rejects as
2728 unknown any "names" that contain non-letters, because all the known class
2729 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2730 for example, whereas PCRE did not - it did not recognize a POSIX character
2731 class. This seemed a bit dangerous, so the code has been changed to be
2732 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2733 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2734 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2735 Perl does, and where it didn't before.
2736
2737 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2738 Windows environments %n is disabled by default.
2739
2740
2741 Version 7.4 21-Sep-07
2742 ---------------------
2743
2744 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2745 means that a class such as [\s] counted as "explicit reference to CR or
2746 LF". That isn't really right - the whole point of the change was to try to
2747 help when there was an actual mention of one of the two characters. So now
2748 the change happens only if \r or \n (or a literal CR or LF) character is
2749 encountered.
2750
2751 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2752 of both had grown to the point where there were only 3 bits left.
2753 Fortunately, there was spare space in the data structure, and so I have
2754 moved the internal flags into a new 16-bit field to free up more option
2755 bits.
2756
2757 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2758 but did not set the internal JCHANGED flag - either of these is enough to
2759 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2760 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2761 start sets both bits.
2762
2763 4. Added options (at build time, compile time, exec time) to change \R from
2764 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2765
2766 5. doc/pcresyntax.html was missing from the distribution.
2767
2768 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2769 compatibility, even though it is no longer used.
2770
2771 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2772 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2773 windows.h file is present (where different names are used). [This was
2774 reversed later after testing - see 16 below.]
2775
2776 8. Changed all #include <config.h> to #include "config.h". There were also
2777 some further <pcre.h> cases that I changed to "pcre.h".
2778
2779 9. When pcregrep was used with the --colour option, it missed the line ending
2780 sequence off the lines that it output.
2781
2782 10. It was pointed out to me that arrays of string pointers cause lots of
2783 relocations when a shared library is dynamically loaded. A technique of
2784 using a single long string with a table of offsets can drastically reduce
2785 these. I have refactored PCRE in four places to do this. The result is
2786 dramatic:
2787
2788 Originally: 290
2789 After changing UCP table: 187
2790 After changing error message table: 43
2791 After changing table of "verbs" 36
2792 After changing table of Posix names 22
2793
2794 Thanks to the folks working on Gregex for glib for this insight.
2795
2796 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2797 unicode-properties was also set.
2798
2799 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2800
2801 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2802 checked only for CRLF.
2803
2804 14. Added casts to pcretest.c to avoid compiler warnings.
2805
2806 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2807
2808 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2809 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2810 entirely. This removes changes made in 7 above.
2811
2812 17. The CMake files have been updated, and there is now more information about
2813 building with CMake in the NON-UNIX-USE document.
2814
2815
2816 Version 7.3 28-Aug-07
2817 ---------------------
2818
2819 1. In the rejigging of the build system that eventually resulted in 7.1, the
2820 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2821 brackets there is not right, since it causes compilers to look for an
2822 installed pcre.h, not the version that is in the source that is being
2823 compiled (which of course may be different). I have changed it back to:
2824
2825 #include "pcre.h"
2826
2827 I have a vague recollection that the change was concerned with compiling in
2828 different directories, but in the new build system, that is taken care of
2829 by the VPATH setting the Makefile.
2830
2831 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2832 when the subject happened to end in the byte 0x85 (e.g. if the last
2833 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2834 characters but of course it shouldn't be taken as a newline when it is part
2835 of another character. The bug was that, for an unlimited repeat of . in
2836 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2837 characters when looking for a newline.
2838
2839 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2840
2841 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2842 in debug output.
2843
2844 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2845 long printing in the pcrecpp unittest when running under MinGW.
2846
2847 6. ESC_K was left out of the EBCDIC table.
2848
2849 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2850 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2851 limit also applies to "virtual nesting" when a pattern is recursive, and in
2852 this case 1000 isn't so big. I have been able to remove this limit at the
2853 expense of backing off one optimization in certain circumstances. Normally,
2854 when pcre_exec() would call its internal match() function recursively and
2855 immediately return the result unconditionally, it uses a "tail recursion"
2856 feature to save stack. However, when a subpattern that can match an empty
2857 string has an unlimited repetition quantifier, it no longer makes this
2858 optimization. That gives it a stack frame in which to save the data for
2859 checking that an empty string has been matched. Previously this was taken
2860 from the 1000-entry workspace that had been reserved. So now there is no
2861 explicit limit, but more stack is used.
2862
2863 8. Applied Daniel's patches to solve problems with the import/export magic
2864 syntax that is required for Windows, and which was going wrong for the
2865 pcreposix and pcrecpp parts of the library. These were overlooked when this
2866 problem was solved for the main library.
2867
2868 9. There were some crude static tests to avoid integer overflow when computing
2869 the size of patterns that contain repeated groups with explicit upper
2870 limits. As the maximum quantifier is 65535, the maximum group length was
2871 set at 30,000 so that the product of these two numbers did not overflow a
2872 32-bit integer. However, it turns out that people want to use groups that
2873 are longer than 30,000 bytes (though not repeat them that many times).
2874 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2875 made it possible to implement the integer overflow checks in a much more
2876 dynamic way, which I have now done. The artificial limitation on group
2877 length has been removed - we now have only the limit on the total length of
2878 the compiled pattern, which depends on the LINK_SIZE setting.
2879
2880 10. Fixed a bug in the documentation for get/copy named substring when
2881 duplicate names are permitted. If none of the named substrings are set, the
2882 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2883 empty string.
2884
2885 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2886 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2887 because the ] is interpreted as the first data character and the
2888 terminating ] is not found. PCRE has been made compatible with Perl in this
2889 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2890 cause memory overwriting.
2891
2892 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2893 string has been matched (to stop an infinite loop). It was not recognizing
2894 a conditional subpattern that could match an empty string if that
2895 subpattern was within another subpattern. For example, it looped when
2896 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2897 condition was not nested. This bug has been fixed.
2898
2899 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2900 past the start of the subject in the presence of bytes with the top bit
2901 set, for example "\x8aBCD".
2902
2903 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2904 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2905
2906 14. Optimized (?!) to (*FAIL).
2907
2908 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2909 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2910 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2911 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2912 does: it's just the validity check that is more restrictive.
2913
2914 16. Inserted checks for integer overflows during escape sequence (backslash)
2915 processing, and also fixed erroneous offset values for syntax errors during
2916 backslash processing.
2917
2918 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2919 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2920
2921 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2922 caused an overrun.
2923
2924 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2925 something other than just ASCII characters) inside a group that had an
2926 unlimited repeat caused a loop at compile time (while checking to see
2927 whether the group could match an empty string).
2928
2929 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2930 [\P{Any}] did so. (Error in the code for printing property names.)
2931
2932 21. An orphan \E inside a character class could cause a crash.
2933
2934 22. A repeated capturing bracket such as (A)? could cause a wild memory
2935 reference during compilation.
2936
2937 23. There are several functions in pcre_compile() that scan along a compiled
2938 expression for various reasons (e.g. to see if it's fixed length for look
2939 behind). There were bugs in these functions when a repeated \p or \P was
2940 present in the pattern. These operators have additional parameters compared
2941 with \d, etc, and these were not being taken into account when moving along
2942 the compiled data. Specifically:
2943
2944 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2945 length.
2946
2947 (b) An item such as \pL+ within a repeated group could cause crashes or
2948 loops.
2949
2950 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2951 "reference to non-existent subpattern" error.
2952
2953 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2954
2955 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2956 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2957
2958 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2959
2960 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2961 character were causing crashes (broken optimization).
2962
2963 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2964 \p or \P) caused a compile-time loop.
2965
2966 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2967 break. For example, the unstudied pattern [\r\n]A does not match the string
2968 "\r\nA" because change 7.0/46 below moves the current point on by two
2969 characters after failing to match at the start. However, the pattern \nA
2970 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2971 the same is true. There doesn't seem any very clean way out of this, but
2972 what I have chosen to do makes the common cases work: PCRE now takes note
2973 of whether there can be an explicit match for \r or \n anywhere in the
2974 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2975 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2976 pattern has explicit CR or LF references.
2977
2978 29. Added (*CR) etc for changing newline setting at start of pattern.
2979
2980
2981 Version 7.2 19-Jun-07
2982 ---------------------
2983
2984 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2985 which is apparently normally available under Windows.
2986
2987 2. Re-jig the pcregrep tests with different newline settings in an attempt
2988 to make them independent of the local environment's newline setting.
2989
2990 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2991
2992 4. Some of the "internals" tests were previously cut out when the link size
2993 was not 2, because the output contained actual offsets. The recent new
2994 "Z" feature of pcretest means that these can be cut out, making the tests
2995 usable with all link sizes.
2996
2997 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2998 stack recursion. This gives a massive performance boost under BSD, but just
2999 a small improvement under Linux. However, it saves one field in the frame
3000 in all cases.
3001
3002 6. Added more features from the forthcoming Perl 5.10:
3003
3004 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3005 recursion call. It refers to the nth most recently opened parentheses.
3006
3007 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3008 to be opened parentheses.
3009
3010 (c) Conditions that refer to capturing parentheses can be specified
3011 relatively, for example, (?(-2)... or (?(+3)...
3012
3013 (d) \K resets the start of the current match so that everything before
3014 is not part of it.
3015
3016 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3017
3018 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3019 reference syntax.
3020
3021 (g) (?| introduces a group in which the numbering of parentheses in each
3022 alternative starts with the same number.
3023
3024 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3025
3026 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3027 PCRE_INFO_JCHANGED.
3028
3029 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3030 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3031 for detecting groups that can match an empty string.
3032
3033 9. A pattern with a very large number of alternatives (more than several
3034 hundred) was running out of internal workspace during the pre-compile
3035 phase, where pcre_compile() figures out how much memory will be needed. A
3036 bit of new cunning has reduced the workspace needed for groups with
3037 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3038 workspace instead of running out of the 4096 that are available.
3039
3040 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3041
3042 11. Applied patch from Google to remove an optimization that didn't quite work.
3043 The report of the bug said:
3044
3045 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3046 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3047 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3048
3049 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3050 it matched the wrong number of bytes.
3051
3052
3053 Version 7.1 24-Apr-07
3054 ---------------------
3055
3056 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3057 that is more "standard", making use of automake and other Autotools. There
3058 is some re-arrangement of the files and adjustment of comments consequent
3059 on this.
3060
3061 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3062 for recursive directory scanning broke on some systems because the files
3063 are not scanned in any specific order and on different systems the order
3064 was different. A call to "sort" has been inserted into RunGrepTest for the
3065 approprate test as a short-term fix. In the longer term there may be an
3066 alternative.
3067
3068 3. I had an email from Eric Raymond about problems translating some of PCRE's
3069 man pages to HTML (despite the fact that I distribute HTML pages, some
3070 people do their own conversions for various reasons). The problems
3071 concerned the use of low-level troff macros .br and .in. I have therefore
3072 removed all such uses from the man pages (some were redundant, some could
3073 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3074 HTML has been updated to handle .nf/.fi and to complain if it encounters
3075 .br or .in.
3076
3077 4. Updated comments in configure.ac that get placed in config.h.in and also
3078 arranged for config.h to be included in the distribution, with the name
3079 config.h.generic, for the benefit of those who have to compile without
3080 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3081
3082 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3083 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3084 makevp.bat for the current PCRE, using the additional files
3085 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3086
3087 6. A Windows user reported a minor discrepancy with test 2, which turned out
3088 to be caused by a trailing space on an input line that had got lost in his
3089 copy. The trailing space was an accident, so I've just removed it.
3090
3091 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3092 that is needed.
3093
3094 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3095 as "const" (a) because they are and (b) because it helps the PHP
3096 maintainers who have recently made a script to detect big data structures
3097 in the php code that should be moved to the .rodata section. I remembered
3098 to update Builducptable as well, so it won't revert if ucptable.h is ever
3099 re-created.
3100
3101 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3102 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3103 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3104 support is not required. This saves 1.5-2K of code, which is important in
3105 some applications.
3106
3107 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3108 so as not to refer to the tables, even though these functions will never be
3109 called when UTF-8 support is disabled. Otherwise there are problems with a
3110 shared library.
3111
3112 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3113
3114 (a) It was defining its arguments as char * instead of void *.
3115
3116 (b) It was assuming that all moves were upwards in memory; this was true
3117 a long time ago when I wrote it, but is no longer the case.
3118
3119 The emulated memove() is provided for those environments that have neither
3120 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3121 is clearly not the case, as these two bugs were recently reported.
3122
3123 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3124 and Detrail to create the HTML documentation, the .txt form of the man
3125 pages, and it removes trailing spaces from listed files. It also creates
3126 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3127 case, it wraps all the #defines with #ifndefs. This script should be run
3128 before "make dist".
3129
3130 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3131 with Unicode property support.
3132
3133 (a) For a maximizing quantifier, if the two different cases of the
3134 character were of different lengths in their UTF-8 codings (there are
3135 some cases like this - I found 11), and the matching function had to
3136 back up over a mixture of the two cases, it incorrectly assumed they
3137 were both the same length.
3138
3139 (b) When PCRE was configured to use the heap rather than the stack for
3140 recursion during matching, it was not correctly preserving the data for
3141 the other case of a UTF-8 character when checking ahead for a match
3142 while processing a minimizing repeat. If the check also involved
3143 matching a wide character, but failed, corruption could cause an
3144 erroneous result when trying to check for a repeat of the original
3145 character.
3146
3147 13. Some tidying changes to the testing mechanism:
3148
3149 (a) The RunTest script now detects the internal link size and whether there
3150 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3151 values substituted by "configure". (The RunGrepTest script already did
3152 this for UTF-8.) The configure.ac script no longer substitutes the
3153 relevant variables.
3154
3155 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3156 with length and offset values. This means that the output is different
3157 for different internal link sizes. Test 2 is skipped for link sizes
3158 other than 2 because of this, bypassing the problem. Unfortunately,
3159 there was also a test in test 3 (the locale tests) that used /B and
3160 failed for link sizes other than 2. Rather than cut the whole test out,
3161 I have added a new /Z option to pcretest that replaces the length and
3162 offset values with spaces. This is now used to make test 3 independent
3163 of link size. (Test 2 will be tidied up later.)
3164
3165 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3166 segmentation fault instead of returning the appropriate error message.
3167
3168 15. In multiline mode when the newline sequence was set to "any", the pattern
3169 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3170 This doesn't seem right; it now treats the CRLF combination as the line
3171 ending, and so does not match in that case. It's only a pattern such as ^$
3172 that would hit this one: something like ^ABC$ would have failed after \r
3173 and then tried again after \r\n.
3174
3175 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3176 in an attempt to make files that differ only in their line terminators
3177 compare equal. This works on Linux.
3178
3179 17. Under certain error circumstances pcregrep might try to free random memory
3180 as it exited. This is now fixed, thanks to valgrind.
3181
3182 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3183 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3184 was because its rules for how to advance for /g after matching an empty
3185 string at the end of a line did not allow for this case. They now check for
3186 it specially.
3187
3188 20. pcretest is supposed to handle patterns and data of any length, by
3189 extending its buffers when necessary. It was getting this wrong when the
3190 buffer for a data line had to be extended.
3191
3192 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3193 CRLF as a newline sequence.
3194
3195 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3196 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3197 I have nevertheless tidied it up.
3198
3199 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3200
3201 24. Added a man page for pcre-config.
3202
3203
3204 Version 7.0 19-Dec-06
3205 ---------------------
3206
3207 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3208 moving to gcc 4.1.1.
3209
3210 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3211 sys/time.h, which is documented as needed for this function. It doesn't
3212 seem to matter on Linux, but it showed up on some releases of OS X.
3213
3214 3. It seems that there are systems where bytes whose values are greater than
3215 127 match isprint() in the "C" locale. The "C" locale should be the
3216 default when a C program starts up. In most systems, only ASCII printing
3217 characters match isprint(). This difference caused the output from pcretest
3218 to vary, making some of the tests fail. I have changed pcretest so that:
3219
3220 (a) When it is outputting text in the compiled version of a pattern, bytes
3221 other than 32-126 are always shown as hex escapes.
3222
3223 (b) When it is outputting text that is a matched part of a subject string,
3224 it does the same, unless a different locale has been set for the match
3225 (using the /L modifier). In this case, it uses isprint() to decide.
3226
3227 4. Fixed a major bug that caused incorrect computation of the amount of memory
3228 required for a compiled pattern when options that changed within the
3229 pattern affected the logic of the preliminary scan that determines the
3230 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3231 that the computed length was too small. The symptoms of this bug were
3232 either the PCRE error "internal error: code overflow" from pcre_compile(),
3233 or a glibc crash with a message such as "pcretest: free(): invalid next
3234 size (fast)". Examples of patterns that provoked this bug (shown in
3235 pcretest format) are:
3236
3237 /(?-x: )/x
3238 /(?x)(?-x: \s*#\s*)/
3239 /((?i)[\x{c0}])/8
3240 /(?i:[\x{c0}])/8
3241
3242 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3243 is now done differently.
3244
3245 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3246 wrapper classes; (b) implement a new function in the C++ scanner that is
3247 more efficient than the old way of doing things because it avoids levels of
3248 recursion in the regex matching; (c) add a paragraph to the documentation
3249 for the FullMatch() function.
3250
3251 6. The escape sequence \n was being treated as whatever was defined as
3252 "newline". Not only was this contrary to the documentation, which states
3253 that \n is character 10 (hex 0A), but it also went horribly wrong when
3254 "newline" was defined as CRLF. This has been fixed.
3255
3256 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3257 was being set to -1 for the "end of line" case (supposedly a value that no
3258 character can have). Though this value is never used (the check for end of
3259 line is "zero bytes in current character"), it caused compiler complaints.
3260 I've changed it to 0xffffffff.
3261
3262 8. In pcre_version.c, the version string was being built by a sequence of
3263 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3264 string (as it is for production releases) called a macro with an empty
3265 argument. The C standard says the result of this is undefined. The gcc
3266 compiler treats it as an empty string (which was what was wanted) but it is
3267 reported that Visual C gives an error. The source has been hacked around to
3268 avoid this problem.
3269
3270 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3271 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3272 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3273 of them did).
3274
3275 10. Originally, pcretest opened its input and output without "b"; then I was
3276 told that "b" was needed in some environments, so it was added for release
3277 5.0 to both the input and output. (It makes no difference on Unix-like
3278 systems.) Later I was told that it is wrong for the input on Windows. I've
3279 now abstracted the modes into two macros, to make it easier to fiddle with
3280 them, and removed "b" from the input mode under Windows.
3281
3282 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3283
3284 12. Added -help and --help to pcretest as an official way of being reminded
3285 of the options.
3286
3287 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3288 and pcrecpp.cc because they annoy compilers at high warning levels.
3289
3290 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3291
3292 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3293 scripts are not C programs :-) and which was not noticed because it works
3294 on Linux.
3295
3296 16. pcretest is supposed to handle any length of pattern and data line (as one
3297 line or as a continued sequence of lines) by extending its input buffer if
3298 necessary. This feature was broken for very long pattern lines, leading to
3299 a string of junk being passed to pcre_compile() if the pattern was longer
3300 than about 50K.
3301
3302 17. I have done a major re-factoring of the way pcre_compile() computes the
3303 amount of memory needed for a compiled pattern. Previously, there was code
3304 that made a preliminary scan of the pattern in order to do this. That was
3305 OK when PCRE was new, but as the facilities have expanded, it has become
3306 harder and harder to keep it in step with the real compile phase, and there
3307 have been a number of bugs (see for example, 4 above). I have now found a
3308 cunning way of running the real compile function in a "fake" mode that
3309 enables it to compute how much memory it would need, while actually only
3310 ever using a few hundred bytes of working memory and without too many
3311 tests of the mode. This should make future maintenance and development
3312 easier. A side effect of this work is that the limit of 200 on the nesting
3313 depth of parentheses has been removed (though this was never a serious
3314 limitation, I suspect). However, there is a downside: pcre_compile() now
3315 runs more slowly than before (30% or more, depending on the pattern). I
3316 hope this isn't a big issue. There is no effect on runtime performance.
3317
3318 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3319 newline (only possible for the last line of a file) and it was a
3320 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3321
3322 19. Added additional timing features to pcretest. (1) The -tm option now times
3323 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3324 separate command line item, by a number that specifies the number of
3325 repeats to use when timing. The default is 50000; this gives better
3326 precision, but takes uncomfortably long for very large patterns.
3327
3328 20. Extended pcre_study() to be more clever in cases where a branch of a
3329 subpattern has no definite first character. For example, (a*|b*)[cd] would
3330 previously give no result from pcre_study(). Now it recognizes that the
3331 first character must be a, b, c, or d.
3332
3333 21. There was an incorrect error "recursive call could loop indefinitely" if
3334 a subpattern (or the entire pattern) that was being tested for matching an
3335 empty string contained only one non-empty item after a nested subpattern.
3336 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3337 incorrectly, because the \d was being skipped in the check.
3338
3339 22. The pcretest program now has a new pattern option /B and a command line
3340 option -b, which is equivalent to adding /B to every pattern. This causes
3341 it to show the compiled bytecode, without the additional information that
3342 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3343 is the same as /B/I).
3344
3345 23. A new optimization is now able automatically to treat some sequences such
3346 as a*b as a*+b. More specifically, if something simple (such as a character
3347 or a simple class like \d) has an unlimited quantifier, and is followed by
3348 something that cannot possibly match the quantified thing, the quantifier
3349 is automatically "possessified".
3350
3351 24. A recursive reference to a subpattern whose number was greater than 39
3352 went wrong under certain circumstances in UTF-8 mode. This bug could also
3353 have affected the operation of pcre_study().
3354
3355 25. Realized that a little bit of performance could be had by replacing
3356 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3357
3358 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3359
3360 27. Possessive quantifiers such as a++ were previously implemented by turning
3361 them into atomic groups such as ($>a+). Now they have their own opcodes,
3362 which improves performance. This includes the automatically created ones
3363 from 23 above.
3364
3365 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3366 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3367 the first matched character to be a colon. This applied both to named and
3368 numbered groups.
3369
3370 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3371
3372 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3373 building PCRE on Windows easier, so I have included it in the distribution.
3374
3375 31. There is now a check in pcretest against a ridiculously large number being
3376 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3377 loop, the loop is abandoned.
3378
3379 32. Forward references to subpatterns in conditions such as (?(2)...) where
3380 subpattern 2 is defined later cause pcre_compile() to search forwards in
3381 the pattern for the relevant set of parentheses. This search went wrong
3382 when there were unescaped parentheses in a character class, parentheses
3383 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3384
3385 33. "Subroutine" calls and backreferences were previously restricted to
3386 referencing subpatterns earlier in the regex. This restriction has now
3387 been removed.
3388
3389 34. Added a number of extra features that are going to be in Perl 5.10. On the
3390 whole, these are just syntactic alternatives for features that PCRE had
3391 previously implemented using the Python syntax or my own invention. The
3392 other formats are all retained for compatibility.
3393
3394 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3395 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3396 also .NET compatible.
3397
3398 (b) A recursion or subroutine call to a named group can now be defined as
3399 (?&name) as well as (?P>name).
3400
3401 (c) A backreference to a named group can now be defined as \k<name> or
3402 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3403 5.10, are also .NET compatible.
3404
3405 (d) A conditional reference to a named group can now use the syntax
3406 (?(<name>) or (?('name') as well as (?(name).
3407
3408 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3409 groups (named and numbered) that are never evaluated inline, but can be
3410 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3411 is always false. There may be only one alternative in such a group.
3412
3413 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3414 as the simple (?(R). The condition is true only if the most recent
3415 recursion is that of the given number or name. It does not search out
3416 through the entire recursion stack.
3417
3418 (g) The escape \gN or \g{N} has been added, where N is a positive or
3419 negative number, specifying an absolute or relative reference.
3420
3421 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3422 some "unreachable code" warnings.
3423
3424 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3425 things, this adds five new scripts.
3426
3427 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3428 There were also incompatibilities regarding the handling of \Q..\E inside
3429 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3430 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3431
3432 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3433 matches an empty string, and forcibly breaks the loop. There were bugs in
3434 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3435 against aaaa the result was just "a" rather than "aaaa", for example. Two
3436 separate and independent bugs (that affected different cases) have been
3437 fixed.
3438
3439 39. Refactored the code to abolish the use of different opcodes for small
3440 capturing bracket numbers. This is a tidy that I avoided doing when I
3441 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3442 The new approach is not only tidier, it makes it possible to reduce the
3443 memory needed to fix the previous bug (38).
3444
3445 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3446 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3447 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3448 mode.
3449
3450 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3451 report.
3452
3453 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3454 copy construction and assignment in the C++ wrapper.
3455
3456 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3457 couple of bugs that could have given wrong results in the "--newline=crlf"
3458 case.
3459
3460 44. Added a number of casts and did some reorganization of signed/unsigned int
3461 variables following suggestions from Dair Grant. Also renamed the variable
3462 "this" as "item" because it is a C++ keyword.
3463
3464 45. Arranged for dftables to add
3465
3466 #include "pcre_internal.h"
3467
3468 to pcre_chartables.c because without it, gcc 4.x may remove the array
3469 definition from the final binary if PCRE is built into a static library and
3470 dead code stripping is activated.
3471
3472 46. For an unanchored pattern, if a match attempt fails at the start of a
3473 newline sequence, and the newline setting is CRLF or ANY, and the next two
3474 characters are CRLF, advance by two characters instead of one.
3475
3476
3477 Version 6.7 04-Jul-06
3478 ---------------------
3479
3480 1. In order to handle tests when input lines are enormously long, pcretest has
3481 been re-factored so that it automatically extends its buffers when
3482 necessary. The code is crude, but this _is_ just a test program. The
3483 default size has been increased from 32K to 50K.
3484
3485 2. The code in pcre_study() was using the value of the re argument before
3486 testing it for NULL. (Of course, in any sensible call of the function, it
3487 won't be NULL.)
3488
3489 3. The memmove() emulation function in pcre_internal.h, which is used on
3490 systems that lack both memmove() and bcopy() - that is, hardly ever -
3491 was missing a "static" storage class specifier.
3492
3493 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3494 containing an extended class (one that cannot be represented by a bitmap
3495 because it contains high-valued characters or Unicode property items, e.g.
3496 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3497 pattern, but PCRE should not loop if you do not (it no longer does).
3498 [Detail: two cases were found: (a) a repeated subpattern containing an
3499 extended class; (b) a recursive reference to a subpattern that followed a
3500 previous extended class. It wasn't skipping over the extended class
3501 correctly when UTF-8 mode was not set.]
3502
3503 5. A negated single-character class was not being recognized as fixed-length
3504 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3505 compile error "lookbehind assertion is not fixed length".
3506
3507 6. The RunPerlTest auxiliary script was showing an unexpected difference
3508 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3509 write a Perl script that can interpret lines of an input file either as
3510 byte characters or as UTF-8, which is what "perltest" was being required to
3511 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3512 can't do is switch easily at run time between having the "use utf8;" pragma
3513 or not. In the end, I fudged it by using the RunPerlTest script to insert
3514 "use utf8;" explicitly for the UTF-8 tests.
3515
3516 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3517 the end of the subject string, contrary to the documentation and to what
3518 Perl does. This was true of both matching functions. Now it matches only at
3519 the start of the subject and immediately after *internal* newlines.
3520
3521 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3522 a pointer to an int instead of a pointer to an unsigned long int. This
3523 caused problems on 64-bit systems.
3524
3525 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3526 instance of the 'standard' template library not being so standard".
3527
3528 10. There was no check on the number of named subpatterns nor the maximum
3529 length of a subpattern name. The product of these values is used to compute
3530 the size of the memory block for a compiled pattern. By supplying a very
3531 long subpattern name and a large number of named subpatterns, the size
3532 computation could be caused to overflow. This is now prevented by limiting
3533 the length of names to 32 characters, and the number of named subpatterns
3534 to 10,000.
3535
3536 11. Subpatterns that are repeated with specific counts have to be replicated in
3537 the compiled pattern. The size of memory for this was computed from the
3538 length of the subpattern and the repeat count. The latter is limited to
3539 65535, but there was no limit on the former, meaning that integer overflow
3540 could in principle occur. The compiled length of a repeated subpattern is
3541 now limited to 30,000 bytes in order to prevent this.
3542
3543 12. Added the optional facility to have named substrings with the same name.
3544
3545 13. Added the ability to use a named substring as a condition, using the
3546 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3547 are numbers (not recommended). Forward references are permitted.
3548
3549 14. Added forward references in named backreferences (if you see what I mean).
3550
3551 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3552 pattern could run off the end of the subject. For example, the pattern
3553 "(?s)(.{1,5})"8 did this with the subject "ab".
3554
3555 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3556 PCRE_CASELESS was set when matching characters that were quantified with ?
3557 or *.
3558
3559 17. A character class other than a single negated character that had a minimum
3560 but no maximum quantifier - for example [ab]{6,} - was not handled
3561 correctly by pce_dfa_exec(). It would match only one character.
3562
3563 18. A valid (though odd) pattern that looked like a POSIX character
3564 class but used an invalid character after [ (for example [[,abc,]]) caused
3565 pcre_compile() to give the error "Failed: internal error: code overflow" or
3566 in some cases to crash with a glibc free() error. This could even happen if
3567 the pattern terminated after [[ but there just happened to be a sequence of
3568 letters, a binary zero, and a closing ] in the memory that followed.
3569
3570 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3571 over the years. Originally (before any Unicode support), just the bottom 8
3572 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3573 output from "man perlunicode" includes this:
3574
3575 The regular expression compiler produces polymorphic opcodes. That
3576 is, the pattern adapts to the data and automatically switches to
3577 the Unicode character scheme when presented with Unicode data--or
3578 instead uses a traditional byte scheme when presented with byte
3579 data.
3580
3581 Sadly, a wide octal escape does not cause a switch, and in a string with
3582 no other multibyte characters, these octal escapes are treated as before.
3583 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3584 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3585 Unicode string.
3586
3587 I have not perpetrated such confusion in PCRE. Up till now, it took just
3588 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3589 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3590 translate to the appropriate multibyte character.
3591
3592 29. Applied some refactoring to reduce the number of warnings from Microsoft
3593 and Borland compilers. This has included removing the fudge introduced
3594 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3595 a warning about an unused variable.
3596
3597 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3598 characters since release 4.0, because Perl (from release 5.004) does not.
3599 [Or at least, is documented not to: some releases seem to be in conflict
3600 with the documentation.] However, when a pattern was studied with
3601 pcre_study() and all its branches started with \s, PCRE still included VT
3602 as a possible starting character. Of course, this did no harm; it just
3603 caused an unnecessary match attempt.
3604
3605 22. Removed a now-redundant internal flag bit that recorded the fact that case
3606 dependency changed within the pattern. This was once needed for "required
3607 byte" processing, but is no longer used. This recovers a now-scarce options
3608 bit. Also moved the least significant internal flag bit to the most-
3609 significant bit of the word, which was not previously used (hangover from
3610 the days when it was an int rather than a uint) to free up another bit for
3611 the future.
3612
3613 23. Added support for CRLF line endings as well as CR and LF. As well as the
3614 default being selectable at build time, it can now be changed at runtime
3615 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3616 specify that it is scanning data with non-default line endings.
3617
3618 24. Changed the definition of CXXLINK to make it agree with the definition of
3619 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3620
3621 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3622 recursions. This makes a big different to stack usage for some patterns.
3623
3624 26. If a subpattern containing a named recursion or subroutine reference such
3625 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3626 the space required for the compiled pattern went wrong and gave too small a
3627 value. Depending on the environment, this could lead to "Failed: internal
3628 error: code overflow at offset 49" or "glibc detected double free or
3629 corruption" errors.
3630
3631 27. Applied patches from Google (a) to support the new newline modes and (b) to
3632 advance over multibyte UTF-8 characters in GlobalReplace.
3633
3634 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3635 difference for some implementation of PCRE in some Windows version.
3636
3637 29. Added some extra testing facilities to pcretest:
3638
3639 \q<number> in a data line sets the "match limit" value
3640 \Q<number> in a data line sets the "match recursion limt" value
3641 -S <number> sets the stack size, where <number> is in megabytes
3642
3643 The -S option isn't available for Windows.
3644
3645
3646 Version 6.6 06-Feb-06
3647 ---------------------
3648
3649 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3650 in pcreposix.h. I have copied the definition from pcre.h.
3651
3652 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3653 because pcre.h is no longer a built file.
3654
3655 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3656 not normally included in the compiled code.
3657
3658
3659 Version 6.5 01-Feb-06
3660 ---------------------
3661
3662 1. When using the partial match feature with pcre_dfa_exec(), it was not
3663 anchoring the second and subsequent partial matches at the new starting
3664 point. This could lead to incorrect results. For example, with the pattern
3665 /1234/, partially matching against "123" and then "a4" gave a match.
3666
3667 2. Changes to pcregrep:
3668
3669 (a) All non-match returns from pcre_exec() were being treated as failures
3670 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3671 error message is output. Some extra information is given for the
3672 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3673 probably the only errors that are likely to be caused by users (by
3674 specifying a regex that has nested indefinite repeats, for instance).
3675 If there are more than 20 of these errors, pcregrep is abandoned.
3676
3677 (b) A binary zero was treated as data while matching, but terminated the
3678 output line if it was written out. This has been fixed: binary zeroes
3679 are now no different to any other data bytes.
3680
3681 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3682 used to set a locale for matching. The --locale=xxxx long option has
3683 been added (no short equivalent) to specify a locale explicitly on the
3684 pcregrep command, overriding the environment variables.
3685
3686 (d) When -B was used with -n, some line numbers in the output were one less
3687 than they should have been.
3688
3689 (e) Added the -o (--only-matching) option.
3690
3691 (f) If -A or -C was used with -c (count only), some lines of context were
3692 accidentally printed for the final match.
3693
3694 (g) Added the -H (--with-filename) option.
3695
3696 (h) The combination of options -rh failed to suppress file names for files
3697 that were found from directory arguments.
3698
3699 (i) Added the -D (--devices) and -d (--directories) options.
3700
3701 (j) Added the -F (--fixed-strings) option.
3702
3703 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3704
3705 (l) Added the --colo(u)r option.
3706
3707 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3708 is not present by default.
3709
3710 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3711 items such as (?R) or (?1), when the recursion could match a number of
3712 alternatives. If it matched one of the alternatives, but subsequently,
3713 outside the recursion, there was a failure, the code tried to back up into
3714 the recursion. However, because of the way PCRE is implemented, this is not
3715 possible, and the result was an incorrect result from the match.
3716
3717 In order to prevent this happening, the specification of recursion has
3718 been changed so that all such subpatterns are automatically treated as
3719 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3720
3721 4. I had overlooked the fact that, in some locales, there are characters for
3722 which isalpha() is true but neither isupper() nor islower() are true. In
3723 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3724 and ordfeminine) are like this. This affected the treatment of \w and \W
3725 when they appeared in character classes, but not when they appeared outside
3726 a character class. The bit map for "word" characters is now created
3727 separately from the results of isalnum() instead of just taking it from the
3728 upper, lower, and digit maps. (Plus the underscore character, of course.)
3729
3730 5. The above bug also affected the handling of POSIX character classes such as
3731 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3732 permanent tables. Instead, the bit maps for such a class were previously
3733 created as the appropriate unions of the upper, lower, and digit bitmaps.
3734 Now they are created by subtraction from the [[:word:]] class, which has
3735 its own bitmap.
3736
3737 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3738 It is created by subtracting the vertical space characters (\x09, \x0a,
3739 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3740 subtraction was done in the overall bitmap for a character class, meaning
3741 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3742 be recognized. This bug has been fixed.
3743
3744 7. Patches from the folks at Google:
3745
3746 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3747 real life, but is still worth protecting against".
3748
3749 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3750 regular expressions".
3751
3752 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3753 have it.
3754
3755 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3756 "configure" and the latter not, in order to fix a problem somebody had
3757 with compiling the Arg class on HP-UX.
3758
3759 (e) Improve the error-handling of the C++ wrapper a little bit.
3760
3761 (f) New tests for checking recursion limiting.
3762
3763 8. The pcre_memmove() function, which is used only if the environment does not
3764 have a standard memmove() function (and is therefore rarely compiled),
3765 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3766 returning a result (though PCRE never actually uses the result).
3767
3768 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3769 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3770 returned instead of calling malloc() with an overflowing number that would
3771 most likely cause subsequent chaos.
3772
3773 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3774
3775 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3776 with this option is matched, the nmatch and pmatch options of regexec() are
3777 ignored.
3778
3779 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3780 provided in case anyone wants to the the POSIX interface with UTF-8
3781 strings.
3782
3783 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3784 C++ linking (needed for some HP-UX environments).
3785
3786 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3787 (unused parameter) and in the pcre_printint() function (omitted "default"
3788 switch label when the default is to do nothing).
3789
3790 15. Added some code to make it possible, when PCRE is compiled as a C++
3791 library, to replace subject pointers for pcre_exec() with a smart pointer
3792 class, thus making it possible to process discontinuous strings.
3793
3794 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3795 much the same function. They were added by different people who were trying
3796 to make PCRE easy to compile on non-Unix systems. It has been suggested
3797 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3798 for compiling on Windows systems. I have therefore replaced it with
3799 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3800 defaults to "extern" for C or "extern C" for C++, which works fine on
3801 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3802 SCOPE with something explicit in config.h. In addition:
3803
3804 (a) pcreposix.h still had just "extern" instead of either of these macros;
3805 I have replaced it with PCRE_DATA_SCOPE.
3806
3807 (b) Functions such as _pcre_xclass(), which are internal to the library,
3808 but external in the C sense, all had PCRE_EXPORT in their definitions.
3809 This is apparently wrong for the Windows case, so I have removed it.
3810 (It makes no difference on Unix-like systems.)
3811
3812 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3813 of recursive calls to match(). This is different to MATCH_LIMIT because
3814 that limits the total number of calls to match(), not all of which increase
3815 the depth of recursion. Limiting the recursion depth limits the amount of
3816 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3817 when PCRE is compiled, and changed at run time. A patch from Google adds
3818 this functionality to the C++ interface.
3819
3820 18. Changes to the handling of Unicode character properties:
3821
3822 (a) Updated the table to Unicode 4.1.0.
3823
3824 (b) Recognize characters that are not in the table as "Cn" (undefined).
3825
3826 (c) I revised the way the table is implemented to a much improved format
3827 which includes recognition of ranges. It now supports the ranges that
3828 are defined in UnicodeData.txt, and it also amalgamates other
3829 characters into ranges. This has reduced the number of entries in the
3830 table from around 16,000 to around 3,000, thus reducing its size
3831 considerably. I realized I did not need to use a tree structure after
3832 all - a binary chop search is just as efficient. Having reduced the
3833 number of entries, I extended their size from 6 bytes to 8 bytes to
3834 allow for more data.
3835
3836 (d) Added support for Unicode script names via properties such as \p{Han}.
3837
3838 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3839 matching that character.
3840
3841 20. When matching a repeated Unicode property with a minimum greater than zero,
3842 (for example \pL{2,}), PCRE could look past the end of the subject if it
3843 reached it while seeking the minimum number of characters. This could
3844 happen only if some of the characters were more than one byte long, because
3845 there is a check for at least the minimum number of bytes.
3846
3847 21. Refactored the implementation of \p and \P so as to be more general, to
3848 allow for more different types of property in future. This has changed the
3849 compiled form incompatibly. Anybody with saved compiled patterns that use
3850 \p or \P will have to recompile them.
3851
3852 22. Added "Any" and "L&" to the supported property types.
3853
3854 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3855 but give a compile time error if the value is greater than 0xff.
3856
3857 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3858 accidentally not being installed or uninstalled.
3859
3860 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3861 made were to insert the current release number. This seemed silly, because
3862 it made things harder for people building PCRE on systems that don't run
3863 "configure". I have turned pcre.h into a distributed file, no longer built
3864 by "configure", with the version identification directly included. There is
3865 no longer a pcre.h.in file.
3866
3867 However, this change necessitated a change to the pcre-config script as
3868 well. It is built from pcre-config.in, and one of the substitutions was the
3869 release number. I have updated configure.ac so that ./configure now finds
3870 the release number by grepping pcre.h.
3871
3872 26. Added the ability to run the tests under valgrind.
3873
3874
3875 Version 6.4 05-Sep-05
3876 ---------------------
3877
3878 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3879 "--" to be printed when multiple files were scanned, even when none of the
3880 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3881 consider it to be a bug, and have restored the previous behaviour.
3882
3883 2. A couple of code tidies to get rid of compiler warnings.
3884
3885 3. The pcretest program used to cheat by referring to symbols in the library
3886 whose names begin with _pcre_. These are internal symbols that are not
3887 really supposed to be visible externally, and in some environments it is
3888 possible to suppress them. The cheating is now confined to including
3889 certain files from the library's source, which is a bit cleaner.
3890
3891 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3892 file's purpose clearer.
3893
3894 5. Reorganized pcre_ucp_findchar().
3895
3896
3897 Version 6.3 15-Aug-05
3898 ---------------------
3899
3900 1. The file libpcre.pc.in did not have general read permission in the tarball.
3901
3902 2. There were some problems when building without C++ support:
3903
3904 (a) If C++ support was not built, "make install" and "make test" still
3905 tried to test it.
3906
3907 (b) There were problems when the value of CXX was explicitly set. Some
3908 changes have been made to try to fix these, and ...
3909
3910 (c) --disable-cpp can now be used to explicitly disable C++ support.
3911
3912 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3913 backslash in a target when C++ was disabled. This confuses some
3914 versions of "make", apparently. Using an intermediate variable solves
3915 this. (Same for CPP_LOBJ.)
3916
3917 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3918 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3919 necessary on certain architectures.
3920
3921 4. Added a setting of -export-symbols-regex to the link command to remove
3922 those symbols that are exported in the C sense, but actually are local
3923 within the library, and not documented. Their names all begin with
3924 "_pcre_". This is not a perfect job, because (a) we have to except some
3925 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3926 available (and never for static libraries). I have made a note to try to
3927 find a way round (a) in the future.
3928
3929
3930 Version 6.2 01-Aug-05
3931 ---------------------
3932
3933 1. There was no test for integer overflow of quantifier values. A construction
3934 such as {1111111111111111} would give undefined results. What is worse, if
3935 a minimum quantifier for a parenthesized subpattern overflowed and became
3936 negative, the calculation of the memory size went wrong. This could have
3937 led to memory overwriting.
3938
3939 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3940
3941 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3942 operating environments where this matters.
3943
3944 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3945 PCRE options from within the C++ wrapper.
3946
3947 5. Named capturing subpatterns were not being correctly counted when a pattern
3948 was compiled. This caused two problems: (a) If there were more than 100
3949 such subpatterns, the calculation of the memory needed for the whole
3950 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3951 back references of the form \12, where the number was greater than 9, were
3952 not recognized as back references, even though there were sufficient
3953 previous subpatterns.
3954
3955 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3956 versions of gcc, e.g. 2.95.4.
3957
3958
3959 Version 6.1 21-Jun-05
3960 ---------------------
3961
3962 1. There was one reference to the variable "posix" in pcretest.c that was not
3963 surrounded by "#if !defined NOPOSIX".
3964
3965 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3966 the cross-check on the old pcre_info() function, for the benefit of the
3967 cut-down version of PCRE that is currently imported into Exim.
3968
3969 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3970 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3971 patterns that start (?i) but I don't think that matters. The use of (?i) is
3972 just an example; this all applies to the other options as well.
3973
3974 4. Since libtool seems to echo the compile commands it is issuing, the output
3975 from "make" can be reduced a bit by putting "@" in front of each libtool
3976 compile command.
3977
3978 5. Patch from the folks at Google for configure.in to be a bit more thorough
3979 in checking for a suitable C++ installation before trying to compile the
3980 C++ stuff. This should fix a reported problem when a compiler was present,
3981 but no suitable headers.
3982
3983 6. The man pages all had just "PCRE" as their title. I have changed them to
3984 be the relevant file name. I have also arranged that these names are
3985 retained in the file doc/pcre.txt, which is a concatenation in text format
3986 of all the man pages except the little individual ones for each function.
3987
3988 7. The NON-UNIX-USE file had not been updated for the different set of source
3989 files that come with release 6. I also added a few comments about the C++
3990 wrapper.
3991
3992
3993 Version 6.0 07-Jun-05
3994 ---------------------
3995
3996 1. Some minor internal re-organization to help with my DFA experiments.
3997
3998 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3999 didn't matter for the library itself when fully configured, but did matter
4000 when compiling without UCP support, or within Exim, where the ucp files are
4001 not imported.
4002
4003 3. Refactoring of the library code to split up the various functions into
4004 different source modules. The addition of the new DFA matching code (see
4005 below) to a single monolithic source would have made it really too
4006 unwieldy, quite apart from causing all the code to be include in a
4007 statically linked application, when only some functions are used. This is
4008 relevant even without the DFA addition now that patterns can be compiled in
4009 one application and matched in another.
4010
4011 The downside of splitting up is that there have to be some external
4012 functions and data tables that are used internally in different modules of
4013 the library but which are not part of the API. These have all had their
4014 names changed to start with "_pcre_" so that they are unlikely to clash
4015 with other external names.
4016
4017 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4018 a different (DFA) algorithm. Although it is slower than the original
4019 function, it does have some advantages for certain types of matching
4020 problem.
4021
4022 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4023 including restarting after a partial match.
4024
4025 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4026 defined when compiling for Windows was sent to me. I have put it into the
4027 code, though I have no means of testing or verifying it.
4028
4029 7. Added the pcre_refcount() auxiliary function.
4030
4031 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4032 match before or at the first newline in the subject string. In pcretest,
4033 the /f option on a pattern can be used to set this.
4034
4035 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4036 would behave wrongly. This has been present in PCRE since release 4.0.
4037
4038 10. A number of changes to the pcregrep command:
4039
4040 (a) Refactored how -x works; insert ^(...)$ instead of setting
4041 PCRE_ANCHORED and checking the length, in preparation for adding
4042 something similar for -w.
4043
4044 (b) Added the -w (match as a word) option.
4045
4046 (c) Refactored the way lines are read and buffered so as to have more
4047 than one at a time available.
4048
4049 (d) Implemented a pcregrep test script.
4050
4051 (e) Added the -M (multiline match) option. This allows patterns to match
4052 over several lines of the subject. The buffering ensures that at least
4053 8K, or the rest of the document (whichever is the shorter) is available
4054 for matching (and similarly the previous 8K for lookbehind assertions).
4055
4056 (f) Changed the --help output so that it now says
4057
4058 -w, --word-regex(p)
4059
4060 instead of two lines, one with "regex" and the other with "regexp"
4061 because that confused at least one person since the short forms are the
4062 same. (This required a bit of code, as the output is generated
4063 automatically from a table. It wasn't just a text change.)
4064
4065 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4066 option but starts with a hyphen. Could be a pattern or a path name
4067 starting with a hyphen, for instance.
4068
4069 (h) "-" can be given as a file name to represent stdin.
4070
4071 (i) When file names are being printed, "(standard input)" is used for
4072 the standard input, for compatibility with GNU grep. Previously
4073 "<stdin>" was used.
4074
4075 (j) The option --label=xxx can be used to supply a name to be used for
4076 stdin when file names are being printed. There is no short form.
4077
4078 (k) Re-factored the options decoding logic because we are going to add
4079 two more options that take data. Such options can now be given in four
4080 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4081
4082 (l) Added the -A, -B, and -C options for requesting that lines of context
4083 around matches be printed.
4084
4085 (m) Added the -L option to print the names of files that do not contain
4086 any matching lines, that is, the complement of -l.
4087
4088 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4089 continue to scan other files.
4090
4091 (o) The -s option was incorrectly implemented. For compatibility with other
4092 greps, it now suppresses the error message for a non-existent or non-
4093 accessible file (but not the return code). There is a new option called
4094 -q that suppresses the output of matching lines, which was what -s was
4095 previously doing.
4096
4097 (p) Added --include and --exclude options to specify files for inclusion
4098 and exclusion when recursing.
4099
4100 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4101 Hopefully, it now does.
4102
4103 12. Missing cast in pcre_study().
4104
4105 13. Added an "uninstall" target to the makefile.
4106
4107 14. Replaced "extern" in the function prototypes in Makefile.in with
4108 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4109 world, but is set differently for Windows.
4110
4111 15. Added a second compiling function called pcre_compile2(). The only
4112 difference is that it has an extra argument, which is a pointer to an
4113 integer error code. When there is a compile-time failure, this is set
4114 non-zero, in addition to the error test pointer being set to point to an
4115 error message. The new argument may be NULL if no error number is required
4116 (but then you may as well call pcre_compile(), which is now just a
4117 wrapper). This facility is provided because some applications need a
4118 numeric error indication, but it has also enabled me to tidy up the way
4119 compile-time errors are handled in the POSIX wrapper.
4120
4121 16. Added VPATH=.libs to the makefile; this should help when building with one
4122 prefix path and installing with another. (Or so I'm told by someone who
4123 knows more about this stuff than I do.)
4124
4125 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4126 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4127 match everything, including newlines. This is not POSIX-compatible, but
4128 somebody wanted the feature. From pcretest it can be activated by using
4129 both the P and the s flags.
4130
4131 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4132
4133 19. libpcre.pc was being incorrectly installed as executable.
4134
4135 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4136 it now also looks for '\r' so that it will work unmodified on Windows.
4137
4138 21. Added Google's contributed C++ wrapper to the distribution.
4139
4140 22. Added some untidy missing memory free() calls in pcretest, to keep
4141 Electric Fence happy when testing.
4142
4143
4144
4145 Version 5.0 13-Sep-04
4146 ---------------------
4147
4148 1. Internal change: literal characters are no longer packed up into items
4149 containing multiple characters in a single byte-string. Each character
4150 is now matched using a separate opcode. However, there may be more than one
4151 byte in the character in UTF-8 mode.
4152
4153 2. The pcre_callout_block structure has two new fields: pattern_position and
4154 next_item_length. These contain the offset in the pattern to the next match
4155 item, and its length, respectively.
4156
4157 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4158 insertion of callouts before each pattern item. Added the /C option to
4159 pcretest to make use of this.
4160
4161 4. On the advice of a Windows user, the lines
4162
4163 #if defined(_WIN32) || defined(WIN32)
4164 _setmode( _fileno( stdout ), 0x8000 );
4165 #endif /* defined(_WIN32) || defined(WIN32) */
4166
4167 have been added to the source of pcretest. This apparently does useful
4168 magic in relation to line terminators.
4169
4170 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4171 for the benefit of those environments where the "b" makes a difference.
4172
4173 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4174 to know about it. I have put a hack into configure.in that adds in code
4175 to set GCC=yes if CC=icc. This seems to end up at a point in the
4176 generated configure script that is early enough to affect the setting of
4177 compiler options, which is what is needed, but I have no means of testing
4178 whether it really works. (The user who reported this had patched the
4179 generated configure script, which of course I cannot do.)
4180
4181 LATER: After change 22 below (new libtool files), the configure script
4182 seems to know about icc (and also ecc). Therefore, I have commented out
4183 this hack in configure.in.
4184
4185 7. Added support for pkg-config (2 patches were sent in).
4186
4187 8. Negated POSIX character classes that used a combination of internal tables
4188 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4189 [[:^ascii]]. Typically, they would match almost any characters. The other
4190 POSIX classes were not broken in this way.
4191
4192 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4193 to find the match, as PCRE was deluded into thinking that the match had to
4194 start at the start point or following a newline. The same bug applied to
4195 patterns with negative forward assertions or any backward assertions
4196 preceding ".*" at the start, unless the pattern required a fixed first
4197 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4198
4199 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4200 starting at the last subject character, bytes beyond the end of the subject
4201 string were read.
4202
4203 11. Renamed the variable "class" as "classbits" to make life easier for C++
4204 users. (Previously there was a macro definition, but it apparently wasn't
4205 enough.)
4206
4207 12. Added the new field "tables" to the extra data so that tables can be passed
4208 in at exec time, or the internal tables can be re-selected. This allows
4209 a compiled regex to be saved and re-used at a later time by a different
4210 program that might have everything at different addresses.
4211
4212 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4213 -R library as well as a -L library.
4214
4215 14. The debugging options of pcretest (-d on the command line or D on a
4216 pattern) showed incorrect output for anything following an extended class
4217 that contained multibyte characters and which was followed by a quantifier.
4218
4219 15. Added optional support for general category Unicode character properties
4220 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4221 support. It adds about 90K to the size of the library. The meanings of the
4222 inbuilt class escapes such as \d and \s have NOT been changed.
4223
4224 16. Updated pcredemo.c to include calls to free() to release the memory for the
4225 compiled pattern.
4226
4227 17. The generated file chartables.c was being created in the source directory
4228 instead of in the building directory. This caused the build to fail if the
4229 source directory was different from the building directory, and was
4230 read-only.
4231
4232 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4233 file. No doubt somebody will tell me if they don't make sense... Also added
4234 Dan Mooney's comments about building on OpenVMS.
4235
4236 19. Added support for partial matching via the PCRE_PARTIAL option for
4237 pcre_exec() and the \P data escape in pcretest.
4238
4239 20. Extended pcretest with 3 new pattern features:
4240
4241 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4242 write the compiled pattern to the file whose name is "rest-of-line".
4243 This is a straight binary dump of the data, with the saved pointer to
4244 the character tables forced to be NULL. The study data, if any, is
4245 written too. After writing, pcretest reads a new pattern.
4246
4247 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4248 compiled pattern from the given file. There must not be any
4249 occurrences of "<" in the file name (pretty unlikely); if there are,
4250 pcretest will instead treat the initial "<" as a pattern delimiter.
4251 After reading in the pattern, pcretest goes on to read data lines as
4252 usual.
4253
4254 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4255 and 16-bit fields in a compiled pattern, to simulate a pattern that
4256 was compiled on a host of opposite endianness.
4257
4258 21. The pcre-exec() function can now cope with patterns that were compiled on
4259 hosts of opposite endianness, with this restriction:
4260
4261 As for any compiled expression that is saved and used later, the tables
4262 pointer field cannot be preserved; the extra_data field in the arguments
4263 to pcre_exec() should be used to pass in a tables address if a value
4264 other than the default internal tables were used at compile time.
4265
4266 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4267 now diagnosed as an error. Previously, most of the time, a negative number
4268 would have been treated as zero, but if in addition "ovector" was passed as
4269 NULL, a crash could occur.
4270
4271 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4272 new versions from the libtool 1.5 distribution (the last one is a copy of
4273 a file called libtool.m4). This seems to have fixed the need to patch
4274 "configure" to support Darwin 1.3 (which I used to do). However, I still
4275 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4276 workstation).
4277
4278 24. Changed the PCRE licence to be the more standard "BSD" licence.
4279
4280
4281 Version 4.5 01-Dec-03
4282 ---------------------
4283
4284 1. There has been some re-arrangement of the code for the match() function so
4285 that it can be compiled in a version that does not call itself recursively.
4286 Instead, it keeps those local variables that need separate instances for
4287 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4288 needs to "recurse". Keeping track of where control must go is done by means
4289 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4290 hide most of the details from the main code, and operates only if
4291 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4292 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4293 operating.
4294
4295 To make it easier for callers to provide specially tailored get/free
4296 functions for this usage, two new functions, pcre_stack_malloc, and
4297 pcre_stack_free, are used. They are always called in strict stacking order,
4298 and the size of block requested is always the same.
4299
4300 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4301 PCRE has been compiled to use the stack or the heap for recursion. The
4302 -C option of pcretest uses this to show which version is compiled.
4303
4304 A new data escape \S, is added to pcretest; it causes the amounts of store
4305 obtained and freed by both kinds of malloc/free at match time to be added
4306 to the output.
4307
4308 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4309 what's available on my current Linux desktop machine.
4310
4311 3. When matching a UTF-8 string, the test for a valid string at the start has
4312 been extended. If start_offset is not zero, PCRE now checks that it points
4313 to a byte that is the start of a UTF-8 character. If not, it returns
4314 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4315 this is necessary because there may be backward assertions in the pattern.
4316 When matching the same subject several times, it may save resources to use
4317 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4318
4319 4. The code for checking the validity of UTF-8 strings has been tightened so
4320 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4321 containing "overlong sequences".
4322
4323 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4324 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4325 should have been "&", but it just so happened that all the cases this let
4326 through by mistake were picked up later in the function.
4327
4328 6. I had used a variable called "isblank" - this is a C99 function, causing
4329 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4330
4331 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4332 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4333 so the version is shown; (c) stop "make test" echoing "./RunTest".
4334
4335 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4336
4337 9. The prototype for memmove() for systems that don't have it was using
4338 size_t, but the inclusion of the header that defines size_t was later. I've
4339 moved the #includes for the C headers earlier to avoid this.
4340
4341 10. Added some adjustments to the code to make it easier to compiler on certain
4342 special systems:
4343
4344 (a) Some "const" qualifiers were missing.
4345 (b) Added the macro EXPORT before all exported functions; by default this
4346 is defined to be empty.
4347 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4348 that it reads its output file name as an argument instead of writing
4349 to the standard output and assuming this can be redirected.
4350
4351 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4352 class containing characters with values greater than 255, PCRE compilation
4353 went into a loop.
4354
4355 12. A recursive reference to a subpattern that was within another subpattern
4356 that had a minimum quantifier of zero caused PCRE to crash. For example,
4357 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4358 recursion. If the recursively-called subpattern itself had a zero repeat,
4359 that was OK.
4360
4361 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4362 buffer into which it was copied (for escape processing) was still set at
4363 1024, so long lines caused crashes.
4364
4365 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4366 "internal error: code overflow...". This applied to any character class
4367 that was followed by a possessive quantifier.
4368
4369 15. Modified the Makefile to add libpcre.la as a prerequisite for
4370 libpcreposix.la because I was told this is needed for a parallel build to
4371 work.
4372
4373 16. If a pattern that contained .* following optional items at the start was
4374 studied, the wrong optimizing data was generated, leading to matching
4375 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4376 matching string must start with a or b or c. The correct conclusion for
4377 this pattern is that a match can start with any character.
4378
4379
4380 Version 4.4 13-Aug-03
4381 ---------------------
4382
4383 1. In UTF-8 mode, a character class containing characters with values between
4384 127 and 255 was not handled correctly if the compiled pattern was studied.
4385 In fixing this, I have also improved the studying algorithm for such
4386 classes (slightly).
4387
4388 2. Three internal functions had redundant arguments passed to them. Removal
4389 might give a very teeny performance improvement.
4390
4391 3. Documentation bug: the value of the capture_top field in a callout is *one
4392 more than* the number of the hightest numbered captured substring.
4393
4394 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4395 in incorrectly linking with a previously installed version. They now link
4396 explicitly with libpcre.la.
4397
4398 5. configure.in no longer needs to recognize Cygwin specially.
4399
4400 6. A problem in pcre.in for Windows platforms is fixed.
4401
4402 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4403 pcretest, it used to include the size of the study block as part of its
4404 output. Unfortunately, the structure contains a field that has a different
4405 size on different hardware architectures. This meant that the tests that
4406 showed this size failed. As the block is currently always of a fixed size,
4407 this information isn't actually particularly useful in pcretest output, so
4408 I have just removed it.
4409
4410 8. Three pre-processor statements accidentally did not start in column 1.
4411 Sadly, there are *still* compilers around that complain, even though
4412 standard C has not required this for well over a decade. Sigh.
4413
4414 9. In pcretest, the code for checking callouts passed small integers in the
4415 callout_data field, which is a void * field. However, some picky compilers
4416 complained about the casts involved for this on 64-bit systems. Now
4417 pcretest passes the address of the small integer instead, which should get
4418 rid of the warnings.
4419
4420 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4421 both compile and run time, and gives an error if an invalid UTF-8 sequence
4422 is found. There is a option for disabling this check in cases where the
4423 string is known to be correct and/or the maximum performance is wanted.
4424
4425 11. In response to a bug report, I changed one line in Makefile.in from
4426
4427 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4428 to
4429 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4430
4431 to look similar to other lines, but I have no way of telling whether this
4432 is the right thing to do, as I do not use Windows. No doubt I'll get told
4433 if it's wrong...
4434
4435
4436 Version 4.3 21-May-03
4437 ---------------------
4438
4439 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4440 Makefile.
4441
4442 2. Some refactoring to improve the quality of the code:
4443
4444 (i) The utf8_table... variables are now declared "const".
4445
4446 (ii) The code for \cx, which used the "case flipping" table to upper case
4447 lower case letters, now just substracts 32. This is ASCII-specific,
4448 but the whole concept of \cx is ASCII-specific, so it seems
4449 reasonable.
4450
4451 (iii) PCRE was using its character types table to recognize decimal and
4452 hexadecimal digits in the pattern. This is silly, because it handles
4453 only 0-9, a-f, and A-F, but the character types table is locale-
4454 specific, which means strange things might happen. A private
4455 table is now used for this - though it costs 256 bytes, a table is
4456 much faster than multiple explicit tests. Of course, the standard
4457 character types table is still used for matching digits in subject
4458 strings against \d.
4459
4460 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4461 ending in _t are). So I've renamed it as ESC_tee.
4462
4463 3. The first argument for regexec() in the POSIX wrapper should have been
4464 defined as "const".
4465
4466 4. Changed pcretest to use malloc() for its buffers so that they can be
4467 Electric Fenced for debugging.
4468
4469 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4470 to read one or more bytes before the start of the subject string. Often this
4471 had no effect on PCRE's behaviour, but in some circumstances it could
4472 provoke a segmentation fault.
4473
4474 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4475 to try to read one or more bytes before the start of the subject string.
4476
4477 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4478 UTF-8 support could misbehave in various ways if the subject string
4479 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4480 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4481 back over UTF-8 characters.)
4482
4483
4484 Version 4.2 14-Apr-03
4485 ---------------------
4486
4487 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4488
4489 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4490 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4491 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4492 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4493 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4494 and BUILD_EXEEXT
4495 Note: automatic setting of the BUILD variables is not yet working
4496 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4497 compile-time but not at link-time
4498 [LINK]: use for linking executables only
4499 make different versions for Windows and non-Windows
4500 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4501 libraries
4502 [LINK_FOR_BUILD]: new variable
4503 [OBJEXT]: use throughout
4504 [EXEEXT]: use throughout
4505 <winshared>: new target
4506 <wininstall>: new target
4507 <dftables.o>: use native compiler
4508 <dftables>: use native linker
4509 <install>: handle Windows platform correctly
4510 <clean>: ditto
4511 <check>: ditto
4512 copy DLL to top builddir before testing
4513
4514 As part of these changes, -no-undefined was removed again. This was reported
4515 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4516 in any case.
4517
4518 3. Some tidies to get rid of compiler warnings:
4519
4520 . In the match_data structure, match_limit was an unsigned long int, whereas
4521 match_call_count was an int. I've made them both unsigned long ints.
4522
4523 . In pcretest the fact that a const uschar * doesn't automatically cast to
4524 a void * provoked a warning.
4525
4526 . Turning on some more compiler warnings threw up some "shadow" variables
4527 and a few more missing casts.
4528
4529 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4530 option, a class that contained a single character with a value between 128
4531 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4532
4533 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4534 option, a class that contained several characters, but with at least one
4535 whose value was between 128 and 255 caused PCRE to crash.
4536
4537
4538 Version 4.1 12-Mar-03
4539 ---------------------
4540
4541 1. Compiling with gcc -pedantic found a couple of places where casts were
4542 needed, and a string in dftables.c that was longer than standard compilers are
4543 required to support.
4544
4545 2. Compiling with Sun's compiler found a few more places where the code could
4546 be tidied up in order to avoid warnings.
4547
4548 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4549 first of these names is deprecated in the latest Autoconf in favour of the name
4550 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4551 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4552 analogy I have changed it to CFLAGS_FOR_BUILD.
4553
4554 4. Added -no-undefined to the linking command in the Makefile, because this is
4555 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4556 linking step for the pcreposix library.
4557
4558 5. PCRE was failing to diagnose the case of two named groups with the same
4559 name.
4560
4561 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4562 literal character that is needed in the subject for a match, and scans along to
4563 ensure that it is present before embarking on the full matching process. This
4564 saves time in cases of nested unlimited repeats that are never going to match.
4565 Problem: the scan can take a lot of time if the subject is very long (e.g.
4566 megabytes), thus penalizing straightforward matches. It is now done only if the
4567 amount of subject to be scanned is less than 1000 bytes.
4568
4569 7. A lesser problem with the same optimization is that it was recording the
4570 first character of an anchored pattern as "needed", thus provoking a search
4571 right along the subject, even when the first match of the pattern was going to
4572 fail. The "needed" character is now not set for anchored patterns, unless it
4573 follows something in the pattern that is of non-fixed length. Thus, it still
4574 fulfils its original purpose of finding quick non-matches in cases of nested
4575 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4576
4577
4578 Version 4.0 17-Feb-03
4579 ---------------------
4580
4581 1. If a comment in an extended regex that started immediately after a meta-item
4582 extended to the end of string, PCRE compiled incorrect data. This could lead to
4583 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4584
4585 2. Moved to autoconf 2.53 and libtool 1.4.2.
4586
4587 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4588 the special perltest8 script is no longer needed - all the tests can be run
4589 from a single perltest script.
4590
4591 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4592 by \s. It has now been removed in PCRE. This means it isn't recognized as
4593 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4594 class [:space:] *does* include VT, thereby creating a mess.
4595
4596 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4597 space and tab.
4598
4599 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4600 its new features into the main test script, reducing the number of scripts.
4601
4602 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4603 were backward compatible, and made the (?i) apply to the whole pattern, as if
4604 /i were given. Now it behaves more logically, and applies the option setting
4605 only to what follows. PCRE has been changed to follow suit. However, if it
4606 finds options settings right at the start of the pattern, it extracts them into
4607 the global options, as before. Thus, they show up in the info data.
4608
4609 8. Added support for the \Q...\E escape sequence. Characters in between are
4610 treated as literals. This is slightly different from Perl in that $ and @ are
4611 also handled as literals inside the quotes. In Perl, they will cause variable
4612 interpolation. Note the following examples:
4613
4614 Pattern PCRE matches Perl matches
4615
4616 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4617 \Qabc\$xyz\E abc\$xyz abc\$xyz
4618 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4619
4620 For compatibility with Perl, \Q...\E sequences are recognized inside character
4621 classes as well as outside them.
4622
4623 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4624 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4625 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4626 signed/unsigned warnings.
4627
4628 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4629 option for pcretest, so I've replaced it by a simple function that does just
4630 that job.
4631
4632 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4633 "pcregrep -".
4634
4635 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4636 Java package. This provides some syntactic sugar for simple cases of what my
4637 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4638 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4639 item, you can use this simplified notation. Note that only makes sense with
4640 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4641 greediness, whatever the setting of the PCRE_UNGREEDY option.
4642
4643 13. A change of greediness default within a pattern was not taking effect at
4644 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4645 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4646 was abstracted outside.
4647
4648 14. PCRE now supports the \G assertion. It is true when the current matching
4649 position is at the start point of the match. This differs from \A when the
4650 starting offset is non-zero. Used with the /g option of pcretest (or similar
4651 code), it works in the same way as it does for Perl's /g option. If all
4652 alternatives of a regex begin with \G, the expression is anchored to the start
4653 match position, and the "anchored" flag is set in the compiled expression.
4654
4655 15. Some bugs concerning the handling of certain option changes within patterns
4656 have been fixed. These applied to options other than (?ims). For example,
4657 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4658 been the other way round. Some of this was related to change 7 above.
4659
4660 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4661 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4662 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4663 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4664
4665 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4666 mode. Unlike ".", it always matches newline, whatever the setting of
4667 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4668 assertions. Perl allows it, but it doesn't (in general) work because it can't
4669 calculate the length of the lookbehind. At least, that's the case for Perl
4670 5.8.0 - I've been told they are going to document that it doesn't work in
4671 future.
4672
4673 18. Added an error diagnosis for escapes that PCRE does not support: these are
4674 \L, \l, \N, \P, \p, \U, \u, and \X.
4675
4676 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4677 reading past the end of the pattern in cases such as /[abcd/.
4678
4679 20. PCRE was getting more memory than necessary for patterns with classes that
4680 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4681
4682 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4683 compiling PCRE for use with Virtual Pascal.
4684
4685 22. Small fix to the Makefile to make it work properly if the build is done
4686 outside the source tree.
4687
4688 23. Added a new extension: a condition to go with recursion. If a conditional
4689 subpattern starts with (?(R) the "true" branch is used if recursion has
4690 happened, whereas the "false" branch is used only at the top level.
4691
4692 24. When there was a very long string of literal characters (over 255 bytes
4693 without UTF support, over 250 bytes with UTF support), the computation of how
4694 much memory was required could be incorrect, leading to segfaults or other
4695 strange effects.
4696
4697 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4698 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4699 there was a subsequent back reference to those brackets. This meant that, for
4700 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4701 possible to check for precisely this case. All we can do is abandon the
4702 optimization if .* occurs inside capturing brackets when there are any back
4703 references whatsoever. (See below for a better fix that came later.)
4704
4705 26. The handling of the optimization for finding the first character of a
4706 non-anchored pattern, and for finding a character that is required later in the
4707 match were failing in some cases. This didn't break the matching; it just
4708 failed to optimize when it could. The way this is done has been re-implemented.
4709
4710 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4711
4712 28. Added a new feature that provides some of the functionality that Perl
4713 provides with (?{...}). The facility is termed a "callout". The way it is done
4714 in PCRE is for the caller to provide an optional function, by setting
4715 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4716 global variable. By default it is unset, which disables all calling out. To get
4717 the function called, the regex must include (?C) at appropriate points. This
4718 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4719 This provides a means of identifying different callout points. When PCRE
4720 reaches such a point in the regex, if pcre_callout has been set, the external
4721 function is called. It is provided with data in a structure called
4722 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4723 matching continues; if it returns a non-zero value, the match at the current
4724 point fails. However, backtracking will occur if possible. [This was changed
4725 later and other features added - see item 49 below.]
4726
4727 29. pcretest is upgraded to test the callout functionality. It provides a
4728 callout function that displays information. By default, it shows the start of
4729 the match and the current position in the text. There are some new data escapes
4730 to vary what happens:
4731
4732 \C+ in addition, show current contents of captured substrings
4733 \C- do not supply a callout function
4734 \C!n return 1 when callout number n is reached
4735 \C!n!m return 1 when callout number n is reached for the mth time
4736
4737 30. If pcregrep was called with the -l option and just a single file name, it
4738 output "<stdin>" if a match was found, instead of the file name.
4739
4740 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4741 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4742 pcre_exec(). This saves a malloc/free per call. The default value of
4743 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4744 when configuring.
4745
4746 32. The default maximum size of a compiled pattern is 64K. There have been a
4747 few cases of people hitting this limit. The code now uses macros to handle the
4748 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4749 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4750 configuring. Tests 2 and 5 work only with 2-byte links because they output
4751 debugging information about compiled patterns.
4752
4753 33. Internal code re-arrangements:
4754
4755 (a) Moved the debugging function for printing out a compiled regex into
4756 its own source file (printint.c) and used #include to pull it into
4757 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4758 separate copies.
4759
4760 (b) Defined the list of op-code names for debugging as a macro in
4761 internal.h so that it is next to the definition of the opcodes.
4762
4763 (c) Defined a table of op-code lengths for simpler skipping along compiled
4764 code. This is again a macro in internal.h so that it is next to the
4765 definition of the opcodes.
4766
4767 34. Added support for recursive calls to individual subpatterns, along the
4768 lines of Robin Houston's patch (but implemented somewhat differently).
4769
4770 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4771 allow it to read and process whole directories in Win32. This code was
4772 contributed by Lionel Fourquaux; it has not been tested by me.
4773
4774 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4775 used to name a group. Names consist of alphanumerics and underscores, and must
4776 be unique. Back references use the syntax (?P=name) and recursive calls use
4777 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4778 numbers. The function pcre_fullinfo() can be used after compilation to extract
4779 a name/number map. There are three relevant calls:
4780
4781 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4782 PCRE_INFO_NAMECOUNT yields the number of entries
4783 PCRE_INFO_NAMETABLE yields a pointer to the map.
4784
4785 The map is a vector of fixed-size entries. The size of each entry depends on
4786 the length of the longest name used. The first two bytes of each entry are the
4787 group number, most significant byte first. There follows the corresponding
4788 name, zero terminated. The names are in alphabetical order.
4789
4790 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4791 case instead of 255. Making it the same both with and without UTF-8 support
4792 means that the same test output works with both.
4793
4794 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4795 calling malloc() with a zero argument.
4796
4797 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4798 optimization. I've improved things by keeping a bitmap of backreferences with
4799 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4800 fact referenced, the optimization can be applied. It is unlikely that a
4801 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4802 the match to follow \n) will appear inside brackets with a number greater than
4803 31, but if it does, any back reference > 31 suppresses the optimization.
4804
4805 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4806 of disabling numbered capturing parentheses. Any opening parenthesis that is
4807 not followed by ? behaves as if it were followed by ?: but named parentheses
4808 can still be used for capturing (and they will acquire numbers in the usual
4809 way).
4810
4811 41. Redesigned the return codes from the match() function into yes/no/error so
4812 that errors can be passed back from deep inside the nested calls. A malloc
4813 failure while inside a recursive subpattern call now causes the
4814 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4815
4816 42. It is now possible to set a limit on the number of times the match()
4817 function is called in a call to pcre_exec(). This facility makes it possible to
4818 limit the amount of recursion and backtracking, though not in a directly
4819 obvious way, because the match() function is used in a number of different
4820 circumstances. The count starts from zero for each position in the subject
4821 string (for non-anchored patterns). The default limit is, for compatibility, a
4822 large number, namely 10 000 000. You can change this in two ways:
4823
4824 (a) When configuring PCRE before making, you can use --with-match-limit=n
4825 to set a default value for the compiled library.
4826
4827 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4828 a different value is set. See 45 below.
4829
4830 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4831
4832 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4833 of things that can be changed at compile time. The first argument specifies
4834 what is wanted and the second points to where the information is to be placed.
4835 The current list of available information is:
4836
4837 PCRE_CONFIG_UTF8
4838
4839 The output is an integer that is set to one if UTF-8 support is available;
4840 otherwise it is set to zero.
4841
4842 PCRE_CONFIG_NEWLINE
4843
4844 The output is an integer that it set to the value of the code that is used for
4845 newline. It is either LF (10) or CR (13).
4846
4847 PCRE_CONFIG_LINK_SIZE
4848
4849 The output is an integer that contains the number of bytes used for internal
4850 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4851
4852 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4853
4854 The output is an integer that contains the threshold above which the POSIX
4855 interface uses malloc() for output vectors. See item 31 above.
4856
4857 PCRE_CONFIG_MATCH_LIMIT
4858
4859 The output is an unsigned integer that contains the default limit of the number
4860 of match() calls in a pcre_exec() execution. See 42 above.
4861
4862 44. pcretest has been upgraded by the addition of the -C option. This causes it
4863 to extract all the available output from the new pcre_config() function, and to
4864 output it. The program then exits immediately.
4865
4866 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4867 order to support additional features. One way would have been to define
4868 pcre_exec2() (for example) with extra arguments, but this would not have been
4869 extensible, and would also have required all calls to the original function to
4870 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4871 is used for passing in "extra" data from pcre_study().
4872
4873 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4874 contains the following fields:
4875
4876 flags a bitmap indicating which of the following fields are set
4877 study_data opaque data from pcre_study()
4878 match_limit a way of specifying a limit on match() calls for a specific
4879 call to pcre_exec()
4880 callout_data data for callouts (see 49 below)
4881
4882 The flag bits are also defined in pcre.h, and are
4883
4884 PCRE_EXTRA_STUDY_DATA
4885 PCRE_EXTRA_MATCH_LIMIT
4886 PCRE_EXTRA_CALLOUT_DATA
4887
4888 The pcre_study() function now returns one of these new pcre_extra blocks, with
4889 the actual study data pointed to by the study_data field, and the
4890 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4891 before. That is, this change is entirely upwards-compatible and requires no
4892 change to existing code.
4893
4894 If you want to pass in additional data to pcre_exec(), you can either place it
4895 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4896 block.
4897
4898 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4899 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4900 times with different match limits, until it finds the minimum value needed for
4901 pcre_exec() to complete. The value is then output. This can be instructive; for
4902 most simple matches the number is quite small, but for pathological cases it
4903 gets very large very quickly.
4904
4905 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4906 returns the size of the data block pointed to by the study_data field in a
4907 pcre_extra block, that is, the value that was passed as the argument to
4908 pcre_malloc() when PCRE was getting memory in which to place the information
4909 created by pcre_study(). The fourth argument should point to a size_t variable.
4910 pcretest has been extended so that this information is shown after a successful
4911 pcre_study() call when information about the compiled regex is being displayed.
4912
4913 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4914 because what follows is always an absolute path. (Later: it turns out that this
4915 is more than cosmetic for MinGW, because it doesn't like empty path
4916 components.)
4917
4918 49. Some changes have been made to the callout feature (see 28 above):
4919
4920 (i) A callout function now has three choices for what it returns:
4921
4922 0 => success, carry on matching
4923 > 0 => failure at this point, but backtrack if possible
4924 < 0 => serious error, return this value from pcre_exec()
4925
4926 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4927 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4928 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4929 use by callout functions. It will never be used by PCRE itself.
4930
4931 (ii) The pcre_extra structure (see 45 above) has a void * field called
4932 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4933 pcre_callout_block structure has a field of the same name. The contents of
4934 the field passed in the pcre_extra structure are passed to the callout
4935 function in the corresponding field in the callout block. This makes it
4936 easier to use the same callout-containing regex from multiple threads. For
4937 testing, the pcretest program has a new data escape
4938
4939 \C*n pass the number n (may be negative) as callout_data
4940
4941 If the callout function in pcretest receives a non-zero value as
4942 callout_data, it returns that value.
4943
4944 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4945 there were some redundant $(CFLAGS) in commands that are now specified as
4946 $(LINK), which already includes $(CFLAGS).
4947
4948 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4949 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4950 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4951 one-byte characters throughout. Note that case-insensitive matching applies
4952 only to characters whose values are less than 256. PCRE doesn't support the
4953 notion of cases for higher-valued characters.
4954
4955 (i) A character class whose characters are all within 0-255 is handled as
4956 a bit map, and the map is inverted for negative classes. Previously, a
4957 character > 255 always failed to match such a class; however it should
4958 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4959
4960 (ii) A negated character class with a single character < 255 is coded as
4961 "not this character" (OP_NOT). This wasn't working properly when the test
4962 character was multibyte, either singly or repeated.
4963
4964 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4965 mode, for example: \x{100}{2,3}.
4966
4967 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4968 singly or repeated) now correctly test multibyte characters. However,
4969 PCRE doesn't recognize any characters with values greater than 255 as
4970 digits, spaces, or word characters. Such characters always match \D, \S,
4971 and \W, and never match \d, \s, or \w.
4972
4973 (v) Classes may now contain characters and character ranges with values
4974 greater than 255. For example: [ab\x{100}-\x{400}].
4975
4976 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4977 PCRE in UTF-8 mode.
4978
4979 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4980 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4981 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4982 value.)
4983
4984 53. The single man page has become too large. I have therefore split it up into
4985 a number of separate man pages. These also give rise to individual HTML pages;
4986 these are now put in a separate directory, and there is an index.html page that
4987 lists them all. Some hyperlinking between the pages has been installed.
4988
4989 54. Added convenience functions for handling named capturing parentheses.
4990
4991 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4992 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4993 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4994 are faulted.
4995
4996 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4997 calling configure. These values are used when compiling the dftables.c program
4998 which is run to generate the source of the default character tables. They
4999 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5000 you will need to set these values.
5001
5002 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5003
5004
5005 Version 3.9 02-Jan-02
5006 ---------------------
5007
5008 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5009
5010 2. If --disable-static was given, the building process failed when trying to
5011 build pcretest and pcregrep. (For some reason it was using libtool to compile
5012 them, which is not right, as they aren't part of the library.)
5013
5014
5015 Version 3.8 18-Dec-01
5016 ---------------------
5017
5018 1. The experimental UTF-8 code was completely screwed up. It was packing the
5019 bytes in the wrong order. How dumb can you get?
5020
5021
5022 Version 3.7 29-Oct-01
5023 ---------------------
5024
5025 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5026 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5027 this didn't happen under Solaris 8, where I normally test things.
5028
5029 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5030 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5031 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5032 this fix an hour or so after the initial 3.7 release.)
5033
5034
5035 Version 3.6 23-Oct-01
5036 ---------------------
5037
5038 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5039 offsets passed as NULL with zero offset count.
5040
5041 2. The config.guess and config.sub files had not been updated when I moved to
5042 the latest autoconf.
5043
5044
5045 Version 3.5 15-Aug-01
5046 ---------------------
5047
5048 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5049 had been forgotten.
5050
5051 2. By using declared but undefined structures, we can avoid using "void"
5052 definitions in pcre.h while keeping the internal definitions of the structures
5053 private.
5054
5055 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5056 user point of view, this means that both static and shared libraries are built
5057 by default, but this can be individually controlled. More of the work of
5058 handling this static/shared cases is now inside libtool instead of PCRE's make
5059 file.
5060
5061 4. The pcretest utility is now installed along with pcregrep because it is
5062 useful for users (to test regexs) and by doing this, it automatically gets
5063 relinked by libtool. The documentation has been turned into a man page, so
5064 there are now .1, .txt, and .html versions in /doc.
5065
5066 5. Upgrades to pcregrep:
5067 (i) Added long-form option names like gnu grep.
5068 (ii) Added --help to list all options with an explanatory phrase.
5069 (iii) Added -r, --recursive to recurse into sub-directories.
5070 (iv) Added -f, --file to read patterns from a file.
5071
5072 6. pcre_exec() was referring to its "code" argument before testing that
5073 argument for NULL (and giving an error if it was NULL).
5074
5075 7. Upgraded Makefile.in to allow for compiling in a different directory from
5076 the source directory.
5077
5078 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5079 options bits, the pointer it was passed was to an int instead of to an unsigned
5080 long int. This mattered only on 64-bit systems.
5081
5082 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5083 generated) instead of pcre.in, which it its source. Also made the same change
5084 in several of the .c files.
5085
5086 10. A new release of gcc defines printf() as a macro, which broke pcretest
5087 because it had an ifdef in the middle of a string argument for printf(). Fixed
5088 by using separate calls to printf().
5089
5090 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5091 script, to force use of CR or LF instead of \n in the source. On non-Unix
5092 systems, the value can be set in config.h.
5093
5094 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5095 absolute limit. Changed the text of the error message to make this clear, and
5096 likewise updated the man page.
5097
5098 13. The limit of 99 on the number of capturing subpatterns has been removed.
5099 The new limit is 65535, which I hope will not be a "real" limit.
5100
5101
5102 Version 3.4 22-Aug-00
5103 ---------------------
5104
5105 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5106
5107 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5108
5109
5110 Version 3.3 01-Aug-00
5111 ---------------------
5112
5113 1. If an octal character was given, but the value was greater than \377, it
5114 was not getting masked to the least significant bits, as documented. This could
5115 lead to crashes in some systems.
5116
5117 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5118 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5119
5120 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5121 These just pass their arguments on to (pcre_free)(), but they are provided
5122 because some uses of PCRE bind it to non-C systems that can call its functions,
5123 but cannot call free() or pcre_free() directly.
5124
5125 4. Add "make test" as a synonym for "make check". Corrected some comments in
5126 the Makefile.
5127
5128 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5129 Makefile.
5130
5131 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5132 command called pgrep for grepping around the active processes.
5133
5134 7. Added the beginnings of support for UTF-8 character strings.
5135
5136 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5137 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5138 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5139 out for the ar command.)
5140
5141
5142 Version 3.2 12-May-00
5143 ---------------------
5144
5145 This is purely a bug fixing release.
5146
5147 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5148 of ZA. This was just one example of several cases that could provoke this bug,
5149 which was introduced by change 9 of version 2.00. The code for breaking
5150 infinite loops after an iteration that matches an empty string was't working
5151 correctly.
5152
5153 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5154 when matched against abbab (for example). After matching an empty string, it
5155 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5156 caused it to match further down the string than it should.
5157
5158 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5159 was there because it doesn't seem to be needed, and it causes trouble on some
5160 systems, as it is not a Standard C header. It has been removed.
5161
5162 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5163 were reported on the Macintosh. The changes were from
5164
5165 while ((c = *(++ptr)) != 0 && c != '\n');
5166 to
5167 while ((c = *(++ptr)) != 0 && c != '\n') ;
5168
5169 Totally extraordinary, but if that's what it takes...
5170
5171 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5172 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5173 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5174 assumes the way PCRE uses memmove() (always moving upwards).
5175
5176 6. PCRE is being used in one environment where strchr() is not available. There
5177 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5178 faster code anyway.
5179
5180
5181 Version 3.1 09-Feb-00
5182 ---------------------
5183
5184 The only change in this release is the fixing of some bugs in Makefile.in for
5185 the "install" target:
5186
5187 (1) It was failing to install pcreposix.h.
5188
5189 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5190
5191
5192 Version 3.0 01-Feb-00
5193 ---------------------
5194
5195 1. Add support for the /+ modifier to perltest (to output $` like it does in
5196 pcretest).
5197
5198 2. Add support for the /g modifier to perltest.
5199
5200 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5201 matches null strings.
5202
5203 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5204 pattern. Perl treats empty patterns specially - it reuses the most recent
5205 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5206 effect.
5207
5208 5. The POSIX interface was broken in that it was just handing over the POSIX
5209 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5210 required a bigger vector, with some working space on the end. This means that
5211 the POSIX wrapper now has to get and free some memory, and copy the results.
5212
5213 6. Added some simple autoconf support, placing the test data and the
5214 documentation in separate directories, re-organizing some of the
5215 information files, and making it build pcre-config (a GNU standard). Also added
5216 libtool support for building PCRE as a shared library, which is now the
5217 default.
5218
5219 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5220 09 are not valid octal constants. Single digits will be used for minor values
5221 less than 10.
5222
5223 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5224 existing programs that set these in the POSIX interface can use PCRE without
5225 modification.
5226
5227 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5228 return all that pcre_info() returns, plus additional data. The pcre_info()
5229 function is retained for compatibility, but is considered to be obsolete.
5230
5231 10. Added experimental recursion feature (?R) to handle one common case that
5232 Perl 5.6 will be able to do with (?p{...}).
5233
5234 11. Added support for POSIX character classes like [:alpha:], which Perl is
5235 adopting.
5236
5237
5238 Version 2.08 31-Aug-99
5239 ----------------------
5240
5241 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5242 trying to match at the startoffset position, but instead was moving forward to
5243 the next newline as if a previous match had failed.
5244
5245 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5246 and could get into a loop if a null string was matched other than at the start
5247 of the subject.
5248
5249 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5250 be distinguished at compile time, and for completeness also added PCRE_DATE.
5251
5252 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5253 in GnuWin32 environments.
5254
5255
5256 Version 2.07 29-Jul-99
5257 ----------------------
5258
5259 1. The documentation is now supplied in plain text form and HTML as well as in
5260 the form of man page sources.
5261
5262 2. C++ compilers don't like assigning (void *) values to other pointer types.
5263 In particular this affects malloc(). Although there is no problem in Standard
5264 C, I've put in casts to keep C++ compilers happy.
5265
5266 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5267 should be (const char *).
5268
5269 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5270 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5271 However, I haven't made this a standard facility. The documentation doesn't
5272 mention it, and the Makefile doesn't support it.
5273
5274 5. The Makefile now contains an "install" target, with editable destinations at
5275 the top of the file. The pcretest program is not installed.
5276
5277 6. pgrep -V now gives the PCRE version number and date.
5278
5279 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5280 causing the entire string to be ignored, instead of just the last character.
5281
5282 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5283 non-matching string, it can take a very, very long time, even for strings of
5284 quite modest length, because of the nested recursion. PCRE now does better in
5285 some of these cases. It does this by remembering the last required literal
5286 character in the pattern, and pre-searching the subject to ensure it is present
5287 before running the real match. In other words, it applies a heuristic to detect
5288 some types of certain failure quickly, and in the above example, if presented
5289 with a string that has no trailing " it gives "no match" very quickly.
5290
5291 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5292 other alternatives are tried instead.
5293
5294
5295 Version 2.06 09-Jun-99
5296 ----------------------
5297
5298 1. Change pcretest's output for amount of store used to show just the code
5299 space, because the remainder (the data block) varies in size between 32-bit and
5300 64-bit systems.
5301
5302 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5303 start matching at. This allows lookbehinds to work when searching for multiple
5304 occurrences in a string.
5305
5306 3. Added additional options to pcretest for testing multiple occurrences:
5307
5308 /+ outputs the rest of the string that follows a match
5309 /g loops for multiple occurrences, using the new startoffset argument
5310 /G loops for multiple occurrences by passing an incremented pointer
5311
5312 4. PCRE wasn't doing the "first character" optimization for patterns starting
5313 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5314 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5315 the letter 'x'. On long subject strings, this gives a significant speed-up.
5316
5317
5318 Version 2.05 21-Apr-99
5319 ----------------------
5320
5321 1. Changed the type of magic_number from int to long int so that it works
5322 properly on 16-bit systems.
5323
5324 2. Fixed a bug which caused patterns starting with .* not to work correctly
5325 when the subject string contained newline characters. PCRE was assuming
5326 anchoring for such patterns in all cases, which is not correct because .* will
5327 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5328 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5329 must be retried after every newline in the subject.
5330
5331
5332 Version 2.04 18-Feb-99
5333 ----------------------
5334
5335 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5336 computation of the store needed to hold the pattern was incorrect (too large).
5337 If such patterns were nested a few deep, this could multiply and become a real
5338 problem.
5339
5340 2. Added /M option to pcretest to show the memory requirement of a specific
5341 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5342
5343 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5344 compiled in such a way that the backtracking after subsequent failure was
5345 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5346 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5347
5348
5349 Version 2.03 02-Feb-99
5350 ----------------------
5351
5352 1. Fixed typo and small mistake in man page.
5353
5354 2. Added 4th condition (GPL supersedes if conflict) and created separate
5355 LICENCE file containing the conditions.
5356
5357 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5358 Perl, that is the internal \ allows the delimiter to be included in the
5359 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5360 the final delimiter, add \ to the end of the pattern (to test the error).
5361
5362 4. Added the convenience functions for extracting substrings after a successful
5363 match. Updated pcretest to make it able to test these functions.
5364
5365
5366 Version 2.02 14-Jan-99
5367 ----------------------
5368
5369 1. Initialized the working variables associated with each extraction so that
5370 their saving and restoring doesn't refer to uninitialized store.
5371
5372 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5373 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5374 fix the problem.
5375
5376 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5377 calls, and wasn't printing the correct value for compiling calls. Increased the
5378 default value of LOOPREPEAT, and the number of significant figures in the
5379 times.
5380
5381 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5382
5383 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5384 a building problem on Windows NT with a FAT file system.
5385
5386
5387 Version 2.01 21-Oct-98
5388 ----------------------
5389
5390 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5391 to character tables built by pcre_maketables() in the current locale. If NULL
5392 is passed, the default tables are used.
5393
5394
5395 Version 2.00 24-Sep-98
5396 ----------------------
5397
5398 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5399 it any more.
5400
5401 2. Allow quantification of (?>) groups, and make it work correctly.
5402
5403 3. The first character computation wasn't working for (?>) groups.
5404
5405 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5406 end of the subject) and add 5.005's \z, which really does match only at the
5407 very end of the subject.
5408
5409 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5410
5411 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5412 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5413 localized options. All options to pcre_study() were also removed.
5414
5415 7. Add other new features from 5.005:
5416
5417 $(?<= positive lookbehind
5418 $(?<! negative lookbehind
5419 (?imsx-imsx) added the unsetting capability
5420 such a setting is global if at outer level; local otherwise
5421 (?imsx-imsx:) non-capturing groups with option setting
5422 (?(cond)re|re) conditional pattern matching
5423
5424 A backreference to itself in a repeated group matches the previous
5425 captured string.
5426
5427 8. General tidying up of studying (both automatic and via "study")
5428 consequential on the addition of new assertions.
5429
5430 9. As in 5.005, unlimited repeated groups that could match an empty substring
5431 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5432 runtime if any iteration does actually match an empty substring.
5433
5434 10. Include the RunTest script in the distribution.
5435
5436 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5437 discrepancies, some of which were old and were also with respect to 5.004. They
5438 have now been fixed.
5439
5440
5441 Version 1.09 28-Apr-98
5442 ----------------------
5443
5444 1. A negated single character class followed by a quantifier with a minimum
5445 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5446 program crashes, or just wrong answers. This did not apply to negated classes
5447 containing more than one character, or to minima other than one.
5448
5449
5450 Version 1.08 27-Mar-98
5451 ----------------------
5452
5453 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5454
5455 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5456 latter must appear before anything that relies on it in the pattern.
5457
5458
5459 Version 1.07 16-Feb-98
5460 ----------------------
5461
5462 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5463 repeat of a potentially empty string).
5464
5465
5466 Version 1.06 23-Jan-98
5467 ----------------------
5468
5469 1. Added Markus Oberhumer's little patches for C++.
5470
5471 2. Literal strings longer than 255 characters were broken.
5472
5473
5474 Version 1.05 23-Dec-97
5475 ----------------------
5476
5477 1. Negated character classes containing more than one character were failing if
5478 PCRE_CASELESS was set at run time.
5479
5480
5481 Version 1.04 19-Dec-97
5482 ----------------------
5483
5484 1. Corrected the man page, where some "const" qualifiers had been omitted.
5485
5486 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5487 input syntax.
5488
5489 3. Fixed memory leak which occurred when a regex with back references was
5490 matched with an offsets vector that wasn't big enough. The temporary memory
5491 that is used in this case wasn't being freed if the match failed.
5492
5493 4. Tidied pcretest to ensure it frees memory that it gets.
5494
5495 5. Temporary memory was being obtained in the case where the passed offsets
5496 vector was exactly big enough.
5497
5498 6. Corrected definition of offsetof() from change 5 below.
5499
5500 7. I had screwed up change 6 below and broken the rules for the use of
5501 setjmp(). Now fixed.
5502
5503
5504 Version 1.03 18-Dec-97
5505 ----------------------
5506
5507 1. A erroneous regex with a missing opening parenthesis was correctly
5508 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5509 on some systems.
5510
5511 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5512 it was reported that one broken compiler failed on the former because "code" is
5513 also an independent variable.
5514
5515 3. The erroneous regex a[]b caused an array overrun reference.
5516
5517 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5518 fail on data ending with that character. (It was going on too far, and checking
5519 the next character, typically a binary zero.) This was specific to the
5520 optimized code for single-character negative classes.
5521
5522 5. Added a contributed patch from the TIN world which does the following:
5523
5524 + Add an undef for memmove, in case the the system defines a macro for it.
5525
5526 + Add a definition of offsetof(), in case there isn't one. (I don't know
5527 the reason behind this - offsetof() is part of the ANSI standard - but
5528 it does no harm).
5529
5530 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5531 most of the places where whitespace preceded '#'. I have given up and
5532 allowed the remaining 2 cases to be at the margin.
5533
5534 + Rename some variables in pcre to eliminate shadowing. This seems very
5535 pedantic, but does no harm, of course.
5536
5537 6. Moved the call to setjmp() into its own function, to get rid of warnings
5538 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5539
5540 7. Constructs such as \d{8,} were compiling into the equivalent of
5541 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5542 outcome, but in this particular case used more store than had been allocated,
5543 which caused the bug to be discovered because it threw up an internal error.
5544
5545 8. The debugging code in both pcre and pcretest for outputting the compiled
5546 form of a regex was going wrong in the case of back references followed by
5547 curly-bracketed repeats.
5548
5549
5550 Version 1.02 12-Dec-97
5551 ----------------------
5552
5553 1. Typos in pcre.3 and comments in the source fixed.
5554
5555 2. Applied a contributed patch to get rid of places where it used to remove
5556 'const' from variables, and fixed some signed/unsigned and uninitialized
5557 variable warnings.
5558
5559 3. Added the "runtest" target to Makefile.
5560
5561 4. Set default compiler flag to -O2 rather than just -O.
5562
5563
5564 Version 1.01 19-Nov-97
5565 ----------------------
5566
5567 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5568 like /([ab]*)*/, that is, for classes with more than one character in them.
5569
5570 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5571 as /((?>a*))*/ (a PCRE_EXTRA facility).
5572
5573
5574 Version 1.00 18-Nov-97
5575 ----------------------
5576
5577 1. Added compile-time macros to support systems such as SunOS4 which don't have
5578 memmove() or strerror() but have other things that can be used instead.
5579
5580 2. Arranged that "make clean" removes the executables.
5581
5582
5583 Version 0.99 27-Oct-97
5584 ----------------------
5585
5586 1. Fixed bug in code for optimizing classes with only one character. It was
5587 initializing a 32-byte map regardless, which could cause it to run off the end
5588 of the memory it had got.
5589
5590 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5591
5592
5593 Version 0.98 22-Oct-97
5594 ----------------------
5595
5596 1. Fixed bug in code for handling temporary memory usage when there are more
5597 back references than supplied space in the ovector. This could cause segfaults.
5598
5599
5600 Version 0.97 21-Oct-97
5601 ----------------------
5602
5603 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5604
5605 2. Optimized negated single characters not to use a bit map.
5606
5607 3. Brought error texts together as macro definitions; clarified some of them;
5608 fixed one that was wrong - it said "range out of order" when it meant "invalid
5609 escape sequence".
5610
5611 4. Changed some char * arguments to const char *.
5612
5613 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5614
5615 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5616 pcretest.
5617
5618
5619 Version 0.96 16-Oct-97
5620 ----------------------
5621
5622 1. Added a simple "pgrep" utility to the distribution.
5623
5624 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5625 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5626 where "ddd" means "one or more decimal digits".
5627
5628 3. Fixed serious bug. If a pattern had a back reference, but the call to
5629 pcre_exec() didn't supply a large enough ovector to record the related
5630 identifying subpattern, the match always failed. PCRE now remembers the number
5631 of the largest back reference, and gets some temporary memory in which to save
5632 the offsets during matching if necessary, in order to ensure that
5633 backreferences always work.
5634
5635 4. Increased the compatibility with Perl in a number of ways:
5636
5637 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5638 to request this handling. The option can be set at compile or exec time.
5639
5640 (b) $ matches before a terminating newline by default; an option
5641 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5642 mode). The option can be set at compile or exec time.
5643
5644 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5645 the same as Perl's. If the decimal number it represents is less than 10
5646 or there aren't that many previous left capturing parentheses, an octal
5647 escape is read. Inside a character class, it's always an octal escape,
5648 even if it is a single digit.
5649
5650 (d) An escaped but undefined alphabetic character is taken as a literal,
5651 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5652 escapes.
5653
5654 (e) {0} is now permitted. (The previous item is removed from the compiled
5655 pattern).
5656
5657 5. Changed all the names of code files so that the basic parts are no longer
5658 than 10 characters, and abolished the teeny "globals.c" file.
5659
5660 6. Changed the handling of character classes; they are now done with a 32-byte
5661 bit map always.
5662
5663 7. Added the -d and /D options to pcretest to make it possible to look at the
5664 internals of compilation without having to recompile pcre.
5665
5666
5667 Version 0.95 23-Sep-97
5668 ----------------------
5669
5670 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5671 \x20 at the start of a run of normal characters. These were being treated as
5672 real characters, instead of the source characters being re-checked.
5673
5674
5675 Version 0.94 18-Sep-97
5676 ----------------------
5677
5678 1. The functions are now thread-safe, with the caveat that the global variables
5679 containing pointers to malloc() and free() or alternative functions are the
5680 same for all threads.
5681
5682 2. Get pcre_study() to generate a bitmap of initial characters for non-
5683 anchored patterns when this is possible, and use it if passed to pcre_exec().
5684
5685
5686 Version 0.93 15-Sep-97
5687 ----------------------
5688
5689 1. /(b)|(:+)/ was computing an incorrect first character.
5690
5691 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5692 but not actually doing anything yet.
5693
5694 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5695 as Perl does (e.g. [-az] or [az-]).
5696
5697 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5698 all possible positions.
5699
5700 5. Split up into different modules to avoid including unneeded functions in a
5701 compiled binary. However, compile and exec are still in one module. The "study"
5702 function is split off.
5703
5704 6. The character tables are now in a separate module whose source is generated
5705 by an auxiliary program - but can then be edited by hand if required. There are
5706 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5707 toupper() in the code.
5708
5709 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5710 make them global. Abolish the function for setting them, as the caller can now
5711 set them directly.
5712
5713
5714 Version 0.92 11-Sep-97
5715 ----------------------
5716
5717 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5718 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5719
5720 2. Caseless matching was not working in character classes if the characters in
5721 the pattern were in upper case.
5722
5723 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5724
5725 4. Make PCRE_ANCHORED public and accept as a compile option.
5726
5727 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5728 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5729 pass them.
5730
5731 6. Give an error if bad option bits passed at compile or run time.
5732
5733 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5734 pcretest to cause it to pass that flag.
5735
5736 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5737 options, and the first character, if set.
5738
5739 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5740
5741
5742 Version 0.91 10-Sep-97
5743 ----------------------
5744
5745 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5746 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5747
5748 2. PCRE was looping on encountering an indefinitely repeated back reference to
5749 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5750 Perl does - treats the match as successful.
5751
5752 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5