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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1628 - (show annotations)
Tue Feb 9 09:46:52 2016 UTC (18 hours, 55 minutes ago) by ph10
File size: 287661 byte(s)
Fix typo.
1 ChangeLog for PCRE
2 ------------------
3
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
6
7 Version 8.39 xx-xxxxxx-201x
8 ---------------------------
9
10 1. If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11 an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12 misbehaved. This bug was found by the LLVM fuzzer.
13
14 2. Similar to the above, if an isolated \E was present between an item and its
15 qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16 bug was found by the LLVM fuzzer.
17
18 3. Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19 working correctly in UCP mode.
20
21 4. The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22 was set when the pmatch argument was NULL. It now returns REG_INVARG.
23
24 5. Allow for up to 32-bit numbers in the ordin() function in pcregrep.
25
26 6. An empty \Q\E sequence between an item and its qualifier caused
27 pcre_compile() to misbehave when auto callouts were enabled. This bug was
28 found by the LLVM fuzzer.
29
30 7. If a pattern that was compiled with PCRE_EXTENDED started with white
31 space or a #-type comment that was followed by (?-x), which turns off
32 PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
33 pcre_compile() assumed that (?-x) applied to the whole pattern and
34 consequently mis-compiled it. This bug was found by the LLVM fuzzer.
35
36 8. An call of pcre_copy_named_substring() for a named substring whose number
37 was greater than the space in the ovector could cause a crash.
38
39 9. Yet another buffer overflow bug involved duplicate named groups with a
40 group that reset capture numbers (compare 8.38/7 below). Once again, I have
41 just allowed for more memory, even if not needed. (A proper fix is
42 implemented in PCRE2, but it involves a lot of refactoring.)
43
44 10. pcre_get_substring_list() crashed if the use of \K in a match caused the
45 start of the match to be earlier than the end.
46
47 11. Migrating appropriate PCRE2 JIT improvements to PCRE.
48
49 12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
50 assertion, caused pcretest to generate incorrect output, and also to read
51 uninitialized memory (detected by ASAN or valgrind).
52
53
54 Version 8.38 23-November-2015
55 -----------------------------
56
57 1. If a group that contained a recursive back reference also contained a
58 forward reference subroutine call followed by a non-forward-reference
59 subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
60 compile correct code, leading to undefined behaviour or an internally
61 detected error. This bug was discovered by the LLVM fuzzer.
62
63 2. Quantification of certain items (e.g. atomic back references) could cause
64 incorrect code to be compiled when recursive forward references were
65 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
66 This bug was discovered by the LLVM fuzzer.
67
68 3. A repeated conditional group whose condition was a reference by name caused
69 a buffer overflow if there was more than one group with the given name.
70 This bug was discovered by the LLVM fuzzer.
71
72 4. A recursive back reference by name within a group that had the same name as
73 another group caused a buffer overflow. For example:
74 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
75
76 5. A forward reference by name to a group whose number is the same as the
77 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
78 a buffer overflow at compile time. This bug was discovered by the LLVM
79 fuzzer.
80
81 6. A lookbehind assertion within a set of mutually recursive subpatterns could
82 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
83
84 7. Another buffer overflow bug involved duplicate named groups with a
85 reference between their definition, with a group that reset capture
86 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
87 fixed by always allowing for more memory, even if not needed. (A proper fix
88 is implemented in PCRE2, but it involves more refactoring.)
89
90 8. There was no check for integer overflow in subroutine calls such as (?123).
91
92 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
93 being treated as a literal 'l' instead of causing an error.
94
95 10. There was a buffer overflow if pcre_exec() was called with an ovector of
96 size 1. This bug was found by american fuzzy lop.
97
98 11. If a non-capturing group containing a conditional group that could match
99 an empty string was repeated, it was not identified as matching an empty
100 string itself. For example: /^(?:(?(1)x|)+)+$()/.
101
102 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
103 \a and \e in test subject lines.
104
105 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
106 instead of the EBCDIC value.
107
108 14. The handling of \c in an EBCDIC environment has been revised so that it is
109 now compatible with the specification in Perl's perlebcdic page.
110
111 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
112 ASCII/Unicode. This has now been added to the list of characters that are
113 recognized as white space in EBCDIC.
114
115 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
116 error (correctly) when used outside a class, but did not give an error
117 within a class.
118
119 17. \h within a class was incorrectly compiled in EBCDIC environments.
120
121 18. A pattern with an unmatched closing parenthesis that contained a backward
122 assertion which itself contained a forward reference caused buffer
123 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
124
125 19. JIT should return with error when the compiled pattern requires more stack
126 space than the maximum.
127
128 20. A possessively repeated conditional group that could match an empty string,
129 for example, /(?(R))*+/, was incorrectly compiled.
130
131 21. Fix infinite recursion in the JIT compiler when certain patterns such as
132 /(?:|a|){100}x/ are analysed.
133
134 22. Some patterns with character classes involving [: and \\ were incorrectly
135 compiled and could cause reading from uninitialized memory or an incorrect
136 error diagnosis.
137
138 23. Pathological patterns containing many nested occurrences of [: caused
139 pcre_compile() to run for a very long time.
140
141 24. A conditional group with only one branch has an implicit empty alternative
142 branch and must therefore be treated as potentially matching an empty
143 string.
144
145 25. If (?R was followed by - or + incorrect behaviour happened instead of a
146 diagnostic.
147
148 26. Arrange to give up on finding the minimum matching length for overly
149 complex patterns.
150
151 27. Similar to (4) above: in a pattern with duplicated named groups and an
152 occurrence of (?| it is possible for an apparently non-recursive back
153 reference to become recursive if a later named group with the relevant
154 number is encountered. This could lead to a buffer overflow. Wen Guanxing
155 from Venustech ADLAB discovered this bug.
156
157 28. If pcregrep was given the -q option with -c or -l, or when handling a
158 binary file, it incorrectly wrote output to stdout.
159
160 29. The JIT compiler did not restore the control verb head in case of *THEN
161 control verbs. This issue was found by Karl Skomski with a custom LLVM
162 fuzzer.
163
164 30. Error messages for syntax errors following \g and \k were giving inaccurate
165 offsets in the pattern.
166
167 31. Added a check for integer overflow in conditions (?(<digits>) and
168 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
169 fuzzer.
170
171 32. Handling recursive references such as (?2) when the reference is to a group
172 later in the pattern uses code that is very hacked about and error-prone.
173 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
174 give an internal error if it is obvious that compiling has gone wrong.
175
176 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
177 This issue was found by Karl Skomski with a custom LLVM fuzzer.
178
179 34. The JIT compiler should restore the control chain for empty possessive
180 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
181
182 35. Match limit check added to JIT recursion. This issue was found by Karl
183 Skomski with a custom LLVM fuzzer.
184
185 36. Yet another case similar to 27 above has been circumvented by an
186 unconditional allocation of extra memory. This issue is fixed "properly" in
187 PCRE2 by refactoring the way references are handled. Wen Guanxing
188 from Venustech ADLAB discovered this bug.
189
190 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
191 with a custom LLVM fuzzer.
192
193 38. Fixed a corner case of range optimization in JIT.
194
195 39. An incorrect error "overran compiling workspace" was given if there were
196 exactly enough group forward references such that the last one extended
197 into the workspace safety margin. The next one would have expanded the
198 workspace. The test for overflow was not including the safety margin.
199
200 40. A match limit issue is fixed in JIT which was found by Karl Skomski
201 with a custom LLVM fuzzer.
202
203 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
204 work under Windows. (Why has it taken so long for anyone to notice?)
205
206 42. In a character class such as [\W\p{Any}] where both a negative-type escape
207 ("not a word character") and a property escape were present, the property
208 escape was being ignored.
209
210 43. Fix crash caused by very long (*MARK) or (*THEN) names.
211
212 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
213 by a single ASCII character in a class item, was incorrectly compiled in
214 UCP mode. The POSIX class got lost, but only if the single character
215 followed it.
216
217 45. [:punct:] in UCP mode was matching some characters in the range 128-255
218 that should not have been matched.
219
220 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
221 class, all characters with code points greater than 255 are in the class.
222 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
223 such as \w are turned into Unicode properties), wide characters were not
224 correctly handled, and could fail to match.
225
226
227 Version 8.37 28-April-2015
228 --------------------------
229
230 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
231 for those parentheses to be closed with whatever has been captured so far.
232 However, it was failing to mark any other groups between the hightest
233 capture so far and the currrent group as "unset". Thus, the ovector for
234 those groups contained whatever was previously there. An example is the
235 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
236
237 2. If an assertion condition was quantified with a minimum of zero (an odd
238 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
239
240 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
241 unrecognized modifier, a crash could occur.
242
243 4. An attempt to do global matching in pcretest with a zero-length ovector
244 caused a crash.
245
246 5. Fixed a memory leak during matching that could occur for a subpattern
247 subroutine call (recursive or otherwise) if the number of captured groups
248 that had to be saved was greater than ten.
249
250 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
251 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
252 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
253
254 7. A UTF pattern containing a "not" match of a non-ASCII character and a
255 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
256
257 8. When a pattern is compiled, it remembers the highest back reference so that
258 when matching, if the ovector is too small, extra memory can be obtained to
259 use instead. A conditional subpattern whose condition is a check on a
260 capture having happened, such as, for example in the pattern
261 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
262 setting the highest backreference number. This mattered only if pcre_exec()
263 was called with an ovector that was too small to hold the capture, and there
264 was no other kind of back reference (a situation which is probably quite
265 rare). The effect of the bug was that the condition was always treated as
266 FALSE when the capture could not be consulted, leading to a incorrect
267 behaviour by pcre_exec(). This bug has been fixed.
268
269 9. A reference to a duplicated named group (either a back reference or a test
270 for being set in a conditional) that occurred in a part of the pattern where
271 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
272 to be incorrectly calculated, leading to overwriting.
273
274 10. A mutually recursive set of back references such as (\2)(\1) caused a
275 segfault at study time (while trying to find the minimum matching length).
276 The infinite loop is now broken (with the minimum length unset, that is,
277 zero).
278
279 11. If an assertion that was used as a condition was quantified with a minimum
280 of zero, matching went wrong. In particular, if the whole group had
281 unlimited repetition and could match an empty string, a segfault was
282 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
283 assertions to be quantified, but not if they are being used as conditions,
284 so the above pattern is faulted by Perl. PCRE has now been changed so that
285 it also rejects such patterns.
286
287 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
288 failed to allow the zero-repeat case if pcre2_exec() was called with an
289 ovector too small to capture the group.
290
291 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
292 Red Hat Product Security:
293
294 (a) A crash if /K and /F were both set with the option to save the compiled
295 pattern.
296
297 (b) Another crash if the option to print captured substrings in a callout
298 was combined with setting a null ovector, for example \O\C+ as a subject
299 string.
300
301 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
302 forward reference repeated a large (but limited) number of times within a
303 repeated outer group that has a zero minimum quantifier, caused incorrect
304 code to be compiled, leading to the error "internal error:
305 previously-checked referenced subpattern not found" when an incorrect
306 memory address was read. This bug was reported as "heap overflow",
307 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
308 CVE-2015-2325.
309
310 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
311 call within a group that also contained a recursive back reference caused
312 incorrect code to be compiled. This bug was reported as "heap overflow",
313 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
314 number CVE-2015-2326.
315
316 24. Computing the size of the JIT read-only data in advance has been a source
317 of various issues, and new ones are still appear unfortunately. To fix
318 existing and future issues, size computation is eliminated from the code,
319 and replaced by on-demand memory allocation.
320
321 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
322 adjacent to the end of the range, and the range contained characters with
323 more than one other case, caused incorrect behaviour when compiled in UTF
324 mode. In that example, the range a-j was left out of the class.
325
326 26. Fix JIT compilation of conditional blocks, which assertion
327 is converted to (*FAIL). E.g: /(?(?!))/.
328
329 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
330 discovered by the LLVM fuzzer.
331
332 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
333 when this assertion was used as a condition, for example (?(?!)a|b). In
334 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
335 error about an unsupported item.
336
337 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
338 possessification code could take exponential time to complete. A recursion
339 depth limit of 1000 has been imposed to limit the resources used by this
340 optimization.
341
342 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
343 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
344 because \S ensures they are all in the class. The code for doing this was
345 interacting badly with the code for computing the amount of space needed to
346 compile the pattern, leading to a buffer overflow. This bug was discovered
347 by the LLVM fuzzer.
348
349 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
350 other kinds of group caused stack overflow at compile time. This bug was
351 discovered by the LLVM fuzzer.
352
353 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
354 between a subroutine call and its quantifier was incorrectly compiled,
355 leading to buffer overflow or other errors. This bug was discovered by the
356 LLVM fuzzer.
357
358 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
359 assertion after (?(. The code was failing to check the character after
360 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
361 was discovered by the LLVM fuzzer.
362
363 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
364 a fixed maximum following a group that contains a subroutine reference was
365 incorrectly compiled and could trigger buffer overflow. This bug was
366 discovered by the LLVM fuzzer.
367
368 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
369 caused a stack overflow instead of the diagnosis of a non-fixed length
370 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
371
372 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
373 (e.g. /(?<=\Ka)/) could make pcregrep loop.
374
375 37. There was a similar problem to 36 in pcretest for global matches.
376
377 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
378 and a subsequent item in the pattern caused a non-match, backtracking over
379 the repeated \X did not stop, but carried on past the start of the subject,
380 causing reference to random memory and/or a segfault. There were also some
381 other cases where backtracking after \C could crash. This set of bugs was
382 discovered by the LLVM fuzzer.
383
384 39. The function for finding the minimum length of a matching string could take
385 a very long time if mutual recursion was present many times in a pattern,
386 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
387 method has been implemented. This infelicity was discovered by the LLVM
388 fuzzer.
389
390 40. Static linking against the PCRE library using the pkg-config module was
391 failing on missing pthread symbols.
392
393
394 Version 8.36 26-September-2014
395 ------------------------------
396
397 1. Got rid of some compiler warnings in the C++ modules that were shown up by
398 -Wmissing-field-initializers and -Wunused-parameter.
399
400 2. The tests for quantifiers being too big (greater than 65535) were being
401 applied after reading the number, and stupidly assuming that integer
402 overflow would give a negative number. The tests are now applied as the
403 numbers are read.
404
405 3. Tidy code in pcre_exec.c where two branches that used to be different are
406 now the same.
407
408 4. The JIT compiler did not generate match limit checks for certain
409 bracketed expressions with quantifiers. This may lead to exponential
410 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
411 issue should be resolved now.
412
413 5. Fixed an issue, which occures when nested alternatives are optimized
414 with table jumps.
415
416 6. Inserted two casts and changed some ints to size_t in the light of some
417 reported 64-bit compiler warnings (Bugzilla 1477).
418
419 7. Fixed a bug concerned with zero-minimum possessive groups that could match
420 an empty string, which sometimes were behaving incorrectly in the
421 interpreter (though correctly in the JIT matcher). This pcretest input is
422 an example:
423
424 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
425 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
426
427 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
428 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
429 for an empty string was breaking the inner loop and carrying on at a lower
430 level, when possessive repeated groups should always return to a higher
431 level as they have no backtrack points in them. The empty string test now
432 occurs at the outer level.
433
434 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
435 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
436
437 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
438 doesn't).
439
440 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
441 to hold a pattern to be incorrectly computed (too small) when there were
442 named back references to duplicated names. This could cause "internal
443 error: code overflow" or "double free or corruption" or other memory
444 handling errors.
445
446 11. When named subpatterns had the same prefixes, back references could be
447 confused. For example, in this pattern:
448
449 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
450
451 the reference to 'Name' was incorrectly treated as a reference to a
452 duplicate name.
453
454 12. A pattern such as /^s?c/mi8 where the optional character has more than
455 one "other case" was incorrectly compiled such that it would only try to
456 match starting at "c".
457
458 13. When a pattern starting with \s was studied, VT was not included in the
459 list of possible starting characters; this should have been part of the
460 8.34/18 patch.
461
462 14. If a character class started [\Qx]... where x is any character, the class
463 was incorrectly terminated at the ].
464
465 15. If a pattern that started with a caseless match for a character with more
466 than one "other case" was studied, PCRE did not set up the starting code
467 unit bit map for the list of possible characters. Now it does. This is an
468 optimization improvement, not a bug fix.
469
470 16. The Unicode data tables have been updated to Unicode 7.0.0.
471
472 17. Fixed a number of memory leaks in pcregrep.
473
474 18. Avoid a compiler warning (from some compilers) for a function call with
475 a cast that removes "const" from an lvalue by using an intermediate
476 variable (to which the compiler does not object).
477
478 19. Incorrect code was compiled if a group that contained an internal recursive
479 back reference was optional (had quantifier with a minimum of zero). This
480 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
481 caused segmentation faults because of stack overflows at compile time.
482
483 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
484 group that is quantified with an indefinite repeat, caused a compile-time
485 loop which used up all the system stack and provoked a segmentation fault.
486 This was not the same bug as 19 above.
487
488 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
489 Patch by Mike Frysinger.
490
491
492 Version 8.35 04-April-2014
493 --------------------------
494
495 1. A new flag is set, when property checks are present in an XCLASS.
496 When this flag is not set, PCRE can perform certain optimizations
497 such as studying these XCLASS-es.
498
499 2. The auto-possessification of character sets were improved: a normal
500 and an extended character set can be compared now. Furthermore
501 the JIT compiler optimizes more character set checks.
502
503 3. Got rid of some compiler warnings for potentially uninitialized variables
504 that show up only when compiled with -O2.
505
506 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
507 of a match later then the end of the match. The pcretest program was not
508 handling the case sensibly - it was outputting from the start to the next
509 binary zero. It now reports this situation in a message, and outputs the
510 text from the end to the start.
511
512 5. Fast forward search is improved in JIT. Instead of the first three
513 characters, any three characters with fixed position can be searched.
514 Search order: first, last, middle.
515
516 6. Improve character range checks in JIT. Characters are read by an inprecise
517 function now, which returns with an unknown value if the character code is
518 above a certain threshold (e.g: 256). The only limitation is that the value
519 must be bigger than the threshold as well. This function is useful when
520 the characters above the threshold are handled in the same way.
521
522 7. The macros whose names start with RAWUCHAR are placeholders for a future
523 mode in which only the bottom 21 bits of 32-bit data items are used. To
524 make this more memorable for those maintaining the code, the names have
525 been changed to start with UCHAR21, and an extensive comment has been added
526 to their definition.
527
528 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
529 to the export list in Makefile.am (they were accidentally omitted from the
530 8.34 tarball).
531
532 9. The informational output from pcretest used the phrase "starting byte set"
533 which is inappropriate for the 16-bit and 32-bit libraries. As the output
534 for "first char" and "need char" really means "non-UTF-char", I've changed
535 "byte" to "char", and slightly reworded the output. The documentation about
536 these values has also been (I hope) clarified.
537
538 10. Another JIT related optimization: use table jumps for selecting the correct
539 backtracking path, when more than four alternatives are present inside a
540 bracket.
541
542 11. Empty match is not possible, when the minimum length is greater than zero,
543 and there is no \K in the pattern. JIT should avoid empty match checks in
544 such cases.
545
546 12. In a caseless character class with UCP support, when a character with more
547 than one alternative case was not the first character of a range, not all
548 the alternative cases were added to the class. For example, s and \x{17f}
549 are both alternative cases for S: the class [RST] was handled correctly,
550 but [R-T] was not.
551
552 13. The configure.ac file always checked for pthread support when JIT was
553 enabled. This is not used in Windows, so I have put this test inside a
554 check for the presence of windows.h (which was already tested for).
555
556 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
557 The algorithm provides a way to skip certain starting offsets, and usually
558 faster than linear prefix searches.
559
560 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
561 as for 'fr_FR' and 'french'. For some reason, however, it then used the
562 Windows-specific input and output files, which have 'french' screwed in.
563 So this could never have worked. One of the problems with locales is that
564 they aren't always the same. I have now updated RunTest so that it checks
565 the output of the locale test (test 3) against three different output
566 files, and it allows the test to pass if any one of them matches. With luck
567 this should make the test pass on some versions of Solaris where it was
568 failing. Because of the uncertainty, the script did not used to stop if
569 test 3 failed; it now does. If further versions of a French locale ever
570 come to light, they can now easily be added.
571
572 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
573 there was a message during ./configure, but it did not stop. This now
574 provokes an error. The invalid example in README has been corrected.
575 If a value less than the minimum is given, the minimum value has always
576 been used, but now a warning is given.
577
578 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
579 was a bug in the test system, which is now fixed. Also, the list of various
580 configurations that are tested for each release did not have one with both
581 16/32 bits and --enable-bar-anycrlf. It now does.
582
583 18. pcretest was missing "-C bsr" for displaying the \R default setting.
584
585 19. Little endian PowerPC systems are supported now by the JIT compiler.
586
587 20. The fast forward newline mechanism could enter to an infinite loop on
588 certain invalid UTF-8 input. Although we don't support these cases
589 this issue can be fixed by a performance optimization.
590
591 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
592 not take account if existing stack usage. There is now a new global
593 variable called pcre_stack_guard that can be set to point to an external
594 function to check stack availability. It is called at the start of
595 processing every parenthesized group.
596
597 22. A typo in the code meant that in ungreedy mode the max/min qualifier
598 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
599 not match "ab".
600
601 23. When UTF was disabled, the JIT program reported some incorrect compile
602 errors. These messages are silenced now.
603
604 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
605 compiler.
606
607 25. Change all the temporary files used in RunGrepTest to be different to those
608 used by RunTest so that the tests can be run simultaneously, for example by
609 "make -j check".
610
611
612 Version 8.34 15-December-2013
613 -----------------------------
614
615 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
616 executable memory. Patch inspired by Carsten Klein.
617
618 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
619 this macro is never tested and has no effect, because the work to support
620 coverage involves only compiling and linking options and special targets in
621 the Makefile. The comment in config.h implied that defining the macro would
622 enable coverage support, which is totally false. There was also support for
623 setting this macro in the CMake files (my fault, I just copied it from
624 configure). SUPPORT_GCOV has now been removed.
625
626 3. Make a small performance improvement in strlen16() and strlen32() in
627 pcretest.
628
629 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
630 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
631 these cases). There was also one in pcretest.c.
632
633 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
634
635 6. In UTF mode, the code for checking whether a group could match an empty
636 string (which is used for indefinitely repeated groups to allow for
637 breaking an infinite loop) was broken when the group contained a repeated
638 negated single-character class with a character that occupied more than one
639 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
640 UTF-8 mode). The effect was undefined: the group might or might not be
641 deemed as matching an empty string, or the program might have crashed.
642
643 7. The code for checking whether a group could match an empty string was not
644 recognizing that \h, \H, \v, \V, and \R must match a character.
645
646 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
647 an empty string. If it can, pcretest shows this in its information output.
648
649 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
650 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
651 matched by pcre_exec() without using JIT:
652
653 (a) If the rest of the pattern did not match after a maximal run of
654 grapheme clusters, the code for backing up to try with fewer of them
655 did not always back up over a full grapheme when characters that do not
656 have the modifier quality were involved, e.g. Hangul syllables.
657
658 (b) If the match point in a subject started with modifier character, and
659 there was no match, the code could incorrectly back up beyond the match
660 point, and potentially beyond the first character in the subject,
661 leading to a segfault or an incorrect match result.
662
663 10. A conditional group with an assertion condition could lead to PCRE
664 recording an incorrect first data item for a match if no other first data
665 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
666 first data item, and therefore matched "ca" after "c" instead of at the
667 start.
668
669 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
670 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
671
672 12. The source of pcregrep now includes z/OS-specific code so that it can be
673 compiled for z/OS as part of the special z/OS distribution.
674
675 13. Added the -T and -TM options to pcretest.
676
677 14. The code in pcre_compile.c for creating the table of named capturing groups
678 has been refactored. Instead of creating the table dynamically during the
679 actual compiling pass, the information is remembered during the pre-compile
680 pass (on the stack unless there are more than 20 named groups, in which
681 case malloc() is used) and the whole table is created before the actual
682 compile happens. This has simplified the code (it is now nearly 150 lines
683 shorter) and prepared the way for better handling of references to groups
684 with duplicate names.
685
686 15. A back reference to a named subpattern when there is more than one of the
687 same name now checks them in the order in which they appear in the pattern.
688 The first one that is set is used for the reference. Previously only the
689 first one was inspected. This change makes PCRE more compatible with Perl.
690
691 16. Unicode character properties were updated from Unicode 6.3.0.
692
693 17. The compile-time code for auto-possessification has been refactored, based
694 on a patch by Zoltan Herczeg. It now happens after instead of during
695 compilation. The code is cleaner, and more cases are handled. The option
696 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
697 options in pcretest are provided to set it. It can also be set by
698 (*NO_AUTO_POSSESS) at the start of a pattern.
699
700 18. The character VT has been added to the default ("C" locale) set of
701 characters that match \s and are generally treated as white space,
702 following this same change in Perl 5.18. There is now no difference between
703 "Perl space" and "POSIX space". Whether VT is treated as white space in
704 other locales depends on the locale.
705
706 19. The code for checking named groups as conditions, either for being set or
707 for being recursed, has been refactored (this is related to 14 and 15
708 above). Processing unduplicated named groups should now be as fast at
709 numerical groups, and processing duplicated groups should be faster than
710 before.
711
712 20. Two patches to the CMake build system, by Alexander Barkov:
713
714 (1) Replace the "source" command by "." in CMakeLists.txt because
715 "source" is a bash-ism.
716
717 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
718 without these the CMake build does not work on Solaris.
719
720 21. Perl has changed its handling of \8 and \9. If there is no previously
721 encountered capturing group of those numbers, they are treated as the
722 literal characters 8 and 9 instead of a binary zero followed by the
723 literals. PCRE now does the same.
724
725 22. Following Perl, added \o{} to specify codepoints in octal, making it
726 possible to specify values greater than 0777 and also making them
727 unambiguous.
728
729 23. Perl now gives an error for missing closing braces after \x{... instead of
730 treating the string as literal. PCRE now does the same.
731
732 24. RunTest used to grumble if an inappropriate test was selected explicitly,
733 but just skip it when running all tests. This make it awkward to run ranges
734 of tests when one of them was inappropriate. Now it just skips any
735 inappropriate tests, as it always did when running all tests.
736
737 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
738 character types such as \d or \w, too many callouts were inserted, and the
739 data that they returned was rubbish.
740
741 26. In UCP mode, \s was not matching two of the characters that Perl matches,
742 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
743 were matched by \h. The code has now been refactored so that the lists of
744 the horizontal and vertical whitespace characters used for \h and \v (which
745 are defined only in one place) are now also used for \s.
746
747 27. Add JIT support for the 64 bit TileGX architecture.
748 Patch by Jiong Wang (Tilera Corporation).
749
750 28. Possessive quantifiers for classes (both explicit and automatically
751 generated) now use special opcodes instead of wrapping in ONCE brackets.
752
753 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
754 (because it's meaningless), this was not happening when PCRE_CASELESS was
755 set. Not wrong, but inefficient.
756
757 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
758 properties for \w, \d, etc) is present in a test regex. Otherwise if the
759 test contains no characters greater than 255, Perl doesn't realise it
760 should be using Unicode semantics.
761
762 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
763 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
764 does in Unicode mode.
765
766 32. Added the "forbid" facility to pcretest so that putting tests into the
767 wrong test files can sometimes be quickly detected.
768
769 33. There is now a limit (default 250) on the depth of nesting of parentheses.
770 This limit is imposed to control the amount of system stack used at compile
771 time. It can be changed at build time by --with-parens-nest-limit=xxx or
772 the equivalent in CMake.
773
774 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
775 errors. Perl warns for these when in warning mode, but PCRE has no facility
776 for giving warnings.
777
778 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
779 However, this was not working for (?!) because it is optimized to (*FAIL),
780 for which PCRE does not allow quantifiers. The optimization is now disabled
781 when a quantifier follows (?!). I can't see any use for this, but it makes
782 things uniform.
783
784 36. Perl no longer allows group names to start with digits, so I have made this
785 change also in PCRE. It simplifies the code a bit.
786
787 37. In extended mode, Perl ignores spaces before a + that indicates a
788 possessive quantifier. PCRE allowed a space before the quantifier, but not
789 before the possessive +. It now does.
790
791 38. The use of \K (reset reported match start) within a repeated possessive
792 group such as (a\Kb)*+ was not working.
793
794 40. Document that the same character tables must be used at compile time and
795 run time, and that the facility to pass tables to pcre_exec() and
796 pcre_dfa_exec() is for use only with saved/restored patterns.
797
798 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
799 features for Builds with MSVC:
800
801 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
802 builds). The .rc files can be used to set FileDescription and many other
803 attributes.
804
805 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
806 This allows higher-level build scripts which want .pdb files to avoid
807 hard-coding the exact files needed."
808
809 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
810 mean "start of word" and "end of word", respectively, as a transition aid.
811
812 43. A minimizing repeat of a class containing codepoints greater than 255 in
813 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
814 compiled to use the heap for recursion.
815
816 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
817 is configured.
818
819
820 Version 8.33 28-May-2013
821 ------------------------
822
823 1. Added 'U' to some constants that are compared to unsigned integers, to
824 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
825 variables that are added to signed variables, to ensure the result is
826 signed and can be negated.
827
828 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
829 CMake config files.
830
831 3. Revise the creation of config.h.generic so that all boolean macros are
832 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
833 overriding via -D on the command line possible.
834
835 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
836 to unsigned int is reported to make a quite noticeable speed difference in
837 a specific Windows environment. Testing on Linux did also appear to show
838 some benefit (and it is clearly not harmful). Also fixed the definition of
839 Xop which should be unsigned.
840
841 5. Related to (4), changing the definition of the intermediate variable cc
842 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
843 improvements.
844
845 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
846 unnecessary spaces.
847
848 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
849 and later.
850
851 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
852
853 9. Optimizing fast_forward_start_bits in JIT.
854
855 10. Adding support for callouts in JIT, and fixing some issues revealed
856 during this work. Namely:
857
858 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
859
860 (b) Minimum length was not checked before the matching is started.
861
862 11. The value of capture_last that is passed to callouts was incorrect in some
863 cases when there was a capture on one path that was subsequently abandoned
864 after a backtrack. Also, the capture_last value is now reset after a
865 recursion, since all captures are also reset in this case.
866
867 12. The interpreter no longer returns the "too many substrings" error in the
868 case when an overflowing capture is in a branch that is subsequently
869 abandoned after a backtrack.
870
871 13. In the pathological case when an offset vector of size 2 is used, pcretest
872 now prints out the matched string after a yield of 0 or 1.
873
874 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
875 Only supported by the JIT compiler at the moment.
876
877 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
878
879 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
880 offset of the starting point of the matching process, provided the offsets
881 vector is large enough.
882
883 17. The \A escape now records a lookbehind value of 1, though its execution
884 does not actually inspect the previous character. This is to ensure that,
885 in partial multi-segment matching, at least one character from the old
886 segment is retained when a new segment is processed. Otherwise, if there
887 are no lookbehinds in the pattern, \A might match incorrectly at the start
888 of a new segment.
889
890 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
891
892 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
893 gives some modest performance improvement in 8-bit mode.
894
895 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
896 be expressed in certain programming languages using Universal Character
897 Names.
898
899 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
900 which points out that "non characters" are not "characters that may not
901 appear in Unicode strings" but rather "characters that are reserved for
902 internal use and have only local meaning".
903
904 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
905 there was a conditional group that depended on an assertion, if the
906 assertion was false, the callout that immediately followed the alternation
907 in the condition was skipped when pcre_exec() was used for matching.
908
909 23. Allow an explicit callout to be inserted before an assertion that is the
910 condition for a conditional group, for compatibility with automatic
911 callouts, which always insert a callout at this point.
912
913 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
914 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
915
916 25. (*PRUNE) is now supported by the JIT compiler.
917
918 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
919
920 27. Fix the case where there are two or more SKIPs with arguments that may be
921 ignored.
922
923 28. (*SKIP) is now supported by the JIT compiler.
924
925 29. (*THEN) is now supported by the JIT compiler.
926
927 30. Update RunTest with additional test selector options.
928
929 31. The way PCRE handles backtracking verbs has been changed in two ways.
930
931 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
932 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
933 backtracking. In some cases this makes it more Perl-compatible, but Perl's
934 rather obscure rules do not always do the same thing.
935
936 (2) Previously, backtracking verbs were confined within assertions. This is
937 no longer the case for positive assertions, except for (*ACCEPT). Again,
938 this sometimes improves Perl compatibility, and sometimes does not.
939
940 32. A number of tests that were in test 2 because Perl did things differently
941 have been moved to test 1, because either Perl or PCRE has changed, and
942 these tests are now compatible.
943
944 32. Backtracking control verbs are now handled in the same way in JIT and
945 interpreter.
946
947 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
948 contained a forward subroutine reference caused a compile error.
949
950 34. Auto-detect and optimize limited repetitions in JIT.
951
952 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
953 blocking (*UTF) etc.
954
955 36. In the interpreter, maximizing pattern repetitions for characters and
956 character types now use tail recursion, which reduces stack usage.
957
958 37. The value of the max lookbehind was not correctly preserved if a compiled
959 and saved regex was reloaded on a host of different endianness.
960
961 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
962 of the compiled pattern block, expand the flags field from 16 to 32 bits
963 because it was almost full.
964
965 39. Try madvise first before posix_madvise.
966
967 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
968 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
969 original change locked out all matches of empty strings. This has been
970 changed so that one match of an empty string per line is recognized.
971 Subsequent searches on the same line (for colouring or for --only-matching,
972 for example) do not recognize empty strings.
973
974 41. Applied a user patch to fix a number of spelling mistakes in comments.
975
976 42. Data lines longer than 65536 caused pcretest to crash.
977
978 43. Clarified the data type for length and startoffset arguments for pcre_exec
979 and pcre_dfa_exec in the function-specific man pages, where they were
980 explicitly stated to be in bytes, never having been updated. I also added
981 some clarification to the pcreapi man page.
982
983 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
984 a segmentation fault.
985
986
987 Version 8.32 30-November-2012
988 -----------------------------
989
990 1. Improved JIT compiler optimizations for first character search and single
991 character iterators.
992
993 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
994 Patch by Daniel Richard G.
995
996 3. Single character iterator optimizations in the JIT compiler.
997
998 4. Improved JIT compiler optimizations for character ranges.
999
1000 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1001 Reported by Giuseppe D'Angelo.
1002
1003 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1004 of a line, was being set incorrectly in cases where .* appeared inside
1005 atomic brackets at the start of a pattern, or where there was a subsequent
1006 *PRUNE or *SKIP.
1007
1008 7. Improved instruction cache flush for POWER/PowerPC.
1009 Patch by Daniel Richard G.
1010
1011 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
1012 grep:
1013
1014 (a) There is now no limit to the number of patterns to be matched.
1015
1016 (b) An error is given if a pattern is too long.
1017
1018 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1019 are now supported.
1020
1021 (d) --exclude-from and --include-from (multiple use) have been added.
1022
1023 (e) Exclusions and inclusions now apply to all files and directories, not
1024 just to those obtained from scanning a directory recursively.
1025
1026 (f) Multiple uses of -f and --file-list are now supported.
1027
1028 (g) In a Windows environment, the default for -d has been changed from
1029 "read" (the GNU grep default) to "skip", because otherwise the presence
1030 of a directory in the file list provokes an error.
1031
1032 (h) The documentation has been revised and clarified in places.
1033
1034 9. Improve the matching speed of capturing brackets.
1035
1036 10. Changed the meaning of \X so that it now matches a Unicode extended
1037 grapheme cluster.
1038
1039 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1040 out POSIX threads when JIT support is configured.
1041
1042 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1043
1044 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1045 structure could go wrong in environments where size_t is not the same size
1046 as int.
1047
1048 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1049 set.
1050
1051 15. The EBCDIC support had decayed; later updates to the code had included
1052 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1053 general tidy up of EBCDIC-related issues, and the documentation was also
1054 not quite right. There is now a test that can be run on ASCII systems to
1055 check some of the EBCDIC-related things (but is it not a full test).
1056
1057 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1058 in a small tidy to the code.
1059
1060 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1061
1062 18. If the --only-matching (-o) option in pcregrep is specified multiple
1063 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1064 substrings matched by the 1st and 2nd capturing parentheses. A separating
1065 string can be specified by --om-separator (default empty).
1066
1067 19. Improving the first n character searches.
1068
1069 20. Turn case lists for horizontal and vertical white space into macros so that
1070 they are defined only once.
1071
1072 21. This set of changes together give more compatible Unicode case-folding
1073 behaviour for characters that have more than one other case when UCP
1074 support is available.
1075
1076 (a) The Unicode property table now has offsets into a new table of sets of
1077 three or more characters that are case-equivalent. The MultiStage2.py
1078 script that generates these tables (the pcre_ucd.c file) now scans
1079 CaseFolding.txt instead of UnicodeData.txt for character case
1080 information.
1081
1082 (b) The code for adding characters or ranges of characters to a character
1083 class has been abstracted into a generalized function that also handles
1084 case-independence. In UTF-mode with UCP support, this uses the new data
1085 to handle characters with more than one other case.
1086
1087 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1088 whose other case is greater than 256 are now correctly matched
1089 caselessly. Previously, the high codepoint matched the low one, but not
1090 vice versa.
1091
1092 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1093 of the new class addition function, using character lists defined as
1094 macros alongside the case definitions of 20 above.
1095
1096 (e) Caseless back references now work with characters that have more than
1097 one other case.
1098
1099 (f) General caseless matching of characters with more than one other case
1100 is supported.
1101
1102 22. Unicode character properties were updated from Unicode 6.2.0
1103
1104 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1105
1106 24. Add support for 32-bit character strings, and UTF-32
1107
1108 25. Major JIT compiler update (code refactoring and bugfixing).
1109 Experimental Sparc 32 support is added.
1110
1111 26. Applied a modified version of Daniel Richard G's patch to create
1112 pcre.h.generic and config.h.generic by "make" instead of in the
1113 PrepareRelease script.
1114
1115 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1116 pcre_compile.c when checking for a zero character.
1117
1118 28. Introducing a native interface for JIT. Through this interface, the compiled
1119 machine code can be directly executed. The purpose of this interface is to
1120 provide fast pattern matching, so several sanity checks are not performed.
1121 However, feature tests are still performed. The new interface provides
1122 1.4x speedup compared to the old one.
1123
1124 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1125 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1126 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1127
1128 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1129 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1130 the "old" RFC 2279). Instead, it ended up passing a negative length to
1131 pcre_exec().
1132
1133 31. Add support for GCC's visibility feature to hide internal functions.
1134
1135 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1136 "unknown -C option" after outputting 0 or 1.
1137
1138 33. There is now support for generating a code coverage report for the test
1139 suite in environments where gcc is the compiler and lcov is installed. This
1140 is mainly for the benefit of the developers.
1141
1142 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1143 unaddressable using valgrind annotations, allowing valgrind to detect
1144 invalid memory accesses. This is mainly for the benefit of the developers.
1145
1146 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1147
1148 26. Give configure error if --enable-cpp but no C++ compiler found.
1149
1150
1151 Version 8.31 06-July-2012
1152 -------------------------
1153
1154 1. Fixing a wrong JIT test case and some compiler warnings.
1155
1156 2. Removed a bashism from the RunTest script.
1157
1158 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1159 to unsigned type, result still unsigned" that was given by an MS compiler
1160 on encountering the code "-sizeof(xxx)".
1161
1162 4. Partial matching support is added to the JIT compiler.
1163
1164 5. Fixed several bugs concerned with partial matching of items that consist
1165 of more than one character:
1166
1167 (a) /^(..)\1/ did not partially match "aba" because checking references was
1168 done on an "all or nothing" basis. This also applied to repeated
1169 references.
1170
1171 (b) \R did not give a hard partial match if \r was found at the end of the
1172 subject.
1173
1174 (c) \X did not give a hard partial match after matching one or more
1175 characters at the end of the subject.
1176
1177 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1178 a partial match for the string "\r".
1179
1180 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1181 a partial match for a CR character at the end of the subject string.
1182
1183 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1184 running pcretest, the text "(JIT)" added to the output whenever JIT is
1185 actually used to run the match.
1186
1187 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1188 or /S+[+] with a digit between 1 and 7.
1189
1190 8. OP_NOT now supports any UTF character not just single-byte ones.
1191
1192 9. (*MARK) control verb is now supported by the JIT compiler.
1193
1194 10. The command "./RunTest list" lists the available tests without actually
1195 running any of them. (Because I keep forgetting what they all are.)
1196
1197 11. Add PCRE_INFO_MAXLOOKBEHIND.
1198
1199 12. Applied a (slightly modified) user-supplied patch that improves performance
1200 when the heap is used for recursion (compiled with --disable-stack-for-
1201 recursion). Instead of malloc and free for each heap frame each time a
1202 logical recursion happens, frames are retained on a chain and re-used where
1203 possible. This sometimes gives as much as 30% improvement.
1204
1205 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1206 call.
1207
1208 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1209
1210 15. It is now possible to link pcretest with libedit as an alternative to
1211 libreadline.
1212
1213 16. (*COMMIT) control verb is now supported by the JIT compiler.
1214
1215 17. The Unicode data tables have been updated to Unicode 6.1.0.
1216
1217 18. Added --file-list option to pcregrep.
1218
1219 19. Added binary file support to pcregrep, including the -a, --binary-files,
1220 -I, and --text options.
1221
1222 20. The madvise function is renamed for posix_madvise for QNX compatibility
1223 reasons. Fixed by Giuseppe D'Angelo.
1224
1225 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1226
1227 22. Changed the option for creating symbolic links for 16-bit man pages from
1228 -s to -sf so that re-installing does not cause issues.
1229
1230 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1231
1232 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1233 matches in certain environments (the workspace was not being correctly
1234 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1235 some of the workspace data at the beginning of a restart.
1236
1237 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1238 was not doing so when it should - probably a typo introduced by SVN 528
1239 (change 8.10/14).
1240
1241 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1242 \w+ when the character tables indicated that \x{c4} was a word character.
1243 There were several related cases, all because the tests for doing a table
1244 lookup were testing for characters less than 127 instead of 255.
1245
1246 27. If a pattern contains capturing parentheses that are not used in a match,
1247 their slots in the ovector are set to -1. For those that are higher than
1248 any matched groups, this happens at the end of processing. In the case when
1249 there were back references that the ovector was too small to contain
1250 (causing temporary malloc'd memory to be used during matching), and the
1251 highest capturing number was not used, memory off the end of the ovector
1252 was incorrectly being set to -1. (It was using the size of the temporary
1253 memory instead of the true size.)
1254
1255 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1256 ovector size, it uses memory at the end of the block that it has got.
1257
1258 29. Check for an overlong MARK name and give an error at compile time. The
1259 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1260
1261 30. JIT compiler update.
1262
1263 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1264 Rill for the patch.
1265
1266 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1267
1268 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1269
1270 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1271 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1272 was enabled.
1273
1274 35. Improve JIT code generation for greedy plus quantifier.
1275
1276 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1277 1 to "aa" instead of to an empty string. The bug affected repeated groups
1278 that could potentially match an empty string.
1279
1280 37. Optimizing single character iterators in JIT.
1281
1282 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1283 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1284 codepoints that are too big for the mode are faulted, and in a UTF mode,
1285 disallowed codepoints are also faulted.
1286
1287 39. If PCRE was compiled with UTF support, in three places in the DFA
1288 matcher there was code that should only have been obeyed in UTF mode, but
1289 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1290 processing when bytes with values greater than 127 were present. In 16-bit
1291 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1292 both cases the values are those that cannot be the first data item in a UTF
1293 character. The three items that might have provoked this were recursions,
1294 possessively repeated groups, and atomic groups.
1295
1296 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1297 and pcregrep, because some OS require shared objects to be explicitly
1298 passed to ld, causing the link step to fail if they are not.
1299
1300 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1301 mode, patterns that started with \h* or \R* might be incorrectly matched.
1302
1303
1304 Version 8.30 04-February-2012
1305 -----------------------------
1306
1307 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1308 name is defined in ctype.h.
1309
1310 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1311 only in quite long subpatterns.
1312
1313 3. Removed the function pcre_info(), which has been obsolete and deprecated
1314 since it was replaced by pcre_fullinfo() in February 2000.
1315
1316 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1317 match a (*MARK), and the match failed at the start of the subject, a
1318 reference to memory before the start of the subject could occur. This bug
1319 was introduced by fix 17 of release 8.21.
1320
1321 5. A reference to an unset group with zero minimum repetition was giving
1322 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1323 /(another)?(\1?)test/ matched against "hello world test". This bug was
1324 introduced in release 8.13.
1325
1326 6. Add support for 16-bit character strings (a large amount of work involving
1327 many changes and refactorings).
1328
1329 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1330 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1331 from a file.
1332
1333 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1334 rounding is not applied in this particular case).
1335
1336 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1337 if they appear, or are escaped, in patterns.
1338
1339 10. Get rid of a number of -Wunused-but-set-variable warnings.
1340
1341 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1342 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1343 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1344 also returns the mark "x". This bug applied to capturing parentheses,
1345 non-capturing parentheses, and atomic parentheses. It also applied to some
1346 assertions.
1347
1348 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1349 information out of configure.ac instead of relying on pcre.h.generic, which
1350 is not stored in the repository.
1351
1352 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1353 -lreadline.
1354
1355 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1356
1357 15. Applied Graycode's patch to put the top-level frame on the stack rather
1358 than the heap when not using the stack for recursion. This gives a
1359 performance improvement in many cases when recursion is not deep.
1360
1361 16. Experimental code added to "pcretest -C" to output the stack frame size.
1362
1363
1364 Version 8.21 12-Dec-2011
1365 ------------------------
1366
1367 1. Updating the JIT compiler.
1368
1369 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1370 are added as well.
1371
1372 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1373 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1374 calling _pcre_jit_exec. Some extra comments are added.
1375
1376 4. (*MARK) settings inside atomic groups that do not contain any capturing
1377 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1378 was introduced by change 18 for 8.20.
1379
1380 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1381 ECMA-262 standard.
1382
1383 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1384 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1385 This bug was probably introduced by change 9 of 8.13.
1386
1387 7. While fixing 6 above, I noticed that a number of other items were being
1388 incorrectly rejected as "not fixed length". This arose partly because newer
1389 opcodes had not been added to the fixed-length checking code. I have (a)
1390 corrected the bug and added tests for these items, and (b) arranged for an
1391 error to occur if an unknown opcode is encountered while checking for fixed
1392 length instead of just assuming "not fixed length". The items that were
1393 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1394 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1395 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1396
1397 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1398 being incorrectly compiled and would have given unpredicatble results.
1399
1400 9. A possessively repeated subpattern with minimum repeat count greater than
1401 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1402 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1403 the first (A) could occur when it should not.
1404
1405 10. Add a cast and remove a redundant test from the code.
1406
1407 11. JIT should use pcre_malloc/pcre_free for allocation.
1408
1409 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1410 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1411 is anything other than /usr, -L is still shown).
1412
1413 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1414
1415 14. Perl does not support \N without a following name in a [] class; PCRE now
1416 also gives an error.
1417
1418 15. If a forward reference was repeated with an upper limit of around 2000,
1419 it caused the error "internal error: overran compiling workspace". The
1420 maximum number of forward references (including repeats) was limited by the
1421 internal workspace, and dependent on the LINK_SIZE. The code has been
1422 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1423 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1424 of around 200,000 forward references. While doing this, I also speeded up
1425 the filling in of repeated forward references.
1426
1427 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1428 incorrectly expecting the subject to contain another "a" after the start.
1429
1430 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1431 in the match, the SKIP should be ignored. This was not happening; instead
1432 the SKIP was being treated as NOMATCH. For patterns such as
1433 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1434 tested.
1435
1436 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1437 now much more compatible with Perl, in particular in cases where the result
1438 is a non-match for a non-anchored pattern. For example, if
1439 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1440 "m", where previously it did not return a name. A side effect of this
1441 change is that for partial matches, the last encountered mark name is
1442 returned, as for non matches. A number of tests that were previously not
1443 Perl-compatible have been moved into the Perl-compatible test files. The
1444 refactoring has had the pleasing side effect of removing one argument from
1445 the match() function, thus reducing its stack requirements.
1446
1447 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1448 subsequent uses of /S (without +) incorrectly behaved like /S+.
1449
1450 21. Retrieve executable code size support for the JIT compiler and fixing
1451 some warnings.
1452
1453 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1454 not work when the shorter character appeared right at the end of the
1455 subject string.
1456
1457 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1458 systems.
1459
1460 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1461 output it when the /M option is used in pcretest.
1462
1463 25. The CheckMan script was not being included in the distribution. Also, added
1464 an explicit "perl" to run Perl scripts from the PrepareRelease script
1465 because this is reportedly needed in Windows.
1466
1467 26. If study data was being save in a file and studying had not found a set of
1468 "starts with" bytes for the pattern, the data written to the file (though
1469 never used) was taken from uninitialized memory and so caused valgrind to
1470 complain.
1471
1472 27. Updated RunTest.bat as provided by Sheri Pierce.
1473
1474 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1475
1476 29. Computation of memory usage for the table of capturing group names was
1477 giving an unnecessarily large value.
1478
1479
1480 Version 8.20 21-Oct-2011
1481 ------------------------
1482
1483 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1484 a POSIX class. After further experiments with Perl, which convinced me that
1485 Perl has bugs and confusions, a closing square bracket is no longer allowed
1486 in a POSIX name. This bug also affected patterns with classes that started
1487 with full stops.
1488
1489 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1490 captured substring, but while checking the failing first alternative,
1491 substring 1 is temporarily captured. If the output vector supplied to
1492 pcre_exec() was not big enough for this capture, the yield of the function
1493 was still zero ("insufficient space for captured substrings"). This cannot
1494 be totally fixed without adding another stack variable, which seems a lot
1495 of expense for a edge case. However, I have improved the situation in cases
1496 such as /(a)(b)x|abc/ matched against "abc", where the return code
1497 indicates that fewer than the maximum number of slots in the ovector have
1498 been set.
1499
1500 3. Related to (2) above: when there are more back references in a pattern than
1501 slots in the output vector, pcre_exec() uses temporary memory during
1502 matching, and copies in the captures as far as possible afterwards. It was
1503 using the entire output vector, but this conflicts with the specification
1504 that only 2/3 is used for passing back captured substrings. Now it uses
1505 only the first 2/3, for compatibility. This is, of course, another edge
1506 case.
1507
1508 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1509 main code base, and can be used by building with --enable-jit. When this is
1510 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1511 runtime --no-jit option is given.
1512
1513 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1514 ovector, the return from the function was zero, implying that there were
1515 other matches that did not fit. The correct "exactly full" value is now
1516 returned.
1517
1518 6. If a subpattern that was called recursively or as a subroutine contained
1519 (*PRUNE) or any other control that caused it to give a non-standard return,
1520 invalid errors such as "Error -26 (nested recursion at the same subject
1521 position)" or even infinite loops could occur.
1522
1523 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1524 computing the minimum length on reaching *ACCEPT, and so ended up with the
1525 wrong value of 1 rather than 0. Further investigation indicates that
1526 computing a minimum subject length in the presence of *ACCEPT is difficult
1527 (think back references, subroutine calls), and so I have changed the code
1528 so that no minimum is registered for a pattern that contains *ACCEPT.
1529
1530 8. If (*THEN) was present in the first (true) branch of a conditional group,
1531 it was not handled as intended. [But see 16 below.]
1532
1533 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1534 Sheri Pierce.
1535
1536 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1537 the first byte in a match must be "a".
1538
1539 11. Change 17 for 8.13 increased the recursion depth for patterns like
1540 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1541 pattern contains any instances of (*THEN). If it does not, the old
1542 optimizations are restored. It would be nice to do this on a per-group
1543 basis, but at the moment that is not feasible.
1544
1545 12. In some environments, the output of pcretest -C is CRLF terminated. This
1546 broke RunTest's code that checks for the link size. A single white space
1547 character after the value is now allowed for.
1548
1549 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1550 For "fr", it uses the Windows-specific input and output files.
1551
1552 14. If (*THEN) appeared in a group that was called recursively or as a
1553 subroutine, it did not work as intended. [But see next item.]
1554
1555 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1556 pattern fragments (but not containing any | characters). If A and B are
1557 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1558 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1559 D. In other words, Perl considers parentheses that do not contain any |
1560 characters to be part of a surrounding alternative, whereas PCRE was
1561 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1562 differently. PCRE now behaves in the same way as Perl, except in the case
1563 of subroutine/recursion calls such as (?1) which have in any case always
1564 been different (but PCRE had them first :-).
1565
1566 16. Related to 15 above: Perl does not treat the | in a conditional group as
1567 creating alternatives. Such a group is treated in the same way as an
1568 ordinary group without any | characters when processing (*THEN). PCRE has
1569 been changed to match Perl's behaviour.
1570
1571 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1572 RunGrepTest script failed.
1573
1574 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1575 inevitable for groups that contain captures, but it can lead to a lot of
1576 stack use in large patterns. The old behaviour has been restored for atomic
1577 groups that do not contain any capturing parentheses.
1578
1579 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1580 suppress the check for a minimum subject length at run time. (If it was
1581 given to pcre_exec() or pcre_dfa_exec() it did work.)
1582
1583 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1584 fail to work when decoding hex characters in data strings in EBCDIC
1585 environments.
1586
1587 21. It appears that in at least one Mac OS environment, the isxdigit() function
1588 is implemented as a macro that evaluates to its argument more than once,
1589 contravening the C 90 Standard (I haven't checked a later standard). There
1590 was an instance in pcretest which caused it to go wrong when processing
1591 \x{...} escapes in subject strings. The has been rewritten to avoid using
1592 things like p++ in the argument of isxdigit().
1593
1594
1595 Version 8.13 16-Aug-2011
1596 ------------------------
1597
1598 1. The Unicode data tables have been updated to Unicode 6.0.0.
1599
1600 2. Two minor typos in pcre_internal.h have been fixed.
1601
1602 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1603 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1604 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1605
1606 4. There were a number of related bugs in the code for matching backrefences
1607 caselessly in UTF-8 mode when codes for the characters concerned were
1608 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1609 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1610 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1611 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1612 2-byte code at the end of the subject (it thought there wasn't enough data
1613 left).
1614
1615 5. Comprehensive information about what went wrong is now returned by
1616 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1617 as the output vector has at least 2 elements. The offset of the start of
1618 the failing character and a reason code are placed in the vector.
1619
1620 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1621 now returned is for the first byte of the failing character, instead of the
1622 last byte inspected. This is an incompatible change, but I hope it is small
1623 enough not to be a problem. It makes the returned offset consistent with
1624 pcre_exec() and pcre_dfa_exec().
1625
1626 7. pcretest now gives a text phrase as well as the error number when
1627 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1628 failure, the offset and reason code are output.
1629
1630 8. When \R was used with a maximizing quantifier it failed to skip backwards
1631 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1632 back over a single character (\n). This seems wrong (because it treated the
1633 two characters as a single entity when going forwards), conflicts with the
1634 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1635 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1636 has been changed.
1637
1638 9. Some internal refactoring has changed the processing so that the handling
1639 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1640 time (the PCRE_DOTALL option was changed this way some time ago: version
1641 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1642 which was always a bit of a fudge. It also means that there is one less
1643 argument for the match() function, which reduces its stack requirements
1644 slightly. This change also fixes an incompatibility with Perl: the pattern
1645 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1646
1647 10. More internal refactoring has drastically reduced the number of recursive
1648 calls to match() for possessively repeated groups such as (abc)++ when
1649 using pcre_exec().
1650
1651 11. While implementing 10, a number of bugs in the handling of groups were
1652 discovered and fixed:
1653
1654 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1655 (a|)*(?1) gave a compile-time internal error.
1656 ((a|)+)+ did not notice that the outer group could match an empty string.
1657 (^a|^)+ was not marked as anchored.
1658 (.*a|.*)+ was not marked as matching at start or after a newline.
1659
1660 12. Yet more internal refactoring has removed another argument from the match()
1661 function. Special calls to this function are now indicated by setting a
1662 value in a variable in the "match data" data block.
1663
1664 13. Be more explicit in pcre_study() instead of relying on "default" for
1665 opcodes that mean there is no starting character; this means that when new
1666 ones are added and accidentally left out of pcre_study(), testing should
1667 pick them up.
1668
1669 14. The -s option of pcretest has been documented for ages as being an old
1670 synonym of -m (show memory usage). I have changed it to mean "force study
1671 for every regex", that is, assume /S for every regex. This is similar to -i
1672 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1673 using it. It makes it easier to run collections of tests with and without
1674 study enabled, and thereby test pcre_study() more easily. All the standard
1675 tests are now run with and without -s (but some patterns can be marked as
1676 "never study" - see 20 below).
1677
1678 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1679 restoration of the capturing data to the outer values was not happening
1680 correctly.
1681
1682 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1683 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1684 pattern had matched an empty string, and so incorrectly returned a no
1685 match.
1686
1687 17. There was optimizing code for the last branch of non-capturing parentheses,
1688 and also for the obeyed branch of a conditional subexpression, which used
1689 tail recursion to cut down on stack usage. Unfortunately, now that there is
1690 the possibility of (*THEN) occurring in these branches, tail recursion is
1691 no longer possible because the return has to be checked for (*THEN). These
1692 two optimizations have therefore been removed. [But see 8.20/11 above.]
1693
1694 18. If a pattern containing \R was studied, it was assumed that \R always
1695 matched two bytes, thus causing the minimum subject length to be
1696 incorrectly computed because \R can also match just one byte.
1697
1698 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1699 was incorrectly computed.
1700
1701 20. If /S is present twice on a test pattern in pcretest input, it now
1702 *disables* studying, thereby overriding the use of -s on the command line
1703 (see 14 above). This is necessary for one or two tests to keep the output
1704 identical in both cases.
1705
1706 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1707 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1708
1709 22. When an atomic group that contained a capturing parenthesis was
1710 successfully matched, but the branch in which it appeared failed, the
1711 capturing was not being forgotten if a higher numbered group was later
1712 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1713 group 1 to "a", when in fact it should be unset. This applied to multi-
1714 branched capturing and non-capturing groups, repeated or not, and also to
1715 positive assertions (capturing in negative assertions does not happen
1716 in PCRE) and also to nested atomic groups.
1717
1718 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1719 subject after a captured substring, to make it easier to tell which of a
1720 number of identical substrings has been captured.
1721
1722 24. The way atomic groups are processed by pcre_exec() has been changed so that
1723 if they are repeated, backtracking one repetition now resets captured
1724 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1725 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1726 "aaa". Previously, it would have been "a". As part of this code
1727 refactoring, the way recursive calls are handled has also been changed.
1728
1729 25. If an assertion condition captured any substrings, they were not passed
1730 back unless some other capturing happened later. For example, if
1731 (?(?=(a))a) was matched against "a", no capturing was returned.
1732
1733 26. When studying a pattern that contained subroutine calls or assertions,
1734 the code for finding the minimum length of a possible match was handling
1735 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1736 group 1 called group 2 while simultaneously a separate group 2 called group
1737 1). A stack overflow occurred in this case. I have fixed this by limiting
1738 the recursion depth to 10.
1739
1740 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1741 Fortmann. This supports explicit test numbers on the command line, and has
1742 argument validation and error reporting.
1743
1744 28. An instance of \X with an unlimited repeat could fail if at any point the
1745 first character it looked at was a mark character.
1746
1747 29. Some minor code refactoring concerning Unicode properties and scripts
1748 should reduce the stack requirement of match() slightly.
1749
1750 30. Added the '=' option to pcretest to check the setting of unused capturing
1751 slots at the end of the pattern, which are documented as being -1, but are
1752 not included in the return count.
1753
1754 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1755 compiled something random. Now it gives a compile-time error (as does
1756 Perl).
1757
1758 32. A *MARK encountered during the processing of a positive assertion is now
1759 recorded and passed back (compatible with Perl).
1760
1761 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1762 had alternative anchored branches, the search for a second match in a line
1763 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1764 matched the line "0102" twice. The same bug affected patterns that started
1765 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1766 twice.
1767
1768 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1769 does, and because of capturing effects, quantifying parenthesized
1770 assertions may at times be useful. Quantifiers are now allowed for
1771 parenthesized assertions.
1772
1773 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1774
1775 36. \g was being checked for fancy things in a character class, when it should
1776 just be a literal "g".
1777
1778 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1779 appearance of a nested POSIX class supersedes an apparent external class.
1780 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1781 unescaped square brackets may also appear as part of class names. For
1782 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1783 more like Perl. (But see 8.20/1 above.)
1784
1785 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1786 was because it thought it was \N{name}, which is not supported).
1787
1788 39. Add minix to OS list not supporting the -S option in pcretest.
1789
1790 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1791 cannot analyze patterns in sufficient detail to catch mutual recursions
1792 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1793 subgroup is called recursively as a subpattern for a second time at the
1794 same position in the subject string. In previous releases this might have
1795 been caught by the recursion limit, or it might have run out of stack.
1796
1797 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1798 happen only once. PCRE was, however incorrectly giving a compile time error
1799 "recursive call could loop indefinitely" because it cannot analyze the
1800 pattern in sufficient detail. The compile time test no longer happens when
1801 PCRE is compiling a conditional subpattern, but actual runaway loops are
1802 now caught at runtime (see 40 above).
1803
1804 42. It seems that Perl allows any characters other than a closing parenthesis
1805 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1806 has been changed to be the same.
1807
1808 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1809 as not to get warnings when autogen.sh is called. Also changed
1810 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1811
1812 44. To help people who use pcregrep to scan files containing exceedingly long
1813 lines, the following changes have been made:
1814
1815 (a) The default value of the buffer size parameter has been increased from
1816 8K to 20K. (The actual buffer used is three times this size.)
1817
1818 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1819 PCRE is built.
1820
1821 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1822 to be set at run time.
1823
1824 (d) Numerical values in pcregrep options can be followed by K or M, for
1825 example --buffer-size=50K.
1826
1827 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1828 given and the return code is set to 2.
1829
1830 45. Add a pointer to the latest mark to the callout data block.
1831
1832 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1833 partial match of an empty string instead of no match. This was specific to
1834 the use of ".".
1835
1836 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1837 complete match instead of a partial match. This bug was dependent on both
1838 the PCRE_UTF8 and PCRE_DOTALL options being set.
1839
1840 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1841 starting byte set, because \b was not being ignored.
1842
1843
1844 Version 8.12 15-Jan-2011
1845 ------------------------
1846
1847 1. Fixed some typos in the markup of the man pages, and wrote a script that
1848 checks for such things as part of the documentation building process.
1849
1850 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1851 --match-limit and --recursion-limit options (added for 8.11). In
1852 particular, this made one of the standard tests fail. (The integer value
1853 went into the wrong half of a long int.)
1854
1855 3. If the --colour option was given to pcregrep with -v (invert match), it
1856 did strange things, either producing crazy output, or crashing. It should,
1857 of course, ignore a request for colour when reporting lines that do not
1858 match.
1859
1860 4. Another pcregrep bug caused similar problems if --colour was specified with
1861 -M (multiline) and the pattern match finished with a line ending.
1862
1863 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1864 matched in multiline mode, the following line was shown as part of the
1865 match. This seems wrong, so I have changed it.
1866
1867 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1868 the check for further matches in the same line (so they could be coloured)
1869 to overrun the end of the current line. If another match was found, it was
1870 incorrectly shown (and then shown again when found in the next line).
1871
1872 7. If pcregrep was compiled under Windows, there was a reference to the
1873 function pcregrep_exit() before it was defined. I am assuming this was
1874 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1875 reported by a user. I've moved the definition above the reference.
1876
1877
1878 Version 8.11 10-Dec-2010
1879 ------------------------
1880
1881 1. (*THEN) was not working properly if there were untried alternatives prior
1882 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1883 backtracked to try for "b" instead of moving to the next alternative branch
1884 at the same level (in this case, to look for "c"). The Perl documentation
1885 is clear that when (*THEN) is backtracked onto, it goes to the "next
1886 alternative in the innermost enclosing group".
1887
1888 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1889 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1890 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1891 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1892 (*THEN).
1893
1894 3. If \s appeared in a character class, it removed the VT character from
1895 the class, even if it had been included by some previous item, for example
1896 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1897 of \s, but is part of the POSIX "space" class.)
1898
1899 4. A partial match never returns an empty string (because you can always
1900 match an empty string at the end of the subject); however the checking for
1901 an empty string was starting at the "start of match" point. This has been
1902 changed to the "earliest inspected character" point, because the returned
1903 data for a partial match starts at this character. This means that, for
1904 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1905 (previously it gave "no match").
1906
1907 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1908 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1909 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1910 has an implication that the given string is incomplete (because a partial
1911 match is preferred over a full match). For this reason, these items now
1912 give a partial match in this situation. [Aside: previously, the one case
1913 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1914 match rather than a full match, which was wrong by the old rules, but is
1915 now correct.]
1916
1917 6. There was a bug in the handling of #-introduced comments, recognized when
1918 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1919 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1920 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1921 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1922 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1923 places in pcre_compile().
1924
1925 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1926 comments when looking ahead for named forward references to subpatterns,
1927 the only newline sequence it recognized was NL. It now handles newlines
1928 according to the set newline convention.
1929
1930 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1931 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1932 cater for a lack of strerror(). These oversights have been fixed.
1933
1934 9. Added --match-limit and --recursion-limit to pcregrep.
1935
1936 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1937
1938 11. When the -o option was used, pcregrep was setting a return code of 1, even
1939 when matches were found, and --line-buffered was not being honoured.
1940
1941 12. Added an optional parentheses number to the -o and --only-matching options
1942 of pcregrep.
1943
1944 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1945 can match an empty string. The code to do it in pcretest and pcredemo
1946 needed fixing:
1947
1948 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1949 only one byte after an empty string match just before CRLF (this case
1950 just got forgotten; "any" and "anycrlf" were OK).
1951
1952 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1953 mode when an empty string match preceded an ASCII character followed by
1954 a non-ASCII character. (The code for advancing by one character rather
1955 than one byte was nonsense.)
1956
1957 (c) The pcredemo.c sample program did not have any code at all to handle
1958 the cases when CRLF is a valid newline sequence.
1959
1960 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1961 as a starting offset was within the subject string. There is now a new
1962 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1963 negative or greater than the length of the string. In order to test this,
1964 pcretest is extended to allow the setting of negative starting offsets.
1965
1966 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1967 starting offset points to the beginning of a UTF-8 character was
1968 unnecessarily clumsy. I tidied it up.
1969
1970 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1971 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1972
1973 17. Nobody had reported that the --include_dir option, which was added in
1974 release 7.7 should have been called --include-dir (hyphen, not underscore)
1975 for compatibility with GNU grep. I have changed it to --include-dir, but
1976 left --include_dir as an undocumented synonym, and the same for
1977 --exclude-dir, though that is not available in GNU grep, at least as of
1978 release 2.5.4.
1979
1980 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1981 characters from a string of bytes) have been redefined so as not to use
1982 loops, in order to improve performance in some environments. At the same
1983 time, I abstracted some of the common code into auxiliary macros to save
1984 repetition (this should not affect the compiled code).
1985
1986 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1987 compile-time error is now given if \c is not followed by an ASCII
1988 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1989 different, and any byte value is allowed.)
1990
1991 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1992 START_OPTIMIZE option, which is now allowed at compile time - but just
1993 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1994 to pcregrep and other applications that have no direct access to PCRE
1995 options. The new /Y option in pcretest sets this option when calling
1996 pcre_compile().
1997
1998 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1999 back references. Groups containing recursive back references were forced to
2000 be atomic by that change, but in the case of named groups, the amount of
2001 memory required was incorrectly computed, leading to "Failed: internal
2002 error: code overflow". This has been fixed.
2003
2004 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2005 pcretest.c, to avoid build problems in some Borland environments.
2006
2007
2008 Version 8.10 25-Jun-2010
2009 ------------------------
2010
2011 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2012 THEN.
2013
2014 2. (*ACCEPT) was not working when inside an atomic group.
2015
2016 3. Inside a character class, \B is treated as a literal by default, but
2017 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2018 causes the error). The code is unchanged, but I tidied the documentation.
2019
2020 4. Inside a character class, PCRE always treated \R and \X as literals,
2021 whereas Perl faults them if its -w option is set. I have changed PCRE so
2022 that it faults them when PCRE_EXTRA is set.
2023
2024 5. Added support for \N, which always matches any character other than
2025 newline. (It is the same as "." when PCRE_DOTALL is not set.)
2026
2027 6. When compiling pcregrep with newer versions of gcc which may have
2028 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2029 declared with attribute warn_unused_result" were given. Just casting the
2030 result to (void) does not stop the warnings; a more elaborate fudge is
2031 needed. I've used a macro to implement this.
2032
2033 7. Minor change to pcretest.c to avoid a compiler warning.
2034
2035 8. Added four artifical Unicode properties to help with an option to make
2036 \s etc use properties (see next item). The new properties are: Xan
2037 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2038
2039 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2040 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2041 this option. Modified pcretest to add /W to test this facility. Added
2042 REG_UCP to make it available via the POSIX interface.
2043
2044 10. Added --line-buffered to pcregrep.
2045
2046 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2047 studied, and the match started with a letter with a code point greater than
2048 127 whose first byte was different to the first byte of the other case of
2049 the letter, the other case of this starting letter was not recognized
2050 (#976).
2051
2052 12. If a pattern that was studied started with a repeated Unicode property
2053 test, for example, \p{Nd}+, there was the theoretical possibility of
2054 setting up an incorrect bitmap of starting bytes, but fortunately it could
2055 not have actually happened in practice until change 8 above was made (it
2056 added property types that matched character-matching opcodes).
2057
2058 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2059 possible starting bytes for non-anchored patterns.
2060
2061 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2062 \R, and also a number of cases that involve Unicode properties, both
2063 explicit and implicit when PCRE_UCP is set.
2064
2065 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2066 input, it could crash or give wrong results if characters with values
2067 greater than 0xc0 were present in the subject string. (Detail: it assumed
2068 UTF-8 input when processing these items.)
2069
2070 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2071 size_t is 64-bit (#991).
2072
2073 17. Added a check for running out of memory when PCRE is compiled with
2074 --disable-stack-for-recursion (#990).
2075
2076 18. If the last data line in a file for pcretest does not have a newline on
2077 the end, a newline was missing in the output.
2078
2079 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2080 less than 128) in its various bitmaps. However, there is a facility for
2081 generating tables according to the current locale when PCRE is compiled. It
2082 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2083 characters, are recognized by isspace() and therefore were getting set in
2084 these tables, and indeed these tables seem to approximate to ISO 8859. This
2085 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2086 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2087 which of course cannot start UTF-8 characters. I have changed the code so
2088 that only real ASCII characters (less than 128) and the correct starting
2089 bytes for UTF-8 encodings are set for characters greater than 127 when in
2090 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2091 altogether.)
2092
2093 20. Added the /T option to pcretest so as to be able to run tests with non-
2094 standard character tables, thus making it possible to include the tests
2095 used for 19 above in the standard set of tests.
2096
2097 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2098 reference to a subpattern the other side of a comment that contains an
2099 opening parenthesis caused either an internal compiling error, or a
2100 reference to the wrong subpattern.
2101
2102
2103 Version 8.02 19-Mar-2010
2104 ------------------------
2105
2106 1. The Unicode data tables have been updated to Unicode 5.2.0.
2107
2108 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2109 configured.
2110
2111 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2112 original author of that file, following a query about its status.
2113
2114 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2115 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2116
2117 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2118 quantifier applied to a forward-referencing subroutine call, could compile
2119 incorrect code or give the error "internal error: previously-checked
2120 referenced subpattern not found".
2121
2122 6. Both MS Visual Studio and Symbian OS have problems with initializing
2123 variables to point to external functions. For these systems, therefore,
2124 pcre_malloc etc. are now initialized to local functions that call the
2125 relevant global functions.
2126
2127 7. There were two entries missing in the vectors called coptable and poptable
2128 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2129 I've fixed the data, and added a kludgy way of testing at compile time that
2130 the lengths are correct (equal to the number of opcodes).
2131
2132 8. Following on from 7, I added a similar kludge to check the length of the
2133 eint vector in pcreposix.c.
2134
2135 9. Error texts for pcre_compile() are held as one long string to avoid too
2136 much relocation at load time. To find a text, the string is searched,
2137 counting zeros. There was no check for running off the end of the string,
2138 which could happen if a new error number was added without updating the
2139 string.
2140
2141 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2142
2143 11. \K was not working if it appeared in an atomic group or in a group that
2144 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2145 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2146 the assertion is positive, but not if it is negative.
2147
2148 12. Change 11 fortuitously reduced the size of the stack frame used in the
2149 "match()" function of pcre_exec.c by one pointer. Forthcoming
2150 implementation of support for (*MARK) will need an extra pointer on the
2151 stack; I have reserved it now, so that the stack frame size does not
2152 decrease.
2153
2154 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2155 item in branch that calls a recursion is a subroutine call - as in the
2156 second branch in the above example - was incorrectly given the compile-
2157 time error "recursive call could loop indefinitely" because pcre_compile()
2158 was not correctly checking the subroutine for matching a non-empty string.
2159
2160 14. The checks for overrunning compiling workspace could trigger after an
2161 overrun had occurred. This is a "should never occur" error, but it can be
2162 triggered by pathological patterns such as hundreds of nested parentheses.
2163 The checks now trigger 100 bytes before the end of the workspace.
2164
2165 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2166
2167
2168 Version 8.01 19-Jan-2010
2169 ------------------------
2170
2171 1. If a pattern contained a conditional subpattern with only one branch (in
2172 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2173 computed the wrong minimum data length (which is of course zero for such
2174 subpatterns). This could cause incorrect "no match" results.
2175
2176 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2177 the pattern is reset in the first branch, pcre_compile() failed with
2178 "internal error: code overflow at offset...". This happened only when
2179 the reset was to the original external option setting. (An optimization
2180 abstracts leading options settings into an external setting, which was the
2181 cause of this.)
2182
2183 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2184 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2185 assertion pattern did not match (meaning that the assertion was true), it
2186 was incorrectly treated as false if the SKIP had been reached during the
2187 matching. This also applied to assertions used as conditions.
2188
2189 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2190 assertion subpattern, including such a pattern used as a condition,
2191 unpredictable results occurred, instead of the error return
2192 PCRE_ERROR_DFA_UITEM.
2193
2194 5. The C++ GlobalReplace function was not working like Perl for the special
2195 situation when an empty string is matched. It now does the fancy magic
2196 stuff that is necessary.
2197
2198 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2199 removed. (These were left over from very, very early versions of PCRE.)
2200
2201 7. Some cosmetic changes to the code to make life easier when compiling it
2202 as part of something else:
2203
2204 (a) Change DEBUG to PCRE_DEBUG.
2205
2206 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2207 called "current" as "current_branch", to prevent a collision with the
2208 Linux macro when compiled as a kernel module.
2209
2210 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2211 prevent a collision with the Linux macro when compiled as a kernel
2212 module.
2213
2214 8. In pcre_compile() there are some checks for integer overflows that used to
2215 cast potentially large values to (double). This has been changed to that
2216 when building, a check for int64_t is made, and if it is found, it is used
2217 instead, thus avoiding the use of floating point arithmetic. (There is no
2218 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2219 double.
2220
2221 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2222 2005 (difference between two addresses compared to an unsigned value).
2223
2224 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2225 custom one, because of the following reported problem in Windows:
2226
2227 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2228 under Win32.
2229 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2230 therefore missing the function definition.
2231 - The compiler thus generates a "C" signature for the test function.
2232 - The linker fails to find the "C" function.
2233 - PCRE fails to configure if asked to do so against libbz2.
2234
2235 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2236 messages were output:
2237
2238 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2239 rerunning libtoolize, to keep the correct libtool macros in-tree.
2240 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2241
2242 I have done both of these things.
2243
2244 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2245 most of the time, it *can* run out if it is given a pattern that contains a
2246 runaway infinite recursion. I updated the discussion in the pcrestack man
2247 page.
2248
2249 13. Now that we have gone to the x.xx style of version numbers, the minor
2250 version may start with zero. Using 08 or 09 is a bad idea because users
2251 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2252 interpreted as invalid octal numbers. I've updated the previous comment in
2253 configure.ac, and also added a check that gives an error if 08 or 09 are
2254 used.
2255
2256 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2257 causing partial matching to fail when the end of the subject matched \W
2258 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2259
2260 15. There were some discrepancies between the declarations in pcre_internal.h
2261 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2262 their definitions. The declarations used "const uschar *" and the
2263 definitions used USPTR. Even though USPTR is normally defined as "const
2264 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2265 reported that: "This difference in casting confuses some C++ compilers, for
2266 example, SunCC recognizes above declarations as different functions and
2267 generates broken code for hbpcre." I have changed the declarations to use
2268 USPTR.
2269
2270 16. GNU libtool is named differently on some systems. The autogen.sh script now
2271 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2272 (FreeBSD).
2273
2274 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2275 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2276 comment: "Figure out how to create a longlong from a string: strtoll and
2277 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2278 instance, but it only takes 2 args instead of 3!"
2279
2280 18. A subtle bug concerned with back references has been fixed by a change of
2281 specification, with a corresponding code fix. A pattern such as
2282 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2283 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2284 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2285 same bug. Such groups have to be quantified to be useful, or contained
2286 inside another quantified group. (If there's no repetition, the reference
2287 can never match.) The problem arises because, having left the group and
2288 moved on to the rest of the pattern, a later failure that backtracks into
2289 the group uses the captured value from the final iteration of the group
2290 rather than the correct earlier one. I have fixed this in PCRE by forcing
2291 any group that contains a reference to itself to be an atomic group; that
2292 is, there cannot be any backtracking into it once it has completed. This is
2293 similar to recursive and subroutine calls.
2294
2295
2296 Version 8.00 19-Oct-09
2297 ----------------------
2298
2299 1. The table for translating pcre_compile() error codes into POSIX error codes
2300 was out-of-date, and there was no check on the pcre_compile() error code
2301 being within the table. This could lead to an OK return being given in
2302 error.
2303
2304 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2305 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2306 in a Windows environment.
2307
2308 3. The pcregrep --count option prints the count for each file even when it is
2309 zero, as does GNU grep. However, pcregrep was also printing all files when
2310 --files-with-matches was added. Now, when both options are given, it prints
2311 counts only for those files that have at least one match. (GNU grep just
2312 prints the file name in this circumstance, but including the count seems
2313 more useful - otherwise, why use --count?) Also ensured that the
2314 combination -clh just lists non-zero counts, with no names.
2315
2316 4. The long form of the pcregrep -F option was incorrectly implemented as
2317 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2318 but it seems right to fix it, and I didn't think it was worth preserving
2319 the old behaviour.
2320
2321 5. The command line items --regex=pattern and --regexp=pattern were not
2322 recognized by pcregrep, which required --regex pattern or --regexp pattern
2323 (with a space rather than an '='). The man page documented the '=' forms,
2324 which are compatible with GNU grep; these now work.
2325
2326 6. No libpcreposix.pc file was created for pkg-config; there was just
2327 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2328
2329 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2330 when UCP support is not needed, by modifying the Python script that
2331 generates it from Unicode data files. This should not matter if the module
2332 is correctly used as a library, but I received one complaint about 50K of
2333 unwanted data. My guess is that the person linked everything into his
2334 program rather than using a library. Anyway, it does no harm.
2335
2336 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2337 was a minimum greater than 1 for a wide character in a possessive
2338 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2339 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2340 character. Chaos in the form of incorrect output or a compiling loop could
2341 result.
2342
2343 9. The restrictions on what a pattern can contain when partial matching is
2344 requested for pcre_exec() have been removed. All patterns can now be
2345 partially matched by this function. In addition, if there are at least two
2346 slots in the offset vector, the offset of the earliest inspected character
2347 for the match and the offset of the end of the subject are set in them when
2348 PCRE_ERROR_PARTIAL is returned.
2349
2350 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2351 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2352 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2353 and may be more useful for multi-segment matching.
2354
2355 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2356 used to be given if ever the end of the subject was reached; now it is
2357 given only if matching could not proceed because another character was
2358 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2359 string "Z", which now yields "no match" instead of "partial match". In the
2360 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2361 final character ended with (*FAIL).
2362
2363 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2364 if the pattern had a "must contain" character that was already found in the
2365 earlier partial match, unless partial matching was again requested. For
2366 example, with the pattern /dog.(body)?/, the "must contain" character is
2367 "g". If the first part-match was for the string "dog", restarting with
2368 "sbody" failed. This bug has been fixed.
2369
2370 13. The string returned by pcre_dfa_exec() after a partial match has been
2371 changed so that it starts at the first inspected character rather than the
2372 first character of the match. This makes a difference only if the pattern
2373 starts with a lookbehind assertion or \b or \B (\K is not supported by
2374 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2375 matching functions compatible, and I think it's the right thing to do.
2376
2377 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2378 so that the demonstration program is easily available in environments where
2379 PCRE has not been installed from source.
2380
2381 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2382 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2383 library.
2384
2385 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2386 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2387 is not the first non-POSIX option to be added. Clearly some people find
2388 these options useful.
2389
2390 17. If a caller to the POSIX matching function regexec() passes a non-zero
2391 value for nmatch with a NULL value for pmatch, the value of
2392 nmatch is forced to zero.
2393
2394 18. RunGrepTest did not have a test for the availability of the -u option of
2395 the diff command, as RunTest does. It now checks in the same way as
2396 RunTest, and also checks for the -b option.
2397
2398 19. If an odd number of negated classes containing just a single character
2399 interposed, within parentheses, between a forward reference to a named
2400 subpattern and the definition of the subpattern, compilation crashed with
2401 an internal error, complaining that it could not find the referenced
2402 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2403 [The bug was that it was starting one character too far in when skipping
2404 over the character class, thus treating the ] as data rather than
2405 terminating the class. This meant it could skip too much.]
2406
2407 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2408 /g option in pcretest when the pattern contains \K, which makes it possible
2409 to have an empty string match not at the start, even when the pattern is
2410 anchored. Updated pcretest and pcredemo to use this option.
2411
2412 21. If the maximum number of capturing subpatterns in a recursion was greater
2413 than the maximum at the outer level, the higher number was returned, but
2414 with unset values at the outer level. The correct (outer level) value is
2415 now given.
2416
2417 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2418 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2419 make it do so. The string so far is captured, making this feature
2420 compatible with Perl.
2421
2422 23. The tests have been re-organized, adding tests 11 and 12, to make it
2423 possible to check the Perl 5.10 features against Perl 5.10.
2424
2425 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2426 pattern matches a fixed length string. PCRE did not allow this; now it
2427 does. Neither allows recursion.
2428
2429 25. I finally figured out how to implement a request to provide the minimum
2430 length of subject string that was needed in order to match a given pattern.
2431 (It was back references and recursion that I had previously got hung up
2432 on.) This code has now been added to pcre_study(); it finds a lower bound
2433 to the length of subject needed. It is not necessarily the greatest lower
2434 bound, but using it to avoid searching strings that are too short does give
2435 some useful speed-ups. The value is available to calling programs via
2436 pcre_fullinfo().
2437
2438 26. While implementing 25, I discovered to my embarrassment that pcretest had
2439 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2440 study optimizations had never been tested with that matching function.
2441 Oops. What is worse, even when it was passed study data, there was a bug in
2442 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2443 were also very few tests of studied patterns with pcre_dfa_exec().
2444
2445 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2446 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2447 on the other side of the coin, they are no longer allowed to have different
2448 names, because these cannot be distinguished in PCRE, and this has caused
2449 confusion. (This is a difference from Perl.)
2450
2451 28. When duplicate subpattern names are present (necessarily with different
2452 numbers, as required by 27 above), and a test is made by name in a
2453 conditional pattern, either for a subpattern having been matched, or for
2454 recursion in such a pattern, all the associated numbered subpatterns are
2455 tested, and the overall condition is true if the condition is true for any
2456 one of them. This is the way Perl works, and is also more like the way
2457 testing by number works.
2458
2459
2460 Version 7.9 11-Apr-09
2461 ---------------------
2462
2463 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2464 (pcretest), all targets were linked against these libraries. This included
2465 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2466 libraries. This caused unwanted dependencies to be created. This problem
2467 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2468 pcretest is linked with readline.
2469
2470 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2471 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2472 moved outside it again, because FALSE and TRUE are already defined in AIX,
2473 but BOOL is not.
2474
2475 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2476 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2477
2478 4. The pcregrep documentation said spaces were inserted as well as colons (or
2479 hyphens) following file names and line numbers when outputting matching
2480 lines. This is not true; no spaces are inserted. I have also clarified the
2481 wording for the --colour (or --color) option.
2482
2483 5. In pcregrep, when --colour was used with -o, the list of matching strings
2484 was not coloured; this is different to GNU grep, so I have changed it to be
2485 the same.
2486
2487 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2488 each matching line was coloured. Now it goes on to look for further matches
2489 of any of the test patterns, which is the same behaviour as GNU grep.
2490
2491 7. A pattern that could match an empty string could cause pcregrep to loop; it
2492 doesn't make sense to accept an empty string match in pcregrep, so I have
2493 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2494 seems to be how GNU grep behaves. [But see later change 40 for release
2495 8.33.]
2496
2497 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2498 start or after a newline", because the conditional assertion was not being
2499 correctly handled. The rule now is that both the assertion and what follows
2500 in the first alternative must satisfy the test.
2501
2502 9. If auto-callout was enabled in a pattern with a conditional group whose
2503 condition was an assertion, PCRE could crash during matching, both with
2504 pcre_exec() and pcre_dfa_exec().
2505
2506 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2507 used for matching.
2508
2509 11. Unicode property support in character classes was not working for
2510 characters (bytes) greater than 127 when not in UTF-8 mode.
2511
2512 12. Added the -M command line option to pcretest.
2513
2514 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2515
2516 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2517
2518 16. Added comments and documentation about mis-use of no_arg in the C++
2519 wrapper.
2520
2521 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2522 from Martin Jerabek that uses macro names for all relevant character and
2523 string constants.
2524
2525 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2526 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2527 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2528 these, but not everybody uses configure.
2529
2530 19. A conditional group that had only one branch was not being correctly
2531 recognized as an item that could match an empty string. This meant that an
2532 enclosing group might also not be so recognized, causing infinite looping
2533 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2534 with the subject "ab", where knowledge that the repeated group can match
2535 nothing is needed in order to break the loop.
2536
2537 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2538 exec(), but without supplying a callout function, matching went wrong.
2539
2540 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2541 leak if the size of the offset vector was greater than 30. When the vector
2542 is smaller, the saved offsets during recursion go onto a local stack
2543 vector, but for larger vectors malloc() is used. It was failing to free
2544 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2545 error, in fact).
2546
2547 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2548 heapframe that is used only when UTF-8 support is enabled. This caused no
2549 problem, but was untidy.
2550
2551 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2552 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2553 included within another project.
2554
2555 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2556 slightly modified by me:
2557
2558 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2559 not building pcregrep.
2560
2561 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2562 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2563
2564 25. Forward references, both numeric and by name, in patterns that made use of
2565 duplicate group numbers, could behave incorrectly or give incorrect errors,
2566 because when scanning forward to find the reference group, PCRE was not
2567 taking into account the duplicate group numbers. A pattern such as
2568 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2569
2570 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2571 the feature of a custom pointer more persuasive (as requested by a user).
2572
2573 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2574 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2575 pre-defined.
2576
2577 28. Added support for (*UTF8) at the start of a pattern.
2578
2579 29. Arrange for flags added by the "release type" setting in CMake to be shown
2580 in the configuration summary.
2581
2582
2583 Version 7.8 05-Sep-08
2584 ---------------------
2585
2586 1. Replaced UCP searching code with optimized version as implemented for Ad
2587 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2588 stage table and inline lookup instead of a function, giving speed ups of 2
2589 to 5 times on some simple patterns that I tested. Permission was given to
2590 distribute the MultiStage2.py script that generates the tables (it's not in
2591 the tarball, but is in the Subversion repository).
2592
2593 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2594 scripts.
2595
2596 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2597 a group with a zero qualifier. The result of the study could be incorrect,
2598 or the function might crash, depending on the pattern.
2599
2600 4. Caseless matching was not working for non-ASCII characters in back
2601 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2602 It now works when Unicode Property Support is available.
2603
2604 5. In pcretest, an escape such as \x{de} in the data was always generating
2605 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2606 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2607 truncation.
2608
2609 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2610
2611 7. Added two (int) casts to pcregrep when printing the difference of two
2612 pointers, in case they are 64-bit values.
2613
2614 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2615 test 2 if it fails.
2616
2617 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2618 and a #define of that name to empty if it is not externally set. This is to
2619 allow users of MSVC to set it if necessary.
2620
2621 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2622 the convenience functions in the pcre_get.c source file.
2623
2624 11. An option change at the start of a pattern that had top-level alternatives
2625 could cause overwriting and/or a crash. This command provoked a crash in
2626 some environments:
2627
2628 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2629
2630 This potential security problem was recorded as CVE-2008-2371.
2631
2632 12. For a pattern where the match had to start at the beginning or immediately
2633 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2634 pcre_dfa_exec() could read past the end of the passed subject if there was
2635 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2636 pcretest so that it places the subject at the end of its malloc-ed buffer.
2637
2638 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2639 exec() might read past the end of the data buffer in UTF-8 mode.
2640
2641 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2642 the data contained the byte 0x85 as part of a UTF-8 character within its
2643 first line. This applied both to normal and DFA matching.
2644
2645 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2646 /^[^d]*?$/8 failed to match "abc".
2647
2648 16. Added a missing copyright notice to pcrecpp_internal.h.
2649
2650 17. Make it more clear in the documentation that values returned from
2651 pcre_exec() in ovector are byte offsets, not character counts.
2652
2653 18. Tidied a few places to stop certain compilers from issuing warnings.
2654
2655 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2656 supplied by Stefan Weber. I made a further small update for 7.8 because
2657 there is a change of source arrangements: the pcre_searchfuncs.c module is
2658 replaced by pcre_ucd.c.
2659
2660
2661 Version 7.7 07-May-08
2662 ---------------------
2663
2664 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2665 a string to a long long, pretend we don't even have a long long." This is
2666 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2667
2668 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2669 pre-7.6 versions, which defined a global no_arg variable instead of putting
2670 it in the RE class. (See also #8 below.)
2671
2672 3. Remove a line of dead code, identified by coverity and reported by Nuno
2673 Lopes.
2674
2675 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2676
2677 (1) The include/exclude patterns were being applied to the whole pathnames
2678 of files, instead of just to the final components.
2679
2680 (2) If there was more than one level of directory, the subdirectories were
2681 skipped unless they satisfied the include/exclude conditions. This is
2682 inconsistent with GNU grep (and could even be seen as contrary to the
2683 pcregrep specification - which I improved to make it absolutely clear).
2684 The action now is always to scan all levels of directory, and just
2685 apply the include/exclude patterns to regular files.
2686
2687 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2688 --exclude_dir in the tests to avoid scanning .svn directories.
2689
2690 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2691 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2692 doesn't support NULs in patterns.
2693
2694 7. Added some missing "const"s to declarations of static tables in
2695 pcre_compile.c and pcre_dfa_exec.c.
2696
2697 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2698 caused by fix #2 above. (Subsequently also a second patch to fix the
2699 first patch. And a third patch - this was a messy problem.)
2700
2701 9. Applied Craig's patch to remove the use of push_back().
2702
2703 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2704 matching function regexec().
2705
2706 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2707 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2708 references. PCRE supports relative numbers with this syntax (I don't think
2709 Oniguruma does).
2710
2711 12. Previously, a group with a zero repeat such as (...){0} was completely
2712 omitted from the compiled regex. However, this means that if the group
2713 was called as a subroutine from elsewhere in the pattern, things went wrong
2714 (an internal error was given). Such groups are now left in the compiled
2715 pattern, with a new opcode that causes them to be skipped at execution
2716 time.
2717
2718 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2719 to the way PCRE behaves:
2720
2721 (a) A lone ] character is dis-allowed (Perl treats it as data).
2722
2723 (b) A back reference to an unmatched subpattern matches an empty string
2724 (Perl fails the current match path).
2725
2726 (c) A data ] in a character class must be notated as \] because if the
2727 first data character in a class is ], it defines an empty class. (In
2728 Perl it is not possible to have an empty class.) The empty class []
2729 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2730 The negative empty class [^] matches any one character, independently
2731 of the DOTALL setting.
2732
2733 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2734 non-existent subpattern following a character class starting with ']' and
2735 containing () gave an internal compiling error instead of "reference to
2736 non-existent subpattern". Fortunately, when the pattern did exist, the
2737 compiled code was correct. (When scanning forwards to check for the
2738 existence of the subpattern, it was treating the data ']' as terminating
2739 the class, so got the count wrong. When actually compiling, the reference
2740 was subsequently set up correctly.)
2741
2742 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2743 it was being rejected as not supported by pcre_dfa_exec(), even though
2744 other assertions are supported. I have made pcre_dfa_exec() support
2745 (*FAIL).
2746
2747 16. The implementation of 13c above involved the invention of a new opcode,
2748 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2749 cannot be changed at match time, I realized I could make a small
2750 improvement to matching performance by compiling OP_ALLANY instead of
2751 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2752 on the OP_ANY path.
2753
2754 17. Compiling pcretest on Windows with readline support failed without the
2755 following two fixes: (1) Make the unistd.h include conditional on
2756 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2757
2758 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2759 ncurses library to be included for pcretest when ReadLine support is
2760 requested, but also to allow for it to be overridden. This patch came from
2761 Daniel Bergström.
2762
2763 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2764 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2765 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2766 spotting this.
2767
2768
2769 Version 7.6 28-Jan-08
2770 ---------------------
2771
2772 1. A character class containing a very large number of characters with
2773 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2774 overflow.
2775
2776 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2777 HAVE_LONG_LONG is not defined.
2778
2779 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2780 bring them up to date and include new features. This patch includes:
2781
2782 - Fixed PH's badly added libz and libbz2 support.
2783 - Fixed a problem with static linking.
2784 - Added pcredemo. [But later removed - see 7 below.]
2785 - Fixed dftables problem and added an option.
2786 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2787 HAVE_LONG_LONG.
2788 - Added readline support for pcretest.
2789 - Added an listing of the option settings after cmake has run.
2790
2791 4. A user submitted a patch to Makefile that makes it easy to create
2792 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2793 Makefile.am that cause it to include this special target, without
2794 affecting anything else. Note that the same mingw target plus all
2795 the other distribution libraries and programs are now supported
2796 when configuring with CMake (see 6 below) instead of with
2797 Configure/Make.
2798
2799 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2800 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2801 exported in the Windows port". It has not yet been confirmed that the patch
2802 solves the problem, but it does no harm.
2803
2804 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2805 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2806 with CMake, and also correct the comment about stack recursion.
2807
2808 7. Remove the automatic building of pcredemo from the ./configure system and
2809 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2810 of a program that users should build themselves after PCRE is installed, so
2811 building it automatically is not really right. What is more, it gave
2812 trouble in some build environments.
2813
2814 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2815
2816
2817 Version 7.5 10-Jan-08
2818 ---------------------
2819
2820 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2821 values in parens when parsing an RE using the C++ wrapper."
2822
2823 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2824 Characters greater than 255 were excluded from the class instead of being
2825 included.
2826
2827 3. The same bug as (2) above applied to negated POSIX classes such as
2828 [:^space:].
2829
2830 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2831 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2832 I have changed it.
2833
2834 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2835 first named subpattern) and a construct such as (?&a) would reference the
2836 first named subpattern whose name started with "a" (in other words, the
2837 length check was missing). Both these problems are fixed. "Subpattern name
2838 expected" is now given for (?&) (a zero-length name), and this patch also
2839 makes it give the same error for \k'' (previously it complained that that
2840 was a reference to a non-existent subpattern).
2841
2842 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2843 this is right because (?- can be followed by option settings as well as by
2844 digits. I have, however, made the messages clearer.
2845
2846 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2847 than the number used in the conditional) now cause a compile-time error.
2848 This is actually not compatible with Perl, which accepts such patterns, but
2849 treats the conditional as always being FALSE (as PCRE used to), but it
2850 seems to me that giving a diagnostic is better.
2851
2852 8. Change "alphameric" to the more common word "alphanumeric" in comments
2853 and messages.
2854
2855 9. Fix two occurrences of "backslash" in comments that should have been
2856 "backspace".
2857
2858 10. Remove two redundant lines of code that can never be obeyed (their function
2859 was moved elsewhere).
2860
2861 11. The program that makes PCRE's Unicode character property table had a bug
2862 which caused it to generate incorrect table entries for sequences of
2863 characters that have the same character type, but are in different scripts.
2864 It amalgamated them into a single range, with the script of the first of
2865 them. In other words, some characters were in the wrong script. There were
2866 thirteen such cases, affecting characters in the following ranges:
2867
2868 U+002b0 - U+002c1
2869 U+0060c - U+0060d
2870 U+0061e - U+00612
2871 U+0064b - U+0065e
2872 U+0074d - U+0076d
2873 U+01800 - U+01805
2874 U+01d00 - U+01d77
2875 U+01d9b - U+01dbf
2876 U+0200b - U+0200f
2877 U+030fc - U+030fe
2878 U+03260 - U+0327f
2879 U+0fb46 - U+0fbb1
2880 U+10450 - U+1049d
2881
2882 12. The -o option (show only the matching part of a line) for pcregrep was not
2883 compatible with GNU grep in that, if there was more than one match in a
2884 line, it showed only the first of them. It now behaves in the same way as
2885 GNU grep.
2886
2887 13. If the -o and -v options were combined for pcregrep, it printed a blank
2888 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2889 does the same. The return code can be used to tell if there were any
2890 non-matching lines.
2891
2892 14. Added --file-offsets and --line-offsets to pcregrep.
2893
2894 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2895 infinitely looping recursion. The bug was that positive lookaheads were not
2896 being skipped when checking for a possible empty match (negative lookaheads
2897 and both kinds of lookbehind were skipped).
2898
2899 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2900 inclusion of <windows.h> to before rather than after the definition of
2901 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2902
2903 17. Specifying a possessive quantifier with a specific limit for a Unicode
2904 character property caused pcre_compile() to compile bad code, which led at
2905 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2906 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2907 caused the error; without that there was no problem.
2908
2909 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2910
2911 19. Added --enable-pcretest-libreadline.
2912
2913 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2914 RE::GlobalReplace(). As a result, the number of replacements returned was
2915 double what it should be. I removed one of the increments, but Craig sent a
2916 later patch that removed the other one (the right fix) and added unit tests
2917 that check the return values (which was not done before).
2918
2919 21. Several CMake things:
2920
2921 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2922 the names libpcre and libpcreposix, not just pcre and pcreposix.
2923
2924 (2) The above change means that pcretest and pcregrep are now correctly
2925 linked with the newly-built libraries, not previously installed ones.
2926
2927 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2928
2929 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2930 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2931 UTF-8 newline character). The key issue is that the pattern starts .*;
2932 this means that the match must be either at the beginning, or after a
2933 newline. The bug was in the code for advancing after a failed match and
2934 checking that the new position followed a newline. It was not taking
2935 account of UTF-8 characters correctly.
2936
2937 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2938 character classes. PCRE was not treating the sequence [:...:] as a
2939 character class unless the ... were all letters. Perl, however, seems to
2940 allow any characters between [: and :], though of course it rejects as
2941 unknown any "names" that contain non-letters, because all the known class
2942 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2943 for example, whereas PCRE did not - it did not recognize a POSIX character
2944 class. This seemed a bit dangerous, so the code has been changed to be
2945 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2946 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2947 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2948 Perl does, and where it didn't before.
2949
2950 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2951 Windows environments %n is disabled by default.
2952
2953
2954 Version 7.4 21-Sep-07
2955 ---------------------
2956
2957 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2958 means that a class such as [\s] counted as "explicit reference to CR or
2959 LF". That isn't really right - the whole point of the change was to try to
2960 help when there was an actual mention of one of the two characters. So now
2961 the change happens only if \r or \n (or a literal CR or LF) character is
2962 encountered.
2963
2964 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2965 of both had grown to the point where there were only 3 bits left.
2966 Fortunately, there was spare space in the data structure, and so I have
2967 moved the internal flags into a new 16-bit field to free up more option
2968 bits.
2969
2970 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2971 but did not set the internal JCHANGED flag - either of these is enough to
2972 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2973 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2974 start sets both bits.
2975
2976 4. Added options (at build time, compile time, exec time) to change \R from
2977 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2978
2979 5. doc/pcresyntax.html was missing from the distribution.
2980
2981 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2982 compatibility, even though it is no longer used.
2983
2984 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2985 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2986 windows.h file is present (where different names are used). [This was
2987 reversed later after testing - see 16 below.]
2988
2989 8. Changed all #include <config.h> to #include "config.h". There were also
2990 some further <pcre.h> cases that I changed to "pcre.h".
2991
2992 9. When pcregrep was used with the --colour option, it missed the line ending
2993 sequence off the lines that it output.
2994
2995 10. It was pointed out to me that arrays of string pointers cause lots of
2996 relocations when a shared library is dynamically loaded. A technique of
2997 using a single long string with a table of offsets can drastically reduce
2998 these. I have refactored PCRE in four places to do this. The result is
2999 dramatic:
3000
3001 Originally: 290
3002 After changing UCP table: 187
3003 After changing error message table: 43
3004 After changing table of "verbs" 36
3005 After changing table of Posix names 22
3006
3007 Thanks to the folks working on Gregex for glib for this insight.
3008
3009 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3010 unicode-properties was also set.
3011
3012 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3013
3014 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3015 checked only for CRLF.
3016
3017 14. Added casts to pcretest.c to avoid compiler warnings.
3018
3019 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3020
3021 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3022 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3023 entirely. This removes changes made in 7 above.
3024
3025 17. The CMake files have been updated, and there is now more information about
3026 building with CMake in the NON-UNIX-USE document.
3027
3028
3029 Version 7.3 28-Aug-07
3030 ---------------------
3031
3032 1. In the rejigging of the build system that eventually resulted in 7.1, the
3033 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3034 brackets there is not right, since it causes compilers to look for an
3035 installed pcre.h, not the version that is in the source that is being
3036 compiled (which of course may be different). I have changed it back to:
3037
3038 #include "pcre.h"
3039
3040 I have a vague recollection that the change was concerned with compiling in
3041 different directories, but in the new build system, that is taken care of
3042 by the VPATH setting the Makefile.
3043
3044 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3045 when the subject happened to end in the byte 0x85 (e.g. if the last
3046 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3047 characters but of course it shouldn't be taken as a newline when it is part
3048 of another character. The bug was that, for an unlimited repeat of . in
3049 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3050 characters when looking for a newline.
3051
3052 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3053
3054 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3055 in debug output.
3056
3057 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3058 long printing in the pcrecpp unittest when running under MinGW.
3059
3060 6. ESC_K was left out of the EBCDIC table.
3061
3062 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3063 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3064 limit also applies to "virtual nesting" when a pattern is recursive, and in
3065 this case 1000 isn't so big. I have been able to remove this limit at the
3066 expense of backing off one optimization in certain circumstances. Normally,
3067 when pcre_exec() would call its internal match() function recursively and
3068 immediately return the result unconditionally, it uses a "tail recursion"
3069 feature to save stack. However, when a subpattern that can match an empty
3070 string has an unlimited repetition quantifier, it no longer makes this
3071 optimization. That gives it a stack frame in which to save the data for
3072 checking that an empty string has been matched. Previously this was taken
3073 from the 1000-entry workspace that had been reserved. So now there is no
3074 explicit limit, but more stack is used.
3075
3076 8. Applied Daniel's patches to solve problems with the import/export magic
3077 syntax that is required for Windows, and which was going wrong for the
3078 pcreposix and pcrecpp parts of the library. These were overlooked when this
3079 problem was solved for the main library.
3080
3081 9. There were some crude static tests to avoid integer overflow when computing
3082 the size of patterns that contain repeated groups with explicit upper
3083 limits. As the maximum quantifier is 65535, the maximum group length was
3084 set at 30,000 so that the product of these two numbers did not overflow a
3085 32-bit integer. However, it turns out that people want to use groups that
3086 are longer than 30,000 bytes (though not repeat them that many times).
3087 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3088 made it possible to implement the integer overflow checks in a much more
3089 dynamic way, which I have now done. The artificial limitation on group
3090 length has been removed - we now have only the limit on the total length of
3091 the compiled pattern, which depends on the LINK_SIZE setting.
3092
3093 10. Fixed a bug in the documentation for get/copy named substring when
3094 duplicate names are permitted. If none of the named substrings are set, the
3095 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3096 empty string.
3097
3098 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3099 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3100 because the ] is interpreted as the first data character and the
3101 terminating ] is not found. PCRE has been made compatible with Perl in this
3102 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3103 cause memory overwriting.
3104
3105 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3106 string has been matched (to stop an infinite loop). It was not recognizing
3107 a conditional subpattern that could match an empty string if that
3108 subpattern was within another subpattern. For example, it looped when
3109 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3110 condition was not nested. This bug has been fixed.
3111
3112 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3113 past the start of the subject in the presence of bytes with the top bit
3114 set, for example "\x8aBCD".
3115
3116 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3117 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3118
3119 14. Optimized (?!) to (*FAIL).
3120
3121 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3122 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3123 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3124 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3125 does: it's just the validity check that is more restrictive.
3126
3127 16. Inserted checks for integer overflows during escape sequence (backslash)
3128 processing, and also fixed erroneous offset values for syntax errors during
3129 backslash processing.
3130
3131 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3132 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3133
3134 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3135 caused an overrun.
3136
3137 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3138 something other than just ASCII characters) inside a group that had an
3139 unlimited repeat caused a loop at compile time (while checking to see
3140 whether the group could match an empty string).
3141
3142 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3143 [\P{Any}] did so. (Error in the code for printing property names.)
3144
3145 21. An orphan \E inside a character class could cause a crash.
3146
3147 22. A repeated capturing bracket such as (A)? could cause a wild memory
3148 reference during compilation.
3149
3150 23. There are several functions in pcre_compile() that scan along a compiled
3151 expression for various reasons (e.g. to see if it's fixed length for look
3152 behind). There were bugs in these functions when a repeated \p or \P was
3153 present in the pattern. These operators have additional parameters compared
3154 with \d, etc, and these were not being taken into account when moving along
3155 the compiled data. Specifically:
3156
3157 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3158 length.
3159
3160 (b) An item such as \pL+ within a repeated group could cause crashes or
3161 loops.
3162
3163 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3164 "reference to non-existent subpattern" error.
3165
3166 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3167
3168 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3169 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3170
3171 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3172
3173 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3174 character were causing crashes (broken optimization).
3175
3176 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3177 \p or \P) caused a compile-time loop.
3178
3179 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3180 break. For example, the unstudied pattern [\r\n]A does not match the string
3181 "\r\nA" because change 7.0/46 below moves the current point on by two
3182 characters after failing to match at the start. However, the pattern \nA
3183 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3184 the same is true. There doesn't seem any very clean way out of this, but
3185 what I have chosen to do makes the common cases work: PCRE now takes note
3186 of whether there can be an explicit match for \r or \n anywhere in the
3187 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3188 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3189 pattern has explicit CR or LF references.
3190
3191 29. Added (*CR) etc for changing newline setting at start of pattern.
3192
3193
3194 Version 7.2 19-Jun-07
3195 ---------------------
3196
3197 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3198 which is apparently normally available under Windows.
3199
3200 2. Re-jig the pcregrep tests with different newline settings in an attempt
3201 to make them independent of the local environment's newline setting.
3202
3203 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3204
3205 4. Some of the "internals" tests were previously cut out when the link size
3206 was not 2, because the output contained actual offsets. The recent new
3207 "Z" feature of pcretest means that these can be cut out, making the tests
3208 usable with all link sizes.
3209
3210 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3211 stack recursion. This gives a massive performance boost under BSD, but just
3212 a small improvement under Linux. However, it saves one field in the frame
3213 in all cases.
3214
3215 6. Added more features from the forthcoming Perl 5.10:
3216
3217 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3218 recursion call. It refers to the nth most recently opened parentheses.
3219
3220 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3221 to be opened parentheses.
3222
3223 (c) Conditions that refer to capturing parentheses can be specified
3224 relatively, for example, (?(-2)... or (?(+3)...
3225
3226 (d) \K resets the start of the current match so that everything before
3227 is not part of it.
3228
3229 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3230
3231 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3232 reference syntax.
3233
3234 (g) (?| introduces a group in which the numbering of parentheses in each
3235 alternative starts with the same number.
3236
3237 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3238
3239 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3240 PCRE_INFO_JCHANGED.
3241
3242 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3243 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3244 for detecting groups that can match an empty string.
3245
3246 9. A pattern with a very large number of alternatives (more than several
3247 hundred) was running out of internal workspace during the pre-compile
3248 phase, where pcre_compile() figures out how much memory will be needed. A
3249 bit of new cunning has reduced the workspace needed for groups with
3250 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3251 workspace instead of running out of the 4096 that are available.
3252
3253 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3254
3255 11. Applied patch from Google to remove an optimization that didn't quite work.
3256 The report of the bug said:
3257
3258 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3259 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3260 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3261
3262 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3263 it matched the wrong number of bytes.
3264
3265
3266 Version 7.1 24-Apr-07
3267 ---------------------
3268
3269 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3270 that is more "standard", making use of automake and other Autotools. There
3271 is some re-arrangement of the files and adjustment of comments consequent
3272 on this.
3273
3274 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3275 for recursive directory scanning broke on some systems because the files
3276 are not scanned in any specific order and on different systems the order
3277 was different. A call to "sort" has been inserted into RunGrepTest for the
3278 approprate test as a short-term fix. In the longer term there may be an
3279 alternative.
3280
3281 3. I had an email from Eric Raymond about problems translating some of PCRE's
3282 man pages to HTML (despite the fact that I distribute HTML pages, some
3283 people do their own conversions for various reasons). The problems
3284 concerned the use of low-level troff macros .br and .in. I have therefore
3285 removed all such uses from the man pages (some were redundant, some could
3286 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3287 HTML has been updated to handle .nf/.fi and to complain if it encounters
3288 .br or .in.
3289
3290 4. Updated comments in configure.ac that get placed in config.h.in and also
3291 arranged for config.h to be included in the distribution, with the name
3292 config.h.generic, for the benefit of those who have to compile without
3293 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3294
3295 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3296 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3297 makevp.bat for the current PCRE, using the additional files
3298 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3299
3300 6. A Windows user reported a minor discrepancy with test 2, which turned out
3301 to be caused by a trailing space on an input line that had got lost in his
3302 copy. The trailing space was an accident, so I've just removed it.
3303
3304 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3305 that is needed.
3306
3307 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3308 as "const" (a) because they are and (b) because it helps the PHP
3309 maintainers who have recently made a script to detect big data structures
3310 in the php code that should be moved to the .rodata section. I remembered
3311 to update Builducptable as well, so it won't revert if ucptable.h is ever
3312 re-created.
3313
3314 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3315 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3316 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3317 support is not required. This saves 1.5-2K of code, which is important in
3318 some applications.
3319
3320 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3321 so as not to refer to the tables, even though these functions will never be
3322 called when UTF-8 support is disabled. Otherwise there are problems with a
3323 shared library.
3324
3325 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3326
3327 (a) It was defining its arguments as char * instead of void *.
3328
3329 (b) It was assuming that all moves were upwards in memory; this was true
3330 a long time ago when I wrote it, but is no longer the case.
3331
3332 The emulated memove() is provided for those environments that have neither
3333 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3334 is clearly not the case, as these two bugs were recently reported.
3335
3336 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3337 and Detrail to create the HTML documentation, the .txt form of the man
3338 pages, and it removes trailing spaces from listed files. It also creates
3339 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3340 case, it wraps all the #defines with #ifndefs. This script should be run
3341 before "make dist".
3342
3343 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3344 with Unicode property support.
3345
3346 (a) For a maximizing quantifier, if the two different cases of the
3347 character were of different lengths in their UTF-8 codings (there are
3348 some cases like this - I found 11), and the matching function had to
3349 back up over a mixture of the two cases, it incorrectly assumed they
3350 were both the same length.
3351
3352 (b) When PCRE was configured to use the heap rather than the stack for
3353 recursion during matching, it was not correctly preserving the data for
3354 the other case of a UTF-8 character when checking ahead for a match
3355 while processing a minimizing repeat. If the check also involved
3356 matching a wide character, but failed, corruption could cause an
3357 erroneous result when trying to check for a repeat of the original
3358 character.
3359
3360 13. Some tidying changes to the testing mechanism:
3361
3362 (a) The RunTest script now detects the internal link size and whether there
3363 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3364 values substituted by "configure". (The RunGrepTest script already did
3365 this for UTF-8.) The configure.ac script no longer substitutes the
3366 relevant variables.
3367
3368 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3369 with length and offset values. This means that the output is different
3370 for different internal link sizes. Test 2 is skipped for link sizes
3371 other than 2 because of this, bypassing the problem. Unfortunately,
3372 there was also a test in test 3 (the locale tests) that used /B and
3373 failed for link sizes other than 2. Rather than cut the whole test out,
3374 I have added a new /Z option to pcretest that replaces the length and
3375 offset values with spaces. This is now used to make test 3 independent
3376 of link size. (Test 2 will be tidied up later.)
3377
3378 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3379 segmentation fault instead of returning the appropriate error message.
3380
3381 15. In multiline mode when the newline sequence was set to "any", the pattern
3382 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3383 This doesn't seem right; it now treats the CRLF combination as the line
3384 ending, and so does not match in that case. It's only a pattern such as ^$
3385 that would hit this one: something like ^ABC$ would have failed after \r
3386 and then tried again after \r\n.
3387
3388 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3389 in an attempt to make files that differ only in their line terminators
3390 compare equal. This works on Linux.
3391
3392 17. Under certain error circumstances pcregrep might try to free random memory
3393 as it exited. This is now fixed, thanks to valgrind.
3394
3395 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3396 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3397 was because its rules for how to advance for /g after matching an empty
3398 string at the end of a line did not allow for this case. They now check for
3399 it specially.
3400
3401 20. pcretest is supposed to handle patterns and data of any length, by
3402 extending its buffers when necessary. It was getting this wrong when the
3403 buffer for a data line had to be extended.
3404
3405 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3406 CRLF as a newline sequence.
3407
3408 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3409 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3410 I have nevertheless tidied it up.
3411
3412 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3413
3414 24. Added a man page for pcre-config.
3415
3416
3417 Version 7.0 19-Dec-06
3418 ---------------------
3419
3420 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3421 moving to gcc 4.1.1.
3422
3423 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3424 sys/time.h, which is documented as needed for this function. It doesn't
3425 seem to matter on Linux, but it showed up on some releases of OS X.
3426
3427 3. It seems that there are systems where bytes whose values are greater than
3428 127 match isprint() in the "C" locale. The "C" locale should be the
3429 default when a C program starts up. In most systems, only ASCII printing
3430 characters match isprint(). This difference caused the output from pcretest
3431 to vary, making some of the tests fail. I have changed pcretest so that:
3432
3433 (a) When it is outputting text in the compiled version of a pattern, bytes
3434 other than 32-126 are always shown as hex escapes.
3435
3436 (b) When it is outputting text that is a matched part of a subject string,
3437 it does the same, unless a different locale has been set for the match
3438 (using the /L modifier). In this case, it uses isprint() to decide.
3439
3440 4. Fixed a major bug that caused incorrect computation of the amount of memory
3441 required for a compiled pattern when options that changed within the
3442 pattern affected the logic of the preliminary scan that determines the
3443 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3444 that the computed length was too small. The symptoms of this bug were
3445 either the PCRE error "internal error: code overflow" from pcre_compile(),
3446 or a glibc crash with a message such as "pcretest: free(): invalid next
3447 size (fast)". Examples of patterns that provoked this bug (shown in
3448 pcretest format) are:
3449
3450 /(?-x: )/x
3451 /(?x)(?-x: \s*#\s*)/
3452 /((?i)[\x{c0}])/8
3453 /(?i:[\x{c0}])/8
3454
3455 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3456 is now done differently.
3457
3458 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3459 wrapper classes; (b) implement a new function in the C++ scanner that is
3460 more efficient than the old way of doing things because it avoids levels of
3461 recursion in the regex matching; (c) add a paragraph to the documentation
3462 for the FullMatch() function.
3463
3464 6. The escape sequence \n was being treated as whatever was defined as
3465 "newline". Not only was this contrary to the documentation, which states
3466 that \n is character 10 (hex 0A), but it also went horribly wrong when
3467 "newline" was defined as CRLF. This has been fixed.
3468
3469 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3470 was being set to -1 for the "end of line" case (supposedly a value that no
3471 character can have). Though this value is never used (the check for end of
3472 line is "zero bytes in current character"), it caused compiler complaints.
3473 I've changed it to 0xffffffff.
3474
3475 8. In pcre_version.c, the version string was being built by a sequence of
3476 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3477 string (as it is for production releases) called a macro with an empty
3478 argument. The C standard says the result of this is undefined. The gcc
3479 compiler treats it as an empty string (which was what was wanted) but it is
3480 reported that Visual C gives an error. The source has been hacked around to
3481 avoid this problem.
3482
3483 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3484 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3485 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3486 of them did).
3487
3488 10. Originally, pcretest opened its input and output without "b"; then I was
3489 told that "b" was needed in some environments, so it was added for release
3490 5.0 to both the input and output. (It makes no difference on Unix-like
3491 systems.) Later I was told that it is wrong for the input on Windows. I've
3492 now abstracted the modes into two macros, to make it easier to fiddle with
3493 them, and removed "b" from the input mode under Windows.
3494
3495 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3496
3497 12. Added -help and --help to pcretest as an official way of being reminded
3498 of the options.
3499
3500 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3501 and pcrecpp.cc because they annoy compilers at high warning levels.
3502
3503 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3504
3505 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3506 scripts are not C programs :-) and which was not noticed because it works
3507 on Linux.
3508
3509 16. pcretest is supposed to handle any length of pattern and data line (as one
3510 line or as a continued sequence of lines) by extending its input buffer if
3511 necessary. This feature was broken for very long pattern lines, leading to
3512 a string of junk being passed to pcre_compile() if the pattern was longer
3513 than about 50K.
3514
3515 17. I have done a major re-factoring of the way pcre_compile() computes the
3516 amount of memory needed for a compiled pattern. Previously, there was code
3517 that made a preliminary scan of the pattern in order to do this. That was
3518 OK when PCRE was new, but as the facilities have expanded, it has become
3519 harder and harder to keep it in step with the real compile phase, and there
3520 have been a number of bugs (see for example, 4 above). I have now found a
3521 cunning way of running the real compile function in a "fake" mode that
3522 enables it to compute how much memory it would need, while actually only
3523 ever using a few hundred bytes of working memory and without too many
3524 tests of the mode. This should make future maintenance and development
3525 easier. A side effect of this work is that the limit of 200 on the nesting
3526 depth of parentheses has been removed (though this was never a serious
3527 limitation, I suspect). However, there is a downside: pcre_compile() now
3528 runs more slowly than before (30% or more, depending on the pattern). I
3529 hope this isn't a big issue. There is no effect on runtime performance.
3530
3531 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3532 newline (only possible for the last line of a file) and it was a
3533 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3534
3535 19. Added additional timing features to pcretest. (1) The -tm option now times
3536 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3537 separate command line item, by a number that specifies the number of
3538 repeats to use when timing. The default is 50000; this gives better
3539 precision, but takes uncomfortably long for very large patterns.
3540
3541 20. Extended pcre_study() to be more clever in cases where a branch of a
3542 subpattern has no definite first character. For example, (a*|b*)[cd] would
3543 previously give no result from pcre_study(). Now it recognizes that the
3544 first character must be a, b, c, or d.
3545
3546 21. There was an incorrect error "recursive call could loop indefinitely" if
3547 a subpattern (or the entire pattern) that was being tested for matching an
3548 empty string contained only one non-empty item after a nested subpattern.
3549 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3550 incorrectly, because the \d was being skipped in the check.
3551
3552 22. The pcretest program now has a new pattern option /B and a command line
3553 option -b, which is equivalent to adding /B to every pattern. This causes
3554 it to show the compiled bytecode, without the additional information that
3555 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3556 is the same as /B/I).
3557
3558 23. A new optimization is now able automatically to treat some sequences such
3559 as a*b as a*+b. More specifically, if something simple (such as a character
3560 or a simple class like \d) has an unlimited quantifier, and is followed by
3561 something that cannot possibly match the quantified thing, the quantifier
3562 is automatically "possessified".
3563
3564 24. A recursive reference to a subpattern whose number was greater than 39
3565 went wrong under certain circumstances in UTF-8 mode. This bug could also
3566 have affected the operation of pcre_study().
3567
3568 25. Realized that a little bit of performance could be had by replacing
3569 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3570
3571 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3572
3573 27. Possessive quantifiers such as a++ were previously implemented by turning
3574 them into atomic groups such as ($>a+). Now they have their own opcodes,
3575 which improves performance. This includes the automatically created ones
3576 from 23 above.
3577
3578 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3579 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3580 the first matched character to be a colon. This applied both to named and
3581 numbered groups.
3582
3583 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3584
3585 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3586 building PCRE on Windows easier, so I have included it in the distribution.
3587
3588 31. There is now a check in pcretest against a ridiculously large number being
3589 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3590 loop, the loop is abandoned.
3591
3592 32. Forward references to subpatterns in conditions such as (?(2)...) where
3593 subpattern 2 is defined later cause pcre_compile() to search forwards in
3594 the pattern for the relevant set of parentheses. This search went wrong
3595 when there were unescaped parentheses in a character class, parentheses
3596 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3597
3598 33. "Subroutine" calls and backreferences were previously restricted to
3599 referencing subpatterns earlier in the regex. This restriction has now
3600 been removed.
3601
3602 34. Added a number of extra features that are going to be in Perl 5.10. On the
3603 whole, these are just syntactic alternatives for features that PCRE had
3604 previously implemented using the Python syntax or my own invention. The
3605 other formats are all retained for compatibility.
3606
3607 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3608 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3609 also .NET compatible.
3610
3611 (b) A recursion or subroutine call to a named group can now be defined as
3612 (?&name) as well as (?P>name).
3613
3614 (c) A backreference to a named group can now be defined as \k<name> or
3615 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3616 5.10, are also .NET compatible.
3617
3618 (d) A conditional reference to a named group can now use the syntax
3619 (?(<name>) or (?('name') as well as (?(name).
3620
3621 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3622 groups (named and numbered) that are never evaluated inline, but can be
3623 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3624 is always false. There may be only one alternative in such a group.
3625
3626 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3627 as the simple (?(R). The condition is true only if the most recent
3628 recursion is that of the given number or name. It does not search out
3629 through the entire recursion stack.
3630
3631 (g) The escape \gN or \g{N} has been added, where N is a positive or
3632 negative number, specifying an absolute or relative reference.
3633
3634 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3635 some "unreachable code" warnings.
3636
3637 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3638 things, this adds five new scripts.
3639
3640 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3641 There were also incompatibilities regarding the handling of \Q..\E inside
3642 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3643 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3644
3645 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3646 matches an empty string, and forcibly breaks the loop. There were bugs in
3647 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3648 against aaaa the result was just "a" rather than "aaaa", for example. Two
3649 separate and independent bugs (that affected different cases) have been
3650 fixed.
3651
3652 39. Refactored the code to abolish the use of different opcodes for small
3653 capturing bracket numbers. This is a tidy that I avoided doing when I
3654 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3655 The new approach is not only tidier, it makes it possible to reduce the
3656 memory needed to fix the previous bug (38).
3657
3658 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3659 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3660 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3661 mode.
3662
3663 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3664 report.
3665
3666 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3667 copy construction and assignment in the C++ wrapper.
3668
3669 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3670 couple of bugs that could have given wrong results in the "--newline=crlf"
3671 case.
3672
3673 44. Added a number of casts and did some reorganization of signed/unsigned int
3674 variables following suggestions from Dair Grant. Also renamed the variable
3675 "this" as "item" because it is a C++ keyword.
3676
3677 45. Arranged for dftables to add
3678
3679 #include "pcre_internal.h"
3680
3681 to pcre_chartables.c because without it, gcc 4.x may remove the array
3682 definition from the final binary if PCRE is built into a static library and
3683 dead code stripping is activated.
3684
3685 46. For an unanchored pattern, if a match attempt fails at the start of a
3686 newline sequence, and the newline setting is CRLF or ANY, and the next two
3687 characters are CRLF, advance by two characters instead of one.
3688
3689
3690 Version 6.7 04-Jul-06
3691 ---------------------
3692
3693 1. In order to handle tests when input lines are enormously long, pcretest has
3694 been re-factored so that it automatically extends its buffers when
3695 necessary. The code is crude, but this _is_ just a test program. The
3696 default size has been increased from 32K to 50K.
3697
3698 2. The code in pcre_study() was using the value of the re argument before
3699 testing it for NULL. (Of course, in any sensible call of the function, it
3700 won't be NULL.)
3701
3702 3. The memmove() emulation function in pcre_internal.h, which is used on
3703 systems that lack both memmove() and bcopy() - that is, hardly ever -
3704 was missing a "static" storage class specifier.
3705
3706 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3707 containing an extended class (one that cannot be represented by a bitmap
3708 because it contains high-valued characters or Unicode property items, e.g.
3709 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3710 pattern, but PCRE should not loop if you do not (it no longer does).
3711 [Detail: two cases were found: (a) a repeated subpattern containing an
3712 extended class; (b) a recursive reference to a subpattern that followed a
3713 previous extended class. It wasn't skipping over the extended class
3714 correctly when UTF-8 mode was not set.]
3715
3716 5. A negated single-character class was not being recognized as fixed-length
3717 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3718 compile error "lookbehind assertion is not fixed length".
3719
3720 6. The RunPerlTest auxiliary script was showing an unexpected difference
3721 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3722 write a Perl script that can interpret lines of an input file either as
3723 byte characters or as UTF-8, which is what "perltest" was being required to
3724 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3725 can't do is switch easily at run time between having the "use utf8;" pragma
3726 or not. In the end, I fudged it by using the RunPerlTest script to insert
3727 "use utf8;" explicitly for the UTF-8 tests.
3728
3729 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3730 the end of the subject string, contrary to the documentation and to what
3731 Perl does. This was true of both matching functions. Now it matches only at
3732 the start of the subject and immediately after *internal* newlines.
3733
3734 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3735 a pointer to an int instead of a pointer to an unsigned long int. This
3736 caused problems on 64-bit systems.
3737
3738 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3739 instance of the 'standard' template library not being so standard".
3740
3741 10. There was no check on the number of named subpatterns nor the maximum
3742 length of a subpattern name. The product of these values is used to compute
3743 the size of the memory block for a compiled pattern. By supplying a very
3744 long subpattern name and a large number of named subpatterns, the size
3745 computation could be caused to overflow. This is now prevented by limiting
3746 the length of names to 32 characters, and the number of named subpatterns
3747 to 10,000.
3748
3749 11. Subpatterns that are repeated with specific counts have to be replicated in
3750 the compiled pattern. The size of memory for this was computed from the
3751 length of the subpattern and the repeat count. The latter is limited to
3752 65535, but there was no limit on the former, meaning that integer overflow
3753 could in principle occur. The compiled length of a repeated subpattern is
3754 now limited to 30,000 bytes in order to prevent this.
3755
3756 12. Added the optional facility to have named substrings with the same name.
3757
3758 13. Added the ability to use a named substring as a condition, using the
3759 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3760 are numbers (not recommended). Forward references are permitted.
3761
3762 14. Added forward references in named backreferences (if you see what I mean).
3763
3764 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3765 pattern could run off the end of the subject. For example, the pattern
3766 "(?s)(.{1,5})"8 did this with the subject "ab".
3767
3768 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3769 PCRE_CASELESS was set when matching characters that were quantified with ?
3770 or *.
3771
3772 17. A character class other than a single negated character that had a minimum
3773 but no maximum quantifier - for example [ab]{6,} - was not handled
3774 correctly by pce_dfa_exec(). It would match only one character.
3775
3776 18. A valid (though odd) pattern that looked like a POSIX character
3777 class but used an invalid character after [ (for example [[,abc,]]) caused
3778 pcre_compile() to give the error "Failed: internal error: code overflow" or
3779 in some cases to crash with a glibc free() error. This could even happen if
3780 the pattern terminated after [[ but there just happened to be a sequence of
3781 letters, a binary zero, and a closing ] in the memory that followed.
3782
3783 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3784 over the years. Originally (before any Unicode support), just the bottom 8
3785 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3786 output from "man perlunicode" includes this:
3787
3788 The regular expression compiler produces polymorphic opcodes. That
3789 is, the pattern adapts to the data and automatically switches to
3790 the Unicode character scheme when presented with Unicode data--or
3791 instead uses a traditional byte scheme when presented with byte
3792 data.
3793
3794 Sadly, a wide octal escape does not cause a switch, and in a string with
3795 no other multibyte characters, these octal escapes are treated as before.
3796 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3797 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3798 Unicode string.
3799
3800 I have not perpetrated such confusion in PCRE. Up till now, it took just
3801 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3802 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3803 translate to the appropriate multibyte character.
3804
3805 29. Applied some refactoring to reduce the number of warnings from Microsoft
3806 and Borland compilers. This has included removing the fudge introduced
3807 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3808 a warning about an unused variable.
3809
3810 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3811 characters since release 4.0, because Perl (from release 5.004) does not.
3812 [Or at least, is documented not to: some releases seem to be in conflict
3813 with the documentation.] However, when a pattern was studied with
3814 pcre_study() and all its branches started with \s, PCRE still included VT
3815 as a possible starting character. Of course, this did no harm; it just
3816 caused an unnecessary match attempt.
3817
3818 22. Removed a now-redundant internal flag bit that recorded the fact that case
3819 dependency changed within the pattern. This was once needed for "required
3820 byte" processing, but is no longer used. This recovers a now-scarce options
3821 bit. Also moved the least significant internal flag bit to the most-
3822 significant bit of the word, which was not previously used (hangover from
3823 the days when it was an int rather than a uint) to free up another bit for
3824 the future.
3825
3826 23. Added support for CRLF line endings as well as CR and LF. As well as the
3827 default being selectable at build time, it can now be changed at runtime
3828 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3829 specify that it is scanning data with non-default line endings.
3830
3831 24. Changed the definition of CXXLINK to make it agree with the definition of
3832 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3833
3834 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3835 recursions. This makes a big different to stack usage for some patterns.
3836
3837 26. If a subpattern containing a named recursion or subroutine reference such
3838 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3839 the space required for the compiled pattern went wrong and gave too small a
3840 value. Depending on the environment, this could lead to "Failed: internal
3841 error: code overflow at offset 49" or "glibc detected double free or
3842 corruption" errors.
3843
3844 27. Applied patches from Google (a) to support the new newline modes and (b) to
3845 advance over multibyte UTF-8 characters in GlobalReplace.
3846
3847 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3848 difference for some implementation of PCRE in some Windows version.
3849
3850 29. Added some extra testing facilities to pcretest:
3851
3852 \q<number> in a data line sets the "match limit" value
3853 \Q<number> in a data line sets the "match recursion limt" value
3854 -S <number> sets the stack size, where <number> is in megabytes
3855
3856 The -S option isn't available for Windows.
3857
3858
3859 Version 6.6 06-Feb-06
3860 ---------------------
3861
3862 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3863 in pcreposix.h. I have copied the definition from pcre.h.
3864
3865 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3866 because pcre.h is no longer a built file.
3867
3868 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3869 not normally included in the compiled code.
3870
3871
3872 Version 6.5 01-Feb-06
3873 ---------------------
3874
3875 1. When using the partial match feature with pcre_dfa_exec(), it was not
3876 anchoring the second and subsequent partial matches at the new starting
3877 point. This could lead to incorrect results. For example, with the pattern
3878 /1234/, partially matching against "123" and then "a4" gave a match.
3879
3880 2. Changes to pcregrep:
3881
3882 (a) All non-match returns from pcre_exec() were being treated as failures
3883 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3884 error message is output. Some extra information is given for the
3885 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3886 probably the only errors that are likely to be caused by users (by
3887 specifying a regex that has nested indefinite repeats, for instance).
3888 If there are more than 20 of these errors, pcregrep is abandoned.
3889
3890 (b) A binary zero was treated as data while matching, but terminated the
3891 output line if it was written out. This has been fixed: binary zeroes
3892 are now no different to any other data bytes.
3893
3894 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3895 used to set a locale for matching. The --locale=xxxx long option has
3896 been added (no short equivalent) to specify a locale explicitly on the
3897 pcregrep command, overriding the environment variables.
3898
3899 (d) When -B was used with -n, some line numbers in the output were one less
3900 than they should have been.
3901
3902 (e) Added the -o (--only-matching) option.
3903
3904 (f) If -A or -C was used with -c (count only), some lines of context were
3905 accidentally printed for the final match.
3906
3907 (g) Added the -H (--with-filename) option.
3908
3909 (h) The combination of options -rh failed to suppress file names for files
3910 that were found from directory arguments.
3911
3912 (i) Added the -D (--devices) and -d (--directories) options.
3913
3914 (j) Added the -F (--fixed-strings) option.
3915
3916 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3917
3918 (l) Added the --colo(u)r option.
3919
3920 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3921 is not present by default.
3922
3923 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3924 items such as (?R) or (?1), when the recursion could match a number of
3925 alternatives. If it matched one of the alternatives, but subsequently,
3926 outside the recursion, there was a failure, the code tried to back up into
3927 the recursion. However, because of the way PCRE is implemented, this is not
3928 possible, and the result was an incorrect result from the match.
3929
3930 In order to prevent this happening, the specification of recursion has
3931 been changed so that all such subpatterns are automatically treated as
3932 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3933
3934 4. I had overlooked the fact that, in some locales, there are characters for
3935 which isalpha() is true but neither isupper() nor islower() are true. In
3936 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3937 and ordfeminine) are like this. This affected the treatment of \w and \W
3938 when they appeared in character classes, but not when they appeared outside
3939 a character class. The bit map for "word" characters is now created
3940 separately from the results of isalnum() instead of just taking it from the
3941 upper, lower, and digit maps. (Plus the underscore character, of course.)
3942
3943 5. The above bug also affected the handling of POSIX character classes such as
3944 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3945 permanent tables. Instead, the bit maps for such a class were previously
3946 created as the appropriate unions of the upper, lower, and digit bitmaps.
3947 Now they are created by subtraction from the [[:word:]] class, which has
3948 its own bitmap.
3949
3950 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3951 It is created by subtracting the vertical space characters (\x09, \x0a,
3952 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3953 subtraction was done in the overall bitmap for a character class, meaning
3954 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3955 be recognized. This bug has been fixed.
3956
3957 7. Patches from the folks at Google:
3958
3959 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3960 real life, but is still worth protecting against".
3961
3962 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3963 regular expressions".
3964
3965 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3966 have it.
3967
3968 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3969 "configure" and the latter not, in order to fix a problem somebody had
3970 with compiling the Arg class on HP-UX.
3971
3972 (e) Improve the error-handling of the C++ wrapper a little bit.
3973
3974 (f) New tests for checking recursion limiting.
3975
3976 8. The pcre_memmove() function, which is used only if the environment does not
3977 have a standard memmove() function (and is therefore rarely compiled),
3978 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3979 returning a result (though PCRE never actually uses the result).
3980
3981 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3982 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3983 returned instead of calling malloc() with an overflowing number that would
3984 most likely cause subsequent chaos.
3985
3986 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3987
3988 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3989 with this option is matched, the nmatch and pmatch options of regexec() are
3990 ignored.
3991
3992 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3993 provided in case anyone wants to the the POSIX interface with UTF-8
3994 strings.
3995
3996 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3997 C++ linking (needed for some HP-UX environments).
3998
3999 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
4000 (unused parameter) and in the pcre_printint() function (omitted "default"
4001 switch label when the default is to do nothing).
4002
4003 15. Added some code to make it possible, when PCRE is compiled as a C++
4004 library, to replace subject pointers for pcre_exec() with a smart pointer
4005 class, thus making it possible to process discontinuous strings.
4006
4007 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
4008 much the same function. They were added by different people who were trying
4009 to make PCRE easy to compile on non-Unix systems. It has been suggested
4010 that PCRE_EXPORT be abolished now that there is more automatic apparatus
4011 for compiling on Windows systems. I have therefore replaced it with
4012 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
4013 defaults to "extern" for C or "extern C" for C++, which works fine on
4014 Unix-like systems. It is now possible to override the value of PCRE_DATA_
4015 SCOPE with something explicit in config.h. In addition:
4016
4017 (a) pcreposix.h still had just "extern" instead of either of these macros;
4018 I have replaced it with PCRE_DATA_SCOPE.
4019
4020 (b) Functions such as _pcre_xclass(), which are internal to the library,
4021 but external in the C sense, all had PCRE_EXPORT in their definitions.
4022 This is apparently wrong for the Windows case, so I have removed it.
4023 (It makes no difference on Unix-like systems.)
4024
4025 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
4026 of recursive calls to match(). This is different to MATCH_LIMIT because
4027 that limits the total number of calls to match(), not all of which increase
4028 the depth of recursion. Limiting the recursion depth limits the amount of
4029 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4030 when PCRE is compiled, and changed at run time. A patch from Google adds
4031 this functionality to the C++ interface.
4032
4033 18. Changes to the handling of Unicode character properties:
4034
4035 (a) Updated the table to Unicode 4.1.0.
4036
4037 (b) Recognize characters that are not in the table as "Cn" (undefined).
4038
4039 (c) I revised the way the table is implemented to a much improved format
4040 which includes recognition of ranges. It now supports the ranges that
4041 are defined in UnicodeData.txt, and it also amalgamates other
4042 characters into ranges. This has reduced the number of entries in the
4043 table from around 16,000 to around 3,000, thus reducing its size
4044 considerably. I realized I did not need to use a tree structure after
4045 all - a binary chop search is just as efficient. Having reduced the
4046 number of entries, I extended their size from 6 bytes to 8 bytes to
4047 allow for more data.
4048
4049 (d) Added support for Unicode script names via properties such as \p{Han}.
4050
4051 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4052 matching that character.
4053
4054 20. When matching a repeated Unicode property with a minimum greater than zero,
4055 (for example \pL{2,}), PCRE could look past the end of the subject if it
4056 reached it while seeking the minimum number of characters. This could
4057 happen only if some of the characters were more than one byte long, because
4058 there is a check for at least the minimum number of bytes.
4059
4060 21. Refactored the implementation of \p and \P so as to be more general, to
4061 allow for more different types of property in future. This has changed the
4062 compiled form incompatibly. Anybody with saved compiled patterns that use
4063 \p or \P will have to recompile them.
4064
4065 22. Added "Any" and "L&" to the supported property types.
4066
4067 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4068 but give a compile time error if the value is greater than 0xff.
4069
4070 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4071 accidentally not being installed or uninstalled.
4072
4073 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4074 made were to insert the current release number. This seemed silly, because
4075 it made things harder for people building PCRE on systems that don't run
4076 "configure". I have turned pcre.h into a distributed file, no longer built
4077 by "configure", with the version identification directly included. There is
4078 no longer a pcre.h.in file.
4079
4080 However, this change necessitated a change to the pcre-config script as
4081 well. It is built from pcre-config.in, and one of the substitutions was the
4082 release number. I have updated configure.ac so that ./configure now finds
4083 the release number by grepping pcre.h.
4084
4085 26. Added the ability to run the tests under valgrind.
4086
4087
4088 Version 6.4 05-Sep-05
4089 ---------------------
4090
4091 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4092 "--" to be printed when multiple files were scanned, even when none of the
4093 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4094 consider it to be a bug, and have restored the previous behaviour.
4095
4096 2. A couple of code tidies to get rid of compiler warnings.
4097
4098 3. The pcretest program used to cheat by referring to symbols in the library
4099 whose names begin with _pcre_. These are internal symbols that are not
4100 really supposed to be visible externally, and in some environments it is
4101 possible to suppress them. The cheating is now confined to including
4102 certain files from the library's source, which is a bit cleaner.
4103
4104 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4105 file's purpose clearer.
4106
4107 5. Reorganized pcre_ucp_findchar().
4108
4109
4110 Version 6.3 15-Aug-05
4111 ---------------------
4112
4113 1. The file libpcre.pc.in did not have general read permission in the tarball.
4114
4115 2. There were some problems when building without C++ support:
4116
4117 (a) If C++ support was not built, "make install" and "make test" still
4118 tried to test it.
4119
4120 (b) There were problems when the value of CXX was explicitly set. Some
4121 changes have been made to try to fix these, and ...
4122
4123 (c) --disable-cpp can now be used to explicitly disable C++ support.
4124
4125 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4126 backslash in a target when C++ was disabled. This confuses some
4127 versions of "make", apparently. Using an intermediate variable solves
4128 this. (Same for CPP_LOBJ.)
4129
4130 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4131 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4132 necessary on certain architectures.
4133
4134 4. Added a setting of -export-symbols-regex to the link command to remove
4135 those symbols that are exported in the C sense, but actually are local
4136 within the library, and not documented. Their names all begin with
4137 "_pcre_". This is not a perfect job, because (a) we have to except some
4138 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4139 available (and never for static libraries). I have made a note to try to
4140 find a way round (a) in the future.
4141
4142
4143 Version 6.2 01-Aug-05
4144 ---------------------
4145
4146 1. There was no test for integer overflow of quantifier values. A construction
4147 such as {1111111111111111} would give undefined results. What is worse, if
4148 a minimum quantifier for a parenthesized subpattern overflowed and became
4149 negative, the calculation of the memory size went wrong. This could have
4150 led to memory overwriting.
4151
4152 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4153
4154 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4155 operating environments where this matters.
4156
4157 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4158 PCRE options from within the C++ wrapper.
4159
4160 5. Named capturing subpatterns were not being correctly counted when a pattern
4161 was compiled. This caused two problems: (a) If there were more than 100
4162 such subpatterns, the calculation of the memory needed for the whole
4163 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4164 back references of the form \12, where the number was greater than 9, were
4165 not recognized as back references, even though there were sufficient
4166 previous subpatterns.
4167
4168 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4169 versions of gcc, e.g. 2.95.4.
4170
4171
4172 Version 6.1 21-Jun-05
4173 ---------------------
4174
4175 1. There was one reference to the variable "posix" in pcretest.c that was not
4176 surrounded by "#if !defined NOPOSIX".
4177
4178 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4179 the cross-check on the old pcre_info() function, for the benefit of the
4180 cut-down version of PCRE that is currently imported into Exim.
4181
4182 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4183 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4184 patterns that start (?i) but I don't think that matters. The use of (?i) is
4185 just an example; this all applies to the other options as well.
4186
4187 4. Since libtool seems to echo the compile commands it is issuing, the output
4188 from "make" can be reduced a bit by putting "@" in front of each libtool
4189 compile command.
4190
4191 5. Patch from the folks at Google for configure.in to be a bit more thorough
4192 in checking for a suitable C++ installation before trying to compile the
4193 C++ stuff. This should fix a reported problem when a compiler was present,
4194 but no suitable headers.
4195
4196 6. The man pages all had just "PCRE" as their title. I have changed them to
4197 be the relevant file name. I have also arranged that these names are
4198 retained in the file doc/pcre.txt, which is a concatenation in text format
4199 of all the man pages except the little individual ones for each function.
4200
4201 7. The NON-UNIX-USE file had not been updated for the different set of source
4202 files that come with release 6. I also added a few comments about the C++
4203 wrapper.
4204
4205
4206 Version 6.0 07-Jun-05
4207 ---------------------
4208
4209 1. Some minor internal re-organization to help with my DFA experiments.
4210
4211 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4212 didn't matter for the library itself when fully configured, but did matter
4213 when compiling without UCP support, or within Exim, where the ucp files are
4214 not imported.
4215
4216 3. Refactoring of the library code to split up the various functions into
4217 different source modules. The addition of the new DFA matching code (see
4218 below) to a single monolithic source would have made it really too
4219 unwieldy, quite apart from causing all the code to be include in a
4220 statically linked application, when only some functions are used. This is
4221 relevant even without the DFA addition now that patterns can be compiled in
4222 one application and matched in another.
4223
4224 The downside of splitting up is that there have to be some external
4225 functions and data tables that are used internally in different modules of
4226 the library but which are not part of the API. These have all had their
4227 names changed to start with "_pcre_" so that they are unlikely to clash
4228 with other external names.
4229
4230 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4231 a different (DFA) algorithm. Although it is slower than the original
4232 function, it does have some advantages for certain types of matching
4233 problem.
4234
4235 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4236 including restarting after a partial match.
4237
4238 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4239 defined when compiling for Windows was sent to me. I have put it into the
4240 code, though I have no means of testing or verifying it.
4241
4242 7. Added the pcre_refcount() auxiliary function.
4243
4244 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4245 match before or at the first newline in the subject string. In pcretest,
4246 the /f option on a pattern can be used to set this.
4247
4248 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4249 would behave wrongly. This has been present in PCRE since release 4.0.
4250
4251 10. A number of changes to the pcregrep command:
4252
4253 (a) Refactored how -x works; insert ^(...)$ instead of setting
4254 PCRE_ANCHORED and checking the length, in preparation for adding
4255 something similar for -w.
4256
4257 (b) Added the -w (match as a word) option.
4258
4259 (c) Refactored the way lines are read and buffered so as to have more
4260 than one at a time available.
4261
4262 (d) Implemented a pcregrep test script.
4263
4264 (e) Added the -M (multiline match) option. This allows patterns to match
4265 over several lines of the subject. The buffering ensures that at least
4266 8K, or the rest of the document (whichever is the shorter) is available
4267 for matching (and similarly the previous 8K for lookbehind assertions).
4268
4269 (f) Changed the --help output so that it now says
4270
4271 -w, --word-regex(p)
4272
4273 instead of two lines, one with "regex" and the other with "regexp"
4274 because that confused at least one person since the short forms are the
4275 same. (This required a bit of code, as the output is generated
4276 automatically from a table. It wasn't just a text change.)
4277
4278 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4279 option but starts with a hyphen. Could be a pattern or a path name
4280 starting with a hyphen, for instance.
4281
4282 (h) "-" can be given as a file name to represent stdin.
4283
4284 (i) When file names are being printed, "(standard input)" is used for
4285 the standard input, for compatibility with GNU grep. Previously
4286 "<stdin>" was used.
4287
4288 (j) The option --label=xxx can be used to supply a name to be used for
4289 stdin when file names are being printed. There is no short form.
4290
4291 (k) Re-factored the options decoding logic because we are going to add
4292 two more options that take data. Such options can now be given in four
4293 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4294
4295 (l) Added the -A, -B, and -C options for requesting that lines of context
4296 around matches be printed.
4297
4298 (m) Added the -L option to print the names of files that do not contain
4299 any matching lines, that is, the complement of -l.
4300
4301 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4302 continue to scan other files.
4303
4304 (o) The -s option was incorrectly implemented. For compatibility with other
4305 greps, it now suppresses the error message for a non-existent or non-
4306 accessible file (but not the return code). There is a new option called
4307 -q that suppresses the output of matching lines, which was what -s was
4308 previously doing.
4309
4310 (p) Added --include and --exclude options to specify files for inclusion
4311 and exclusion when recursing.
4312
4313 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4314 Hopefully, it now does.
4315
4316 12. Missing cast in pcre_study().
4317
4318 13. Added an "uninstall" target to the makefile.
4319
4320 14. Replaced "extern" in the function prototypes in Makefile.in with
4321 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4322 world, but is set differently for Windows.
4323
4324 15. Added a second compiling function called pcre_compile2(). The only
4325 difference is that it has an extra argument, which is a pointer to an
4326 integer error code. When there is a compile-time failure, this is set
4327 non-zero, in addition to the error test pointer being set to point to an
4328 error message. The new argument may be NULL if no error number is required
4329 (but then you may as well call pcre_compile(), which is now just a
4330 wrapper). This facility is provided because some applications need a
4331 numeric error indication, but it has also enabled me to tidy up the way
4332 compile-time errors are handled in the POSIX wrapper.
4333
4334 16. Added VPATH=.libs to the makefile; this should help when building with one
4335 prefix path and installing with another. (Or so I'm told by someone who
4336 knows more about this stuff than I do.)
4337
4338 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4339 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4340 match everything, including newlines. This is not POSIX-compatible, but
4341 somebody wanted the feature. From pcretest it can be activated by using
4342 both the P and the s flags.
4343
4344 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4345
4346 19. libpcre.pc was being incorrectly installed as executable.
4347
4348 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4349 it now also looks for '\r' so that it will work unmodified on Windows.
4350
4351 21. Added Google's contributed C++ wrapper to the distribution.
4352
4353 22. Added some untidy missing memory free() calls in pcretest, to keep
4354 Electric Fence happy when testing.
4355
4356
4357
4358 Version 5.0 13-Sep-04
4359 ---------------------
4360
4361 1. Internal change: literal characters are no longer packed up into items
4362 containing multiple characters in a single byte-string. Each character
4363 is now matched using a separate opcode. However, there may be more than one
4364 byte in the character in UTF-8 mode.
4365
4366 2. The pcre_callout_block structure has two new fields: pattern_position and
4367 next_item_length. These contain the offset in the pattern to the next match
4368 item, and its length, respectively.
4369
4370 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4371 insertion of callouts before each pattern item. Added the /C option to
4372 pcretest to make use of this.
4373
4374 4. On the advice of a Windows user, the lines
4375
4376 #if defined(_WIN32) || defined(WIN32)
4377 _setmode( _fileno( stdout ), 0x8000 );
4378 #endif /* defined(_WIN32) || defined(WIN32) */
4379
4380 have been added to the source of pcretest. This apparently does useful
4381 magic in relation to line terminators.
4382
4383 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4384 for the benefit of those environments where the "b" makes a difference.
4385
4386 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4387 to know about it. I have put a hack into configure.in that adds in code
4388 to set GCC=yes if CC=icc. This seems to end up at a point in the
4389 generated configure script that is early enough to affect the setting of
4390 compiler options, which is what is needed, but I have no means of testing
4391 whether it really works. (The user who reported this had patched the
4392 generated configure script, which of course I cannot do.)
4393
4394 LATER: After change 22 below (new libtool files), the configure script
4395 seems to know about icc (and also ecc). Therefore, I have commented out
4396 this hack in configure.in.
4397
4398 7. Added support for pkg-config (2 patches were sent in).
4399
4400 8. Negated POSIX character classes that used a combination of internal tables
4401 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4402 [[:^ascii]]. Typically, they would match almost any characters. The other
4403 POSIX classes were not broken in this way.
4404
4405 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4406 to find the match, as PCRE was deluded into thinking that the match had to
4407 start at the start point or following a newline. The same bug applied to
4408 patterns with negative forward assertions or any backward assertions
4409 preceding ".*" at the start, unless the pattern required a fixed first
4410 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4411
4412 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4413 starting at the last subject character, bytes beyond the end of the subject
4414 string were read.
4415
4416 11. Renamed the variable "class" as "classbits" to make life easier for C++
4417 users. (Previously there was a macro definition, but it apparently wasn't
4418 enough.)
4419
4420 12. Added the new field "tables" to the extra data so that tables can be passed
4421 in at exec time, or the internal tables can be re-selected. This allows
4422 a compiled regex to be saved and re-used at a later time by a different
4423 program that might have everything at different addresses.
4424
4425 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4426 -R library as well as a -L library.
4427
4428 14. The debugging options of pcretest (-d on the command line or D on a
4429 pattern) showed incorrect output for anything following an extended class
4430 that contained multibyte characters and which was followed by a quantifier.
4431
4432 15. Added optional support for general category Unicode character properties
4433 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4434 support. It adds about 90K to the size of the library. The meanings of the
4435 inbuilt class escapes such as \d and \s have NOT been changed.
4436
4437 16. Updated pcredemo.c to include calls to free() to release the memory for the
4438 compiled pattern.
4439
4440 17. The generated file chartables.c was being created in the source directory
4441 instead of in the building directory. This caused the build to fail if the
4442 source directory was different from the building directory, and was
4443 read-only.
4444
4445 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4446 file. No doubt somebody will tell me if they don't make sense... Also added
4447 Dan Mooney's comments about building on OpenVMS.
4448
4449 19. Added support for partial matching via the PCRE_PARTIAL option for
4450 pcre_exec() and the \P data escape in pcretest.
4451
4452 20. Extended pcretest with 3 new pattern features:
4453
4454 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4455 write the compiled pattern to the file whose name is "rest-of-line".
4456 This is a straight binary dump of the data, with the saved pointer to
4457 the character tables forced to be NULL. The study data, if any, is
4458 written too. After writing, pcretest reads a new pattern.
4459
4460 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4461 compiled pattern from the given file. There must not be any
4462 occurrences of "<" in the file name (pretty unlikely); if there are,
4463 pcretest will instead treat the initial "<" as a pattern delimiter.
4464 After reading in the pattern, pcretest goes on to read data lines as
4465 usual.
4466
4467 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4468 and 16-bit fields in a compiled pattern, to simulate a pattern that
4469 was compiled on a host of opposite endianness.
4470
4471 21. The pcre-exec() function can now cope with patterns that were compiled on
4472 hosts of opposite endianness, with this restriction:
4473
4474 As for any compiled expression that is saved and used later, the tables
4475 pointer field cannot be preserved; the extra_data field in the arguments
4476 to pcre_exec() should be used to pass in a tables address if a value
4477 other than the default internal tables were used at compile time.
4478
4479 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4480 now diagnosed as an error. Previously, most of the time, a negative number
4481 would have been treated as zero, but if in addition "ovector" was passed as
4482 NULL, a crash could occur.
4483
4484 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4485 new versions from the libtool 1.5 distribution (the last one is a copy of
4486 a file called libtool.m4). This seems to have fixed the need to patch
4487 "configure" to support Darwin 1.3 (which I used to do). However, I still
4488 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4489 workstation).
4490
4491 24. Changed the PCRE licence to be the more standard "BSD" licence.
4492
4493
4494 Version 4.5 01-Dec-03
4495 ---------------------
4496
4497 1. There has been some re-arrangement of the code for the match() function so
4498 that it can be compiled in a version that does not call itself recursively.
4499 Instead, it keeps those local variables that need separate instances for
4500 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4501 needs to "recurse". Keeping track of where control must go is done by means
4502 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4503 hide most of the details from the main code, and operates only if
4504 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4505 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4506 operating.
4507
4508 To make it easier for callers to provide specially tailored get/free
4509 functions for this usage, two new functions, pcre_stack_malloc, and
4510 pcre_stack_free, are used. They are always called in strict stacking order,
4511 and the size of block requested is always the same.
4512
4513 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4514 PCRE has been compiled to use the stack or the heap for recursion. The
4515 -C option of pcretest uses this to show which version is compiled.
4516
4517 A new data escape \S, is added to pcretest; it causes the amounts of store
4518 obtained and freed by both kinds of malloc/free at match time to be added
4519 to the output.
4520
4521 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4522 what's available on my current Linux desktop machine.
4523
4524 3. When matching a UTF-8 string, the test for a valid string at the start has
4525 been extended. If start_offset is not zero, PCRE now checks that it points
4526 to a byte that is the start of a UTF-8 character. If not, it returns
4527 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4528 this is necessary because there may be backward assertions in the pattern.
4529 When matching the same subject several times, it may save resources to use
4530 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4531
4532 4. The code for checking the validity of UTF-8 strings has been tightened so
4533 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4534 containing "overlong sequences".
4535
4536 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4537 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4538 should have been "&", but it just so happened that all the cases this let
4539 through by mistake were picked up later in the function.
4540
4541 6. I had used a variable called "isblank" - this is a C99 function, causing
4542 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4543
4544 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4545 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4546 so the version is shown; (c) stop "make test" echoing "./RunTest".
4547
4548 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4549
4550 9. The prototype for memmove() for systems that don't have it was using
4551 size_t, but the inclusion of the header that defines size_t was later. I've
4552 moved the #includes for the C headers earlier to avoid this.
4553
4554 10. Added some adjustments to the code to make it easier to compiler on certain
4555 special systems:
4556
4557 (a) Some "const" qualifiers were missing.
4558 (b) Added the macro EXPORT before all exported functions; by default this
4559 is defined to be empty.
4560 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4561 that it reads its output file name as an argument instead of writing
4562 to the standard output and assuming this can be redirected.
4563
4564 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4565 class containing characters with values greater than 255, PCRE compilation
4566 went into a loop.
4567
4568 12. A recursive reference to a subpattern that was within another subpattern
4569 that had a minimum quantifier of zero caused PCRE to crash. For example,
4570 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4571 recursion. If the recursively-called subpattern itself had a zero repeat,
4572 that was OK.
4573
4574 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4575 buffer into which it was copied (for escape processing) was still set at
4576 1024, so long lines caused crashes.
4577
4578 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4579 "internal error: code overflow...". This applied to any character class
4580 that was followed by a possessive quantifier.
4581
4582 15. Modified the Makefile to add libpcre.la as a prerequisite for
4583 libpcreposix.la because I was told this is needed for a parallel build to
4584 work.
4585
4586 16. If a pattern that contained .* following optional items at the start was
4587 studied, the wrong optimizing data was generated, leading to matching
4588 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4589 matching string must start with a or b or c. The correct conclusion for
4590 this pattern is that a match can start with any character.
4591
4592
4593 Version 4.4 13-Aug-03
4594 ---------------------
4595
4596 1. In UTF-8 mode, a character class containing characters with values between
4597 127 and 255 was not handled correctly if the compiled pattern was studied.
4598 In fixing this, I have also improved the studying algorithm for such
4599 classes (slightly).
4600
4601 2. Three internal functions had redundant arguments passed to them. Removal
4602 might give a very teeny performance improvement.
4603
4604 3. Documentation bug: the value of the capture_top field in a callout is *one
4605 more than* the number of the hightest numbered captured substring.
4606
4607 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4608 in incorrectly linking with a previously installed version. They now link
4609 explicitly with libpcre.la.
4610
4611 5. configure.in no longer needs to recognize Cygwin specially.
4612
4613 6. A problem in pcre.in for Windows platforms is fixed.
4614
4615 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4616 pcretest, it used to include the size of the study block as part of its
4617 output. Unfortunately, the structure contains a field that has a different
4618 size on different hardware architectures. This meant that the tests that
4619 showed this size failed. As the block is currently always of a fixed size,
4620 this information isn't actually particularly useful in pcretest output, so
4621 I have just removed it.
4622
4623 8. Three pre-processor statements accidentally did not start in column 1.
4624 Sadly, there are *still* compilers around that complain, even though
4625 standard C has not required this for well over a decade. Sigh.
4626
4627 9. In pcretest, the code for checking callouts passed small integers in the
4628 callout_data field, which is a void * field. However, some picky compilers
4629 complained about the casts involved for this on 64-bit systems. Now
4630 pcretest passes the address of the small integer instead, which should get
4631 rid of the warnings.
4632
4633 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4634 both compile and run time, and gives an error if an invalid UTF-8 sequence
4635 is found. There is a option for disabling this check in cases where the
4636 string is known to be correct and/or the maximum performance is wanted.
4637
4638 11. In response to a bug report, I changed one line in Makefile.in from
4639
4640 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4641 to
4642 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4643
4644 to look similar to other lines, but I have no way of telling whether this
4645 is the right thing to do, as I do not use Windows. No doubt I'll get told
4646 if it's wrong...
4647
4648
4649 Version 4.3 21-May-03
4650 ---------------------
4651
4652 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4653 Makefile.
4654
4655 2. Some refactoring to improve the quality of the code:
4656
4657 (i) The utf8_table... variables are now declared "const".
4658
4659 (ii) The code for \cx, which used the "case flipping" table to upper case
4660 lower case letters, now just substracts 32. This is ASCII-specific,
4661 but the whole concept of \cx is ASCII-specific, so it seems
4662 reasonable.
4663
4664 (iii) PCRE was using its character types table to recognize decimal and
4665 hexadecimal digits in the pattern. This is silly, because it handles
4666 only 0-9, a-f, and A-F, but the character types table is locale-
4667 specific, which means strange things might happen. A private
4668 table is now used for this - though it costs 256 bytes, a table is
4669 much faster than multiple explicit tests. Of course, the standard
4670 character types table is still used for matching digits in subject
4671 strings against \d.
4672
4673 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4674 ending in _t are). So I've renamed it as ESC_tee.
4675
4676 3. The first argument for regexec() in the POSIX wrapper should have been
4677 defined as "const".
4678
4679 4. Changed pcretest to use malloc() for its buffers so that they can be
4680 Electric Fenced for debugging.
4681
4682 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4683 to read one or more bytes before the start of the subject string. Often this
4684 had no effect on PCRE's behaviour, but in some circumstances it could
4685 provoke a segmentation fault.
4686
4687 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4688 to try to read one or more bytes before the start of the subject string.
4689
4690 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4691 UTF-8 support could misbehave in various ways if the subject string
4692 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4693 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4694 back over UTF-8 characters.)
4695
4696
4697 Version 4.2 14-Apr-03
4698 ---------------------
4699
4700 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4701
4702 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4703 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4704 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4705 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4706 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4707 and BUILD_EXEEXT
4708 Note: automatic setting of the BUILD variables is not yet working
4709 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4710 compile-time but not at link-time
4711 [LINK]: use for linking executables only
4712 make different versions for Windows and non-Windows
4713 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4714 libraries
4715 [LINK_FOR_BUILD]: new variable
4716 [OBJEXT]: use throughout
4717 [EXEEXT]: use throughout
4718 <winshared>: new target
4719 <wininstall>: new target
4720 <dftables.o>: use native compiler
4721 <dftables>: use native linker
4722 <install>: handle Windows platform correctly
4723 <clean>: ditto
4724 <check>: ditto
4725 copy DLL to top builddir before testing
4726
4727 As part of these changes, -no-undefined was removed again. This was reported
4728 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4729 in any case.
4730
4731 3. Some tidies to get rid of compiler warnings:
4732
4733 . In the match_data structure, match_limit was an unsigned long int, whereas
4734 match_call_count was an int. I've made them both unsigned long ints.
4735
4736 . In pcretest the fact that a const uschar * doesn't automatically cast to
4737 a void * provoked a warning.
4738
4739 . Turning on some more compiler warnings threw up some "shadow" variables
4740 and a few more missing casts.
4741
4742 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4743 option, a class that contained a single character with a value between 128
4744 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4745
4746 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4747 option, a class that contained several characters, but with at least one
4748 whose value was between 128 and 255 caused PCRE to crash.
4749
4750
4751 Version 4.1 12-Mar-03
4752 ---------------------
4753
4754 1. Compiling with gcc -pedantic found a couple of places where casts were
4755 needed, and a string in dftables.c that was longer than standard compilers are
4756 required to support.
4757
4758 2. Compiling with Sun's compiler found a few more places where the code could
4759 be tidied up in order to avoid warnings.
4760
4761 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4762 first of these names is deprecated in the latest Autoconf in favour of the name
4763 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4764 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4765 analogy I have changed it to CFLAGS_FOR_BUILD.
4766
4767 4. Added -no-undefined to the linking command in the Makefile, because this is
4768 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4769 linking step for the pcreposix library.
4770
4771 5. PCRE was failing to diagnose the case of two named groups with the same
4772 name.
4773
4774 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4775 literal character that is needed in the subject for a match, and scans along to
4776 ensure that it is present before embarking on the full matching process. This
4777 saves time in cases of nested unlimited repeats that are never going to match.
4778 Problem: the scan can take a lot of time if the subject is very long (e.g.
4779 megabytes), thus penalizing straightforward matches. It is now done only if the
4780 amount of subject to be scanned is less than 1000 bytes.
4781
4782 7. A lesser problem with the same optimization is that it was recording the
4783 first character of an anchored pattern as "needed", thus provoking a search
4784 right along the subject, even when the first match of the pattern was going to
4785 fail. The "needed" character is now not set for anchored patterns, unless it
4786 follows something in the pattern that is of non-fixed length. Thus, it still
4787 fulfils its original purpose of finding quick non-matches in cases of nested
4788 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4789
4790
4791 Version 4.0 17-Feb-03
4792 ---------------------
4793
4794 1. If a comment in an extended regex that started immediately after a meta-item
4795 extended to the end of string, PCRE compiled incorrect data. This could lead to
4796 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4797
4798 2. Moved to autoconf 2.53 and libtool 1.4.2.
4799
4800 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4801 the special perltest8 script is no longer needed - all the tests can be run
4802 from a single perltest script.
4803
4804 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4805 by \s. It has now been removed in PCRE. This means it isn't recognized as
4806 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4807 class [:space:] *does* include VT, thereby creating a mess.
4808
4809 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4810 space and tab.
4811
4812 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4813 its new features into the main test script, reducing the number of scripts.
4814
4815 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4816 were backward compatible, and made the (?i) apply to the whole pattern, as if
4817 /i were given. Now it behaves more logically, and applies the option setting
4818 only to what follows. PCRE has been changed to follow suit. However, if it
4819 finds options settings right at the start of the pattern, it extracts them into
4820 the global options, as before. Thus, they show up in the info data.
4821
4822 8. Added support for the \Q...\E escape sequence. Characters in between are
4823 treated as literals. This is slightly different from Perl in that $ and @ are
4824 also handled as literals inside the quotes. In Perl, they will cause variable
4825 interpolation. Note the following examples:
4826
4827 Pattern PCRE matches Perl matches
4828
4829 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4830 \Qabc\$xyz\E abc\$xyz abc\$xyz
4831 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4832
4833 For compatibility with Perl, \Q...\E sequences are recognized inside character
4834 classes as well as outside them.
4835
4836 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4837 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4838 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4839 signed/unsigned warnings.
4840
4841 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4842 option for pcretest, so I've replaced it by a simple function that does just
4843 that job.
4844
4845 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4846 "pcregrep -".
4847
4848 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4849 Java package. This provides some syntactic sugar for simple cases of what my
4850 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4851 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4852 item, you can use this simplified notation. Note that only makes sense with
4853 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4854 greediness, whatever the setting of the PCRE_UNGREEDY option.
4855
4856 13. A change of greediness default within a pattern was not taking effect at
4857 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4858 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4859 was abstracted outside.
4860
4861 14. PCRE now supports the \G assertion. It is true when the current matching
4862 position is at the start point of the match. This differs from \A when the
4863 starting offset is non-zero. Used with the /g option of pcretest (or similar
4864 code), it works in the same way as it does for Perl's /g option. If all
4865 alternatives of a regex begin with \G, the expression is anchored to the start
4866 match position, and the "anchored" flag is set in the compiled expression.
4867
4868 15. Some bugs concerning the handling of certain option changes within patterns
4869 have been fixed. These applied to options other than (?ims). For example,
4870 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4871 been the other way round. Some of this was related to change 7 above.
4872
4873 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4874 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4875 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4876 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4877
4878 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4879 mode. Unlike ".", it always matches newline, whatever the setting of
4880 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4881 assertions. Perl allows it, but it doesn't (in general) work because it can't
4882 calculate the length of the lookbehind. At least, that's the case for Perl
4883 5.8.0 - I've been told they are going to document that it doesn't work in
4884 future.
4885
4886 18. Added an error diagnosis for escapes that PCRE does not support: these are
4887 \L, \l, \N, \P, \p, \U, \u, and \X.
4888
4889 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4890 reading past the end of the pattern in cases such as /[abcd/.
4891
4892 20. PCRE was getting more memory than necessary for patterns with classes that
4893 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4894
4895 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4896 compiling PCRE for use with Virtual Pascal.
4897
4898 22. Small fix to the Makefile to make it work properly if the build is done
4899 outside the source tree.
4900
4901 23. Added a new extension: a condition to go with recursion. If a conditional
4902 subpattern starts with (?(R) the "true" branch is used if recursion has
4903 happened, whereas the "false" branch is used only at the top level.
4904
4905 24. When there was a very long string of literal characters (over 255 bytes
4906 without UTF support, over 250 bytes with UTF support), the computation of how
4907 much memory was required could be incorrect, leading to segfaults or other
4908 strange effects.
4909
4910 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4911 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4912 there was a subsequent back reference to those brackets. This meant that, for
4913 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4914 possible to check for precisely this case. All we can do is abandon the
4915 optimization if .* occurs inside capturing brackets when there are any back
4916 references whatsoever. (See below for a better fix that came later.)
4917
4918 26. The handling of the optimization for finding the first character of a
4919 non-anchored pattern, and for finding a character that is required later in the
4920 match were failing in some cases. This didn't break the matching; it just
4921 failed to optimize when it could. The way this is done has been re-implemented.
4922
4923 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4924
4925 28. Added a new feature that provides some of the functionality that Perl
4926 provides with (?{...}). The facility is termed a "callout". The way it is done
4927 in PCRE is for the caller to provide an optional function, by setting
4928 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4929 global variable. By default it is unset, which disables all calling out. To get
4930 the function called, the regex must include (?C) at appropriate points. This
4931 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4932 This provides a means of identifying different callout points. When PCRE
4933 reaches such a point in the regex, if pcre_callout has been set, the external
4934 function is called. It is provided with data in a structure called
4935 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4936 matching continues; if it returns a non-zero value, the match at the current
4937 point fails. However, backtracking will occur if possible. [This was changed
4938 later and other features added - see item 49 below.]
4939
4940 29. pcretest is upgraded to test the callout functionality. It provides a
4941 callout function that displays information. By default, it shows the start of
4942 the match and the current position in the text. There are some new data escapes
4943 to vary what happens:
4944
4945 \C+ in addition, show current contents of captured substrings
4946 \C- do not supply a callout function
4947 \C!n return 1 when callout number n is reached
4948 \C!n!m return 1 when callout number n is reached for the mth time
4949
4950 30. If pcregrep was called with the -l option and just a single file name, it
4951 output "<stdin>" if a match was found, instead of the file name.
4952
4953 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4954 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4955 pcre_exec(). This saves a malloc/free per call. The default value of
4956 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4957 when configuring.
4958
4959 32. The default maximum size of a compiled pattern is 64K. There have been a
4960 few cases of people hitting this limit. The code now uses macros to handle the
4961 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4962 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4963 configuring. Tests 2 and 5 work only with 2-byte links because they output
4964 debugging information about compiled patterns.
4965
4966 33. Internal code re-arrangements:
4967
4968 (a) Moved the debugging function for printing out a compiled regex into
4969 its own source file (printint.c) and used #include to pull it into
4970 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4971 separate copies.
4972
4973 (b) Defined the list of op-code names for debugging as a macro in
4974 internal.h so that it is next to the definition of the opcodes.
4975
4976 (c) Defined a table of op-code lengths for simpler skipping along compiled
4977 code. This is again a macro in internal.h so that it is next to the
4978 definition of the opcodes.
4979
4980 34. Added support for recursive calls to individual subpatterns, along the
4981 lines of Robin Houston's patch (but implemented somewhat differently).
4982
4983 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4984 allow it to read and process whole directories in Win32. This code was
4985 contributed by Lionel Fourquaux; it has not been tested by me.
4986
4987 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4988 used to name a group. Names consist of alphanumerics and underscores, and must
4989 be unique. Back references use the syntax (?P=name) and recursive calls use
4990 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4991 numbers. The function pcre_fullinfo() can be used after compilation to extract
4992 a name/number map. There are three relevant calls:
4993
4994 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4995 PCRE_INFO_NAMECOUNT yields the number of entries
4996 PCRE_INFO_NAMETABLE yields a pointer to the map.
4997
4998 The map is a vector of fixed-size entries. The size of each entry depends on
4999 the length of the longest name used. The first two bytes of each entry are the
5000 group number, most significant byte first. There follows the corresponding
5001 name, zero terminated. The names are in alphabetical order.
5002
5003 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
5004 case instead of 255. Making it the same both with and without UTF-8 support
5005 means that the same test output works with both.
5006
5007 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
5008 calling malloc() with a zero argument.
5009
5010 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
5011 optimization. I've improved things by keeping a bitmap of backreferences with
5012 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
5013 fact referenced, the optimization can be applied. It is unlikely that a
5014 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
5015 the match to follow \n) will appear inside brackets with a number greater than
5016 31, but if it does, any back reference > 31 suppresses the optimization.
5017
5018 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
5019 of disabling numbered capturing parentheses. Any opening parenthesis that is
5020 not followed by ? behaves as if it were followed by ?: but named parentheses
5021 can still be used for capturing (and they will acquire numbers in the usual
5022 way).
5023
5024 41. Redesigned the return codes from the match() function into yes/no/error so
5025 that errors can be passed back from deep inside the nested calls. A malloc
5026 failure while inside a recursive subpattern call now causes the
5027 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
5028
5029 42. It is now possible to set a limit on the number of times the match()
5030 function is called in a call to pcre_exec(). This facility makes it possible to
5031 limit the amount of recursion and backtracking, though not in a directly
5032 obvious way, because the match() function is used in a number of different
5033 circumstances. The count starts from zero for each position in the subject
5034 string (for non-anchored patterns). The default limit is, for compatibility, a
5035 large number, namely 10 000 000. You can change this in two ways:
5036
5037 (a) When configuring PCRE before making, you can use --with-match-limit=n
5038 to set a default value for the compiled library.
5039
5040 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5041 a different value is set. See 45 below.
5042
5043 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5044
5045 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5046 of things that can be changed at compile time. The first argument specifies
5047 what is wanted and the second points to where the information is to be placed.
5048 The current list of available information is:
5049
5050 PCRE_CONFIG_UTF8
5051
5052 The output is an integer that is set to one if UTF-8 support is available;
5053 otherwise it is set to zero.
5054
5055 PCRE_CONFIG_NEWLINE
5056
5057 The output is an integer that it set to the value of the code that is used for
5058 newline. It is either LF (10) or CR (13).
5059
5060 PCRE_CONFIG_LINK_SIZE
5061
5062 The output is an integer that contains the number of bytes used for internal
5063 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5064
5065 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
5066
5067 The output is an integer that contains the threshold above which the POSIX
5068 interface uses malloc() for output vectors. See item 31 above.
5069
5070 PCRE_CONFIG_MATCH_LIMIT
5071
5072 The output is an unsigned integer that contains the default limit of the number
5073 of match() calls in a pcre_exec() execution. See 42 above.
5074
5075 44. pcretest has been upgraded by the addition of the -C option. This causes it
5076 to extract all the available output from the new pcre_config() function, and to
5077 output it. The program then exits immediately.
5078
5079 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5080 order to support additional features. One way would have been to define
5081 pcre_exec2() (for example) with extra arguments, but this would not have been
5082 extensible, and would also have required all calls to the original function to
5083 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5084 is used for passing in "extra" data from pcre_study().
5085
5086 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5087 contains the following fields:
5088
5089 flags a bitmap indicating which of the following fields are set
5090 study_data opaque data from pcre_study()
5091 match_limit a way of specifying a limit on match() calls for a specific
5092 call to pcre_exec()
5093 callout_data data for callouts (see 49 below)
5094
5095 The flag bits are also defined in pcre.h, and are
5096
5097 PCRE_EXTRA_STUDY_DATA
5098 PCRE_EXTRA_MATCH_LIMIT
5099 PCRE_EXTRA_CALLOUT_DATA
5100
5101 The pcre_study() function now returns one of these new pcre_extra blocks, with
5102 the actual study data pointed to by the study_data field, and the
5103 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5104 before. That is, this change is entirely upwards-compatible and requires no
5105 change to existing code.
5106
5107 If you want to pass in additional data to pcre_exec(), you can either place it
5108 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5109 block.
5110
5111 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5112 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5113 times with different match limits, until it finds the minimum value needed for
5114 pcre_exec() to complete. The value is then output. This can be instructive; for
5115 most simple matches the number is quite small, but for pathological cases it
5116 gets very large very quickly.
5117
5118 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5119 returns the size of the data block pointed to by the study_data field in a
5120 pcre_extra block, that is, the value that was passed as the argument to
5121 pcre_malloc() when PCRE was getting memory in which to place the information
5122 created by pcre_study(). The fourth argument should point to a size_t variable.
5123 pcretest has been extended so that this information is shown after a successful
5124 pcre_study() call when information about the compiled regex is being displayed.
5125
5126 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5127 because what follows is always an absolute path. (Later: it turns out that this
5128 is more than cosmetic for MinGW, because it doesn't like empty path
5129 components.)
5130
5131 49. Some changes have been made to the callout feature (see 28 above):
5132
5133 (i) A callout function now has three choices for what it returns:
5134
5135 0 => success, carry on matching
5136 > 0 => failure at this point, but backtrack if possible
5137 < 0 => serious error, return this value from pcre_exec()
5138
5139 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5140 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5141 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5142 use by callout functions. It will never be used by PCRE itself.
5143
5144 (ii) The pcre_extra structure (see 45 above) has a void * field called
5145 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5146 pcre_callout_block structure has a field of the same name. The contents of
5147 the field passed in the pcre_extra structure are passed to the callout
5148 function in the corresponding field in the callout block. This makes it
5149 easier to use the same callout-containing regex from multiple threads. For
5150 testing, the pcretest program has a new data escape
5151
5152 \C*n pass the number n (may be negative) as callout_data
5153
5154 If the callout function in pcretest receives a non-zero value as
5155 callout_data, it returns that value.
5156
5157 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5158 there were some redundant $(CFLAGS) in commands that are now specified as
5159 $(LINK), which already includes $(CFLAGS).
5160
5161 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5162 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5163 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5164 one-byte characters throughout. Note that case-insensitive matching applies
5165 only to characters whose values are less than 256. PCRE doesn't support the
5166 notion of cases for higher-valued characters.
5167
5168 (i) A character class whose characters are all within 0-255 is handled as
5169 a bit map, and the map is inverted for negative classes. Previously, a
5170 character > 255 always failed to match such a class; however it should
5171 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5172
5173 (ii) A negated character class with a single character < 255 is coded as
5174 "not this character" (OP_NOT). This wasn't working properly when the test
5175 character was multibyte, either singly or repeated.
5176
5177 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5178 mode, for example: \x{100}{2,3}.
5179
5180 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5181 singly or repeated) now correctly test multibyte characters. However,
5182 PCRE doesn't recognize any characters with values greater than 255 as
5183 digits, spaces, or word characters. Such characters always match \D, \S,
5184 and \W, and never match \d, \s, or \w.
5185
5186 (v) Classes may now contain characters and character ranges with values
5187 greater than 255. For example: [ab\x{100}-\x{400}].
5188
5189 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5190 PCRE in UTF-8 mode.
5191
5192 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5193 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5194 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5195 value.)
5196
5197 53. The single man page has become too large. I have therefore split it up into
5198 a number of separate man pages. These also give rise to individual HTML pages;
5199 these are now put in a separate directory, and there is an index.html page that
5200 lists them all. Some hyperlinking between the pages has been installed.
5201
5202 54. Added convenience functions for handling named capturing parentheses.
5203
5204 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5205 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5206 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5207 are faulted.
5208
5209 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5210 calling configure. These values are used when compiling the dftables.c program
5211 which is run to generate the source of the default character tables. They
5212 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5213 you will need to set these values.
5214
5215 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5216
5217
5218 Version 3.9 02-Jan-02
5219 ---------------------
5220
5221 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5222
5223 2. If --disable-static was given, the building process failed when trying to
5224 build pcretest and pcregrep. (For some reason it was using libtool to compile
5225 them, which is not right, as they aren't part of the library.)
5226
5227
5228 Version 3.8 18-Dec-01
5229 ---------------------
5230
5231 1. The experimental UTF-8 code was completely screwed up. It was packing the
5232 bytes in the wrong order. How dumb can you get?
5233
5234
5235 Version 3.7 29-Oct-01
5236 ---------------------
5237
5238 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5239 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5240 this didn't happen under Solaris 8, where I normally test things.
5241
5242 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5243 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5244 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5245 this fix an hour or so after the initial 3.7 release.)
5246
5247
5248 Version 3.6 23-Oct-01
5249 ---------------------
5250
5251 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5252 offsets passed as NULL with zero offset count.
5253
5254 2. The config.guess and config.sub files had not been updated when I moved to
5255 the latest autoconf.
5256
5257
5258 Version 3.5 15-Aug-01
5259 ---------------------
5260
5261 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5262 had been forgotten.
5263
5264 2. By using declared but undefined structures, we can avoid using "void"
5265 definitions in pcre.h while keeping the internal definitions of the structures
5266 private.
5267
5268 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5269 user point of view, this means that both static and shared libraries are built
5270 by default, but this can be individually controlled. More of the work of
5271 handling this static/shared cases is now inside libtool instead of PCRE's make
5272 file.
5273
5274 4. The pcretest utility is now installed along with pcregrep because it is
5275 useful for users (to test regexs) and by doing this, it automatically gets
5276 relinked by libtool. The documentation has been turned into a man page, so
5277 there are now .1, .txt, and .html versions in /doc.
5278
5279 5. Upgrades to pcregrep:
5280 (i) Added long-form option names like gnu grep.
5281 (ii) Added --help to list all options with an explanatory phrase.
5282 (iii) Added -r, --recursive to recurse into sub-directories.
5283 (iv) Added -f, --file to read patterns from a file.
5284
5285 6. pcre_exec() was referring to its "code" argument before testing that
5286 argument for NULL (and giving an error if it was NULL).
5287
5288 7. Upgraded Makefile.in to allow for compiling in a different directory from
5289 the source directory.
5290
5291 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5292 options bits, the pointer it was passed was to an int instead of to an unsigned
5293 long int. This mattered only on 64-bit systems.
5294
5295 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5296 generated) instead of pcre.in, which it its source. Also made the same change
5297 in several of the .c files.
5298
5299 10. A new release of gcc defines printf() as a macro, which broke pcretest
5300 because it had an ifdef in the middle of a string argument for printf(). Fixed
5301 by using separate calls to printf().
5302
5303 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5304 script, to force use of CR or LF instead of \n in the source. On non-Unix
5305 systems, the value can be set in config.h.
5306
5307 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5308 absolute limit. Changed the text of the error message to make this clear, and
5309 likewise updated the man page.
5310
5311 13. The limit of 99 on the number of capturing subpatterns has been removed.
5312 The new limit is 65535, which I hope will not be a "real" limit.
5313
5314
5315 Version 3.4 22-Aug-00
5316 ---------------------
5317
5318 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5319
5320 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5321
5322
5323 Version 3.3 01-Aug-00
5324 ---------------------
5325
5326 1. If an octal character was given, but the value was greater than \377, it
5327 was not getting masked to the least significant bits, as documented. This could
5328 lead to crashes in some systems.
5329
5330 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5331 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5332
5333 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5334 These just pass their arguments on to (pcre_free)(), but they are provided
5335 because some uses of PCRE bind it to non-C systems that can call its functions,
5336 but cannot call free() or pcre_free() directly.
5337
5338 4. Add "make test" as a synonym for "make check". Corrected some comments in
5339 the Makefile.
5340
5341 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5342 Makefile.
5343
5344 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5345 command called pgrep for grepping around the active processes.
5346
5347 7. Added the beginnings of support for UTF-8 character strings.
5348
5349 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5350 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5351 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5352 out for the ar command.)
5353
5354
5355 Version 3.2 12-May-00
5356 ---------------------
5357
5358 This is purely a bug fixing release.
5359
5360 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5361 of ZA. This was just one example of several cases that could provoke this bug,
5362 which was introduced by change 9 of version 2.00. The code for breaking
5363 infinite loops after an iteration that matches an empty string was't working
5364 correctly.
5365
5366 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5367 when matched against abbab (for example). After matching an empty string, it
5368 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5369 caused it to match further down the string than it should.
5370
5371 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5372 was there because it doesn't seem to be needed, and it causes trouble on some
5373 systems, as it is not a Standard C header. It has been removed.
5374
5375 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5376 were reported on the Macintosh. The changes were from
5377
5378 while ((c = *(++ptr)) != 0 && c != '\n');
5379 to
5380 while ((c = *(++ptr)) != 0 && c != '\n') ;
5381
5382 Totally extraordinary, but if that's what it takes...
5383
5384 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5385 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5386 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5387 assumes the way PCRE uses memmove() (always moving upwards).
5388
5389 6. PCRE is being used in one environment where strchr() is not available. There
5390 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5391 faster code anyway.
5392
5393
5394 Version 3.1 09-Feb-00
5395 ---------------------
5396
5397 The only change in this release is the fixing of some bugs in Makefile.in for
5398 the "install" target:
5399
5400 (1) It was failing to install pcreposix.h.
5401
5402 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5403
5404
5405 Version 3.0 01-Feb-00
5406 ---------------------
5407
5408 1. Add support for the /+ modifier to perltest (to output $` like it does in
5409 pcretest).
5410
5411 2. Add support for the /g modifier to perltest.
5412
5413 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5414 matches null strings.
5415
5416 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5417 pattern. Perl treats empty patterns specially - it reuses the most recent
5418 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5419 effect.
5420
5421 5. The POSIX interface was broken in that it was just handing over the POSIX
5422 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5423 required a bigger vector, with some working space on the end. This means that
5424 the POSIX wrapper now has to get and free some memory, and copy the results.
5425
5426 6. Added some simple autoconf support, placing the test data and the
5427 documentation in separate directories, re-organizing some of the
5428 information files, and making it build pcre-config (a GNU standard). Also added
5429 libtool support for building PCRE as a shared library, which is now the
5430 default.
5431
5432 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5433 09 are not valid octal constants. Single digits will be used for minor values
5434 less than 10.
5435
5436 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5437 existing programs that set these in the POSIX interface can use PCRE without
5438 modification.
5439
5440 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5441 return all that pcre_info() returns, plus additional data. The pcre_info()
5442 function is retained for compatibility, but is considered to be obsolete.
5443
5444 10. Added experimental recursion feature (?R) to handle one common case that
5445 Perl 5.6 will be able to do with (?p{...}).
5446
5447 11. Added support for POSIX character classes like [:alpha:], which Perl is
5448 adopting.
5449
5450
5451 Version 2.08 31-Aug-99
5452 ----------------------
5453
5454 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5455 trying to match at the startoffset position, but instead was moving forward to
5456 the next newline as if a previous match had failed.
5457
5458 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5459 and could get into a loop if a null string was matched other than at the start
5460 of the subject.
5461
5462 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5463 be distinguished at compile time, and for completeness also added PCRE_DATE.
5464
5465 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5466 in GnuWin32 environments.
5467
5468
5469 Version 2.07 29-Jul-99
5470 ----------------------
5471
5472 1. The documentation is now supplied in plain text form and HTML as well as in
5473 the form of man page sources.
5474
5475 2. C++ compilers don't like assigning (void *) values to other pointer types.
5476 In particular this affects malloc(). Although there is no problem in Standard
5477 C, I've put in casts to keep C++ compilers happy.
5478
5479 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5480 should be (const char *).
5481
5482 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5483 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5484 However, I haven't made this a standard facility. The documentation doesn't
5485 mention it, and the Makefile doesn't support it.
5486
5487 5. The Makefile now contains an "install" target, with editable destinations at
5488 the top of the file. The pcretest program is not installed.
5489
5490 6. pgrep -V now gives the PCRE version number and date.
5491
5492 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5493 causing the entire string to be ignored, instead of just the last character.
5494
5495 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5496 non-matching string, it can take a very, very long time, even for strings of
5497 quite modest length, because of the nested recursion. PCRE now does better in
5498 some of these cases. It does this by remembering the last required literal
5499 character in the pattern, and pre-searching the subject to ensure it is present
5500 before running the real match. In other words, it applies a heuristic to detect
5501 some types of certain failure quickly, and in the above example, if presented
5502 with a string that has no trailing " it gives "no match" very quickly.
5503
5504 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5505 other alternatives are tried instead.
5506
5507
5508 Version 2.06 09-Jun-99
5509 ----------------------
5510
5511 1. Change pcretest's output for amount of store used to show just the code
5512 space, because the remainder (the data block) varies in size between 32-bit and
5513 64-bit systems.
5514
5515 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5516 start matching at. This allows lookbehinds to work when searching for multiple
5517 occurrences in a string.
5518
5519 3. Added additional options to pcretest for testing multiple occurrences:
5520
5521 /+ outputs the rest of the string that follows a match
5522 /g loops for multiple occurrences, using the new startoffset argument
5523 /G loops for multiple occurrences by passing an incremented pointer
5524
5525 4. PCRE wasn't doing the "first character" optimization for patterns starting
5526 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5527 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5528 the letter 'x'. On long subject strings, this gives a significant speed-up.
5529
5530
5531 Version 2.05 21-Apr-99
5532 ----------------------
5533
5534 1. Changed the type of magic_number from int to long int so that it works
5535 properly on 16-bit systems.
5536
5537 2. Fixed a bug which caused patterns starting with .* not to work correctly
5538 when the subject string contained newline characters. PCRE was assuming
5539 anchoring for such patterns in all cases, which is not correct because .* will
5540 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5541 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5542 must be retried after every newline in the subject.
5543
5544
5545 Version 2.04 18-Feb-99
5546 ----------------------
5547
5548 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5549 computation of the store needed to hold the pattern was incorrect (too large).
5550 If such patterns were nested a few deep, this could multiply and become a real
5551 problem.
5552
5553 2. Added /M option to pcretest to show the memory requirement of a specific
5554 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5555
5556 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5557 compiled in such a way that the backtracking after subsequent failure was
5558 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5559 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5560
5561
5562 Version 2.03 02-Feb-99
5563 ----------------------
5564
5565 1. Fixed typo and small mistake in man page.
5566
5567 2. Added 4th condition (GPL supersedes if conflict) and created separate
5568 LICENCE file containing the conditions.
5569
5570 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5571 Perl, that is the internal \ allows the delimiter to be included in the
5572 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5573 the final delimiter, add \ to the end of the pattern (to test the error).
5574
5575 4. Added the convenience functions for extracting substrings after a successful
5576 match. Updated pcretest to make it able to test these functions.
5577
5578
5579 Version 2.02 14-Jan-99
5580 ----------------------
5581
5582 1. Initialized the working variables associated with each extraction so that
5583 their saving and restoring doesn't refer to uninitialized store.
5584
5585 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5586 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5587 fix the problem.
5588
5589 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5590 calls, and wasn't printing the correct value for compiling calls. Increased the
5591 default value of LOOPREPEAT, and the number of significant figures in the
5592 times.
5593
5594 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5595
5596 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5597 a building problem on Windows NT with a FAT file system.
5598
5599
5600 Version 2.01 21-Oct-98
5601 ----------------------
5602
5603 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5604 to character tables built by pcre_maketables() in the current locale. If NULL
5605 is passed, the default tables are used.
5606
5607
5608 Version 2.00 24-Sep-98
5609 ----------------------
5610
5611 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5612 it any more.
5613
5614 2. Allow quantification of (?>) groups, and make it work correctly.
5615
5616 3. The first character computation wasn't working for (?>) groups.
5617
5618 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5619 end of the subject) and add 5.005's \z, which really does match only at the
5620 very end of the subject.
5621
5622 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5623
5624 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5625 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5626 localized options. All options to pcre_study() were also removed.
5627
5628 7. Add other new features from 5.005:
5629
5630 $(?<= positive lookbehind
5631 $(?<! negative lookbehind
5632 (?imsx-imsx) added the unsetting capability
5633 such a setting is global if at outer level; local otherwise
5634 (?imsx-imsx:) non-capturing groups with option setting
5635 (?(cond)re|re) conditional pattern matching
5636
5637 A backreference to itself in a repeated group matches the previous
5638 captured string.
5639
5640 8. General tidying up of studying (both automatic and via "study")
5641 consequential on the addition of new assertions.
5642
5643 9. As in 5.005, unlimited repeated groups that could match an empty substring
5644 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5645 runtime if any iteration does actually match an empty substring.
5646
5647 10. Include the RunTest script in the distribution.
5648
5649 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5650 discrepancies, some of which were old and were also with respect to 5.004. They
5651 have now been fixed.
5652
5653
5654 Version 1.09 28-Apr-98
5655 ----------------------
5656
5657 1. A negated single character class followed by a quantifier with a minimum
5658 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5659 program crashes, or just wrong answers. This did not apply to negated classes
5660 containing more than one character, or to minima other than one.
5661
5662
5663 Version 1.08 27-Mar-98
5664 ----------------------
5665
5666 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5667
5668 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5669 latter must appear before anything that relies on it in the pattern.
5670
5671
5672 Version 1.07 16-Feb-98
5673 ----------------------
5674
5675 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5676 repeat of a potentially empty string).
5677
5678
5679 Version 1.06 23-Jan-98
5680 ----------------------
5681
5682 1. Added Markus Oberhumer's little patches for C++.
5683
5684 2. Literal strings longer than 255 characters were broken.
5685
5686
5687 Version 1.05 23-Dec-97
5688 ----------------------
5689
5690 1. Negated character classes containing more than one character were failing if
5691 PCRE_CASELESS was set at run time.
5692
5693
5694 Version 1.04 19-Dec-97
5695 ----------------------
5696
5697 1. Corrected the man page, where some "const" qualifiers had been omitted.
5698
5699 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5700 input syntax.
5701
5702 3. Fixed memory leak which occurred when a regex with back references was
5703 matched with an offsets vector that wasn't big enough. The temporary memory
5704 that is used in this case wasn't being freed if the match failed.
5705
5706 4. Tidied pcretest to ensure it frees memory that it gets.
5707
5708 5. Temporary memory was being obtained in the case where the passed offsets
5709 vector was exactly big enough.
5710
5711 6. Corrected definition of offsetof() from change 5 below.
5712
5713 7. I had screwed up change 6 below and broken the rules for the use of
5714 setjmp(). Now fixed.
5715
5716
5717 Version 1.03 18-Dec-97
5718 ----------------------
5719
5720 1. A erroneous regex with a missing opening parenthesis was correctly
5721 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5722 on some systems.
5723
5724 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5725 it was reported that one broken compiler failed on the former because "code" is
5726 also an independent variable.
5727
5728 3. The erroneous regex a[]b caused an array overrun reference.
5729
5730 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5731 fail on data ending with that character. (It was going on too far, and checking
5732 the next character, typically a binary zero.) This was specific to the
5733 optimized code for single-character negative classes.
5734
5735 5. Added a contributed patch from the TIN world which does the following:
5736
5737 + Add an undef for memmove, in case the the system defines a macro for it.
5738
5739 + Add a definition of offsetof(), in case there isn't one. (I don't know
5740 the reason behind this - offsetof() is part of the ANSI standard - but
5741 it does no harm).
5742
5743 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5744 most of the places where whitespace preceded '#'. I have given up and
5745 allowed the remaining 2 cases to be at the margin.
5746
5747 + Rename some variables in pcre to eliminate shadowing. This seems very
5748 pedantic, but does no harm, of course.
5749
5750 6. Moved the call to setjmp() into its own function, to get rid of warnings
5751 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5752
5753 7. Constructs such as \d{8,} were compiling into the equivalent of
5754 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5755 outcome, but in this particular case used more store than had been allocated,
5756 which caused the bug to be discovered because it threw up an internal error.
5757
5758 8. The debugging code in both pcre and pcretest for outputting the compiled
5759 form of a regex was going wrong in the case of back references followed by
5760 curly-bracketed repeats.
5761
5762
5763 Version 1.02 12-Dec-97
5764 ----------------------
5765
5766 1. Typos in pcre.3 and comments in the source fixed.
5767
5768 2. Applied a contributed patch to get rid of places where it used to remove
5769 'const' from variables, and fixed some signed/unsigned and uninitialized
5770 variable warnings.
5771
5772 3. Added the "runtest" target to Makefile.
5773
5774 4. Set default compiler flag to -O2 rather than just -O.
5775
5776
5777 Version 1.01 19-Nov-97
5778 ----------------------
5779
5780 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5781 like /([ab]*)*/, that is, for classes with more than one character in them.
5782
5783 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5784 as /((?>a*))*/ (a PCRE_EXTRA facility).
5785
5786
5787 Version 1.00 18-Nov-97
5788 ----------------------
5789
5790 1. Added compile-time macros to support systems such as SunOS4 which don't have
5791 memmove() or strerror() but have other things that can be used instead.
5792
5793 2. Arranged that "make clean" removes the executables.
5794
5795
5796 Version 0.99 27-Oct-97
5797 ----------------------
5798
5799 1. Fixed bug in code for optimizing classes with only one character. It was
5800 initializing a 32-byte map regardless, which could cause it to run off the end
5801 of the memory it had got.
5802
5803 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5804
5805
5806 Version 0.98 22-Oct-97
5807 ----------------------
5808
5809 1. Fixed bug in code for handling temporary memory usage when there are more
5810 back references than supplied space in the ovector. This could cause segfaults.
5811
5812
5813 Version 0.97 21-Oct-97
5814 ----------------------
5815
5816 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5817
5818 2. Optimized negated single characters not to use a bit map.
5819
5820 3. Brought error texts together as macro definitions; clarified some of them;
5821 fixed one that was wrong - it said "range out of order" when it meant "invalid
5822 escape sequence".
5823
5824 4. Changed some char * arguments to const char *.
5825
5826 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5827
5828 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5829 pcretest.
5830
5831
5832 Version 0.96 16-Oct-97
5833 ----------------------
5834
5835 1. Added a simple "pgrep" utility to the distribution.
5836
5837 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5838 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5839 where "ddd" means "one or more decimal digits".
5840
5841 3. Fixed serious bug. If a pattern had a back reference, but the call to
5842 pcre_exec() didn't supply a large enough ovector to record the related
5843 identifying subpattern, the match always failed. PCRE now remembers the number
5844 of the largest back reference, and gets some temporary memory in which to save
5845 the offsets during matching if necessary, in order to ensure that
5846 backreferences always work.
5847
5848 4. Increased the compatibility with Perl in a number of ways:
5849
5850 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5851 to request this handling. The option can be set at compile or exec time.
5852
5853 (b) $ matches before a terminating newline by default; an option
5854 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5855 mode). The option can be set at compile or exec time.
5856
5857 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5858 the same as Perl's. If the decimal number it represents is less than 10
5859 or there aren't that many previous left capturing parentheses, an octal
5860 escape is read. Inside a character class, it's always an octal escape,
5861 even if it is a single digit.
5862
5863 (d) An escaped but undefined alphabetic character is taken as a literal,
5864 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5865 escapes.
5866
5867 (e) {0} is now permitted. (The previous item is removed from the compiled
5868 pattern).
5869
5870 5. Changed all the names of code files so that the basic parts are no longer
5871 than 10 characters, and abolished the teeny "globals.c" file.
5872
5873 6. Changed the handling of character classes; they are now done with a 32-byte
5874 bit map always.
5875
5876 7. Added the -d and /D options to pcretest to make it possible to look at the
5877 internals of compilation without having to recompile pcre.
5878
5879
5880 Version 0.95 23-Sep-97
5881 ----------------------
5882
5883 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5884 \x20 at the start of a run of normal characters. These were being treated as
5885 real characters, instead of the source characters being re-checked.
5886
5887
5888 Version 0.94 18-Sep-97
5889 ----------------------
5890
5891 1. The functions are now thread-safe, with the caveat that the global variables
5892 containing pointers to malloc() and free() or alternative functions are the
5893 same for all threads.
5894
5895 2. Get pcre_study() to generate a bitmap of initial characters for non-
5896 anchored patterns when this is possible, and use it if passed to pcre_exec().
5897
5898
5899 Version 0.93 15-Sep-97
5900 ----------------------
5901
5902 1. /(b)|(:+)/ was computing an incorrect first character.
5903
5904 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5905 but not actually doing anything yet.
5906
5907 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5908 as Perl does (e.g. [-az] or [az-]).
5909
5910 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5911 all possible positions.
5912
5913 5. Split up into different modules to avoid including unneeded functions in a
5914 compiled binary. However, compile and exec are still in one module. The "study"
5915 function is split off.
5916
5917 6. The character tables are now in a separate module whose source is generated
5918 by an auxiliary program - but can then be edited by hand if required. There are
5919 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5920 toupper() in the code.
5921
5922 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5923 make them global. Abolish the function for setting them, as the caller can now
5924 set them directly.
5925
5926
5927 Version 0.92 11-Sep-97
5928 ----------------------
5929
5930 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5931 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5932
5933 2. Caseless matching was not working in character classes if the characters in
5934 the pattern were in upper case.
5935
5936 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5937
5938 4. Make PCRE_ANCHORED public and accept as a compile option.
5939
5940 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5941 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5942 pass them.
5943
5944 6. Give an error if bad option bits passed at compile or run time.
5945
5946 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5947 pcretest to cause it to pass that flag.
5948
5949 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5950 options, and the first character, if set.
5951
5952 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5953
5954
5955 Version 0.91 10-Sep-97
5956 ----------------------
5957
5958 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5959 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5960
5961 2. PCRE was looping on encountering an indefinitely repeated back reference to
5962 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5963 Perl does - treats the match as successful.
5964
5965 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5