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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1618 - (show annotations)
Sat Dec 5 16:30:14 2015 UTC (3 years, 11 months ago) by ph10
File size: 286957 byte(s)
Fix copy named substring bug.
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
40 Version 8.38 23-November-2015
41 -----------------------------
42
43 1. If a group that contained a recursive back reference also contained a
44 forward reference subroutine call followed by a non-forward-reference
45 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
46 compile correct code, leading to undefined behaviour or an internally
47 detected error. This bug was discovered by the LLVM fuzzer.
48
49 2. Quantification of certain items (e.g. atomic back references) could cause
50 incorrect code to be compiled when recursive forward references were
51 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
52 This bug was discovered by the LLVM fuzzer.
53
54 3. A repeated conditional group whose condition was a reference by name caused
55 a buffer overflow if there was more than one group with the given name.
56 This bug was discovered by the LLVM fuzzer.
57
58 4. A recursive back reference by name within a group that had the same name as
59 another group caused a buffer overflow. For example:
60 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
61
62 5. A forward reference by name to a group whose number is the same as the
63 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
64 a buffer overflow at compile time. This bug was discovered by the LLVM
65 fuzzer.
66
67 6. A lookbehind assertion within a set of mutually recursive subpatterns could
68 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
69
70 7. Another buffer overflow bug involved duplicate named groups with a
71 reference between their definition, with a group that reset capture
72 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
73 fixed by always allowing for more memory, even if not needed. (A proper fix
74 is implemented in PCRE2, but it involves more refactoring.)
75
76 8. There was no check for integer overflow in subroutine calls such as (?123).
77
78 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
79 being treated as a literal 'l' instead of causing an error.
80
81 10. There was a buffer overflow if pcre_exec() was called with an ovector of
82 size 1. This bug was found by american fuzzy lop.
83
84 11. If a non-capturing group containing a conditional group that could match
85 an empty string was repeated, it was not identified as matching an empty
86 string itself. For example: /^(?:(?(1)x|)+)+$()/.
87
88 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
89 \a and \e in test subject lines.
90
91 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
92 instead of the EBCDIC value.
93
94 14. The handling of \c in an EBCDIC environment has been revised so that it is
95 now compatible with the specification in Perl's perlebcdic page.
96
97 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
98 ASCII/Unicode. This has now been added to the list of characters that are
99 recognized as white space in EBCDIC.
100
101 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
102 error (correctly) when used outside a class, but did not give an error
103 within a class.
104
105 17. \h within a class was incorrectly compiled in EBCDIC environments.
106
107 18. A pattern with an unmatched closing parenthesis that contained a backward
108 assertion which itself contained a forward reference caused buffer
109 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
110
111 19. JIT should return with error when the compiled pattern requires more stack
112 space than the maximum.
113
114 20. A possessively repeated conditional group that could match an empty string,
115 for example, /(?(R))*+/, was incorrectly compiled.
116
117 21. Fix infinite recursion in the JIT compiler when certain patterns such as
118 /(?:|a|){100}x/ are analysed.
119
120 22. Some patterns with character classes involving [: and \\ were incorrectly
121 compiled and could cause reading from uninitialized memory or an incorrect
122 error diagnosis.
123
124 23. Pathological patterns containing many nested occurrences of [: caused
125 pcre_compile() to run for a very long time.
126
127 24. A conditional group with only one branch has an implicit empty alternative
128 branch and must therefore be treated as potentially matching an empty
129 string.
130
131 25. If (?R was followed by - or + incorrect behaviour happened instead of a
132 diagnostic.
133
134 26. Arrange to give up on finding the minimum matching length for overly
135 complex patterns.
136
137 27. Similar to (4) above: in a pattern with duplicated named groups and an
138 occurrence of (?| it is possible for an apparently non-recursive back
139 reference to become recursive if a later named group with the relevant
140 number is encountered. This could lead to a buffer overflow. Wen Guanxing
141 from Venustech ADLAB discovered this bug.
142
143 28. If pcregrep was given the -q option with -c or -l, or when handling a
144 binary file, it incorrectly wrote output to stdout.
145
146 29. The JIT compiler did not restore the control verb head in case of *THEN
147 control verbs. This issue was found by Karl Skomski with a custom LLVM
148 fuzzer.
149
150 30. Error messages for syntax errors following \g and \k were giving inaccurate
151 offsets in the pattern.
152
153 31. Added a check for integer overflow in conditions (?(<digits>) and
154 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
155 fuzzer.
156
157 32. Handling recursive references such as (?2) when the reference is to a group
158 later in the pattern uses code that is very hacked about and error-prone.
159 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
160 give an internal error if it is obvious that compiling has gone wrong.
161
162 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
163 This issue was found by Karl Skomski with a custom LLVM fuzzer.
164
165 34. The JIT compiler should restore the control chain for empty possessive
166 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
167
168 35. Match limit check added to JIT recursion. This issue was found by Karl
169 Skomski with a custom LLVM fuzzer.
170
171 36. Yet another case similar to 27 above has been circumvented by an
172 unconditional allocation of extra memory. This issue is fixed "properly" in
173 PCRE2 by refactoring the way references are handled. Wen Guanxing
174 from Venustech ADLAB discovered this bug.
175
176 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
177 with a custom LLVM fuzzer.
178
179 38. Fixed a corner case of range optimization in JIT.
180
181 39. An incorrect error "overran compiling workspace" was given if there were
182 exactly enough group forward references such that the last one extended
183 into the workspace safety margin. The next one would have expanded the
184 workspace. The test for overflow was not including the safety margin.
185
186 40. A match limit issue is fixed in JIT which was found by Karl Skomski
187 with a custom LLVM fuzzer.
188
189 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
190 work under Windows. (Why has it taken so long for anyone to notice?)
191
192 42. In a character class such as [\W\p{Any}] where both a negative-type escape
193 ("not a word character") and a property escape were present, the property
194 escape was being ignored.
195
196 43. Fix crash caused by very long (*MARK) or (*THEN) names.
197
198 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
199 by a single ASCII character in a class item, was incorrectly compiled in
200 UCP mode. The POSIX class got lost, but only if the single character
201 followed it.
202
203 45. [:punct:] in UCP mode was matching some characters in the range 128-255
204 that should not have been matched.
205
206 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
207 class, all characters with code points greater than 255 are in the class.
208 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
209 such as \w are turned into Unicode properties), wide characters were not
210 correctly handled, and could fail to match.
211
212
213 Version 8.37 28-April-2015
214 --------------------------
215
216 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
217 for those parentheses to be closed with whatever has been captured so far.
218 However, it was failing to mark any other groups between the hightest
219 capture so far and the currrent group as "unset". Thus, the ovector for
220 those groups contained whatever was previously there. An example is the
221 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
222
223 2. If an assertion condition was quantified with a minimum of zero (an odd
224 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
225
226 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
227 unrecognized modifier, a crash could occur.
228
229 4. An attempt to do global matching in pcretest with a zero-length ovector
230 caused a crash.
231
232 5. Fixed a memory leak during matching that could occur for a subpattern
233 subroutine call (recursive or otherwise) if the number of captured groups
234 that had to be saved was greater than ten.
235
236 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
237 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
238 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
239
240 7. A UTF pattern containing a "not" match of a non-ASCII character and a
241 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
242
243 8. When a pattern is compiled, it remembers the highest back reference so that
244 when matching, if the ovector is too small, extra memory can be obtained to
245 use instead. A conditional subpattern whose condition is a check on a
246 capture having happened, such as, for example in the pattern
247 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
248 setting the highest backreference number. This mattered only if pcre_exec()
249 was called with an ovector that was too small to hold the capture, and there
250 was no other kind of back reference (a situation which is probably quite
251 rare). The effect of the bug was that the condition was always treated as
252 FALSE when the capture could not be consulted, leading to a incorrect
253 behaviour by pcre_exec(). This bug has been fixed.
254
255 9. A reference to a duplicated named group (either a back reference or a test
256 for being set in a conditional) that occurred in a part of the pattern where
257 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
258 to be incorrectly calculated, leading to overwriting.
259
260 10. A mutually recursive set of back references such as (\2)(\1) caused a
261 segfault at study time (while trying to find the minimum matching length).
262 The infinite loop is now broken (with the minimum length unset, that is,
263 zero).
264
265 11. If an assertion that was used as a condition was quantified with a minimum
266 of zero, matching went wrong. In particular, if the whole group had
267 unlimited repetition and could match an empty string, a segfault was
268 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
269 assertions to be quantified, but not if they are being used as conditions,
270 so the above pattern is faulted by Perl. PCRE has now been changed so that
271 it also rejects such patterns.
272
273 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
274 failed to allow the zero-repeat case if pcre2_exec() was called with an
275 ovector too small to capture the group.
276
277 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
278 Red Hat Product Security:
279
280 (a) A crash if /K and /F were both set with the option to save the compiled
281 pattern.
282
283 (b) Another crash if the option to print captured substrings in a callout
284 was combined with setting a null ovector, for example \O\C+ as a subject
285 string.
286
287 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
288 forward reference repeated a large (but limited) number of times within a
289 repeated outer group that has a zero minimum quantifier, caused incorrect
290 code to be compiled, leading to the error "internal error:
291 previously-checked referenced subpattern not found" when an incorrect
292 memory address was read. This bug was reported as "heap overflow",
293 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
294 CVE-2015-2325.
295
296 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
297 call within a group that also contained a recursive back reference caused
298 incorrect code to be compiled. This bug was reported as "heap overflow",
299 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
300 number CVE-2015-2326.
301
302 24. Computing the size of the JIT read-only data in advance has been a source
303 of various issues, and new ones are still appear unfortunately. To fix
304 existing and future issues, size computation is eliminated from the code,
305 and replaced by on-demand memory allocation.
306
307 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
308 adjacent to the end of the range, and the range contained characters with
309 more than one other case, caused incorrect behaviour when compiled in UTF
310 mode. In that example, the range a-j was left out of the class.
311
312 26. Fix JIT compilation of conditional blocks, which assertion
313 is converted to (*FAIL). E.g: /(?(?!))/.
314
315 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
316 discovered by the LLVM fuzzer.
317
318 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
319 when this assertion was used as a condition, for example (?(?!)a|b). In
320 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
321 error about an unsupported item.
322
323 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
324 possessification code could take exponential time to complete. A recursion
325 depth limit of 1000 has been imposed to limit the resources used by this
326 optimization.
327
328 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
329 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
330 because \S ensures they are all in the class. The code for doing this was
331 interacting badly with the code for computing the amount of space needed to
332 compile the pattern, leading to a buffer overflow. This bug was discovered
333 by the LLVM fuzzer.
334
335 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
336 other kinds of group caused stack overflow at compile time. This bug was
337 discovered by the LLVM fuzzer.
338
339 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
340 between a subroutine call and its quantifier was incorrectly compiled,
341 leading to buffer overflow or other errors. This bug was discovered by the
342 LLVM fuzzer.
343
344 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
345 assertion after (?(. The code was failing to check the character after
346 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
347 was discovered by the LLVM fuzzer.
348
349 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
350 a fixed maximum following a group that contains a subroutine reference was
351 incorrectly compiled and could trigger buffer overflow. This bug was
352 discovered by the LLVM fuzzer.
353
354 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
355 caused a stack overflow instead of the diagnosis of a non-fixed length
356 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
357
358 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
359 (e.g. /(?<=\Ka)/) could make pcregrep loop.
360
361 37. There was a similar problem to 36 in pcretest for global matches.
362
363 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
364 and a subsequent item in the pattern caused a non-match, backtracking over
365 the repeated \X did not stop, but carried on past the start of the subject,
366 causing reference to random memory and/or a segfault. There were also some
367 other cases where backtracking after \C could crash. This set of bugs was
368 discovered by the LLVM fuzzer.
369
370 39. The function for finding the minimum length of a matching string could take
371 a very long time if mutual recursion was present many times in a pattern,
372 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
373 method has been implemented. This infelicity was discovered by the LLVM
374 fuzzer.
375
376 40. Static linking against the PCRE library using the pkg-config module was
377 failing on missing pthread symbols.
378
379
380 Version 8.36 26-September-2014
381 ------------------------------
382
383 1. Got rid of some compiler warnings in the C++ modules that were shown up by
384 -Wmissing-field-initializers and -Wunused-parameter.
385
386 2. The tests for quantifiers being too big (greater than 65535) were being
387 applied after reading the number, and stupidly assuming that integer
388 overflow would give a negative number. The tests are now applied as the
389 numbers are read.
390
391 3. Tidy code in pcre_exec.c where two branches that used to be different are
392 now the same.
393
394 4. The JIT compiler did not generate match limit checks for certain
395 bracketed expressions with quantifiers. This may lead to exponential
396 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
397 issue should be resolved now.
398
399 5. Fixed an issue, which occures when nested alternatives are optimized
400 with table jumps.
401
402 6. Inserted two casts and changed some ints to size_t in the light of some
403 reported 64-bit compiler warnings (Bugzilla 1477).
404
405 7. Fixed a bug concerned with zero-minimum possessive groups that could match
406 an empty string, which sometimes were behaving incorrectly in the
407 interpreter (though correctly in the JIT matcher). This pcretest input is
408 an example:
409
410 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
411 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
412
413 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
414 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
415 for an empty string was breaking the inner loop and carrying on at a lower
416 level, when possessive repeated groups should always return to a higher
417 level as they have no backtrack points in them. The empty string test now
418 occurs at the outer level.
419
420 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
421 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
422
423 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
424 doesn't).
425
426 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
427 to hold a pattern to be incorrectly computed (too small) when there were
428 named back references to duplicated names. This could cause "internal
429 error: code overflow" or "double free or corruption" or other memory
430 handling errors.
431
432 11. When named subpatterns had the same prefixes, back references could be
433 confused. For example, in this pattern:
434
435 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
436
437 the reference to 'Name' was incorrectly treated as a reference to a
438 duplicate name.
439
440 12. A pattern such as /^s?c/mi8 where the optional character has more than
441 one "other case" was incorrectly compiled such that it would only try to
442 match starting at "c".
443
444 13. When a pattern starting with \s was studied, VT was not included in the
445 list of possible starting characters; this should have been part of the
446 8.34/18 patch.
447
448 14. If a character class started [\Qx]... where x is any character, the class
449 was incorrectly terminated at the ].
450
451 15. If a pattern that started with a caseless match for a character with more
452 than one "other case" was studied, PCRE did not set up the starting code
453 unit bit map for the list of possible characters. Now it does. This is an
454 optimization improvement, not a bug fix.
455
456 16. The Unicode data tables have been updated to Unicode 7.0.0.
457
458 17. Fixed a number of memory leaks in pcregrep.
459
460 18. Avoid a compiler warning (from some compilers) for a function call with
461 a cast that removes "const" from an lvalue by using an intermediate
462 variable (to which the compiler does not object).
463
464 19. Incorrect code was compiled if a group that contained an internal recursive
465 back reference was optional (had quantifier with a minimum of zero). This
466 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
467 caused segmentation faults because of stack overflows at compile time.
468
469 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
470 group that is quantified with an indefinite repeat, caused a compile-time
471 loop which used up all the system stack and provoked a segmentation fault.
472 This was not the same bug as 19 above.
473
474 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
475 Patch by Mike Frysinger.
476
477
478 Version 8.35 04-April-2014
479 --------------------------
480
481 1. A new flag is set, when property checks are present in an XCLASS.
482 When this flag is not set, PCRE can perform certain optimizations
483 such as studying these XCLASS-es.
484
485 2. The auto-possessification of character sets were improved: a normal
486 and an extended character set can be compared now. Furthermore
487 the JIT compiler optimizes more character set checks.
488
489 3. Got rid of some compiler warnings for potentially uninitialized variables
490 that show up only when compiled with -O2.
491
492 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
493 of a match later then the end of the match. The pcretest program was not
494 handling the case sensibly - it was outputting from the start to the next
495 binary zero. It now reports this situation in a message, and outputs the
496 text from the end to the start.
497
498 5. Fast forward search is improved in JIT. Instead of the first three
499 characters, any three characters with fixed position can be searched.
500 Search order: first, last, middle.
501
502 6. Improve character range checks in JIT. Characters are read by an inprecise
503 function now, which returns with an unknown value if the character code is
504 above a certain threshold (e.g: 256). The only limitation is that the value
505 must be bigger than the threshold as well. This function is useful when
506 the characters above the threshold are handled in the same way.
507
508 7. The macros whose names start with RAWUCHAR are placeholders for a future
509 mode in which only the bottom 21 bits of 32-bit data items are used. To
510 make this more memorable for those maintaining the code, the names have
511 been changed to start with UCHAR21, and an extensive comment has been added
512 to their definition.
513
514 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
515 to the export list in Makefile.am (they were accidentally omitted from the
516 8.34 tarball).
517
518 9. The informational output from pcretest used the phrase "starting byte set"
519 which is inappropriate for the 16-bit and 32-bit libraries. As the output
520 for "first char" and "need char" really means "non-UTF-char", I've changed
521 "byte" to "char", and slightly reworded the output. The documentation about
522 these values has also been (I hope) clarified.
523
524 10. Another JIT related optimization: use table jumps for selecting the correct
525 backtracking path, when more than four alternatives are present inside a
526 bracket.
527
528 11. Empty match is not possible, when the minimum length is greater than zero,
529 and there is no \K in the pattern. JIT should avoid empty match checks in
530 such cases.
531
532 12. In a caseless character class with UCP support, when a character with more
533 than one alternative case was not the first character of a range, not all
534 the alternative cases were added to the class. For example, s and \x{17f}
535 are both alternative cases for S: the class [RST] was handled correctly,
536 but [R-T] was not.
537
538 13. The configure.ac file always checked for pthread support when JIT was
539 enabled. This is not used in Windows, so I have put this test inside a
540 check for the presence of windows.h (which was already tested for).
541
542 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
543 The algorithm provides a way to skip certain starting offsets, and usually
544 faster than linear prefix searches.
545
546 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
547 as for 'fr_FR' and 'french'. For some reason, however, it then used the
548 Windows-specific input and output files, which have 'french' screwed in.
549 So this could never have worked. One of the problems with locales is that
550 they aren't always the same. I have now updated RunTest so that it checks
551 the output of the locale test (test 3) against three different output
552 files, and it allows the test to pass if any one of them matches. With luck
553 this should make the test pass on some versions of Solaris where it was
554 failing. Because of the uncertainty, the script did not used to stop if
555 test 3 failed; it now does. If further versions of a French locale ever
556 come to light, they can now easily be added.
557
558 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
559 there was a message during ./configure, but it did not stop. This now
560 provokes an error. The invalid example in README has been corrected.
561 If a value less than the minimum is given, the minimum value has always
562 been used, but now a warning is given.
563
564 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
565 was a bug in the test system, which is now fixed. Also, the list of various
566 configurations that are tested for each release did not have one with both
567 16/32 bits and --enable-bar-anycrlf. It now does.
568
569 18. pcretest was missing "-C bsr" for displaying the \R default setting.
570
571 19. Little endian PowerPC systems are supported now by the JIT compiler.
572
573 20. The fast forward newline mechanism could enter to an infinite loop on
574 certain invalid UTF-8 input. Although we don't support these cases
575 this issue can be fixed by a performance optimization.
576
577 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
578 not take account if existing stack usage. There is now a new global
579 variable called pcre_stack_guard that can be set to point to an external
580 function to check stack availability. It is called at the start of
581 processing every parenthesized group.
582
583 22. A typo in the code meant that in ungreedy mode the max/min qualifier
584 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
585 not match "ab".
586
587 23. When UTF was disabled, the JIT program reported some incorrect compile
588 errors. These messages are silenced now.
589
590 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
591 compiler.
592
593 25. Change all the temporary files used in RunGrepTest to be different to those
594 used by RunTest so that the tests can be run simultaneously, for example by
595 "make -j check".
596
597
598 Version 8.34 15-December-2013
599 -----------------------------
600
601 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
602 executable memory. Patch inspired by Carsten Klein.
603
604 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
605 this macro is never tested and has no effect, because the work to support
606 coverage involves only compiling and linking options and special targets in
607 the Makefile. The comment in config.h implied that defining the macro would
608 enable coverage support, which is totally false. There was also support for
609 setting this macro in the CMake files (my fault, I just copied it from
610 configure). SUPPORT_GCOV has now been removed.
611
612 3. Make a small performance improvement in strlen16() and strlen32() in
613 pcretest.
614
615 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
616 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
617 these cases). There was also one in pcretest.c.
618
619 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
620
621 6. In UTF mode, the code for checking whether a group could match an empty
622 string (which is used for indefinitely repeated groups to allow for
623 breaking an infinite loop) was broken when the group contained a repeated
624 negated single-character class with a character that occupied more than one
625 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
626 UTF-8 mode). The effect was undefined: the group might or might not be
627 deemed as matching an empty string, or the program might have crashed.
628
629 7. The code for checking whether a group could match an empty string was not
630 recognizing that \h, \H, \v, \V, and \R must match a character.
631
632 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
633 an empty string. If it can, pcretest shows this in its information output.
634
635 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
636 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
637 matched by pcre_exec() without using JIT:
638
639 (a) If the rest of the pattern did not match after a maximal run of
640 grapheme clusters, the code for backing up to try with fewer of them
641 did not always back up over a full grapheme when characters that do not
642 have the modifier quality were involved, e.g. Hangul syllables.
643
644 (b) If the match point in a subject started with modifier character, and
645 there was no match, the code could incorrectly back up beyond the match
646 point, and potentially beyond the first character in the subject,
647 leading to a segfault or an incorrect match result.
648
649 10. A conditional group with an assertion condition could lead to PCRE
650 recording an incorrect first data item for a match if no other first data
651 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
652 first data item, and therefore matched "ca" after "c" instead of at the
653 start.
654
655 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
656 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
657
658 12. The source of pcregrep now includes z/OS-specific code so that it can be
659 compiled for z/OS as part of the special z/OS distribution.
660
661 13. Added the -T and -TM options to pcretest.
662
663 14. The code in pcre_compile.c for creating the table of named capturing groups
664 has been refactored. Instead of creating the table dynamically during the
665 actual compiling pass, the information is remembered during the pre-compile
666 pass (on the stack unless there are more than 20 named groups, in which
667 case malloc() is used) and the whole table is created before the actual
668 compile happens. This has simplified the code (it is now nearly 150 lines
669 shorter) and prepared the way for better handling of references to groups
670 with duplicate names.
671
672 15. A back reference to a named subpattern when there is more than one of the
673 same name now checks them in the order in which they appear in the pattern.
674 The first one that is set is used for the reference. Previously only the
675 first one was inspected. This change makes PCRE more compatible with Perl.
676
677 16. Unicode character properties were updated from Unicode 6.3.0.
678
679 17. The compile-time code for auto-possessification has been refactored, based
680 on a patch by Zoltan Herczeg. It now happens after instead of during
681 compilation. The code is cleaner, and more cases are handled. The option
682 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
683 options in pcretest are provided to set it. It can also be set by
684 (*NO_AUTO_POSSESS) at the start of a pattern.
685
686 18. The character VT has been added to the default ("C" locale) set of
687 characters that match \s and are generally treated as white space,
688 following this same change in Perl 5.18. There is now no difference between
689 "Perl space" and "POSIX space". Whether VT is treated as white space in
690 other locales depends on the locale.
691
692 19. The code for checking named groups as conditions, either for being set or
693 for being recursed, has been refactored (this is related to 14 and 15
694 above). Processing unduplicated named groups should now be as fast at
695 numerical groups, and processing duplicated groups should be faster than
696 before.
697
698 20. Two patches to the CMake build system, by Alexander Barkov:
699
700 (1) Replace the "source" command by "." in CMakeLists.txt because
701 "source" is a bash-ism.
702
703 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
704 without these the CMake build does not work on Solaris.
705
706 21. Perl has changed its handling of \8 and \9. If there is no previously
707 encountered capturing group of those numbers, they are treated as the
708 literal characters 8 and 9 instead of a binary zero followed by the
709 literals. PCRE now does the same.
710
711 22. Following Perl, added \o{} to specify codepoints in octal, making it
712 possible to specify values greater than 0777 and also making them
713 unambiguous.
714
715 23. Perl now gives an error for missing closing braces after \x{... instead of
716 treating the string as literal. PCRE now does the same.
717
718 24. RunTest used to grumble if an inappropriate test was selected explicitly,
719 but just skip it when running all tests. This make it awkward to run ranges
720 of tests when one of them was inappropriate. Now it just skips any
721 inappropriate tests, as it always did when running all tests.
722
723 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
724 character types such as \d or \w, too many callouts were inserted, and the
725 data that they returned was rubbish.
726
727 26. In UCP mode, \s was not matching two of the characters that Perl matches,
728 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
729 were matched by \h. The code has now been refactored so that the lists of
730 the horizontal and vertical whitespace characters used for \h and \v (which
731 are defined only in one place) are now also used for \s.
732
733 27. Add JIT support for the 64 bit TileGX architecture.
734 Patch by Jiong Wang (Tilera Corporation).
735
736 28. Possessive quantifiers for classes (both explicit and automatically
737 generated) now use special opcodes instead of wrapping in ONCE brackets.
738
739 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
740 (because it's meaningless), this was not happening when PCRE_CASELESS was
741 set. Not wrong, but inefficient.
742
743 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
744 properties for \w, \d, etc) is present in a test regex. Otherwise if the
745 test contains no characters greater than 255, Perl doesn't realise it
746 should be using Unicode semantics.
747
748 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
749 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
750 does in Unicode mode.
751
752 32. Added the "forbid" facility to pcretest so that putting tests into the
753 wrong test files can sometimes be quickly detected.
754
755 33. There is now a limit (default 250) on the depth of nesting of parentheses.
756 This limit is imposed to control the amount of system stack used at compile
757 time. It can be changed at build time by --with-parens-nest-limit=xxx or
758 the equivalent in CMake.
759
760 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
761 errors. Perl warns for these when in warning mode, but PCRE has no facility
762 for giving warnings.
763
764 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
765 However, this was not working for (?!) because it is optimized to (*FAIL),
766 for which PCRE does not allow quantifiers. The optimization is now disabled
767 when a quantifier follows (?!). I can't see any use for this, but it makes
768 things uniform.
769
770 36. Perl no longer allows group names to start with digits, so I have made this
771 change also in PCRE. It simplifies the code a bit.
772
773 37. In extended mode, Perl ignores spaces before a + that indicates a
774 possessive quantifier. PCRE allowed a space before the quantifier, but not
775 before the possessive +. It now does.
776
777 38. The use of \K (reset reported match start) within a repeated possessive
778 group such as (a\Kb)*+ was not working.
779
780 40. Document that the same character tables must be used at compile time and
781 run time, and that the facility to pass tables to pcre_exec() and
782 pcre_dfa_exec() is for use only with saved/restored patterns.
783
784 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
785 features for Builds with MSVC:
786
787 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
788 builds). The .rc files can be used to set FileDescription and many other
789 attributes.
790
791 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
792 This allows higher-level build scripts which want .pdb files to avoid
793 hard-coding the exact files needed."
794
795 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
796 mean "start of word" and "end of word", respectively, as a transition aid.
797
798 43. A minimizing repeat of a class containing codepoints greater than 255 in
799 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
800 compiled to use the heap for recursion.
801
802 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
803 is configured.
804
805
806 Version 8.33 28-May-2013
807 ------------------------
808
809 1. Added 'U' to some constants that are compared to unsigned integers, to
810 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
811 variables that are added to signed variables, to ensure the result is
812 signed and can be negated.
813
814 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
815 CMake config files.
816
817 3. Revise the creation of config.h.generic so that all boolean macros are
818 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
819 overriding via -D on the command line possible.
820
821 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
822 to unsigned int is reported to make a quite noticeable speed difference in
823 a specific Windows environment. Testing on Linux did also appear to show
824 some benefit (and it is clearly not harmful). Also fixed the definition of
825 Xop which should be unsigned.
826
827 5. Related to (4), changing the definition of the intermediate variable cc
828 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
829 improvements.
830
831 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
832 unnecessary spaces.
833
834 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
835 and later.
836
837 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
838
839 9. Optimizing fast_forward_start_bits in JIT.
840
841 10. Adding support for callouts in JIT, and fixing some issues revealed
842 during this work. Namely:
843
844 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
845
846 (b) Minimum length was not checked before the matching is started.
847
848 11. The value of capture_last that is passed to callouts was incorrect in some
849 cases when there was a capture on one path that was subsequently abandoned
850 after a backtrack. Also, the capture_last value is now reset after a
851 recursion, since all captures are also reset in this case.
852
853 12. The interpreter no longer returns the "too many substrings" error in the
854 case when an overflowing capture is in a branch that is subsequently
855 abandoned after a backtrack.
856
857 13. In the pathological case when an offset vector of size 2 is used, pcretest
858 now prints out the matched string after a yield of 0 or 1.
859
860 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
861 Only supported by the JIT compiler at the moment.
862
863 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
864
865 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
866 offset of the starting point of the matching process, provided the offsets
867 vector is large enough.
868
869 17. The \A escape now records a lookbehind value of 1, though its execution
870 does not actually inspect the previous character. This is to ensure that,
871 in partial multi-segment matching, at least one character from the old
872 segment is retained when a new segment is processed. Otherwise, if there
873 are no lookbehinds in the pattern, \A might match incorrectly at the start
874 of a new segment.
875
876 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
877
878 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
879 gives some modest performance improvement in 8-bit mode.
880
881 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
882 be expressed in certain programming languages using Universal Character
883 Names.
884
885 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
886 which points out that "non characters" are not "characters that may not
887 appear in Unicode strings" but rather "characters that are reserved for
888 internal use and have only local meaning".
889
890 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
891 there was a conditional group that depended on an assertion, if the
892 assertion was false, the callout that immediately followed the alternation
893 in the condition was skipped when pcre_exec() was used for matching.
894
895 23. Allow an explicit callout to be inserted before an assertion that is the
896 condition for a conditional group, for compatibility with automatic
897 callouts, which always insert a callout at this point.
898
899 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
900 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
901
902 25. (*PRUNE) is now supported by the JIT compiler.
903
904 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
905
906 27. Fix the case where there are two or more SKIPs with arguments that may be
907 ignored.
908
909 28. (*SKIP) is now supported by the JIT compiler.
910
911 29. (*THEN) is now supported by the JIT compiler.
912
913 30. Update RunTest with additional test selector options.
914
915 31. The way PCRE handles backtracking verbs has been changed in two ways.
916
917 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
918 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
919 backtracking. In some cases this makes it more Perl-compatible, but Perl's
920 rather obscure rules do not always do the same thing.
921
922 (2) Previously, backtracking verbs were confined within assertions. This is
923 no longer the case for positive assertions, except for (*ACCEPT). Again,
924 this sometimes improves Perl compatibility, and sometimes does not.
925
926 32. A number of tests that were in test 2 because Perl did things differently
927 have been moved to test 1, because either Perl or PCRE has changed, and
928 these tests are now compatible.
929
930 32. Backtracking control verbs are now handled in the same way in JIT and
931 interpreter.
932
933 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
934 contained a forward subroutine reference caused a compile error.
935
936 34. Auto-detect and optimize limited repetitions in JIT.
937
938 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
939 blocking (*UTF) etc.
940
941 36. In the interpreter, maximizing pattern repetitions for characters and
942 character types now use tail recursion, which reduces stack usage.
943
944 37. The value of the max lookbehind was not correctly preserved if a compiled
945 and saved regex was reloaded on a host of different endianness.
946
947 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
948 of the compiled pattern block, expand the flags field from 16 to 32 bits
949 because it was almost full.
950
951 39. Try madvise first before posix_madvise.
952
953 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
954 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
955 original change locked out all matches of empty strings. This has been
956 changed so that one match of an empty string per line is recognized.
957 Subsequent searches on the same line (for colouring or for --only-matching,
958 for example) do not recognize empty strings.
959
960 41. Applied a user patch to fix a number of spelling mistakes in comments.
961
962 42. Data lines longer than 65536 caused pcretest to crash.
963
964 43. Clarified the data type for length and startoffset arguments for pcre_exec
965 and pcre_dfa_exec in the function-specific man pages, where they were
966 explicitly stated to be in bytes, never having been updated. I also added
967 some clarification to the pcreapi man page.
968
969 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
970 a segmentation fault.
971
972
973 Version 8.32 30-November-2012
974 -----------------------------
975
976 1. Improved JIT compiler optimizations for first character search and single
977 character iterators.
978
979 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
980 Patch by Daniel Richard G.
981
982 3. Single character iterator optimizations in the JIT compiler.
983
984 4. Improved JIT compiler optimizations for character ranges.
985
986 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
987 Reported by Giuseppe D'Angelo.
988
989 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
990 of a line, was being set incorrectly in cases where .* appeared inside
991 atomic brackets at the start of a pattern, or where there was a subsequent
992 *PRUNE or *SKIP.
993
994 7. Improved instruction cache flush for POWER/PowerPC.
995 Patch by Daniel Richard G.
996
997 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
998 grep:
999
1000 (a) There is now no limit to the number of patterns to be matched.
1001
1002 (b) An error is given if a pattern is too long.
1003
1004 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1005 are now supported.
1006
1007 (d) --exclude-from and --include-from (multiple use) have been added.
1008
1009 (e) Exclusions and inclusions now apply to all files and directories, not
1010 just to those obtained from scanning a directory recursively.
1011
1012 (f) Multiple uses of -f and --file-list are now supported.
1013
1014 (g) In a Windows environment, the default for -d has been changed from
1015 "read" (the GNU grep default) to "skip", because otherwise the presence
1016 of a directory in the file list provokes an error.
1017
1018 (h) The documentation has been revised and clarified in places.
1019
1020 9. Improve the matching speed of capturing brackets.
1021
1022 10. Changed the meaning of \X so that it now matches a Unicode extended
1023 grapheme cluster.
1024
1025 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1026 out POSIX threads when JIT support is configured.
1027
1028 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1029
1030 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1031 structure could go wrong in environments where size_t is not the same size
1032 as int.
1033
1034 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1035 set.
1036
1037 15. The EBCDIC support had decayed; later updates to the code had included
1038 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1039 general tidy up of EBCDIC-related issues, and the documentation was also
1040 not quite right. There is now a test that can be run on ASCII systems to
1041 check some of the EBCDIC-related things (but is it not a full test).
1042
1043 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1044 in a small tidy to the code.
1045
1046 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1047
1048 18. If the --only-matching (-o) option in pcregrep is specified multiple
1049 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1050 substrings matched by the 1st and 2nd capturing parentheses. A separating
1051 string can be specified by --om-separator (default empty).
1052
1053 19. Improving the first n character searches.
1054
1055 20. Turn case lists for horizontal and vertical white space into macros so that
1056 they are defined only once.
1057
1058 21. This set of changes together give more compatible Unicode case-folding
1059 behaviour for characters that have more than one other case when UCP
1060 support is available.
1061
1062 (a) The Unicode property table now has offsets into a new table of sets of
1063 three or more characters that are case-equivalent. The MultiStage2.py
1064 script that generates these tables (the pcre_ucd.c file) now scans
1065 CaseFolding.txt instead of UnicodeData.txt for character case
1066 information.
1067
1068 (b) The code for adding characters or ranges of characters to a character
1069 class has been abstracted into a generalized function that also handles
1070 case-independence. In UTF-mode with UCP support, this uses the new data
1071 to handle characters with more than one other case.
1072
1073 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1074 whose other case is greater than 256 are now correctly matched
1075 caselessly. Previously, the high codepoint matched the low one, but not
1076 vice versa.
1077
1078 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1079 of the new class addition function, using character lists defined as
1080 macros alongside the case definitions of 20 above.
1081
1082 (e) Caseless back references now work with characters that have more than
1083 one other case.
1084
1085 (f) General caseless matching of characters with more than one other case
1086 is supported.
1087
1088 22. Unicode character properties were updated from Unicode 6.2.0
1089
1090 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1091
1092 24. Add support for 32-bit character strings, and UTF-32
1093
1094 25. Major JIT compiler update (code refactoring and bugfixing).
1095 Experimental Sparc 32 support is added.
1096
1097 26. Applied a modified version of Daniel Richard G's patch to create
1098 pcre.h.generic and config.h.generic by "make" instead of in the
1099 PrepareRelease script.
1100
1101 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1102 pcre_compile.c when checking for a zero character.
1103
1104 28. Introducing a native interface for JIT. Through this interface, the compiled
1105 machine code can be directly executed. The purpose of this interface is to
1106 provide fast pattern matching, so several sanity checks are not performed.
1107 However, feature tests are still performed. The new interface provides
1108 1.4x speedup compared to the old one.
1109
1110 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1111 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1112 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1113
1114 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1115 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1116 the "old" RFC 2279). Instead, it ended up passing a negative length to
1117 pcre_exec().
1118
1119 31. Add support for GCC's visibility feature to hide internal functions.
1120
1121 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1122 "unknown -C option" after outputting 0 or 1.
1123
1124 33. There is now support for generating a code coverage report for the test
1125 suite in environments where gcc is the compiler and lcov is installed. This
1126 is mainly for the benefit of the developers.
1127
1128 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1129 unaddressable using valgrind annotations, allowing valgrind to detect
1130 invalid memory accesses. This is mainly for the benefit of the developers.
1131
1132 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1133
1134 26. Give configure error if --enable-cpp but no C++ compiler found.
1135
1136
1137 Version 8.31 06-July-2012
1138 -------------------------
1139
1140 1. Fixing a wrong JIT test case and some compiler warnings.
1141
1142 2. Removed a bashism from the RunTest script.
1143
1144 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1145 to unsigned type, result still unsigned" that was given by an MS compiler
1146 on encountering the code "-sizeof(xxx)".
1147
1148 4. Partial matching support is added to the JIT compiler.
1149
1150 5. Fixed several bugs concerned with partial matching of items that consist
1151 of more than one character:
1152
1153 (a) /^(..)\1/ did not partially match "aba" because checking references was
1154 done on an "all or nothing" basis. This also applied to repeated
1155 references.
1156
1157 (b) \R did not give a hard partial match if \r was found at the end of the
1158 subject.
1159
1160 (c) \X did not give a hard partial match after matching one or more
1161 characters at the end of the subject.
1162
1163 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1164 a partial match for the string "\r".
1165
1166 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1167 a partial match for a CR character at the end of the subject string.
1168
1169 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1170 running pcretest, the text "(JIT)" added to the output whenever JIT is
1171 actually used to run the match.
1172
1173 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1174 or /S+[+] with a digit between 1 and 7.
1175
1176 8. OP_NOT now supports any UTF character not just single-byte ones.
1177
1178 9. (*MARK) control verb is now supported by the JIT compiler.
1179
1180 10. The command "./RunTest list" lists the available tests without actually
1181 running any of them. (Because I keep forgetting what they all are.)
1182
1183 11. Add PCRE_INFO_MAXLOOKBEHIND.
1184
1185 12. Applied a (slightly modified) user-supplied patch that improves performance
1186 when the heap is used for recursion (compiled with --disable-stack-for-
1187 recursion). Instead of malloc and free for each heap frame each time a
1188 logical recursion happens, frames are retained on a chain and re-used where
1189 possible. This sometimes gives as much as 30% improvement.
1190
1191 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1192 call.
1193
1194 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1195
1196 15. It is now possible to link pcretest with libedit as an alternative to
1197 libreadline.
1198
1199 16. (*COMMIT) control verb is now supported by the JIT compiler.
1200
1201 17. The Unicode data tables have been updated to Unicode 6.1.0.
1202
1203 18. Added --file-list option to pcregrep.
1204
1205 19. Added binary file support to pcregrep, including the -a, --binary-files,
1206 -I, and --text options.
1207
1208 20. The madvise function is renamed for posix_madvise for QNX compatibility
1209 reasons. Fixed by Giuseppe D'Angelo.
1210
1211 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1212
1213 22. Changed the option for creating symbolic links for 16-bit man pages from
1214 -s to -sf so that re-installing does not cause issues.
1215
1216 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1217
1218 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1219 matches in certain environments (the workspace was not being correctly
1220 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1221 some of the workspace data at the beginning of a restart.
1222
1223 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1224 was not doing so when it should - probably a typo introduced by SVN 528
1225 (change 8.10/14).
1226
1227 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1228 \w+ when the character tables indicated that \x{c4} was a word character.
1229 There were several related cases, all because the tests for doing a table
1230 lookup were testing for characters less than 127 instead of 255.
1231
1232 27. If a pattern contains capturing parentheses that are not used in a match,
1233 their slots in the ovector are set to -1. For those that are higher than
1234 any matched groups, this happens at the end of processing. In the case when
1235 there were back references that the ovector was too small to contain
1236 (causing temporary malloc'd memory to be used during matching), and the
1237 highest capturing number was not used, memory off the end of the ovector
1238 was incorrectly being set to -1. (It was using the size of the temporary
1239 memory instead of the true size.)
1240
1241 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1242 ovector size, it uses memory at the end of the block that it has got.
1243
1244 29. Check for an overlong MARK name and give an error at compile time. The
1245 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1246
1247 30. JIT compiler update.
1248
1249 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1250 Rill for the patch.
1251
1252 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1253
1254 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1255
1256 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1257 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1258 was enabled.
1259
1260 35. Improve JIT code generation for greedy plus quantifier.
1261
1262 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1263 1 to "aa" instead of to an empty string. The bug affected repeated groups
1264 that could potentially match an empty string.
1265
1266 37. Optimizing single character iterators in JIT.
1267
1268 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1269 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1270 codepoints that are too big for the mode are faulted, and in a UTF mode,
1271 disallowed codepoints are also faulted.
1272
1273 39. If PCRE was compiled with UTF support, in three places in the DFA
1274 matcher there was code that should only have been obeyed in UTF mode, but
1275 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1276 processing when bytes with values greater than 127 were present. In 16-bit
1277 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1278 both cases the values are those that cannot be the first data item in a UTF
1279 character. The three items that might have provoked this were recursions,
1280 possessively repeated groups, and atomic groups.
1281
1282 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1283 and pcregrep, because some OS require shared objects to be explicitly
1284 passed to ld, causing the link step to fail if they are not.
1285
1286 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1287 mode, patterns that started with \h* or \R* might be incorrectly matched.
1288
1289
1290 Version 8.30 04-February-2012
1291 -----------------------------
1292
1293 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1294 name is defined in ctype.h.
1295
1296 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1297 only in quite long subpatterns.
1298
1299 3. Removed the function pcre_info(), which has been obsolete and deprecated
1300 since it was replaced by pcre_fullinfo() in February 2000.
1301
1302 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1303 match a (*MARK), and the match failed at the start of the subject, a
1304 reference to memory before the start of the subject could occur. This bug
1305 was introduced by fix 17 of release 8.21.
1306
1307 5. A reference to an unset group with zero minimum repetition was giving
1308 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1309 /(another)?(\1?)test/ matched against "hello world test". This bug was
1310 introduced in release 8.13.
1311
1312 6. Add support for 16-bit character strings (a large amount of work involving
1313 many changes and refactorings).
1314
1315 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1316 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1317 from a file.
1318
1319 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1320 rounding is not applied in this particular case).
1321
1322 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1323 if they appear, or are escaped, in patterns.
1324
1325 10. Get rid of a number of -Wunused-but-set-variable warnings.
1326
1327 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1328 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1329 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1330 also returns the mark "x". This bug applied to capturing parentheses,
1331 non-capturing parentheses, and atomic parentheses. It also applied to some
1332 assertions.
1333
1334 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1335 information out of configure.ac instead of relying on pcre.h.generic, which
1336 is not stored in the repository.
1337
1338 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1339 -lreadline.
1340
1341 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1342
1343 15. Applied Graycode's patch to put the top-level frame on the stack rather
1344 than the heap when not using the stack for recursion. This gives a
1345 performance improvement in many cases when recursion is not deep.
1346
1347 16. Experimental code added to "pcretest -C" to output the stack frame size.
1348
1349
1350 Version 8.21 12-Dec-2011
1351 ------------------------
1352
1353 1. Updating the JIT compiler.
1354
1355 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1356 are added as well.
1357
1358 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1359 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1360 calling _pcre_jit_exec. Some extra comments are added.
1361
1362 4. (*MARK) settings inside atomic groups that do not contain any capturing
1363 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1364 was introduced by change 18 for 8.20.
1365
1366 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1367 ECMA-262 standard.
1368
1369 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1370 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1371 This bug was probably introduced by change 9 of 8.13.
1372
1373 7. While fixing 6 above, I noticed that a number of other items were being
1374 incorrectly rejected as "not fixed length". This arose partly because newer
1375 opcodes had not been added to the fixed-length checking code. I have (a)
1376 corrected the bug and added tests for these items, and (b) arranged for an
1377 error to occur if an unknown opcode is encountered while checking for fixed
1378 length instead of just assuming "not fixed length". The items that were
1379 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1380 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1381 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1382
1383 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1384 being incorrectly compiled and would have given unpredicatble results.
1385
1386 9. A possessively repeated subpattern with minimum repeat count greater than
1387 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1388 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1389 the first (A) could occur when it should not.
1390
1391 10. Add a cast and remove a redundant test from the code.
1392
1393 11. JIT should use pcre_malloc/pcre_free for allocation.
1394
1395 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1396 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1397 is anything other than /usr, -L is still shown).
1398
1399 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1400
1401 14. Perl does not support \N without a following name in a [] class; PCRE now
1402 also gives an error.
1403
1404 15. If a forward reference was repeated with an upper limit of around 2000,
1405 it caused the error "internal error: overran compiling workspace". The
1406 maximum number of forward references (including repeats) was limited by the
1407 internal workspace, and dependent on the LINK_SIZE. The code has been
1408 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1409 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1410 of around 200,000 forward references. While doing this, I also speeded up
1411 the filling in of repeated forward references.
1412
1413 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1414 incorrectly expecting the subject to contain another "a" after the start.
1415
1416 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1417 in the match, the SKIP should be ignored. This was not happening; instead
1418 the SKIP was being treated as NOMATCH. For patterns such as
1419 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1420 tested.
1421
1422 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1423 now much more compatible with Perl, in particular in cases where the result
1424 is a non-match for a non-anchored pattern. For example, if
1425 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1426 "m", where previously it did not return a name. A side effect of this
1427 change is that for partial matches, the last encountered mark name is
1428 returned, as for non matches. A number of tests that were previously not
1429 Perl-compatible have been moved into the Perl-compatible test files. The
1430 refactoring has had the pleasing side effect of removing one argument from
1431 the match() function, thus reducing its stack requirements.
1432
1433 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1434 subsequent uses of /S (without +) incorrectly behaved like /S+.
1435
1436 21. Retrieve executable code size support for the JIT compiler and fixing
1437 some warnings.
1438
1439 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1440 not work when the shorter character appeared right at the end of the
1441 subject string.
1442
1443 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1444 systems.
1445
1446 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1447 output it when the /M option is used in pcretest.
1448
1449 25. The CheckMan script was not being included in the distribution. Also, added
1450 an explicit "perl" to run Perl scripts from the PrepareRelease script
1451 because this is reportedly needed in Windows.
1452
1453 26. If study data was being save in a file and studying had not found a set of
1454 "starts with" bytes for the pattern, the data written to the file (though
1455 never used) was taken from uninitialized memory and so caused valgrind to
1456 complain.
1457
1458 27. Updated RunTest.bat as provided by Sheri Pierce.
1459
1460 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1461
1462 29. Computation of memory usage for the table of capturing group names was
1463 giving an unnecessarily large value.
1464
1465
1466 Version 8.20 21-Oct-2011
1467 ------------------------
1468
1469 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1470 a POSIX class. After further experiments with Perl, which convinced me that
1471 Perl has bugs and confusions, a closing square bracket is no longer allowed
1472 in a POSIX name. This bug also affected patterns with classes that started
1473 with full stops.
1474
1475 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1476 captured substring, but while checking the failing first alternative,
1477 substring 1 is temporarily captured. If the output vector supplied to
1478 pcre_exec() was not big enough for this capture, the yield of the function
1479 was still zero ("insufficient space for captured substrings"). This cannot
1480 be totally fixed without adding another stack variable, which seems a lot
1481 of expense for a edge case. However, I have improved the situation in cases
1482 such as /(a)(b)x|abc/ matched against "abc", where the return code
1483 indicates that fewer than the maximum number of slots in the ovector have
1484 been set.
1485
1486 3. Related to (2) above: when there are more back references in a pattern than
1487 slots in the output vector, pcre_exec() uses temporary memory during
1488 matching, and copies in the captures as far as possible afterwards. It was
1489 using the entire output vector, but this conflicts with the specification
1490 that only 2/3 is used for passing back captured substrings. Now it uses
1491 only the first 2/3, for compatibility. This is, of course, another edge
1492 case.
1493
1494 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1495 main code base, and can be used by building with --enable-jit. When this is
1496 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1497 runtime --no-jit option is given.
1498
1499 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1500 ovector, the return from the function was zero, implying that there were
1501 other matches that did not fit. The correct "exactly full" value is now
1502 returned.
1503
1504 6. If a subpattern that was called recursively or as a subroutine contained
1505 (*PRUNE) or any other control that caused it to give a non-standard return,
1506 invalid errors such as "Error -26 (nested recursion at the same subject
1507 position)" or even infinite loops could occur.
1508
1509 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1510 computing the minimum length on reaching *ACCEPT, and so ended up with the
1511 wrong value of 1 rather than 0. Further investigation indicates that
1512 computing a minimum subject length in the presence of *ACCEPT is difficult
1513 (think back references, subroutine calls), and so I have changed the code
1514 so that no minimum is registered for a pattern that contains *ACCEPT.
1515
1516 8. If (*THEN) was present in the first (true) branch of a conditional group,
1517 it was not handled as intended. [But see 16 below.]
1518
1519 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1520 Sheri Pierce.
1521
1522 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1523 the first byte in a match must be "a".
1524
1525 11. Change 17 for 8.13 increased the recursion depth for patterns like
1526 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1527 pattern contains any instances of (*THEN). If it does not, the old
1528 optimizations are restored. It would be nice to do this on a per-group
1529 basis, but at the moment that is not feasible.
1530
1531 12. In some environments, the output of pcretest -C is CRLF terminated. This
1532 broke RunTest's code that checks for the link size. A single white space
1533 character after the value is now allowed for.
1534
1535 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1536 For "fr", it uses the Windows-specific input and output files.
1537
1538 14. If (*THEN) appeared in a group that was called recursively or as a
1539 subroutine, it did not work as intended. [But see next item.]
1540
1541 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1542 pattern fragments (but not containing any | characters). If A and B are
1543 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1544 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1545 D. In other words, Perl considers parentheses that do not contain any |
1546 characters to be part of a surrounding alternative, whereas PCRE was
1547 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1548 differently. PCRE now behaves in the same way as Perl, except in the case
1549 of subroutine/recursion calls such as (?1) which have in any case always
1550 been different (but PCRE had them first :-).
1551
1552 16. Related to 15 above: Perl does not treat the | in a conditional group as
1553 creating alternatives. Such a group is treated in the same way as an
1554 ordinary group without any | characters when processing (*THEN). PCRE has
1555 been changed to match Perl's behaviour.
1556
1557 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1558 RunGrepTest script failed.
1559
1560 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1561 inevitable for groups that contain captures, but it can lead to a lot of
1562 stack use in large patterns. The old behaviour has been restored for atomic
1563 groups that do not contain any capturing parentheses.
1564
1565 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1566 suppress the check for a minimum subject length at run time. (If it was
1567 given to pcre_exec() or pcre_dfa_exec() it did work.)
1568
1569 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1570 fail to work when decoding hex characters in data strings in EBCDIC
1571 environments.
1572
1573 21. It appears that in at least one Mac OS environment, the isxdigit() function
1574 is implemented as a macro that evaluates to its argument more than once,
1575 contravening the C 90 Standard (I haven't checked a later standard). There
1576 was an instance in pcretest which caused it to go wrong when processing
1577 \x{...} escapes in subject strings. The has been rewritten to avoid using
1578 things like p++ in the argument of isxdigit().
1579
1580
1581 Version 8.13 16-Aug-2011
1582 ------------------------
1583
1584 1. The Unicode data tables have been updated to Unicode 6.0.0.
1585
1586 2. Two minor typos in pcre_internal.h have been fixed.
1587
1588 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1589 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1590 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1591
1592 4. There were a number of related bugs in the code for matching backrefences
1593 caselessly in UTF-8 mode when codes for the characters concerned were
1594 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1595 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1596 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1597 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1598 2-byte code at the end of the subject (it thought there wasn't enough data
1599 left).
1600
1601 5. Comprehensive information about what went wrong is now returned by
1602 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1603 as the output vector has at least 2 elements. The offset of the start of
1604 the failing character and a reason code are placed in the vector.
1605
1606 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1607 now returned is for the first byte of the failing character, instead of the
1608 last byte inspected. This is an incompatible change, but I hope it is small
1609 enough not to be a problem. It makes the returned offset consistent with
1610 pcre_exec() and pcre_dfa_exec().
1611
1612 7. pcretest now gives a text phrase as well as the error number when
1613 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1614 failure, the offset and reason code are output.
1615
1616 8. When \R was used with a maximizing quantifier it failed to skip backwards
1617 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1618 back over a single character (\n). This seems wrong (because it treated the
1619 two characters as a single entity when going forwards), conflicts with the
1620 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1621 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1622 has been changed.
1623
1624 9. Some internal refactoring has changed the processing so that the handling
1625 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1626 time (the PCRE_DOTALL option was changed this way some time ago: version
1627 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1628 which was always a bit of a fudge. It also means that there is one less
1629 argument for the match() function, which reduces its stack requirements
1630 slightly. This change also fixes an incompatibility with Perl: the pattern
1631 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1632
1633 10. More internal refactoring has drastically reduced the number of recursive
1634 calls to match() for possessively repeated groups such as (abc)++ when
1635 using pcre_exec().
1636
1637 11. While implementing 10, a number of bugs in the handling of groups were
1638 discovered and fixed:
1639
1640 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1641 (a|)*(?1) gave a compile-time internal error.
1642 ((a|)+)+ did not notice that the outer group could match an empty string.
1643 (^a|^)+ was not marked as anchored.
1644 (.*a|.*)+ was not marked as matching at start or after a newline.
1645
1646 12. Yet more internal refactoring has removed another argument from the match()
1647 function. Special calls to this function are now indicated by setting a
1648 value in a variable in the "match data" data block.
1649
1650 13. Be more explicit in pcre_study() instead of relying on "default" for
1651 opcodes that mean there is no starting character; this means that when new
1652 ones are added and accidentally left out of pcre_study(), testing should
1653 pick them up.
1654
1655 14. The -s option of pcretest has been documented for ages as being an old
1656 synonym of -m (show memory usage). I have changed it to mean "force study
1657 for every regex", that is, assume /S for every regex. This is similar to -i
1658 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1659 using it. It makes it easier to run collections of tests with and without
1660 study enabled, and thereby test pcre_study() more easily. All the standard
1661 tests are now run with and without -s (but some patterns can be marked as
1662 "never study" - see 20 below).
1663
1664 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1665 restoration of the capturing data to the outer values was not happening
1666 correctly.
1667
1668 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1669 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1670 pattern had matched an empty string, and so incorrectly returned a no
1671 match.
1672
1673 17. There was optimizing code for the last branch of non-capturing parentheses,
1674 and also for the obeyed branch of a conditional subexpression, which used
1675 tail recursion to cut down on stack usage. Unfortunately, now that there is
1676 the possibility of (*THEN) occurring in these branches, tail recursion is
1677 no longer possible because the return has to be checked for (*THEN). These
1678 two optimizations have therefore been removed. [But see 8.20/11 above.]
1679
1680 18. If a pattern containing \R was studied, it was assumed that \R always
1681 matched two bytes, thus causing the minimum subject length to be
1682 incorrectly computed because \R can also match just one byte.
1683
1684 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1685 was incorrectly computed.
1686
1687 20. If /S is present twice on a test pattern in pcretest input, it now
1688 *disables* studying, thereby overriding the use of -s on the command line
1689 (see 14 above). This is necessary for one or two tests to keep the output
1690 identical in both cases.
1691
1692 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1693 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1694
1695 22. When an atomic group that contained a capturing parenthesis was
1696 successfully matched, but the branch in which it appeared failed, the
1697 capturing was not being forgotten if a higher numbered group was later
1698 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1699 group 1 to "a", when in fact it should be unset. This applied to multi-
1700 branched capturing and non-capturing groups, repeated or not, and also to
1701 positive assertions (capturing in negative assertions does not happen
1702 in PCRE) and also to nested atomic groups.
1703
1704 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1705 subject after a captured substring, to make it easier to tell which of a
1706 number of identical substrings has been captured.
1707
1708 24. The way atomic groups are processed by pcre_exec() has been changed so that
1709 if they are repeated, backtracking one repetition now resets captured
1710 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1711 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1712 "aaa". Previously, it would have been "a". As part of this code
1713 refactoring, the way recursive calls are handled has also been changed.
1714
1715 25. If an assertion condition captured any substrings, they were not passed
1716 back unless some other capturing happened later. For example, if
1717 (?(?=(a))a) was matched against "a", no capturing was returned.
1718
1719 26. When studying a pattern that contained subroutine calls or assertions,
1720 the code for finding the minimum length of a possible match was handling
1721 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1722 group 1 called group 2 while simultaneously a separate group 2 called group
1723 1). A stack overflow occurred in this case. I have fixed this by limiting
1724 the recursion depth to 10.
1725
1726 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1727 Fortmann. This supports explicit test numbers on the command line, and has
1728 argument validation and error reporting.
1729
1730 28. An instance of \X with an unlimited repeat could fail if at any point the
1731 first character it looked at was a mark character.
1732
1733 29. Some minor code refactoring concerning Unicode properties and scripts
1734 should reduce the stack requirement of match() slightly.
1735
1736 30. Added the '=' option to pcretest to check the setting of unused capturing
1737 slots at the end of the pattern, which are documented as being -1, but are
1738 not included in the return count.
1739
1740 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1741 compiled something random. Now it gives a compile-time error (as does
1742 Perl).
1743
1744 32. A *MARK encountered during the processing of a positive assertion is now
1745 recorded and passed back (compatible with Perl).
1746
1747 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1748 had alternative anchored branches, the search for a second match in a line
1749 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1750 matched the line "0102" twice. The same bug affected patterns that started
1751 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1752 twice.
1753
1754 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1755 does, and because of capturing effects, quantifying parenthesized
1756 assertions may at times be useful. Quantifiers are now allowed for
1757 parenthesized assertions.
1758
1759 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1760
1761 36. \g was being checked for fancy things in a character class, when it should
1762 just be a literal "g".
1763
1764 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1765 appearance of a nested POSIX class supersedes an apparent external class.
1766 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1767 unescaped square brackets may also appear as part of class names. For
1768 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1769 more like Perl. (But see 8.20/1 above.)
1770
1771 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1772 was because it thought it was \N{name}, which is not supported).
1773
1774 39. Add minix to OS list not supporting the -S option in pcretest.
1775
1776 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1777 cannot analyze patterns in sufficient detail to catch mutual recursions
1778 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1779 subgroup is called recursively as a subpattern for a second time at the
1780 same position in the subject string. In previous releases this might have
1781 been caught by the recursion limit, or it might have run out of stack.
1782
1783 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1784 happen only once. PCRE was, however incorrectly giving a compile time error
1785 "recursive call could loop indefinitely" because it cannot analyze the
1786 pattern in sufficient detail. The compile time test no longer happens when
1787 PCRE is compiling a conditional subpattern, but actual runaway loops are
1788 now caught at runtime (see 40 above).
1789
1790 42. It seems that Perl allows any characters other than a closing parenthesis
1791 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1792 has been changed to be the same.
1793
1794 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1795 as not to get warnings when autogen.sh is called. Also changed
1796 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1797
1798 44. To help people who use pcregrep to scan files containing exceedingly long
1799 lines, the following changes have been made:
1800
1801 (a) The default value of the buffer size parameter has been increased from
1802 8K to 20K. (The actual buffer used is three times this size.)
1803
1804 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1805 PCRE is built.
1806
1807 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1808 to be set at run time.
1809
1810 (d) Numerical values in pcregrep options can be followed by K or M, for
1811 example --buffer-size=50K.
1812
1813 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1814 given and the return code is set to 2.
1815
1816 45. Add a pointer to the latest mark to the callout data block.
1817
1818 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1819 partial match of an empty string instead of no match. This was specific to
1820 the use of ".".
1821
1822 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1823 complete match instead of a partial match. This bug was dependent on both
1824 the PCRE_UTF8 and PCRE_DOTALL options being set.
1825
1826 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1827 starting byte set, because \b was not being ignored.
1828
1829
1830 Version 8.12 15-Jan-2011
1831 ------------------------
1832
1833 1. Fixed some typos in the markup of the man pages, and wrote a script that
1834 checks for such things as part of the documentation building process.
1835
1836 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1837 --match-limit and --recursion-limit options (added for 8.11). In
1838 particular, this made one of the standard tests fail. (The integer value
1839 went into the wrong half of a long int.)
1840
1841 3. If the --colour option was given to pcregrep with -v (invert match), it
1842 did strange things, either producing crazy output, or crashing. It should,
1843 of course, ignore a request for colour when reporting lines that do not
1844 match.
1845
1846 4. Another pcregrep bug caused similar problems if --colour was specified with
1847 -M (multiline) and the pattern match finished with a line ending.
1848
1849 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1850 matched in multiline mode, the following line was shown as part of the
1851 match. This seems wrong, so I have changed it.
1852
1853 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1854 the check for further matches in the same line (so they could be coloured)
1855 to overrun the end of the current line. If another match was found, it was
1856 incorrectly shown (and then shown again when found in the next line).
1857
1858 7. If pcregrep was compiled under Windows, there was a reference to the
1859 function pcregrep_exit() before it was defined. I am assuming this was
1860 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1861 reported by a user. I've moved the definition above the reference.
1862
1863
1864 Version 8.11 10-Dec-2010
1865 ------------------------
1866
1867 1. (*THEN) was not working properly if there were untried alternatives prior
1868 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1869 backtracked to try for "b" instead of moving to the next alternative branch
1870 at the same level (in this case, to look for "c"). The Perl documentation
1871 is clear that when (*THEN) is backtracked onto, it goes to the "next
1872 alternative in the innermost enclosing group".
1873
1874 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1875 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1876 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1877 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1878 (*THEN).
1879
1880 3. If \s appeared in a character class, it removed the VT character from
1881 the class, even if it had been included by some previous item, for example
1882 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1883 of \s, but is part of the POSIX "space" class.)
1884
1885 4. A partial match never returns an empty string (because you can always
1886 match an empty string at the end of the subject); however the checking for
1887 an empty string was starting at the "start of match" point. This has been
1888 changed to the "earliest inspected character" point, because the returned
1889 data for a partial match starts at this character. This means that, for
1890 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1891 (previously it gave "no match").
1892
1893 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1894 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1895 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1896 has an implication that the given string is incomplete (because a partial
1897 match is preferred over a full match). For this reason, these items now
1898 give a partial match in this situation. [Aside: previously, the one case
1899 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1900 match rather than a full match, which was wrong by the old rules, but is
1901 now correct.]
1902
1903 6. There was a bug in the handling of #-introduced comments, recognized when
1904 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1905 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1906 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1907 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1908 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1909 places in pcre_compile().
1910
1911 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1912 comments when looking ahead for named forward references to subpatterns,
1913 the only newline sequence it recognized was NL. It now handles newlines
1914 according to the set newline convention.
1915
1916 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1917 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1918 cater for a lack of strerror(). These oversights have been fixed.
1919
1920 9. Added --match-limit and --recursion-limit to pcregrep.
1921
1922 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1923
1924 11. When the -o option was used, pcregrep was setting a return code of 1, even
1925 when matches were found, and --line-buffered was not being honoured.
1926
1927 12. Added an optional parentheses number to the -o and --only-matching options
1928 of pcregrep.
1929
1930 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1931 can match an empty string. The code to do it in pcretest and pcredemo
1932 needed fixing:
1933
1934 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1935 only one byte after an empty string match just before CRLF (this case
1936 just got forgotten; "any" and "anycrlf" were OK).
1937
1938 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1939 mode when an empty string match preceded an ASCII character followed by
1940 a non-ASCII character. (The code for advancing by one character rather
1941 than one byte was nonsense.)
1942
1943 (c) The pcredemo.c sample program did not have any code at all to handle
1944 the cases when CRLF is a valid newline sequence.
1945
1946 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1947 as a starting offset was within the subject string. There is now a new
1948 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1949 negative or greater than the length of the string. In order to test this,
1950 pcretest is extended to allow the setting of negative starting offsets.
1951
1952 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1953 starting offset points to the beginning of a UTF-8 character was
1954 unnecessarily clumsy. I tidied it up.
1955
1956 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1957 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1958
1959 17. Nobody had reported that the --include_dir option, which was added in
1960 release 7.7 should have been called --include-dir (hyphen, not underscore)
1961 for compatibility with GNU grep. I have changed it to --include-dir, but
1962 left --include_dir as an undocumented synonym, and the same for
1963 --exclude-dir, though that is not available in GNU grep, at least as of
1964 release 2.5.4.
1965
1966 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1967 characters from a string of bytes) have been redefined so as not to use
1968 loops, in order to improve performance in some environments. At the same
1969 time, I abstracted some of the common code into auxiliary macros to save
1970 repetition (this should not affect the compiled code).
1971
1972 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1973 compile-time error is now given if \c is not followed by an ASCII
1974 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1975 different, and any byte value is allowed.)
1976
1977 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1978 START_OPTIMIZE option, which is now allowed at compile time - but just
1979 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1980 to pcregrep and other applications that have no direct access to PCRE
1981 options. The new /Y option in pcretest sets this option when calling
1982 pcre_compile().
1983
1984 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1985 back references. Groups containing recursive back references were forced to
1986 be atomic by that change, but in the case of named groups, the amount of
1987 memory required was incorrectly computed, leading to "Failed: internal
1988 error: code overflow". This has been fixed.
1989
1990 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1991 pcretest.c, to avoid build problems in some Borland environments.
1992
1993
1994 Version 8.10 25-Jun-2010
1995 ------------------------
1996
1997 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1998 THEN.
1999
2000 2. (*ACCEPT) was not working when inside an atomic group.
2001
2002 3. Inside a character class, \B is treated as a literal by default, but
2003 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2004 causes the error). The code is unchanged, but I tidied the documentation.
2005
2006 4. Inside a character class, PCRE always treated \R and \X as literals,
2007 whereas Perl faults them if its -w option is set. I have changed PCRE so
2008 that it faults them when PCRE_EXTRA is set.
2009
2010 5. Added support for \N, which always matches any character other than
2011 newline. (It is the same as "." when PCRE_DOTALL is not set.)
2012
2013 6. When compiling pcregrep with newer versions of gcc which may have
2014 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2015 declared with attribute warn_unused_result" were given. Just casting the
2016 result to (void) does not stop the warnings; a more elaborate fudge is
2017 needed. I've used a macro to implement this.
2018
2019 7. Minor change to pcretest.c to avoid a compiler warning.
2020
2021 8. Added four artifical Unicode properties to help with an option to make
2022 \s etc use properties (see next item). The new properties are: Xan
2023 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2024
2025 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2026 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2027 this option. Modified pcretest to add /W to test this facility. Added
2028 REG_UCP to make it available via the POSIX interface.
2029
2030 10. Added --line-buffered to pcregrep.
2031
2032 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2033 studied, and the match started with a letter with a code point greater than
2034 127 whose first byte was different to the first byte of the other case of
2035 the letter, the other case of this starting letter was not recognized
2036 (#976).
2037
2038 12. If a pattern that was studied started with a repeated Unicode property
2039 test, for example, \p{Nd}+, there was the theoretical possibility of
2040 setting up an incorrect bitmap of starting bytes, but fortunately it could
2041 not have actually happened in practice until change 8 above was made (it
2042 added property types that matched character-matching opcodes).
2043
2044 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2045 possible starting bytes for non-anchored patterns.
2046
2047 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2048 \R, and also a number of cases that involve Unicode properties, both
2049 explicit and implicit when PCRE_UCP is set.
2050
2051 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2052 input, it could crash or give wrong results if characters with values
2053 greater than 0xc0 were present in the subject string. (Detail: it assumed
2054 UTF-8 input when processing these items.)
2055
2056 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2057 size_t is 64-bit (#991).
2058
2059 17. Added a check for running out of memory when PCRE is compiled with
2060 --disable-stack-for-recursion (#990).
2061
2062 18. If the last data line in a file for pcretest does not have a newline on
2063 the end, a newline was missing in the output.
2064
2065 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2066 less than 128) in its various bitmaps. However, there is a facility for
2067 generating tables according to the current locale when PCRE is compiled. It
2068 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2069 characters, are recognized by isspace() and therefore were getting set in
2070 these tables, and indeed these tables seem to approximate to ISO 8859. This
2071 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2072 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2073 which of course cannot start UTF-8 characters. I have changed the code so
2074 that only real ASCII characters (less than 128) and the correct starting
2075 bytes for UTF-8 encodings are set for characters greater than 127 when in
2076 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2077 altogether.)
2078
2079 20. Added the /T option to pcretest so as to be able to run tests with non-
2080 standard character tables, thus making it possible to include the tests
2081 used for 19 above in the standard set of tests.
2082
2083 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2084 reference to a subpattern the other side of a comment that contains an
2085 opening parenthesis caused either an internal compiling error, or a
2086 reference to the wrong subpattern.
2087
2088
2089 Version 8.02 19-Mar-2010
2090 ------------------------
2091
2092 1. The Unicode data tables have been updated to Unicode 5.2.0.
2093
2094 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2095 configured.
2096
2097 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2098 original author of that file, following a query about its status.
2099
2100 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2101 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2102
2103 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2104 quantifier applied to a forward-referencing subroutine call, could compile
2105 incorrect code or give the error "internal error: previously-checked
2106 referenced subpattern not found".
2107
2108 6. Both MS Visual Studio and Symbian OS have problems with initializing
2109 variables to point to external functions. For these systems, therefore,
2110 pcre_malloc etc. are now initialized to local functions that call the
2111 relevant global functions.
2112
2113 7. There were two entries missing in the vectors called coptable and poptable
2114 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2115 I've fixed the data, and added a kludgy way of testing at compile time that
2116 the lengths are correct (equal to the number of opcodes).
2117
2118 8. Following on from 7, I added a similar kludge to check the length of the
2119 eint vector in pcreposix.c.
2120
2121 9. Error texts for pcre_compile() are held as one long string to avoid too
2122 much relocation at load time. To find a text, the string is searched,
2123 counting zeros. There was no check for running off the end of the string,
2124 which could happen if a new error number was added without updating the
2125 string.
2126
2127 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2128
2129 11. \K was not working if it appeared in an atomic group or in a group that
2130 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2131 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2132 the assertion is positive, but not if it is negative.
2133
2134 12. Change 11 fortuitously reduced the size of the stack frame used in the
2135 "match()" function of pcre_exec.c by one pointer. Forthcoming
2136 implementation of support for (*MARK) will need an extra pointer on the
2137 stack; I have reserved it now, so that the stack frame size does not
2138 decrease.
2139
2140 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2141 item in branch that calls a recursion is a subroutine call - as in the
2142 second branch in the above example - was incorrectly given the compile-
2143 time error "recursive call could loop indefinitely" because pcre_compile()
2144 was not correctly checking the subroutine for matching a non-empty string.
2145
2146 14. The checks for overrunning compiling workspace could trigger after an
2147 overrun had occurred. This is a "should never occur" error, but it can be
2148 triggered by pathological patterns such as hundreds of nested parentheses.
2149 The checks now trigger 100 bytes before the end of the workspace.
2150
2151 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2152
2153
2154 Version 8.01 19-Jan-2010
2155 ------------------------
2156
2157 1. If a pattern contained a conditional subpattern with only one branch (in
2158 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2159 computed the wrong minimum data length (which is of course zero for such
2160 subpatterns). This could cause incorrect "no match" results.
2161
2162 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2163 the pattern is reset in the first branch, pcre_compile() failed with
2164 "internal error: code overflow at offset...". This happened only when
2165 the reset was to the original external option setting. (An optimization
2166 abstracts leading options settings into an external setting, which was the
2167 cause of this.)
2168
2169 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2170 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2171 assertion pattern did not match (meaning that the assertion was true), it
2172 was incorrectly treated as false if the SKIP had been reached during the
2173 matching. This also applied to assertions used as conditions.
2174
2175 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2176 assertion subpattern, including such a pattern used as a condition,
2177 unpredictable results occurred, instead of the error return
2178 PCRE_ERROR_DFA_UITEM.
2179
2180 5. The C++ GlobalReplace function was not working like Perl for the special
2181 situation when an empty string is matched. It now does the fancy magic
2182 stuff that is necessary.
2183
2184 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2185 removed. (These were left over from very, very early versions of PCRE.)
2186
2187 7. Some cosmetic changes to the code to make life easier when compiling it
2188 as part of something else:
2189
2190 (a) Change DEBUG to PCRE_DEBUG.
2191
2192 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2193 called "current" as "current_branch", to prevent a collision with the
2194 Linux macro when compiled as a kernel module.
2195
2196 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2197 prevent a collision with the Linux macro when compiled as a kernel
2198 module.
2199
2200 8. In pcre_compile() there are some checks for integer overflows that used to
2201 cast potentially large values to (double). This has been changed to that
2202 when building, a check for int64_t is made, and if it is found, it is used
2203 instead, thus avoiding the use of floating point arithmetic. (There is no
2204 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2205 double.
2206
2207 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2208 2005 (difference between two addresses compared to an unsigned value).
2209
2210 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2211 custom one, because of the following reported problem in Windows:
2212
2213 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2214 under Win32.
2215 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2216 therefore missing the function definition.
2217 - The compiler thus generates a "C" signature for the test function.
2218 - The linker fails to find the "C" function.
2219 - PCRE fails to configure if asked to do so against libbz2.
2220
2221 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2222 messages were output:
2223
2224 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2225 rerunning libtoolize, to keep the correct libtool macros in-tree.
2226 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2227
2228 I have done both of these things.
2229
2230 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2231 most of the time, it *can* run out if it is given a pattern that contains a
2232 runaway infinite recursion. I updated the discussion in the pcrestack man
2233 page.
2234
2235 13. Now that we have gone to the x.xx style of version numbers, the minor
2236 version may start with zero. Using 08 or 09 is a bad idea because users
2237 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2238 interpreted as invalid octal numbers. I've updated the previous comment in
2239 configure.ac, and also added a check that gives an error if 08 or 09 are
2240 used.
2241
2242 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2243 causing partial matching to fail when the end of the subject matched \W
2244 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2245
2246 15. There were some discrepancies between the declarations in pcre_internal.h
2247 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2248 their definitions. The declarations used "const uschar *" and the
2249 definitions used USPTR. Even though USPTR is normally defined as "const
2250 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2251 reported that: "This difference in casting confuses some C++ compilers, for
2252 example, SunCC recognizes above declarations as different functions and
2253 generates broken code for hbpcre." I have changed the declarations to use
2254 USPTR.
2255
2256 16. GNU libtool is named differently on some systems. The autogen.sh script now
2257 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2258 (FreeBSD).
2259
2260 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2261 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2262 comment: "Figure out how to create a longlong from a string: strtoll and
2263 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2264 instance, but it only takes 2 args instead of 3!"
2265
2266 18. A subtle bug concerned with back references has been fixed by a change of
2267 specification, with a corresponding code fix. A pattern such as
2268 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2269 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2270 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2271 same bug. Such groups have to be quantified to be useful, or contained
2272 inside another quantified group. (If there's no repetition, the reference
2273 can never match.) The problem arises because, having left the group and
2274 moved on to the rest of the pattern, a later failure that backtracks into
2275 the group uses the captured value from the final iteration of the group
2276 rather than the correct earlier one. I have fixed this in PCRE by forcing
2277 any group that contains a reference to itself to be an atomic group; that
2278 is, there cannot be any backtracking into it once it has completed. This is
2279 similar to recursive and subroutine calls.
2280
2281
2282 Version 8.00 19-Oct-09
2283 ----------------------
2284
2285 1. The table for translating pcre_compile() error codes into POSIX error codes
2286 was out-of-date, and there was no check on the pcre_compile() error code
2287 being within the table. This could lead to an OK return being given in
2288 error.
2289
2290 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2291 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2292 in a Windows environment.
2293
2294 3. The pcregrep --count option prints the count for each file even when it is
2295 zero, as does GNU grep. However, pcregrep was also printing all files when
2296 --files-with-matches was added. Now, when both options are given, it prints
2297 counts only for those files that have at least one match. (GNU grep just
2298 prints the file name in this circumstance, but including the count seems
2299 more useful - otherwise, why use --count?) Also ensured that the
2300 combination -clh just lists non-zero counts, with no names.
2301
2302 4. The long form of the pcregrep -F option was incorrectly implemented as
2303 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2304 but it seems right to fix it, and I didn't think it was worth preserving
2305 the old behaviour.
2306
2307 5. The command line items --regex=pattern and --regexp=pattern were not
2308 recognized by pcregrep, which required --regex pattern or --regexp pattern
2309 (with a space rather than an '='). The man page documented the '=' forms,
2310 which are compatible with GNU grep; these now work.
2311
2312 6. No libpcreposix.pc file was created for pkg-config; there was just
2313 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2314
2315 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2316 when UCP support is not needed, by modifying the Python script that
2317 generates it from Unicode data files. This should not matter if the module
2318 is correctly used as a library, but I received one complaint about 50K of
2319 unwanted data. My guess is that the person linked everything into his
2320 program rather than using a library. Anyway, it does no harm.
2321
2322 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2323 was a minimum greater than 1 for a wide character in a possessive
2324 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2325 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2326 character. Chaos in the form of incorrect output or a compiling loop could
2327 result.
2328
2329 9. The restrictions on what a pattern can contain when partial matching is
2330 requested for pcre_exec() have been removed. All patterns can now be
2331 partially matched by this function. In addition, if there are at least two
2332 slots in the offset vector, the offset of the earliest inspected character
2333 for the match and the offset of the end of the subject are set in them when
2334 PCRE_ERROR_PARTIAL is returned.
2335
2336 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2337 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2338 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2339 and may be more useful for multi-segment matching.
2340
2341 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2342 used to be given if ever the end of the subject was reached; now it is
2343 given only if matching could not proceed because another character was
2344 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2345 string "Z", which now yields "no match" instead of "partial match". In the
2346 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2347 final character ended with (*FAIL).
2348
2349 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2350 if the pattern had a "must contain" character that was already found in the
2351 earlier partial match, unless partial matching was again requested. For
2352 example, with the pattern /dog.(body)?/, the "must contain" character is
2353 "g". If the first part-match was for the string "dog", restarting with
2354 "sbody" failed. This bug has been fixed.
2355
2356 13. The string returned by pcre_dfa_exec() after a partial match has been
2357 changed so that it starts at the first inspected character rather than the
2358 first character of the match. This makes a difference only if the pattern
2359 starts with a lookbehind assertion or \b or \B (\K is not supported by
2360 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2361 matching functions compatible, and I think it's the right thing to do.
2362
2363 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2364 so that the demonstration program is easily available in environments where
2365 PCRE has not been installed from source.
2366
2367 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2368 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2369 library.
2370
2371 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2372 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2373 is not the first non-POSIX option to be added. Clearly some people find
2374 these options useful.
2375
2376 17. If a caller to the POSIX matching function regexec() passes a non-zero
2377 value for nmatch with a NULL value for pmatch, the value of
2378 nmatch is forced to zero.
2379
2380 18. RunGrepTest did not have a test for the availability of the -u option of
2381 the diff command, as RunTest does. It now checks in the same way as
2382 RunTest, and also checks for the -b option.
2383
2384 19. If an odd number of negated classes containing just a single character
2385 interposed, within parentheses, between a forward reference to a named
2386 subpattern and the definition of the subpattern, compilation crashed with
2387 an internal error, complaining that it could not find the referenced
2388 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2389 [The bug was that it was starting one character too far in when skipping
2390 over the character class, thus treating the ] as data rather than
2391 terminating the class. This meant it could skip too much.]
2392
2393 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2394 /g option in pcretest when the pattern contains \K, which makes it possible
2395 to have an empty string match not at the start, even when the pattern is
2396 anchored. Updated pcretest and pcredemo to use this option.
2397
2398 21. If the maximum number of capturing subpatterns in a recursion was greater
2399 than the maximum at the outer level, the higher number was returned, but
2400 with unset values at the outer level. The correct (outer level) value is
2401 now given.
2402
2403 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2404 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2405 make it do so. The string so far is captured, making this feature
2406 compatible with Perl.
2407
2408 23. The tests have been re-organized, adding tests 11 and 12, to make it
2409 possible to check the Perl 5.10 features against Perl 5.10.
2410
2411 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2412 pattern matches a fixed length string. PCRE did not allow this; now it
2413 does. Neither allows recursion.
2414
2415 25. I finally figured out how to implement a request to provide the minimum
2416 length of subject string that was needed in order to match a given pattern.
2417 (It was back references and recursion that I had previously got hung up
2418 on.) This code has now been added to pcre_study(); it finds a lower bound
2419 to the length of subject needed. It is not necessarily the greatest lower
2420 bound, but using it to avoid searching strings that are too short does give
2421 some useful speed-ups. The value is available to calling programs via
2422 pcre_fullinfo().
2423
2424 26. While implementing 25, I discovered to my embarrassment that pcretest had
2425 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2426 study optimizations had never been tested with that matching function.
2427 Oops. What is worse, even when it was passed study data, there was a bug in
2428 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2429 were also very few tests of studied patterns with pcre_dfa_exec().
2430
2431 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2432 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2433 on the other side of the coin, they are no longer allowed to have different
2434 names, because these cannot be distinguished in PCRE, and this has caused
2435 confusion. (This is a difference from Perl.)
2436
2437 28. When duplicate subpattern names are present (necessarily with different
2438 numbers, as required by 27 above), and a test is made by name in a
2439 conditional pattern, either for a subpattern having been matched, or for
2440 recursion in such a pattern, all the associated numbered subpatterns are
2441 tested, and the overall condition is true if the condition is true for any
2442 one of them. This is the way Perl works, and is also more like the way
2443 testing by number works.
2444
2445
2446 Version 7.9 11-Apr-09
2447 ---------------------
2448
2449 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2450 (pcretest), all targets were linked against these libraries. This included
2451 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2452 libraries. This caused unwanted dependencies to be created. This problem
2453 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2454 pcretest is linked with readline.
2455
2456 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2457 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2458 moved outside it again, because FALSE and TRUE are already defined in AIX,
2459 but BOOL is not.
2460
2461 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2462 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2463
2464 4. The pcregrep documentation said spaces were inserted as well as colons (or
2465 hyphens) following file names and line numbers when outputting matching
2466 lines. This is not true; no spaces are inserted. I have also clarified the
2467 wording for the --colour (or --color) option.
2468
2469 5. In pcregrep, when --colour was used with -o, the list of matching strings
2470 was not coloured; this is different to GNU grep, so I have changed it to be
2471 the same.
2472
2473 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2474 each matching line was coloured. Now it goes on to look for further matches
2475 of any of the test patterns, which is the same behaviour as GNU grep.
2476
2477 7. A pattern that could match an empty string could cause pcregrep to loop; it
2478 doesn't make sense to accept an empty string match in pcregrep, so I have
2479 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2480 seems to be how GNU grep behaves. [But see later change 40 for release
2481 8.33.]
2482
2483 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2484 start or after a newline", because the conditional assertion was not being
2485 correctly handled. The rule now is that both the assertion and what follows
2486 in the first alternative must satisfy the test.
2487
2488 9. If auto-callout was enabled in a pattern with a conditional group whose
2489 condition was an assertion, PCRE could crash during matching, both with
2490 pcre_exec() and pcre_dfa_exec().
2491
2492 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2493 used for matching.
2494
2495 11. Unicode property support in character classes was not working for
2496 characters (bytes) greater than 127 when not in UTF-8 mode.
2497
2498 12. Added the -M command line option to pcretest.
2499
2500 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2501
2502 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2503
2504 16. Added comments and documentation about mis-use of no_arg in the C++
2505 wrapper.
2506
2507 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2508 from Martin Jerabek that uses macro names for all relevant character and
2509 string constants.
2510
2511 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2512 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2513 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2514 these, but not everybody uses configure.
2515
2516 19. A conditional group that had only one branch was not being correctly
2517 recognized as an item that could match an empty string. This meant that an
2518 enclosing group might also not be so recognized, causing infinite looping
2519 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2520 with the subject "ab", where knowledge that the repeated group can match
2521 nothing is needed in order to break the loop.
2522
2523 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2524 exec(), but without supplying a callout function, matching went wrong.
2525
2526 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2527 leak if the size of the offset vector was greater than 30. When the vector
2528 is smaller, the saved offsets during recursion go onto a local stack
2529 vector, but for larger vectors malloc() is used. It was failing to free
2530 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2531 error, in fact).
2532
2533 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2534 heapframe that is used only when UTF-8 support is enabled. This caused no
2535 problem, but was untidy.
2536
2537 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2538 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2539 included within another project.
2540
2541 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2542 slightly modified by me:
2543
2544 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2545 not building pcregrep.
2546
2547 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2548 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2549
2550 25. Forward references, both numeric and by name, in patterns that made use of
2551 duplicate group numbers, could behave incorrectly or give incorrect errors,
2552 because when scanning forward to find the reference group, PCRE was not
2553 taking into account the duplicate group numbers. A pattern such as
2554 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2555
2556 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2557 the feature of a custom pointer more persuasive (as requested by a user).
2558
2559 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2560 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2561 pre-defined.
2562
2563 28. Added support for (*UTF8) at the start of a pattern.
2564
2565 29. Arrange for flags added by the "release type" setting in CMake to be shown
2566 in the configuration summary.
2567
2568
2569 Version 7.8 05-Sep-08
2570 ---------------------
2571
2572 1. Replaced UCP searching code with optimized version as implemented for Ad
2573 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2574 stage table and inline lookup instead of a function, giving speed ups of 2
2575 to 5 times on some simple patterns that I tested. Permission was given to
2576 distribute the MultiStage2.py script that generates the tables (it's not in
2577 the tarball, but is in the Subversion repository).
2578
2579 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2580 scripts.
2581
2582 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2583 a group with a zero qualifier. The result of the study could be incorrect,
2584 or the function might crash, depending on the pattern.
2585
2586 4. Caseless matching was not working for non-ASCII characters in back
2587 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2588 It now works when Unicode Property Support is available.
2589
2590 5. In pcretest, an escape such as \x{de} in the data was always generating
2591 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2592 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2593 truncation.
2594
2595 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2596
2597 7. Added two (int) casts to pcregrep when printing the difference of two
2598 pointers, in case they are 64-bit values.
2599
2600 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2601 test 2 if it fails.
2602
2603 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2604 and a #define of that name to empty if it is not externally set. This is to
2605 allow users of MSVC to set it if necessary.
2606
2607 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2608 the convenience functions in the pcre_get.c source file.
2609
2610 11. An option change at the start of a pattern that had top-level alternatives
2611 could cause overwriting and/or a crash. This command provoked a crash in
2612 some environments:
2613
2614 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2615
2616 This potential security problem was recorded as CVE-2008-2371.
2617
2618 12. For a pattern where the match had to start at the beginning or immediately
2619 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2620 pcre_dfa_exec() could read past the end of the passed subject if there was
2621 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2622 pcretest so that it places the subject at the end of its malloc-ed buffer.
2623
2624 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2625 exec() might read past the end of the data buffer in UTF-8 mode.
2626
2627 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2628 the data contained the byte 0x85 as part of a UTF-8 character within its
2629 first line. This applied both to normal and DFA matching.
2630
2631 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2632 /^[^d]*?$/8 failed to match "abc".
2633
2634 16. Added a missing copyright notice to pcrecpp_internal.h.
2635
2636 17. Make it more clear in the documentation that values returned from
2637 pcre_exec() in ovector are byte offsets, not character counts.
2638
2639 18. Tidied a few places to stop certain compilers from issuing warnings.
2640
2641 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2642 supplied by Stefan Weber. I made a further small update for 7.8 because
2643 there is a change of source arrangements: the pcre_searchfuncs.c module is
2644 replaced by pcre_ucd.c.
2645
2646
2647 Version 7.7 07-May-08
2648 ---------------------
2649
2650 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2651 a string to a long long, pretend we don't even have a long long." This is
2652 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2653
2654 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2655 pre-7.6 versions, which defined a global no_arg variable instead of putting
2656 it in the RE class. (See also #8 below.)
2657
2658 3. Remove a line of dead code, identified by coverity and reported by Nuno
2659 Lopes.
2660
2661 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2662
2663 (1) The include/exclude patterns were being applied to the whole pathnames
2664 of files, instead of just to the final components.
2665
2666 (2) If there was more than one level of directory, the subdirectories were
2667 skipped unless they satisfied the include/exclude conditions. This is
2668 inconsistent with GNU grep (and could even be seen as contrary to the
2669 pcregrep specification - which I improved to make it absolutely clear).
2670 The action now is always to scan all levels of directory, and just
2671 apply the include/exclude patterns to regular files.
2672
2673 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2674 --exclude_dir in the tests to avoid scanning .svn directories.
2675
2676 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2677 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2678 doesn't support NULs in patterns.
2679
2680 7. Added some missing "const"s to declarations of static tables in
2681 pcre_compile.c and pcre_dfa_exec.c.
2682
2683 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2684 caused by fix #2 above. (Subsequently also a second patch to fix the
2685 first patch. And a third patch - this was a messy problem.)
2686
2687 9. Applied Craig's patch to remove the use of push_back().
2688
2689 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2690 matching function regexec().
2691
2692 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2693 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2694 references. PCRE supports relative numbers with this syntax (I don't think
2695 Oniguruma does).
2696
2697 12. Previously, a group with a zero repeat such as (...){0} was completely
2698 omitted from the compiled regex. However, this means that if the group
2699 was called as a subroutine from elsewhere in the pattern, things went wrong
2700 (an internal error was given). Such groups are now left in the compiled
2701 pattern, with a new opcode that causes them to be skipped at execution
2702 time.
2703
2704 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2705 to the way PCRE behaves:
2706
2707 (a) A lone ] character is dis-allowed (Perl treats it as data).
2708
2709 (b) A back reference to an unmatched subpattern matches an empty string
2710 (Perl fails the current match path).
2711
2712 (c) A data ] in a character class must be notated as \] because if the
2713 first data character in a class is ], it defines an empty class. (In
2714 Perl it is not possible to have an empty class.) The empty class []
2715 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2716 The negative empty class [^] matches any one character, independently
2717 of the DOTALL setting.
2718
2719 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2720 non-existent subpattern following a character class starting with ']' and
2721 containing () gave an internal compiling error instead of "reference to
2722 non-existent subpattern". Fortunately, when the pattern did exist, the
2723 compiled code was correct. (When scanning forwards to check for the
2724 existence of the subpattern, it was treating the data ']' as terminating
2725 the class, so got the count wrong. When actually compiling, the reference
2726 was subsequently set up correctly.)
2727
2728 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2729 it was being rejected as not supported by pcre_dfa_exec(), even though
2730 other assertions are supported. I have made pcre_dfa_exec() support
2731 (*FAIL).
2732
2733 16. The implementation of 13c above involved the invention of a new opcode,
2734 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2735 cannot be changed at match time, I realized I could make a small
2736 improvement to matching performance by compiling OP_ALLANY instead of
2737 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2738 on the OP_ANY path.
2739
2740 17. Compiling pcretest on Windows with readline support failed without the
2741 following two fixes: (1) Make the unistd.h include conditional on
2742 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2743
2744 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2745 ncurses library to be included for pcretest when ReadLine support is
2746 requested, but also to allow for it to be overridden. This patch came from
2747 Daniel Bergström.
2748
2749 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2750 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2751 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2752 spotting this.
2753
2754
2755 Version 7.6 28-Jan-08
2756 ---------------------
2757
2758 1. A character class containing a very large number of characters with
2759 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2760 overflow.
2761
2762 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2763 HAVE_LONG_LONG is not defined.
2764
2765 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2766 bring them up to date and include new features. This patch includes:
2767
2768 - Fixed PH's badly added libz and libbz2 support.
2769 - Fixed a problem with static linking.
2770 - Added pcredemo. [But later removed - see 7 below.]
2771 - Fixed dftables problem and added an option.
2772 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2773 HAVE_LONG_LONG.
2774 - Added readline support for pcretest.
2775 - Added an listing of the option settings after cmake has run.
2776
2777 4. A user submitted a patch to Makefile that makes it easy to create
2778 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2779 Makefile.am that cause it to include this special target, without
2780 affecting anything else. Note that the same mingw target plus all
2781 the other distribution libraries and programs are now supported
2782 when configuring with CMake (see 6 below) instead of with
2783 Configure/Make.
2784
2785 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2786 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2787 exported in the Windows port". It has not yet been confirmed that the patch
2788 solves the problem, but it does no harm.
2789
2790 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2791 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2792 with CMake, and also correct the comment about stack recursion.
2793
2794 7. Remove the automatic building of pcredemo from the ./configure system and
2795 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2796 of a program that users should build themselves after PCRE is installed, so
2797 building it automatically is not really right. What is more, it gave
2798 trouble in some build environments.
2799
2800 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2801
2802
2803 Version 7.5 10-Jan-08
2804 ---------------------
2805
2806 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2807 values in parens when parsing an RE using the C++ wrapper."
2808
2809 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2810 Characters greater than 255 were excluded from the class instead of being
2811 included.
2812
2813 3. The same bug as (2) above applied to negated POSIX classes such as
2814 [:^space:].
2815
2816 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2817 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2818 I have changed it.
2819
2820 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2821 first named subpattern) and a construct such as (?&a) would reference the
2822 first named subpattern whose name started with "a" (in other words, the
2823 length check was missing). Both these problems are fixed. "Subpattern name
2824 expected" is now given for (?&) (a zero-length name), and this patch also
2825 makes it give the same error for \k'' (previously it complained that that
2826 was a reference to a non-existent subpattern).
2827
2828 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2829 this is right because (?- can be followed by option settings as well as by
2830 digits. I have, however, made the messages clearer.
2831
2832 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2833 than the number used in the conditional) now cause a compile-time error.
2834 This is actually not compatible with Perl, which accepts such patterns, but
2835 treats the conditional as always being FALSE (as PCRE used to), but it
2836 seems to me that giving a diagnostic is better.
2837
2838 8. Change "alphameric" to the more common word "alphanumeric" in comments
2839 and messages.
2840
2841 9. Fix two occurrences of "backslash" in comments that should have been
2842 "backspace".
2843
2844 10. Remove two redundant lines of code that can never be obeyed (their function
2845 was moved elsewhere).
2846
2847 11. The program that makes PCRE's Unicode character property table had a bug
2848 which caused it to generate incorrect table entries for sequences of
2849 characters that have the same character type, but are in different scripts.
2850 It amalgamated them into a single range, with the script of the first of
2851 them. In other words, some characters were in the wrong script. There were
2852 thirteen such cases, affecting characters in the following ranges:
2853
2854 U+002b0 - U+002c1
2855 U+0060c - U+0060d
2856 U+0061e - U+00612
2857 U+0064b - U+0065e
2858 U+0074d - U+0076d
2859 U+01800 - U+01805
2860 U+01d00 - U+01d77
2861 U+01d9b - U+01dbf
2862 U+0200b - U+0200f
2863 U+030fc - U+030fe
2864 U+03260 - U+0327f
2865 U+0fb46 - U+0fbb1
2866 U+10450 - U+1049d
2867
2868 12. The -o option (show only the matching part of a line) for pcregrep was not
2869 compatible with GNU grep in that, if there was more than one match in a
2870 line, it showed only the first of them. It now behaves in the same way as
2871 GNU grep.
2872
2873 13. If the -o and -v options were combined for pcregrep, it printed a blank
2874 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2875 does the same. The return code can be used to tell if there were any
2876 non-matching lines.
2877
2878 14. Added --file-offsets and --line-offsets to pcregrep.
2879
2880 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2881 infinitely looping recursion. The bug was that positive lookaheads were not
2882 being skipped when checking for a possible empty match (negative lookaheads
2883 and both kinds of lookbehind were skipped).
2884
2885 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2886 inclusion of <windows.h> to before rather than after the definition of
2887 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2888
2889 17. Specifying a possessive quantifier with a specific limit for a Unicode
2890 character property caused pcre_compile() to compile bad code, which led at
2891 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2892 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2893 caused the error; without that there was no problem.
2894
2895 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2896
2897 19. Added --enable-pcretest-libreadline.
2898
2899 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2900 RE::GlobalReplace(). As a result, the number of replacements returned was
2901 double what it should be. I removed one of the increments, but Craig sent a
2902 later patch that removed the other one (the right fix) and added unit tests
2903 that check the return values (which was not done before).
2904
2905 21. Several CMake things:
2906
2907 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2908 the names libpcre and libpcreposix, not just pcre and pcreposix.
2909
2910 (2) The above change means that pcretest and pcregrep are now correctly
2911 linked with the newly-built libraries, not previously installed ones.
2912
2913 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2914
2915 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2916 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2917 UTF-8 newline character). The key issue is that the pattern starts .*;
2918 this means that the match must be either at the beginning, or after a
2919 newline. The bug was in the code for advancing after a failed match and
2920 checking that the new position followed a newline. It was not taking
2921 account of UTF-8 characters correctly.
2922
2923 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2924 character classes. PCRE was not treating the sequence [:...:] as a
2925 character class unless the ... were all letters. Perl, however, seems to
2926 allow any characters between [: and :], though of course it rejects as
2927 unknown any "names" that contain non-letters, because all the known class
2928 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2929 for example, whereas PCRE did not - it did not recognize a POSIX character
2930 class. This seemed a bit dangerous, so the code has been changed to be
2931 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2932 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2933 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2934 Perl does, and where it didn't before.
2935
2936 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2937 Windows environments %n is disabled by default.
2938
2939
2940 Version 7.4 21-Sep-07
2941 ---------------------
2942
2943 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2944 means that a class such as [\s] counted as "explicit reference to CR or
2945 LF". That isn't really right - the whole point of the change was to try to
2946 help when there was an actual mention of one of the two characters. So now
2947 the change happens only if \r or \n (or a literal CR or LF) character is
2948 encountered.
2949
2950 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2951 of both had grown to the point where there were only 3 bits left.
2952 Fortunately, there was spare space in the data structure, and so I have
2953 moved the internal flags into a new 16-bit field to free up more option
2954 bits.
2955
2956 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2957 but did not set the internal JCHANGED flag - either of these is enough to
2958 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2959 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2960 start sets both bits.
2961
2962 4. Added options (at build time, compile time, exec time) to change \R from
2963 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2964
2965 5. doc/pcresyntax.html was missing from the distribution.
2966
2967 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2968 compatibility, even though it is no longer used.
2969
2970 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2971 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2972 windows.h file is present (where different names are used). [This was
2973 reversed later after testing - see 16 below.]
2974
2975 8. Changed all #include <config.h> to #include "config.h". There were also
2976 some further <pcre.h> cases that I changed to "pcre.h".
2977
2978 9. When pcregrep was used with the --colour option, it missed the line ending
2979 sequence off the lines that it output.
2980
2981 10. It was pointed out to me that arrays of string pointers cause lots of
2982 relocations when a shared library is dynamically loaded. A technique of
2983 using a single long string with a table of offsets can drastically reduce
2984 these. I have refactored PCRE in four places to do this. The result is
2985 dramatic:
2986
2987 Originally: 290
2988 After changing UCP table: 187
2989 After changing error message table: 43
2990 After changing table of "verbs" 36
2991 After changing table of Posix names 22
2992
2993 Thanks to the folks working on Gregex for glib for this insight.
2994
2995 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2996 unicode-properties was also set.
2997
2998 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2999
3000 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3001 checked only for CRLF.
3002
3003 14. Added casts to pcretest.c to avoid compiler warnings.
3004
3005 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3006
3007 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3008 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3009 entirely. This removes changes made in 7 above.
3010
3011 17. The CMake files have been updated, and there is now more information about
3012 building with CMake in the NON-UNIX-USE document.
3013
3014
3015 Version 7.3 28-Aug-07
3016 ---------------------
3017
3018 1. In the rejigging of the build system that eventually resulted in 7.1, the
3019 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3020 brackets there is not right, since it causes compilers to look for an
3021 installed pcre.h, not the version that is in the source that is being
3022 compiled (which of course may be different). I have changed it back to:
3023
3024 #include "pcre.h"
3025
3026 I have a vague recollection that the change was concerned with compiling in
3027 different directories, but in the new build system, that is taken care of
3028 by the VPATH setting the Makefile.
3029
3030 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3031 when the subject happened to end in the byte 0x85 (e.g. if the last
3032 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3033 characters but of course it shouldn't be taken as a newline when it is part
3034 of another character. The bug was that, for an unlimited repeat of . in
3035 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3036 characters when looking for a newline.
3037
3038 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3039
3040 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3041 in debug output.
3042
3043 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3044 long printing in the pcrecpp unittest when running under MinGW.
3045
3046 6. ESC_K was left out of the EBCDIC table.
3047
3048 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3049 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3050 limit also applies to "virtual nesting" when a pattern is recursive, and in
3051 this case 1000 isn't so big. I have been able to remove this limit at the
3052 expense of backing off one optimization in certain circumstances. Normally,
3053 when pcre_exec() would call its internal match() function recursively and
3054 immediately return the result unconditionally, it uses a "tail recursion"
3055 feature to save stack. However, when a subpattern that can match an empty
3056 string has an unlimited repetition quantifier, it no longer makes this
3057 optimization. That gives it a stack frame in which to save the data for
3058 checking that an empty string has been matched. Previously this was taken
3059 from the 1000-entry workspace that had been reserved. So now there is no
3060 explicit limit, but more stack is used.
3061
3062 8. Applied Daniel's patches to solve problems with the import/export magic
3063 syntax that is required for Windows, and which was going wrong for the
3064 pcreposix and pcrecpp parts of the library. These were overlooked when this
3065 problem was solved for the main library.
3066
3067 9. There were some crude static tests to avoid integer overflow when computing
3068 the size of patterns that contain repeated groups with explicit upper
3069 limits. As the maximum quantifier is 65535, the maximum group length was
3070 set at 30,000 so that the product of these two numbers did not overflow a
3071 32-bit integer. However, it turns out that people want to use groups that
3072 are longer than 30,000 bytes (though not repeat them that many times).
3073 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3074 made it possible to implement the integer overflow checks in a much more
3075 dynamic way, which I have now done. The artificial limitation on group
3076 length has been removed - we now have only the limit on the total length of
3077 the compiled pattern, which depends on the LINK_SIZE setting.
3078
3079 10. Fixed a bug in the documentation for get/copy named substring when
3080 duplicate names are permitted. If none of the named substrings are set, the
3081 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3082 empty string.
3083
3084 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3085 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3086 because the ] is interpreted as the first data character and the
3087 terminating ] is not found. PCRE has been made compatible with Perl in this
3088 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3089 cause memory overwriting.
3090
3091 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3092 string has been matched (to stop an infinite loop). It was not recognizing
3093 a conditional subpattern that could match an empty string if that
3094 subpattern was within another subpattern. For example, it looped when
3095 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3096 condition was not nested. This bug has been fixed.
3097
3098 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3099 past the start of the subject in the presence of bytes with the top bit
3100 set, for example "\x8aBCD".
3101
3102 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3103 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3104
3105 14. Optimized (?!) to (*FAIL).
3106
3107 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3108 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3109 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3110 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3111 does: it's just the validity check that is more restrictive.
3112
3113 16. Inserted checks for integer overflows during escape sequence (backslash)
3114 processing, and also fixed erroneous offset values for syntax errors during
3115 backslash processing.
3116
3117 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3118 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3119
3120 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3121 caused an overrun.
3122
3123 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3124 something other than just ASCII characters) inside a group that had an
3125 unlimited repeat caused a loop at compile time (while checking to see
3126 whether the group could match an empty string).
3127
3128 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3129 [\P{Any}] did so. (Error in the code for printing property names.)
3130
3131 21. An orphan \E inside a character class could cause a crash.
3132
3133 22. A repeated capturing bracket such as (A)? could cause a wild memory
3134 reference during compilation.
3135
3136 23. There are several functions in pcre_compile() that scan along a compiled
3137 expression for various reasons (e.g. to see if it's fixed length for look
3138 behind). There were bugs in these functions when a repeated \p or \P was
3139 present in the pattern. These operators have additional parameters compared
3140 with \d, etc, and these were not being taken into account when moving along
3141 the compiled data. Specifically:
3142
3143 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3144 length.
3145
3146 (b) An item such as \pL+ within a repeated group could cause crashes or
3147 loops.
3148
3149 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3150 "reference to non-existent subpattern" error.
3151
3152 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3153
3154 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3155 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3156
3157 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3158
3159 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3160 character were causing crashes (broken optimization).
3161
3162 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3163 \p or \P) caused a compile-time loop.
3164
3165 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3166 break. For example, the unstudied pattern [\r\n]A does not match the string
3167 "\r\nA" because change 7.0/46 below moves the current point on by two
3168 characters after failing to match at the start. However, the pattern \nA
3169 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3170 the same is true. There doesn't seem any very clean way out of this, but
3171 what I have chosen to do makes the common cases work: PCRE now takes note
3172 of whether there can be an explicit match for \r or \n anywhere in the
3173 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3174 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3175 pattern has explicit CR or LF references.
3176
3177 29. Added (*CR) etc for changing newline setting at start of pattern.
3178
3179
3180 Version 7.2 19-Jun-07
3181 ---------------------
3182
3183 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3184 which is apparently normally available under Windows.
3185
3186 2. Re-jig the pcregrep tests with different newline settings in an attempt
3187 to make them independent of the local environment's newline setting.
3188
3189 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3190
3191 4. Some of the "internals" tests were previously cut out when the link size
3192 was not 2, because the output contained actual offsets. The recent new
3193 "Z" feature of pcretest means that these can be cut out, making the tests
3194 usable with all link sizes.
3195
3196 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3197 stack recursion. This gives a massive performance boost under BSD, but just
3198 a small improvement under Linux. However, it saves one field in the frame
3199 in all cases.
3200
3201 6. Added more features from the forthcoming Perl 5.10:
3202
3203 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3204 recursion call. It refers to the nth most recently opened parentheses.
3205
3206 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3207 to be opened parentheses.
3208
3209 (c) Conditions that refer to capturing parentheses can be specified
3210 relatively, for example, (?(-2)... or (?(+3)...
3211
3212 (d) \K resets the start of the current match so that everything before
3213 is not part of it.
3214
3215 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3216
3217 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3218 reference syntax.
3219
3220 (g) (?| introduces a group in which the numbering of parentheses in each
3221 alternative starts with the same number.
3222
3223 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3224
3225 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3226 PCRE_INFO_JCHANGED.
3227
3228 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3229 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3230 for detecting groups that can match an empty string.
3231
3232 9. A pattern with a very large number of alternatives (more than several
3233 hundred) was running out of internal workspace during the pre-compile
3234 phase, where pcre_compile() figures out how much memory will be needed. A
3235 bit of new cunning has reduced the workspace needed for groups with
3236 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3237 workspace instead of running out of the 4096 that are available.
3238
3239 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3240
3241 11. Applied patch from Google to remove an optimization that didn't quite work.
3242 The report of the bug said:
3243
3244 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3245 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3246 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3247
3248 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3249 it matched the wrong number of bytes.
3250
3251
3252 Version 7.1 24-Apr-07
3253 ---------------------
3254
3255 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3256 that is more "standard", making use of automake and other Autotools. There
3257 is some re-arrangement of the files and adjustment of comments consequent
3258 on this.
3259
3260 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3261 for recursive directory scanning broke on some systems because the files
3262 are not scanned in any specific order and on different systems the order
3263 was different. A call to "sort" has been inserted into RunGrepTest for the
3264 approprate test as a short-term fix. In the longer term there may be an
3265 alternative.
3266
3267 3. I had an email from Eric Raymond about problems translating some of PCRE's
3268 man pages to HTML (despite the fact that I distribute HTML pages, some
3269 people do their own conversions for various reasons). The problems
3270 concerned the use of low-level troff macros .br and .in. I have therefore
3271 removed all such uses from the man pages (some were redundant, some could
3272 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3273 HTML has been updated to handle .nf/.fi and to complain if it encounters
3274 .br or .in.
3275
3276 4. Updated comments in configure.ac that get placed in config.h.in and also
3277 arranged for config.h to be included in the distribution, with the name
3278 config.h.generic, for the benefit of those who have to compile without
3279 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3280
3281 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3282 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3283 makevp.bat for the current PCRE, using the additional files
3284 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3285
3286 6. A Windows user reported a minor discrepancy with test 2, which turned out
3287 to be caused by a trailing space on an input line that had got lost in his
3288 copy. The trailing space was an accident, so I've just removed it.
3289
3290 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3291 that is needed.
3292
3293 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3294 as "const" (a) because they are and (b) because it helps the PHP
3295 maintainers who have recently made a script to detect big data structures
3296 in the php code that should be moved to the .rodata section. I remembered
3297 to update Builducptable as well, so it won't revert if ucptable.h is ever
3298 re-created.
3299
3300 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3301 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3302 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3303 support is not required. This saves 1.5-2K of code, which is important in
3304 some applications.
3305
3306 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3307 so as not to refer to the tables, even though these functions will never be
3308 called when UTF-8 support is disabled. Otherwise there are problems with a
3309 shared library.
3310
3311 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3312
3313 (a) It was defining its arguments as char * instead of void *.
3314
3315 (b) It was assuming that all moves were upwards in memory; this was true
3316 a long time ago when I wrote it, but is no longer the case.
3317
3318 The emulated memove() is provided for those environments that have neither
3319 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3320 is clearly not the case, as these two bugs were recently reported.
3321
3322 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3323 and Detrail to create the HTML documentation, the .txt form of the man
3324 pages, and it removes trailing spaces from listed files. It also creates
3325 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3326 case, it wraps all the #defines with #ifndefs. This script should be run
3327 before "make dist".
3328
3329 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3330 with Unicode property support.
3331
3332 (a) For a maximizing quantifier, if the two different cases of the
3333 character were of different lengths in their UTF-8 codings (there are
3334 some cases like this - I found 11), and the matching function had to
3335 back up over a mixture of the two cases, it incorrectly assumed they
3336 were both the same length.
3337
3338 (b) When PCRE was configured to use the heap rather than the stack for
3339 recursion during matching, it was not correctly preserving the data for
3340 the other case of a UTF-8 character when checking ahead for a match
3341 while processing a minimizing repeat. If the check also involved
3342 matching a wide character, but failed, corruption could cause an
3343 erroneous result when trying to check for a repeat of the original
3344 character.
3345
3346 13. Some tidying changes to the testing mechanism:
3347
3348 (a) The RunTest script now detects the internal link size and whether there
3349 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3350 values substituted by "configure". (The RunGrepTest script already did
3351 this for UTF-8.) The configure.ac script no longer substitutes the
3352 relevant variables.
3353
3354 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3355 with length and offset values. This means that the output is different
3356 for different internal link sizes. Test 2 is skipped for link sizes
3357 other than 2 because of this, bypassing the problem. Unfortunately,
3358 there was also a test in test 3 (the locale tests) that used /B and
3359 failed for link sizes other than 2. Rather than cut the whole test out,
3360 I have added a new /Z option to pcretest that replaces the length and
3361 offset values with spaces. This is now used to make test 3 independent
3362 of link size. (Test 2 will be tidied up later.)
3363
3364 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3365 segmentation fault instead of returning the appropriate error message.
3366
3367 15. In multiline mode when the newline sequence was set to "any", the pattern
3368 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3369 This doesn't seem right; it now treats the CRLF combination as the line
3370 ending, and so does not match in that case. It's only a pattern such as ^$
3371 that would hit this one: something like ^ABC$ would have failed after \r
3372 and then tried again after \r\n.
3373
3374 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3375 in an attempt to make files that differ only in their line terminators
3376 compare equal. This works on Linux.
3377
3378 17. Under certain error circumstances pcregrep might try to free random memory
3379 as it exited. This is now fixed, thanks to valgrind.
3380
3381 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3382 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3383 was because its rules for how to advance for /g after matching an empty
3384 string at the end of a line did not allow for this case. They now check for
3385 it specially.
3386
3387 20. pcretest is supposed to handle patterns and data of any length, by
3388 extending its buffers when necessary. It was getting this wrong when the
3389 buffer for a data line had to be extended.
3390
3391 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3392 CRLF as a newline sequence.
3393
3394 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3395 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3396 I have nevertheless tidied it up.
3397
3398 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3399
3400 24. Added a man page for pcre-config.
3401
3402
3403 Version 7.0 19-Dec-06
3404 ---------------------
3405
3406 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3407 moving to gcc 4.1.1.
3408
3409 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3410 sys/time.h, which is documented as needed for this function. It doesn't
3411 seem to matter on Linux, but it showed up on some releases of OS X.
3412
3413 3. It seems that there are systems where bytes whose values are greater than
3414 127 match isprint() in the "C" locale. The "C" locale should be the
3415 default when a C program starts up. In most systems, only ASCII printing
3416 characters match isprint(). This difference caused the output from pcretest
3417 to vary, making some of the tests fail. I have changed pcretest so that:
3418
3419 (a) When it is outputting text in the compiled version of a pattern, bytes
3420 other than 32-126 are always shown as hex escapes.
3421
3422 (b) When it is outputting text that is a matched part of a subject string,
3423 it does the same, unless a different locale has been set for the match
3424 (using the /L modifier). In this case, it uses isprint() to decide.
3425
3426 4. Fixed a major bug that caused incorrect computation of the amount of memory
3427 required for a compiled pattern when options that changed within the
3428 pattern affected the logic of the preliminary scan that determines the
3429 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3430 that the computed length was too small. The symptoms of this bug were
3431 either the PCRE error "internal error: code overflow" from pcre_compile(),
3432 or a glibc crash with a message such as "pcretest: free(): invalid next
3433 size (fast)". Examples of patterns that provoked this bug (shown in
3434 pcretest format) are:
3435
3436 /(?-x: )/x
3437 /(?x)(?-x: \s*#\s*)/
3438 /((?i)[\x{c0}])/8
3439 /(?i:[\x{c0}])/8
3440
3441 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3442 is now done differently.
3443
3444 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3445 wrapper classes; (b) implement a new function in the C++ scanner that is
3446 more efficient than the old way of doing things because it avoids levels of
3447 recursion in the regex matching; (c) add a paragraph to the documentation
3448 for the FullMatch() function.
3449
3450 6. The escape sequence \n was being treated as whatever was defined as
3451 "newline". Not only was this contrary to the documentation, which states
3452 that \n is character 10 (hex 0A), but it also went horribly wrong when
3453 "newline" was defined as CRLF. This has been fixed.
3454
3455 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3456 was being set to -1 for the "end of line" case (supposedly a value that no
3457 character can have). Though this value is never used (the check for end of
3458 line is "zero bytes in current character"), it caused compiler complaints.
3459 I've changed it to 0xffffffff.
3460
3461 8. In pcre_version.c, the version string was being built by a sequence of
3462 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3463 string (as it is for production releases) called a macro with an empty
3464 argument. The C standard says the result of this is undefined. The gcc
3465 compiler treats it as an empty string (which was what was wanted) but it is
3466 reported that Visual C gives an error. The source has been hacked around to
3467 avoid this problem.
3468
3469 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3470 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3471 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3472 of them did).
3473
3474 10. Originally, pcretest opened its input and output without "b"; then I was
3475 told that "b" was needed in some environments, so it was added for release
3476 5.0 to both the input and output. (It makes no difference on Unix-like
3477 systems.) Later I was told that it is wrong for the input on Windows. I've
3478 now abstracted the modes into two macros, to make it easier to fiddle with
3479 them, and removed "b" from the input mode under Windows.
3480
3481 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3482
3483 12. Added -help and --help to pcretest as an official way of being reminded
3484 of the options.
3485
3486 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3487 and pcrecpp.cc because they annoy compilers at high warning levels.
3488
3489 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3490
3491 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3492 scripts are not C programs :-) and which was not noticed because it works
3493 on Linux.
3494
3495 16. pcretest is supposed to handle any length of pattern and data line (as one
3496 line or as a continued sequence of lines) by extending its input buffer if
3497 necessary. This feature was broken for very long pattern lines, leading to
3498 a string of junk being passed to pcre_compile() if the pattern was longer
3499 than about 50K.
3500
3501 17. I have done a major re-factoring of the way pcre_compile() computes the
3502 amount of memory needed for a compiled pattern. Previously, there was code
3503 that made a preliminary scan of the pattern in order to do this. That was
3504 OK when PCRE was new, but as the facilities have expanded, it has become
3505 harder and harder to keep it in step with the real compile phase, and there
3506 have been a number of bugs (see for example, 4 above). I have now found a
3507 cunning way of running the real compile function in a "fake" mode that
3508 enables it to compute how much memory it would need, while actually only
3509 ever using a few hundred bytes of working memory and without too many
3510 tests of the mode. This should make future maintenance and development
3511 easier. A side effect of this work is that the limit of 200 on the nesting
3512 depth of parentheses has been removed (though this was never a serious
3513 limitation, I suspect). However, there is a downside: pcre_compile() now
3514 runs more slowly than before (30% or more, depending on the pattern). I
3515 hope this isn't a big issue. There is no effect on runtime performance.
3516
3517 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3518 newline (only possible for the last line of a file) and it was a
3519 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3520
3521 19. Added additional timing features to pcretest. (1) The -tm option now times
3522 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3523 separate command line item, by a number that specifies the number of
3524 repeats to use when timing. The default is 50000; this gives better
3525 precision, but takes uncomfortably long for very large patterns.
3526
3527 20. Extended pcre_study() to be more clever in cases where a branch of a
3528 subpattern has no definite first character. For example, (a*|b*)[cd] would
3529 previously give no result from pcre_study(). Now it recognizes that the
3530 first character must be a, b, c, or d.
3531
3532 21. There was an incorrect error "recursive call could loop indefinitely" if
3533 a subpattern (or the entire pattern) that was being tested for matching an
3534 empty string contained only one non-empty item after a nested subpattern.
3535 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3536 incorrectly, because the \d was being skipped in the check.
3537
3538 22. The pcretest program now has a new pattern option /B and a command line
3539 option -b, which is equivalent to adding /B to every pattern. This causes
3540 it to show the compiled bytecode, without the additional information that
3541 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3542 is the same as /B/I).
3543
3544 23. A new optimization is now able automatically to treat some sequences such
3545 as a*b as a*+b. More specifically, if something simple (such as a character
3546 or a simple class like \d) has an unlimited quantifier, and is followed by
3547 something that cannot possibly match the quantified thing, the quantifier
3548 is automatically "possessified".
3549
3550 24. A recursive reference to a subpattern whose number was greater than 39
3551 went wrong under certain circumstances in UTF-8 mode. This bug could also
3552 have affected the operation of pcre_study().
3553
3554 25. Realized that a little bit of performance could be had by replacing
3555 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3556
3557 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3558
3559 27. Possessive quantifiers such as a++ were previously implemented by turning
3560 them into atomic groups such as ($>a+). Now they have their own opcodes,
3561 which improves performance. This includes the automatically created ones
3562 from 23 above.
3563
3564 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3565 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3566 the first matched character to be a colon. This applied both to named and
3567 numbered groups.
3568
3569 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3570
3571 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3572 building PCRE on Windows easier, so I have included it in the distribution.
3573
3574 31. There is now a check in pcretest against a ridiculously large number being
3575 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3576 loop, the loop is abandoned.
3577
3578 32. Forward references to subpatterns in conditions such as (?(2)...) where
3579 subpattern 2 is defined later cause pcre_compile() to search forwards in
3580 the pattern for the relevant set of parentheses. This search went wrong
3581 when there were unescaped parentheses in a character class, parentheses
3582 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3583
3584 33. "Subroutine" calls and backreferences were previously restricted to
3585 referencing subpatterns earlier in the regex. This restriction has now
3586 been removed.
3587
3588 34. Added a number of extra features that are going to be in Perl 5.10. On the
3589 whole, these are just syntactic alternatives for features that PCRE had
3590 previously implemented using the Python syntax or my own invention. The
3591 other formats are all retained for compatibility.
3592
3593 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3594 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3595 also .NET compatible.
3596
3597 (b) A recursion or subroutine call to a named group can now be defined as
3598 (?&name) as well as (?P>name).
3599
3600 (c) A backreference to a named group can now be defined as \k<name> or
3601 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3602 5.10, are also .NET compatible.
3603
3604 (d) A conditional reference to a named group can now use the syntax
3605 (?(<name>) or (?('name') as well as (?(name).
3606
3607 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3608 groups (named and numbered) that are never evaluated inline, but can be
3609 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3610 is always false. There may be only one alternative in such a group.
3611
3612 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3613 as the simple (?(R). The condition is true only if the most recent
3614 recursion is that of the given number or name. It does not search out
3615 through the entire recursion stack.
3616
3617 (g) The escape \gN or \g{N} has been added, where N is a positive or
3618 negative number, specifying an absolute or relative reference.
3619
3620 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3621 some "unreachable code" warnings.
3622
3623 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3624 things, this adds five new scripts.
3625
3626 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3627 There were also incompatibilities regarding the handling of \Q..\E inside
3628 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3629 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3630
3631 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3632 matches an empty string, and forcibly breaks the loop. There were bugs in
3633 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3634 against aaaa the result was just "a" rather than "aaaa", for example. Two
3635 separate and independent bugs (that affected different cases) have been
3636 fixed.
3637
3638 39. Refactored the code to abolish the use of different opcodes for small
3639 capturing bracket numbers. This is a tidy that I avoided doing when I
3640 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3641 The new approach is not only tidier, it makes it possible to reduce the
3642 memory needed to fix the previous bug (38).
3643
3644 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3645 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3646 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3647 mode.
3648
3649 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3650 report.
3651
3652 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3653 copy construction and assignment in the C++ wrapper.
3654
3655 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3656 couple of bugs that could have given wrong results in the "--newline=crlf"
3657 case.
3658
3659 44. Added a number of casts and did some reorganization of signed/unsigned int
3660 variables following suggestions from Dair Grant. Also renamed the variable
3661 "this" as "item" because it is a C++ keyword.
3662
3663 45. Arranged for dftables to add
3664
3665 #include "pcre_internal.h"
3666
3667 to pcre_chartables.c because without it, gcc 4.x may remove the array
3668 definition from the final binary if PCRE is built into a static library and
3669 dead code stripping is activated.
3670
3671 46. For an unanchored pattern, if a match attempt fails at the start of a
3672 newline sequence, and the newline setting is CRLF or ANY, and the next two
3673 characters are CRLF, advance by two characters instead of one.
3674
3675
3676 Version 6.7 04-Jul-06
3677 ---------------------
3678
3679 1. In order to handle tests when input lines are enormously long, pcretest has
3680 been re-factored so that it automatically extends its buffers when
3681 necessary. The code is crude, but this _is_ just a test program. The
3682 default size has been increased from 32K to 50K.
3683
3684 2. The code in pcre_study() was using the value of the re argument before
3685 testing it for NULL. (Of course, in any sensible call of the function, it
3686 won't be NULL.)
3687
3688 3. The memmove() emulation function in pcre_internal.h, which is used on
3689 systems that lack both memmove() and bcopy() - that is, hardly ever -
3690 was missing a "static" storage class specifier.
3691
3692 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3693 containing an extended class (one that cannot be represented by a bitmap
3694 because it contains high-valued characters or Unicode property items, e.g.
3695 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3696 pattern, but PCRE should not loop if you do not (it no longer does).
3697 [Detail: two cases were found: (a) a repeated subpattern containing an
3698 extended class; (b) a recursive reference to a subpattern that followed a
3699 previous extended class. It wasn't skipping over the extended class
3700 correctly when UTF-8 mode was not set.]
3701
3702 5. A negated single-character class was not being recognized as fixed-length
3703 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3704 compile error "lookbehind assertion is not fixed length".
3705
3706 6. The RunPerlTest auxiliary script was showing an unexpected difference
3707 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3708 write a Perl script that can interpret lines of an input file either as
3709 byte characters or as UTF-8, which is what "perltest" was being required to
3710 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3711 can't do is switch easily at run time between having the "use utf8;" pragma
3712 or not. In the end, I fudged it by using the RunPerlTest script to insert
3713 "use utf8;" explicitly for the UTF-8 tests.
3714
3715 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3716 the end of the subject string, contrary to the documentation and to what
3717 Perl does. This was true of both matching functions. Now it matches only at
3718 the start of the subject and immediately after *internal* newlines.
3719
3720 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3721 a pointer to an int instead of a pointer to an unsigned long int. This
3722 caused problems on 64-bit systems.
3723
3724 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3725 instance of the 'standard' template library not being so standard".
3726
3727 10. There was no check on the number of named subpatterns nor the maximum
3728 length of a subpattern name. The product of these values is used to compute
3729 the size of the memory block for a compiled pattern. By supplying a very
3730 long subpattern name and a large number of named subpatterns, the size
3731 computation could be caused to overflow. This is now prevented by limiting
3732 the length of names to 32 characters, and the number of named subpatterns
3733 to 10,000.
3734
3735 11. Subpatterns that are repeated with specific counts have to be replicated in
3736 the compiled pattern. The size of memory for this was computed from the
3737 length of the subpattern and the repeat count. The latter is limited to
3738 65535, but there was no limit on the former, meaning that integer overflow
3739 could in principle occur. The compiled length of a repeated subpattern is
3740 now limited to 30,000 bytes in order to prevent this.
3741
3742 12. Added the optional facility to have named substrings with the same name.
3743
3744 13. Added the ability to use a named substring as a condition, using the
3745 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3746 are numbers (not recommended). Forward references are permitted.
3747
3748 14. Added forward references in named backreferences (if you see what I mean).
3749
3750 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3751 pattern could run off the end of the subject. For example, the pattern
3752 "(?s)(.{1,5})"8 did this with the subject "ab".
3753
3754 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3755 PCRE_CASELESS was set when matching characters that were quantified with ?
3756 or *.
3757
3758 17. A character class other than a single negated character that had a minimum
3759 but no maximum quantifier - for example [ab]{6,} - was not handled
3760 correctly by pce_dfa_exec(). It would match only one character.
3761
3762 18. A valid (though odd) pattern that looked like a POSIX character
3763 class but used an invalid character after [ (for example [[,abc,]]) caused
3764 pcre_compile() to give the error "Failed: internal error: code overflow" or
3765 in some cases to crash with a glibc free() error. This could even happen if
3766 the pattern terminated after [[ but there just happened to be a sequence of
3767 letters, a binary zero, and a closing ] in the memory that followed.
3768
3769 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3770 over the years. Originally (before any Unicode support), just the bottom 8
3771 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3772 output from "man perlunicode" includes this:
3773
3774 The regular expression compiler produces polymorphic opcodes. That
3775 is, the pattern adapts to the data and automatically switches to
3776 the Unicode character scheme when presented with Unicode data--or
3777 instead uses a traditional byte scheme when presented with byte
3778 data.
3779
3780 Sadly, a wide octal escape does not cause a switch, and in a string with
3781 no other multibyte characters, these octal escapes are treated as before.
3782 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3783 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3784 Unicode string.
3785
3786 I have not perpetrated such confusion in PCRE. Up till now, it took just
3787 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3788 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3789 translate to the appropriate multibyte character.
3790
3791 29. Applied some refactoring to reduce the number of warnings from Microsoft
3792 and Borland compilers. This has included removing the fudge introduced
3793 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3794 a warning about an unused variable.
3795
3796 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3797 characters since release 4.0, because Perl (from release 5.004) does not.
3798 [Or at least, is documented not to: some releases seem to be in conflict
3799 with the documentation.] However, when a pattern was studied with
3800 pcre_study() and all its branches started with \s, PCRE still included VT
3801 as a possible starting character. Of course, this did no harm; it just
3802 caused an unnecessary match attempt.
3803
3804 22. Removed a now-redundant internal flag bit that recorded the fact that case
3805 dependency changed within the pattern. This was once needed for "required
3806 byte" processing, but is no longer used. This recovers a now-scarce options
3807 bit. Also moved the least significant internal flag bit to the most-
3808 significant bit of the word, which was not previously used (hangover from
3809 the days when it was an int rather than a uint) to free up another bit for
3810 the future.
3811
3812 23. Added support for CRLF line endings as well as CR and LF. As well as the
3813 default being selectable at build time, it can now be changed at runtime
3814 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3815 specify that it is scanning data with non-default line endings.
3816
3817 24. Changed the definition of CXXLINK to make it agree with the definition of
3818 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3819
3820 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3821 recursions. This makes a big different to stack usage for some patterns.
3822
3823 26. If a subpattern containing a named recursion or subroutine reference such
3824 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3825 the space required for the compiled pattern went wrong and gave too small a
3826 value. Depending on the environment, this could lead to "Failed: internal
3827 error: code overflow at offset 49" or "glibc detected double free or
3828 corruption" errors.
3829
3830 27. Applied patches from Google (a) to support the new newline modes and (b) to
3831 advance over multibyte UTF-8 characters in GlobalReplace.
3832
3833 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3834 difference for some implementation of PCRE in some Windows version.
3835
3836 29. Added some extra testing facilities to pcretest:
3837
3838 \q<number> in a data line sets the "match limit" value
3839 \Q<number> in a data line sets the "match recursion limt" value
3840 -S <number> sets the stack size, where <number> is in megabytes
3841
3842 The -S option isn't available for Windows.
3843
3844
3845 Version 6.6 06-Feb-06
3846 ---------------------
3847
3848 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3849 in pcreposix.h. I have copied the definition from pcre.h.
3850
3851 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3852 because pcre.h is no longer a built file.
3853
3854 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3855 not normally included in the compiled code.
3856
3857
3858 Version 6.5 01-Feb-06
3859 ---------------------
3860
3861 1. When using the partial match feature with pcre_dfa_exec(), it was not
3862 anchoring the second and subsequent partial matches at the new starting
3863 point. This could lead to incorrect results. For example, with the pattern
3864 /1234/, partially matching against "123" and then "a4" gave a match.
3865
3866 2. Changes to pcregrep:
3867
3868 (a) All non-match returns from pcre_exec() were being treated as failures
3869 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3870 error message is output. Some extra information is given for the
3871 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3872 probably the only errors that are likely to be caused by users (by
3873 specifying a regex that has nested indefinite repeats, for instance).
3874 If there are more than 20 of these errors, pcregrep is abandoned.
3875
3876 (b) A binary zero was treated as data while matching, but terminated the
3877 output line if it was written out. This has been fixed: binary zeroes
3878 are now no different to any other data bytes.
3879
3880 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3881 used to set a locale for matching. The --locale=xxxx long option has
3882 been added (no short equivalent) to specify a locale explicitly on the
3883 pcregrep command, overriding the environment variables.
3884
3885 (d) When -B was used with -n, some line numbers in the output were one less
3886 than they should have been.
3887
3888 (e) Added the -o (--only-matching) option.
3889
3890 (f) If -A or -C was used with -c (count only), some lines of context were
3891 accidentally printed for the final match.
3892
3893 (g) Added the -H (--with-filename) option.
3894
3895 (h) The combination of options -rh failed to suppress file names for files
3896 that were found from directory arguments.
3897
3898 (i) Added the -D (--devices) and -d (--directories) options.
3899
3900 (j) Added the -F (--fixed-strings) option.
3901
3902 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3903
3904 (l) Added the --colo(u)r option.
3905
3906 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3907 is not present by default.
3908
3909 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3910 items such as (?R) or (?1), when the recursion could match a number of
3911 alternatives. If it matched one of the alternatives, but subsequently,
3912 outside the recursion, there was a failure, the code tried to back up into
3913 the recursion. However, because of the way PCRE is implemented, this is not
3914 possible, and the result was an incorrect result from the match.
3915
3916 In order to prevent this happening, the specification of recursion has
3917 been changed so that all such subpatterns are automatically treated as
3918 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3919
3920 4. I had overlooked the fact that, in some locales, there are characters for
3921 which isalpha() is true but neither isupper() nor islower() are true. In
3922 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3923 and ordfeminine) are like this. This affected the treatment of \w and \W
3924 when they appeared in character classes, but not when they appeared outside
3925 a character class. The bit map for "word" characters is now created
3926 separately from the results of isalnum() instead of just taking it from the
3927 upper, lower, and digit maps. (Plus the underscore character, of course.)
3928
3929 5. The above bug also affected the handling of POSIX character classes such as
3930 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3931 permanent tables. Instead, the bit maps for such a class were previously
3932 created as the appropriate unions of the upper, lower, and digit bitmaps.
3933 Now they are created by subtraction from the [[:word:]] class, which has
3934 its own bitmap.
3935
3936 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3937 It is created by subtracting the vertical space characters (\x09, \x0a,
3938 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3939 subtraction was done in the overall bitmap for a character class, meaning
3940 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3941 be recognized. This bug has been fixed.
3942
3943 7. Patches from the folks at Google:
3944
3945 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3946 real life, but is still worth protecting against".
3947
3948 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3949 regular expressions".
3950
3951 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3952 have it.
3953
3954 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3955 "configure" and the latter not, in order to fix a problem somebody had
3956 with compiling the Arg class on HP-UX.
3957
3958 (e) Improve the error-handling of the C++ wrapper a little bit.
3959
3960 (f) New tests for checking recursion limiting.
3961
3962 8. The pcre_memmove() function, which is used only if the environment does not
3963 have a standard memmove() function (and is therefore rarely compiled),
3964 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3965 returning a result (though PCRE never actually uses the result).
3966
3967 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3968 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3969 returned instead of calling malloc() with an overflowing number that would
3970 most likely cause subsequent chaos.
3971
3972 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3973
3974 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3975 with this option is matched, the nmatch and pmatch options of regexec() are
3976 ignored.
3977
3978 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3979 provided in case anyone wants to the the POSIX interface with UTF-8
3980 strings.
3981
3982 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3983 C++ linking (needed for some HP-UX environments).
3984
3985 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3986 (unused parameter) and in the pcre_printint() function (omitted "default"
3987 switch label when the default is to do nothing).
3988
3989 15. Added some code to make it possible, when PCRE is compiled as a C++
3990 library, to replace subject pointers for pcre_exec() with a smart pointer
3991 class, thus making it possible to process discontinuous strings.
3992
3993 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3994 much the same function. They were added by different people who were trying
3995 to make PCRE easy to compile on non-Unix systems. It has been suggested
3996 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3997 for compiling on Windows systems. I have therefore replaced it with
3998 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3999 defaults to "extern" for C or "extern C" for C++, which works fine on
4000 Unix-like systems. It is now possible to override the value of PCRE_DATA_
4001 SCOPE with something explicit in config.h. In addition:
4002
4003 (a) pcreposix.h still had just "extern" instead of either of these macros;
4004 I have replaced it with PCRE_DATA_SCOPE.
4005
4006 (b) Functions such as _pcre_xclass(), which are internal to the library,
4007 but external in the C sense, all had PCRE_EXPORT in their definitions.
4008 This is apparently wrong for the Windows case, so I have removed it.
4009 (It makes no difference on Unix-like systems.)
4010
4011 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
4012 of recursive calls to match(). This is different to MATCH_LIMIT because
4013 that limits the total number of calls to match(), not all of which increase
4014 the depth of recursion. Limiting the recursion depth limits the amount of
4015 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4016 when PCRE is compiled, and changed at run time. A patch from Google adds
4017 this functionality to the C++ interface.
4018
4019 18. Changes to the handling of Unicode character properties:
4020
4021 (a) Updated the table to Unicode 4.1.0.
4022
4023 (b) Recognize characters that are not in the table as "Cn" (undefined).
4024
4025 (c) I revised the way the table is implemented to a much improved format
4026 which includes recognition of ranges. It now supports the ranges that
4027 are defined in UnicodeData.txt, and it also amalgamates other
4028 characters into ranges. This has reduced the number of entries in the
4029 table from around 16,000 to around 3,000, thus reducing its size
4030 considerably. I realized I did not need to use a tree structure after
4031 all - a binary chop search is just as efficient. Having reduced the
4032 number of entries, I extended their size from 6 bytes to 8 bytes to
4033 allow for more data.
4034
4035 (d) Added support for Unicode script names via properties such as \p{Han}.
4036
4037 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4038 matching that character.
4039
4040 20. When matching a repeated Unicode property with a minimum greater than zero,
4041 (for example \pL{2,}), PCRE could look past the end of the subject if it
4042 reached it while seeking the minimum number of characters. This could
4043 happen only if some of the characters were more than one byte long, because
4044 there is a check for at least the minimum number of bytes.
4045
4046 21. Refactored the implementation of \p and \P so as to be more general, to
4047 allow for more different types of property in future. This has changed the
4048 compiled form incompatibly. Anybody with saved compiled patterns that use
4049 \p or \P will have to recompile them.
4050
4051 22. Added "Any" and "L&" to the supported property types.
4052
4053 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4054 but give a compile time error if the value is greater than 0xff.
4055
4056 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4057 accidentally not being installed or uninstalled.
4058
4059 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4060 made were to insert the current release number. This seemed silly, because
4061 it made things harder for people building PCRE on systems that don't run
4062 "configure". I have turned pcre.h into a distributed file, no longer built
4063 by "configure", with the version identification directly included. There is
4064 no longer a pcre.h.in file.
4065
4066 However, this change necessitated a change to the pcre-config script as
4067 well. It is built from pcre-config.in, and one of the substitutions was the
4068 release number. I have updated configure.ac so that ./configure now finds
4069 the release number by grepping pcre.h.
4070
4071 26. Added the ability to run the tests under valgrind.
4072
4073
4074 Version 6.4 05-Sep-05
4075 ---------------------
4076
4077 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4078 "--" to be printed when multiple files were scanned, even when none of the
4079 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4080 consider it to be a bug, and have restored the previous behaviour.
4081
4082 2. A couple of code tidies to get rid of compiler warnings.
4083
4084 3. The pcretest program used to cheat by referring to symbols in the library
4085 whose names begin with _pcre_. These are internal symbols that are not
4086 really supposed to be visible externally, and in some environments it is
4087 possible to suppress them. The cheating is now confined to including
4088 certain files from the library's source, which is a bit cleaner.
4089
4090 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4091 file's purpose clearer.
4092
4093 5. Reorganized pcre_ucp_findchar().
4094
4095
4096 Version 6.3 15-Aug-05
4097 ---------------------
4098
4099 1. The file libpcre.pc.in did not have general read permission in the tarball.
4100
4101 2. There were some problems when building without C++ support:
4102
4103 (a) If C++ support was not built, "make install" and "make test" still
4104 tried to test it.
4105
4106 (b) There were problems when the value of CXX was explicitly set. Some
4107 changes have been made to try to fix these, and ...
4108
4109 (c) --disable-cpp can now be used to explicitly disable C++ support.
4110
4111 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4112 backslash in a target when C++ was disabled. This confuses some
4113 versions of "make", apparently. Using an intermediate variable solves
4114 this. (Same for CPP_LOBJ.)
4115
4116 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4117 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4118 necessary on certain architectures.
4119
4120 4. Added a setting of -export-symbols-regex to the link command to remove
4121 those symbols that are exported in the C sense, but actually are local
4122 within the library, and not documented. Their names all begin with
4123 "_pcre_". This is not a perfect job, because (a) we have to except some
4124 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4125 available (and never for static libraries). I have made a note to try to
4126 find a way round (a) in the future.
4127
4128
4129 Version 6.2 01-Aug-05
4130 ---------------------
4131
4132 1. There was no test for integer overflow of quantifier values. A construction
4133 such as {1111111111111111} would give undefined results. What is worse, if
4134 a minimum quantifier for a parenthesized subpattern overflowed and became
4135 negative, the calculation of the memory size went wrong. This could have
4136 led to memory overwriting.
4137
4138 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4139
4140 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4141 operating environments where this matters.
4142
4143 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4144 PCRE options from within the C++ wrapper.
4145
4146 5. Named capturing subpatterns were not being correctly counted when a pattern
4147 was compiled. This caused two problems: (a) If there were more than 100
4148 such subpatterns, the calculation of the memory needed for the whole
4149 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4150 back references of the form \12, where the number was greater than 9, were
4151 not recognized as back references, even though there were sufficient
4152 previous subpatterns.
4153
4154 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4155 versions of gcc, e.g. 2.95.4.
4156
4157
4158 Version 6.1 21-Jun-05
4159 ---------------------
4160
4161 1. There was one reference to the variable "posix" in pcretest.c that was not
4162 surrounded by "#if !defined NOPOSIX".
4163
4164 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4165 the cross-check on the old pcre_info() function, for the benefit of the
4166 cut-down version of PCRE that is currently imported into Exim.
4167
4168 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4169 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4170 patterns that start (?i) but I don't think that matters. The use of (?i) is
4171 just an example; this all applies to the other options as well.
4172
4173 4. Since libtool seems to echo the compile commands it is issuing, the output
4174 from "make" can be reduced a bit by putting "@" in front of each libtool
4175 compile command.
4176
4177 5. Patch from the folks at Google for configure.in to be a bit more thorough
4178 in checking for a suitable C++ installation before trying to compile the
4179 C++ stuff. This should fix a reported problem when a compiler was present,
4180 but no suitable headers.
4181
4182 6. The man pages all had just "PCRE" as their title. I have changed them to
4183 be the relevant file name. I have also arranged that these names are
4184 retained in the file doc/pcre.txt, which is a concatenation in text format
4185 of all the man pages except the little individual ones for each function.
4186
4187 7. The NON-UNIX-USE file had not been updated for the different set of source
4188 files that come with release 6. I also added a few comments about the C++
4189 wrapper.
4190
4191
4192 Version 6.0 07-Jun-05
4193 ---------------------
4194
4195 1. Some minor internal re-organization to help with my DFA experiments.
4196
4197 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4198 didn't matter for the library itself when fully configured, but did matter
4199 when compiling without UCP support, or within Exim, where the ucp files are
4200 not imported.
4201
4202 3. Refactoring of the library code to split up the various functions into
4203 different source modules. The addition of the new DFA matching code (see
4204 below) to a single monolithic source would have made it really too
4205 unwieldy, quite apart from causing all the code to be include in a
4206 statically linked application, when only some functions are used. This is
4207 relevant even without the DFA addition now that patterns can be compiled in
4208 one application and matched in another.
4209
4210 The downside of splitting up is that there have to be some external
4211 functions and data tables that are used internally in different modules of
4212 the library but which are not part of the API. These have all had their
4213 names changed to start with "_pcre_" so that they are unlikely to clash
4214 with other external names.
4215
4216 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4217 a different (DFA) algorithm. Although it is slower than the original
4218 function, it does have some advantages for certain types of matching
4219 problem.
4220
4221 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4222 including restarting after a partial match.
4223
4224 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4225 defined when compiling for Windows was sent to me. I have put it into the
4226 code, though I have no means of testing or verifying it.
4227
4228 7. Added the pcre_refcount() auxiliary function.
4229
4230 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4231 match before or at the first newline in the subject string. In pcretest,
4232 the /f option on a pattern can be used to set this.
4233
4234 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4235 would behave wrongly. This has been present in PCRE since release 4.0.
4236
4237 10. A number of changes to the pcregrep command:
4238
4239 (a) Refactored how -x works; insert ^(...)$ instead of setting
4240 PCRE_ANCHORED and checking the length, in preparation for adding
4241 something similar for -w.
4242
4243 (b) Added the -w (match as a word) option.
4244
4245 (c) Refactored the way lines are read and buffered so as to have more
4246 than one at a time available.
4247
4248 (d) Implemented a pcregrep test script.
4249
4250 (e) Added the -M (multiline match) option. This allows patterns to match
4251 over several lines of the subject. The buffering ensures that at least
4252 8K, or the rest of the document (whichever is the shorter) is available
4253 for matching (and similarly the previous 8K for lookbehind assertions).
4254
4255 (f) Changed the --help output so that it now says
4256
4257 -w, --word-regex(p)
4258
4259 instead of two lines, one with "regex" and the other with "regexp"
4260 because that confused at least one person since the short forms are the
4261 same. (This required a bit of code, as the output is generated
4262 automatically from a table. It wasn't just a text change.)
4263
4264 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4265 option but starts with a hyphen. Could be a pattern or a path name
4266 starting with a hyphen, for instance.
4267
4268 (h) "-" can be given as a file name to represent stdin.
4269
4270 (i) When file names are being printed, "(standard input)" is used for
4271 the standard input, for compatibility with GNU grep. Previously
4272 "<stdin>" was used.
4273
4274 (j) The option --label=xxx can be used to supply a name to be used for
4275 stdin when file names are being printed. There is no short form.
4276
4277 (k) Re-factored the options decoding logic because we are going to add
4278 two more options that take data. Such options can now be given in four
4279 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4280
4281 (l) Added the -A, -B, and -C options for requesting that lines of context
4282 around matches be printed.
4283
4284 (m) Added the -L option to print the names of files that do not contain
4285 any matching lines, that is, the complement of -l.
4286
4287 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4288 continue to scan other files.
4289
4290 (o) The -s option was incorrectly implemented. For compatibility with other
4291 greps, it now suppresses the error message for a non-existent or non-
4292 accessible file (but not the return code). There is a new option called
4293 -q that suppresses the output of matching lines, which was what -s was
4294 previously doing.
4295
4296 (p) Added --include and --exclude options to specify files for inclusion
4297 and exclusion when recursing.
4298
4299 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4300 Hopefully, it now does.
4301
4302 12. Missing cast in pcre_study().
4303
4304 13. Added an "uninstall" target to the makefile.
4305
4306 14. Replaced "extern" in the function prototypes in Makefile.in with
4307 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4308 world, but is set differently for Windows.
4309
4310 15. Added a second compiling function called pcre_compile2(). The only
4311 difference is that it has an extra argument, which is a pointer to an
4312 integer error code. When there is a compile-time failure, this is set
4313 non-zero, in addition to the error test pointer being set to point to an
4314 error message. The new argument may be NULL if no error number is required
4315 (but then you may as well call pcre_compile(), which is now just a
4316 wrapper). This facility is provided because some applications need a
4317 numeric error indication, but it has also enabled me to tidy up the way
4318 compile-time errors are handled in the POSIX wrapper.
4319
4320 16. Added VPATH=.libs to the makefile; this should help when building with one
4321 prefix path and installing with another. (Or so I'm told by someone who
4322 knows more about this stuff than I do.)
4323
4324 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4325 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4326 match everything, including newlines. This is not POSIX-compatible, but
4327 somebody wanted the feature. From pcretest it can be activated by using
4328 both the P and the s flags.
4329
4330 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4331
4332 19. libpcre.pc was being incorrectly installed as executable.
4333
4334 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4335 it now also looks for '\r' so that it will work unmodified on Windows.
4336
4337 21. Added Google's contributed C++ wrapper to the distribution.
4338
4339 22. Added some untidy missing memory free() calls in pcretest, to keep
4340 Electric Fence happy when testing.
4341
4342
4343
4344 Version 5.0 13-Sep-04
4345 ---------------------
4346
4347 1. Internal change: literal characters are no longer packed up into items
4348 containing multiple characters in a single byte-string. Each character
4349 is now matched using a separate opcode. However, there may be more than one
4350 byte in the character in UTF-8 mode.
4351
4352 2. The pcre_callout_block structure has two new fields: pattern_position and
4353 next_item_length. These contain the offset in the pattern to the next match
4354 item, and its length, respectively.
4355
4356 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4357 insertion of callouts before each pattern item. Added the /C option to
4358 pcretest to make use of this.
4359
4360 4. On the advice of a Windows user, the lines
4361
4362 #if defined(_WIN32) || defined(WIN32)
4363 _setmode( _fileno( stdout ), 0x8000 );
4364 #endif /* defined(_WIN32) || defined(WIN32) */
4365
4366 have been added to the source of pcretest. This apparently does useful
4367 magic in relation to line terminators.
4368
4369 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4370 for the benefit of those environments where the "b" makes a difference.
4371
4372 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4373 to know about it. I have put a hack into configure.in that adds in code
4374 to set GCC=yes if CC=icc. This seems to end up at a point in the
4375 generated configure script that is early enough to affect the setting of
4376 compiler options, which is what is needed, but I have no means of testing
4377 whether it really works. (The user who reported this had patched the
4378 generated configure script, which of course I cannot do.)
4379
4380 LATER: After change 22 below (new libtool files), the configure script
4381 seems to know about icc (and also ecc). Therefore, I have commented out
4382 this hack in configure.in.
4383
4384 7. Added support for pkg-config (2 patches were sent in).
4385
4386 8. Negated POSIX character classes that used a combination of internal tables
4387 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4388 [[:^ascii]]. Typically, they would match almost any characters. The other
4389 POSIX classes were not broken in this way.
4390
4391 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4392 to find the match, as PCRE was deluded into thinking that the match had to
4393 start at the start point or following a newline. The same bug applied to
4394 patterns with negative forward assertions or any backward assertions
4395 preceding ".*" at the start, unless the pattern required a fixed first
4396 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4397
4398 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4399 starting at the last subject character, bytes beyond the end of the subject
4400 string were read.
4401
4402 11. Renamed the variable "class" as "classbits" to make life easier for C++
4403 users. (Previously there was a macro definition, but it apparently wasn't
4404 enough.)
4405
4406 12. Added the new field "tables" to the extra data so that tables can be passed
4407 in at exec time, or the internal tables can be re-selected. This allows
4408 a compiled regex to be saved and re-used at a later time by a different
4409 program that might have everything at different addresses.
4410
4411 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4412 -R library as well as a -L library.
4413
4414 14. The debugging options of pcretest (-d on the command line or D on a
4415 pattern) showed incorrect output for anything following an extended class
4416 that contained multibyte characters and which was followed by a quantifier.
4417
4418 15. Added optional support for general category Unicode character properties
4419 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4420 support. It adds about 90K to the size of the library. The meanings of the
4421 inbuilt class escapes such as \d and \s have NOT been changed.
4422
4423 16. Updated pcredemo.c to include calls to free() to release the memory for the
4424 compiled pattern.
4425
4426 17. The generated file chartables.c was being created in the source directory
4427 instead of in the building directory. This caused the build to fail if the
4428 source directory was different from the building directory, and was
4429 read-only.
4430
4431 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4432 file. No doubt somebody will tell me if they don't make sense... Also added
4433 Dan Mooney's comments about building on OpenVMS.
4434
4435 19. Added support for partial matching via the PCRE_PARTIAL option for
4436 pcre_exec() and the \P data escape in pcretest.
4437
4438 20. Extended pcretest with 3 new pattern features:
4439
4440 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4441 write the compiled pattern to the file whose name is "rest-of-line".
4442 This is a straight binary dump of the data, with the saved pointer to
4443 the character tables forced to be NULL. The study data, if any, is
4444 written too. After writing, pcretest reads a new pattern.
4445
4446 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4447 compiled pattern from the given file. There must not be any
4448 occurrences of "<" in the file name (pretty unlikely); if there are,
4449 pcretest will instead treat the initial "<" as a pattern delimiter.
4450 After reading in the pattern, pcretest goes on to read data lines as
4451 usual.
4452
4453 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4454 and 16-bit fields in a compiled pattern, to simulate a pattern that
4455 was compiled on a host of opposite endianness.
4456
4457 21. The pcre-exec() function can now cope with patterns that were compiled on
4458 hosts of opposite endianness, with this restriction:
4459
4460 As for any compiled expression that is saved and used later, the tables
4461 pointer field cannot be preserved; the extra_data field in the arguments
4462 to pcre_exec() should be used to pass in a tables address if a value
4463 other than the default internal tables were used at compile time.
4464
4465 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4466 now diagnosed as an error. Previously, most of the time, a negative number
4467 would have been treated as zero, but if in addition "ovector" was passed as
4468 NULL, a crash could occur.
4469
4470 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4471 new versions from the libtool 1.5 distribution (the last one is a copy of
4472 a file called libtool.m4). This seems to have fixed the need to patch
4473 "configure" to support Darwin 1.3 (which I used to do). However, I still
4474 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4475 workstation).
4476
4477 24. Changed the PCRE licence to be the more standard "BSD" licence.
4478
4479
4480 Version 4.5 01-Dec-03
4481 ---------------------
4482
4483 1. There has been some re-arrangement of the code for the match() function so
4484 that it can be compiled in a version that does not call itself recursively.
4485 Instead, it keeps those local variables that need separate instances for
4486 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4487 needs to "recurse". Keeping track of where control must go is done by means
4488 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4489 hide most of the details from the main code, and operates only if
4490 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4491 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4492 operating.
4493
4494 To make it easier for callers to provide specially tailored get/free
4495 functions for this usage, two new functions, pcre_stack_malloc, and
4496 pcre_stack_free, are used. They are always called in strict stacking order,
4497 and the size of block requested is always the same.
4498
4499 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4500 PCRE has been compiled to use the stack or the heap for recursion. The
4501 -C option of pcretest uses this to show which version is compiled.
4502
4503 A new data escape \S, is added to pcretest; it causes the amounts of store
4504 obtained and freed by both kinds of malloc/free at match time to be added
4505 to the output.
4506
4507 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4508 what's available on my current Linux desktop machine.
4509
4510 3. When matching a UTF-8 string, the test for a valid string at the start has
4511 been extended. If start_offset is not zero, PCRE now checks that it points
4512 to a byte that is the start of a UTF-8 character. If not, it returns
4513 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4514 this is necessary because there may be backward assertions in the pattern.
4515 When matching the same subject several times, it may save resources to use
4516 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4517
4518 4. The code for checking the validity of UTF-8 strings has been tightened so
4519 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4520 containing "overlong sequences".
4521
4522 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4523 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4524 should have been "&", but it just so happened that all the cases this let
4525 through by mistake were picked up later in the function.
4526
4527 6. I had used a variable called "isblank" - this is a C99 function, causing
4528 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4529
4530 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4531 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4532 so the version is shown; (c) stop "make test" echoing "./RunTest".
4533
4534 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4535
4536 9. The prototype for memmove() for systems that don't have it was using
4537 size_t, but the inclusion of the header that defines size_t was later. I've
4538 moved the #includes for the C headers earlier to avoid this.
4539
4540 10. Added some adjustments to the code to make it easier to compiler on certain
4541 special systems:
4542
4543 (a) Some "const" qualifiers were missing.
4544 (b) Added the macro EXPORT before all exported functions; by default this
4545 is defined to be empty.
4546 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4547 that it reads its output file name as an argument instead of writing
4548 to the standard output and assuming this can be redirected.
4549
4550 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4551 class containing characters with values greater than 255, PCRE compilation
4552 went into a loop.
4553
4554 12. A recursive reference to a subpattern that was within another subpattern
4555 that had a minimum quantifier of zero caused PCRE to crash. For example,
4556 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4557 recursion. If the recursively-called subpattern itself had a zero repeat,
4558 that was OK.
4559
4560 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4561 buffer into which it was copied (for escape processing) was still set at
4562 1024, so long lines caused crashes.
4563
4564 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4565 "internal error: code overflow...". This applied to any character class
4566 that was followed by a possessive quantifier.
4567
4568 15. Modified the Makefile to add libpcre.la as a prerequisite for
4569 libpcreposix.la because I was told this is needed for a parallel build to
4570 work.
4571
4572 16. If a pattern that contained .* following optional items at the start was
4573 studied, the wrong optimizing data was generated, leading to matching
4574 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4575 matching string must start with a or b or c. The correct conclusion for
4576 this pattern is that a match can start with any character.
4577
4578
4579 Version 4.4 13-Aug-03
4580 ---------------------
4581
4582 1. In UTF-8 mode, a character class containing characters with values between
4583 127 and 255 was not handled correctly if the compiled pattern was studied.
4584 In fixing this, I have also improved the studying algorithm for such
4585 classes (slightly).
4586
4587 2. Three internal functions had redundant arguments passed to them. Removal
4588 might give a very teeny performance improvement.
4589
4590 3. Documentation bug: the value of the capture_top field in a callout is *one
4591 more than* the number of the hightest numbered captured substring.
4592
4593 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4594 in incorrectly linking with a previously installed version. They now link
4595 explicitly with libpcre.la.
4596
4597 5. configure.in no longer needs to recognize Cygwin specially.
4598
4599 6. A problem in pcre.in for Windows platforms is fixed.
4600
4601 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4602 pcretest, it used to include the size of the study block as part of its
4603 output. Unfortunately, the structure contains a field that has a different
4604 size on different hardware architectures. This meant that the tests that
4605 showed this size failed. As the block is currently always of a fixed size,
4606 this information isn't actually particularly useful in pcretest output, so
4607 I have just removed it.
4608
4609 8. Three pre-processor statements accidentally did not start in column 1.
4610 Sadly, there are *still* compilers around that complain, even though
4611 standard C has not required this for well over a decade. Sigh.
4612
4613 9. In pcretest, the code for checking callouts passed small integers in the
4614 callout_data field, which is a void * field. However, some picky compilers
4615 complained about the casts involved for this on 64-bit systems. Now
4616 pcretest passes the address of the small integer instead, which should get
4617 rid of the warnings.
4618
4619 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4620 both compile and run time, and gives an error if an invalid UTF-8 sequence
4621 is found. There is a option for disabling this check in cases where the
4622 string is known to be correct and/or the maximum performance is wanted.
4623
4624 11. In response to a bug report, I changed one line in Makefile.in from
4625
4626 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4627 to
4628 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4629
4630 to look similar to other lines, but I have no way of telling whether this
4631 is the right thing to do, as I do not use Windows. No doubt I'll get told
4632 if it's wrong...
4633
4634
4635 Version 4.3 21-May-03
4636 ---------------------
4637
4638 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4639 Makefile.
4640
4641 2. Some refactoring to improve the quality of the code:
4642
4643 (i) The utf8_table... variables are now declared "const".
4644
4645 (ii) The code for \cx, which used the "case flipping" table to upper case
4646 lower case letters, now just substracts 32. This is ASCII-specific,
4647 but the whole concept of \cx is ASCII-specific, so it seems
4648 reasonable.
4649
4650 (iii) PCRE was using its character types table to recognize decimal and
4651 hexadecimal digits in the pattern. This is silly, because it handles
4652 only 0-9, a-f, and A-F, but the character types table is locale-
4653 specific, which means strange things might happen. A private
4654 table is now used for this - though it costs 256 bytes, a table is
4655 much faster than multiple explicit tests. Of course, the standard
4656 character types table is still used for matching digits in subject
4657 strings against \d.
4658
4659 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4660 ending in _t are). So I've renamed it as ESC_tee.
4661
4662 3. The first argument for regexec() in the POSIX wrapper should have been
4663 defined as "const".
4664
4665 4. Changed pcretest to use malloc() for its buffers so that they can be
4666 Electric Fenced for debugging.
4667
4668 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4669 to read one or more bytes before the start of the subject string. Often this
4670 had no effect on PCRE's behaviour, but in some circumstances it could
4671 provoke a segmentation fault.
4672
4673 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4674 to try to read one or more bytes before the start of the subject string.
4675
4676 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4677 UTF-8 support could misbehave in various ways if the subject string
4678 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4679 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4680 back over UTF-8 characters.)
4681
4682
4683 Version 4.2 14-Apr-03
4684 ---------------------
4685
4686 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4687
4688 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4689 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4690 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4691 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4692 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4693 and BUILD_EXEEXT
4694 Note: automatic setting of the BUILD variables is not yet working
4695 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4696 compile-time but not at link-time
4697 [LINK]: use for linking executables only
4698 make different versions for Windows and non-Windows
4699 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4700 libraries
4701 [LINK_FOR_BUILD]: new variable
4702 [OBJEXT]: use throughout
4703 [EXEEXT]: use throughout
4704 <winshared>: new target
4705 <wininstall>: new target
4706 <dftables.o>: use native compiler
4707 <dftables>: use native linker
4708 <install>: handle Windows platform correctly
4709 <clean>: ditto
4710 <check>: ditto
4711 copy DLL to top builddir before testing
4712
4713 As part of these changes, -no-undefined was removed again. This was reported
4714 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4715 in any case.
4716
4717 3. Some tidies to get rid of compiler warnings:
4718
4719 . In the match_data structure, match_limit was an unsigned long int, whereas
4720 match_call_count was an int. I've made them both unsigned long ints.
4721
4722 . In pcretest the fact that a const uschar * doesn't automatically cast to
4723 a void * provoked a warning.
4724
4725 . Turning on some more compiler warnings threw up some "shadow" variables
4726 and a few more missing casts.
4727
4728 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4729 option, a class that contained a single character with a value between 128
4730 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4731
4732 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4733 option, a class that contained several characters, but with at least one
4734 whose value was between 128 and 255 caused PCRE to crash.
4735
4736
4737 Version 4.1 12-Mar-03
4738 ---------------------
4739
4740 1. Compiling with gcc -pedantic found a couple of places where casts were
4741 needed, and a string in dftables.c that was longer than standard compilers are
4742 required to support.
4743
4744 2. Compiling with Sun's compiler found a few more places where the code could
4745 be tidied up in order to avoid warnings.
4746
4747 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4748 first of these names is deprecated in the latest Autoconf in favour of the name
4749 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4750 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4751 analogy I have changed it to CFLAGS_FOR_BUILD.
4752
4753 4. Added -no-undefined to the linking command in the Makefile, because this is
4754 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4755 linking step for the pcreposix library.
4756
4757 5. PCRE was failing to diagnose the case of two named groups with the same
4758 name.
4759
4760 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4761 literal character that is needed in the subject for a match, and scans along to
4762 ensure that it is present before embarking on the full matching process. This
4763 saves time in cases of nested unlimited repeats that are never going to match.
4764 Problem: the scan can take a lot of time if the subject is very long (e.g.
4765 megabytes), thus penalizing straightforward matches. It is now done only if the
4766 amount of subject to be scanned is less than 1000 bytes.
4767
4768 7. A lesser problem with the same optimization is that it was recording the
4769 first character of an anchored pattern as "needed", thus provoking a search
4770 right along the subject, even when the first match of the pattern was going to
4771 fail. The "needed" character is now not set for anchored patterns, unless it
4772 follows something in the pattern that is of non-fixed length. Thus, it still
4773 fulfils its original purpose of finding quick non-matches in cases of nested
4774 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4775
4776
4777 Version 4.0 17-Feb-03
4778 ---------------------
4779
4780 1. If a comment in an extended regex that started immediately after a meta-item
4781 extended to the end of string, PCRE compiled incorrect data. This could lead to
4782 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4783
4784 2. Moved to autoconf 2.53 and libtool 1.4.2.
4785
4786 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4787 the special perltest8 script is no longer needed - all the tests can be run
4788 from a single perltest script.
4789
4790 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4791 by \s. It has now been removed in PCRE. This means it isn't recognized as
4792 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4793 class [:space:] *does* include VT, thereby creating a mess.
4794
4795 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4796 space and tab.
4797
4798 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4799 its new features into the main test script, reducing the number of scripts.
4800
4801 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4802 were backward compatible, and made the (?i) apply to the whole pattern, as if
4803 /i were given. Now it behaves more logically, and applies the option setting
4804 only to what follows. PCRE has been changed to follow suit. However, if it
4805 finds options settings right at the start of the pattern, it extracts them into
4806 the global options, as before. Thus, they show up in the info data.
4807
4808 8. Added support for the \Q...\E escape sequence. Characters in between are
4809 treated as literals. This is slightly different from Perl in that $ and @ are
4810 also handled as literals inside the quotes. In Perl, they will cause variable
4811 interpolation. Note the following examples:
4812
4813 Pattern PCRE matches Perl matches
4814
4815 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4816 \Qabc\$xyz\E abc\$xyz abc\$xyz
4817 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4818
4819 For compatibility with Perl, \Q...\E sequences are recognized inside character
4820 classes as well as outside them.
4821
4822 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4823 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4824 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4825 signed/unsigned warnings.
4826
4827 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4828 option for pcretest, so I've replaced it by a simple function that does just
4829 that job.
4830
4831 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4832 "pcregrep -".
4833
4834 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4835 Java package. This provides some syntactic sugar for simple cases of what my
4836 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4837 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4838 item, you can use this simplified notation. Note that only makes sense with
4839 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4840 greediness, whatever the setting of the PCRE_UNGREEDY option.
4841
4842 13. A change of greediness default within a pattern was not taking effect at
4843 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4844 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4845 was abstracted outside.
4846
4847 14. PCRE now supports the \G assertion. It is true when the current matching
4848 position is at the start point of the match. This differs from \A when the
4849 starting offset is non-zero. Used with the /g option of pcretest (or similar
4850 code), it works in the same way as it does for Perl's /g option. If all
4851 alternatives of a regex begin with \G, the expression is anchored to the start
4852 match position, and the "anchored" flag is set in the compiled expression.
4853
4854 15. Some bugs concerning the handling of certain option changes within patterns
4855 have been fixed. These applied to options other than (?ims). For example,
4856 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4857 been the other way round. Some of this was related to change 7 above.
4858
4859 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4860 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4861 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4862 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4863
4864 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4865 mode. Unlike ".", it always matches newline, whatever the setting of
4866 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4867 assertions. Perl allows it, but it doesn't (in general) work because it can't
4868 calculate the length of the lookbehind. At least, that's the case for Perl
4869 5.8.0 - I've been told they are going to document that it doesn't work in
4870 future.
4871
4872 18. Added an error diagnosis for escapes that PCRE does not support: these are
4873 \L, \l, \N, \P, \p, \U, \u, and \X.
4874
4875 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4876 reading past the end of the pattern in cases such as /[abcd/.
4877
4878 20. PCRE was getting more memory than necessary for patterns with classes that
4879 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4880
4881 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4882 compiling PCRE for use with Virtual Pascal.
4883
4884 22. Small fix to the Makefile to make it work properly if the build is done
4885 outside the source tree.
4886
4887 23. Added a new extension: a condition to go with recursion. If a conditional
4888 subpattern starts with (?(R) the "true" branch is used if recursion has
4889 happened, whereas the "false" branch is used only at the top level.
4890
4891 24. When there was a very long string of literal characters (over 255 bytes
4892 without UTF support, over 250 bytes with UTF support), the computation of how
4893 much memory was required could be incorrect, leading to segfaults or other
4894 strange effects.
4895
4896 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4897 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4898 there was a subsequent back reference to those brackets. This meant that, for
4899 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4900 possible to check for precisely this case. All we can do is abandon the
4901 optimization if .* occurs inside capturing brackets when there are any back
4902 references whatsoever. (See below for a better fix that came later.)
4903
4904 26. The handling of the optimization for finding the first character of a
4905 non-anchored pattern, and for finding a character that is required later in the
4906 match were failing in some cases. This didn't break the matching; it just
4907 failed to optimize when it could. The way this is done has been re-implemented.
4908
4909 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4910
4911 28. Added a new feature that provides some of the functionality that Perl
4912 provides with (?{...}). The facility is termed a "callout". The way it is done
4913 in PCRE is for the caller to provide an optional function, by setting
4914 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4915 global variable. By default it is unset, which disables all calling out. To get
4916 the function called, the regex must include (?C) at appropriate points. This
4917 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4918 This provides a means of identifying different callout points. When PCRE
4919 reaches such a point in the regex, if pcre_callout has been set, the external
4920 function is called. It is provided with data in a structure called
4921 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4922 matching continues; if it returns a non-zero value, the match at the current
4923 point fails. However, backtracking will occur if possible. [This was changed
4924 later and other features added - see item 49 below.]
4925
4926 29. pcretest is upgraded to test the callout functionality. It provides a
4927 callout function that displays information. By default, it shows the start of
4928 the match and the current position in the text. There are some new data escapes
4929 to vary what happens:
4930
4931 \C+ in addition, show current contents of captured substrings
4932 \C- do not supply a callout function
4933 \C!n return 1 when callout number n is reached
4934 \C!n!m return 1 when callout number n is reached for the mth time
4935
4936 30. If pcregrep was called with the -l option and just a single file name, it
4937 output "<stdin>" if a match was found, instead of the file name.
4938
4939 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4940 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4941 pcre_exec(). This saves a malloc/free per call. The default value of
4942 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4943 when configuring.
4944
4945 32. The default maximum size of a compiled pattern is 64K. There have been a
4946 few cases of people hitting this limit. The code now uses macros to handle the
4947 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4948 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4949 configuring. Tests 2 and 5 work only with 2-byte links because they output
4950 debugging information about compiled patterns.
4951
4952 33. Internal code re-arrangements:
4953
4954 (a) Moved the debugging function for printing out a compiled regex into
4955 its own source file (printint.c) and used #include to pull it into
4956 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4957 separate copies.
4958
4959 (b) Defined the list of op-code names for debugging as a macro in
4960 internal.h so that it is next to the definition of the opcodes.
4961
4962 (c) Defined a table of op-code lengths for simpler skipping along compiled
4963 code. This is again a macro in internal.h so that it is next to the
4964 definition of the opcodes.
4965
4966 34. Added support for recursive calls to individual subpatterns, along the
4967 lines of Robin Houston's patch (but implemented somewhat differently).
4968
4969 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4970 allow it to read and process whole directories in Win32. This code was
4971 contributed by Lionel Fourquaux; it has not been tested by me.
4972
4973 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4974 used to name a group. Names consist of alphanumerics and underscores, and must
4975 be unique. Back references use the syntax (?P=name) and recursive calls use
4976 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4977 numbers. The function pcre_fullinfo() can be used after compilation to extract
4978 a name/number map. There are three relevant calls:
4979
4980 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4981 PCRE_INFO_NAMECOUNT yields the number of entries
4982 PCRE_INFO_NAMETABLE yields a pointer to the map.
4983
4984 The map is a vector of fixed-size entries. The size of each entry depends on
4985 the length of the longest name used. The first two bytes of each entry are the
4986 group number, most significant byte first. There follows the corresponding
4987 name, zero terminated. The names are in alphabetical order.
4988
4989 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4990 case instead of 255. Making it the same both with and without UTF-8 support
4991 means that the same test output works with both.
4992
4993 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4994 calling malloc() with a zero argument.
4995
4996 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4997 optimization. I've improved things by keeping a bitmap of backreferences with
4998 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4999 fact referenced, the optimization can be applied. It is unlikely that a
5000 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
5001 the match to follow \n) will appear inside brackets with a number greater than
5002 31, but if it does, any back reference > 31 suppresses the optimization.
5003
5004 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
5005 of disabling numbered capturing parentheses. Any opening parenthesis that is
5006 not followed by ? behaves as if it were followed by ?: but named parentheses
5007 can still be used for capturing (and they will acquire numbers in the usual
5008 way).
5009
5010 41. Redesigned the return codes from the match() function into yes/no/error so
5011 that errors can be passed back from deep inside the nested calls. A malloc
5012 failure while inside a recursive subpattern call now causes the
5013 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
5014
5015 42. It is now possible to set a limit on the number of times the match()
5016 function is called in a call to pcre_exec(). This facility makes it possible to
5017 limit the amount of recursion and backtracking, though not in a directly
5018 obvious way, because the match() function is used in a number of different
5019 circumstances. The count starts from zero for each position in the subject
5020 string (for non-anchored patterns). The default limit is, for compatibility, a
5021 large number, namely 10 000 000. You can change this in two ways:
5022
5023 (a) When configuring PCRE before making, you can use --with-match-limit=n
5024 to set a default value for the compiled library.
5025
5026 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5027 a different value is set. See 45 below.
5028
5029 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5030
5031 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5032 of things that can be changed at compile time. The first argument specifies
5033 what is wanted and the second points to where the information is to be placed.
5034 The current list of available information is:
5035
5036 PCRE_CONFIG_UTF8
5037
5038 The output is an integer that is set to one if UTF-8 support is available;
5039 otherwise it is set to zero.
5040
5041 PCRE_CONFIG_NEWLINE
5042
5043 The output is an integer that it set to the value of the code that is used for
5044 newline. It is either LF (10) or CR (13).
5045
5046 PCRE_CONFIG_LINK_SIZE
5047
5048 The output is an integer that contains the number of bytes used for internal
5049 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5050
5051 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
5052
5053 The output is an integer that contains the threshold above which the POSIX
5054 interface uses malloc() for output vectors. See item 31 above.
5055
5056 PCRE_CONFIG_MATCH_LIMIT
5057
5058 The output is an unsigned integer that contains the default limit of the number
5059 of match() calls in a pcre_exec() execution. See 42 above.
5060
5061 44. pcretest has been upgraded by the addition of the -C option. This causes it
5062 to extract all the available output from the new pcre_config() function, and to
5063 output it. The program then exits immediately.
5064
5065 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5066 order to support additional features. One way would have been to define
5067 pcre_exec2() (for example) with extra arguments, but this would not have been
5068 extensible, and would also have required all calls to the original function to
5069 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5070 is used for passing in "extra" data from pcre_study().
5071
5072 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5073 contains the following fields:
5074
5075 flags a bitmap indicating which of the following fields are set
5076 study_data opaque data from pcre_study()
5077 match_limit a way of specifying a limit on match() calls for a specific
5078 call to pcre_exec()
5079 callout_data data for callouts (see 49 below)
5080
5081 The flag bits are also defined in pcre.h, and are
5082
5083 PCRE_EXTRA_STUDY_DATA
5084 PCRE_EXTRA_MATCH_LIMIT
5085 PCRE_EXTRA_CALLOUT_DATA
5086
5087 The pcre_study() function now returns one of these new pcre_extra blocks, with
5088 the actual study data pointed to by the study_data field, and the
5089 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5090 before. That is, this change is entirely upwards-compatible and requires no
5091 change to existing code.
5092
5093 If you want to pass in additional data to pcre_exec(), you can either place it
5094 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5095 block.
5096
5097 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5098 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5099 times with different match limits, until it finds the minimum value needed for
5100 pcre_exec() to complete. The value is then output. This can be instructive; for
5101 most simple matches the number is quite small, but for pathological cases it
5102 gets very large very quickly.
5103
5104 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5105 returns the size of the data block pointed to by the study_data field in a
5106 pcre_extra block, that is, the value that was passed as the argument to
5107 pcre_malloc() when PCRE was getting memory in which to place the information
5108 created by pcre_study(). The fourth argument should point to a size_t variable.
5109 pcretest has been extended so that this information is shown after a successful
5110 pcre_study() call when information about the compiled regex is being displayed.
5111
5112 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5113 because what follows is always an absolute path. (Later: it turns out that this
5114 is more than cosmetic for MinGW, because it doesn't like empty path
5115 components.)
5116
5117 49. Some changes have been made to the callout feature (see 28 above):
5118
5119 (i) A callout function now has three choices for what it returns:
5120
5121 0 => success, carry on matching
5122 > 0 => failure at this point, but backtrack if possible
5123 < 0 => serious error, return this value from pcre_exec()
5124
5125 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5126 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5127 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5128 use by callout functions. It will never be used by PCRE itself.
5129
5130 (ii) The pcre_extra structure (see 45 above) has a void * field called
5131 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5132 pcre_callout_block structure has a field of the same name. The contents of
5133 the field passed in the pcre_extra structure are passed to the callout
5134 function in the corresponding field in the callout block. This makes it
5135 easier to use the same callout-containing regex from multiple threads. For
5136 testing, the pcretest program has a new data escape
5137
5138 \C*n pass the number n (may be negative) as callout_data
5139
5140 If the callout function in pcretest receives a non-zero value as
5141 callout_data, it returns that value.
5142
5143 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5144 there were some redundant $(CFLAGS) in commands that are now specified as
5145 $(LINK), which already includes $(CFLAGS).
5146
5147 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5148 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5149 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5150 one-byte characters throughout. Note that case-insensitive matching applies
5151 only to characters whose values are less than 256. PCRE doesn't support the
5152 notion of cases for higher-valued characters.
5153
5154 (i) A character class whose characters are all within 0-255 is handled as
5155 a bit map, and the map is inverted for negative classes. Previously, a
5156 character > 255 always failed to match such a class; however it should
5157 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5158
5159 (ii) A negated character class with a single character < 255 is coded as
5160 "not this character" (OP_NOT). This wasn't working properly when the test
5161 character was multibyte, either singly or repeated.
5162
5163 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5164 mode, for example: \x{100}{2,3}.
5165
5166 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5167 singly or repeated) now correctly test multibyte characters. However,
5168 PCRE doesn't recognize any characters with values greater than 255 as
5169 digits, spaces, or word characters. Such characters always match \D, \S,
5170 and \W, and never match \d, \s, or \w.
5171
5172 (v) Classes may now contain characters and character ranges with values
5173 greater than 255. For example: [ab\x{100}-\x{400}].
5174
5175 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5176 PCRE in UTF-8 mode.
5177
5178 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5179 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5180 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5181 value.)
5182
5183 53. The single man page has become too large. I have therefore split it up into
5184 a number of separate man pages. These also give rise to individual HTML pages;
5185 these are now put in a separate directory, and there is an index.html page that
5186 lists them all. Some hyperlinking between the pages has been installed.
5187
5188 54. Added convenience functions for handling named capturing parentheses.
5189
5190 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5191 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5192 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5193 are faulted.
5194
5195 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5196 calling configure. These values are used when compiling the dftables.c program
5197 which is run to generate the source of the default character tables. They
5198 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5199 you will need to set these values.
5200
5201 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5202
5203
5204 Version 3.9 02-Jan-02
5205 ---------------------
5206
5207 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5208
5209 2. If --disable-static was given, the building process failed when trying to
5210 build pcretest and pcregrep. (For some reason it was using libtool to compile
5211 them, which is not right, as they aren't part of the library.)
5212
5213
5214 Version 3.8 18-Dec-01
5215 ---------------------
5216
5217 1. The experimental UTF-8 code was completely screwed up. It was packing the
5218 bytes in the wrong order. How dumb can you get?
5219
5220
5221 Version 3.7 29-Oct-01
5222 ---------------------
5223
5224 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5225 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5226 this didn't happen under Solaris 8, where I normally test things.
5227
5228 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5229 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5230 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5231 this fix an hour or so after the initial 3.7 release.)
5232
5233
5234 Version 3.6 23-Oct-01
5235 ---------------------
5236
5237 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5238 offsets passed as NULL with zero offset count.
5239
5240 2. The config.guess and config.sub files had not been updated when I moved to
5241 the latest autoconf.
5242
5243
5244 Version 3.5 15-Aug-01
5245 ---------------------
5246
5247 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5248 had been forgotten.
5249
5250 2. By using declared but undefined structures, we can avoid using "void"
5251 definitions in pcre.h while keeping the internal definitions of the structures
5252 private.
5253
5254 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5255 user point of view, this means that both static and shared libraries are built
5256 by default, but this can be individually controlled. More of the work of
5257 handling this static/shared cases is now inside libtool instead of PCRE's make
5258 file.
5259
5260 4. The pcretest utility is now installed along with pcregrep because it is
5261 useful for users (to test regexs) and by doing this, it automatically gets
5262 relinked by libtool. The documentation has been turned into a man page, so
5263 there are now .1, .txt, and .html versions in /doc.
5264
5265 5. Upgrades to pcregrep:
5266 (i) Added long-form option names like gnu grep.
5267 (ii) Added --help to list all options with an explanatory phrase.
5268 (iii) Added -r, --recursive to recurse into sub-directories.
5269 (iv) Added -f, --file to read patterns from a file.
5270
5271 6. pcre_exec() was referring to its "code" argument before testing that
5272 argument for NULL (and giving an error if it was NULL).
5273
5274 7. Upgraded Makefile.in to allow for compiling in a different directory from
5275 the source directory.
5276
5277 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5278 options bits, the pointer it was passed was to an int instead of to an unsigned
5279 long int. This mattered only on 64-bit systems.
5280
5281 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5282 generated) instead of pcre.in, which it its source. Also made the same change
5283 in several of the .c files.
5284
5285 10. A new release of gcc defines printf() as a macro, which broke pcretest
5286 because it had an ifdef in the middle of a string argument for printf(). Fixed
5287 by using separate calls to printf().
5288
5289 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5290 script, to force use of CR or LF instead of \n in the source. On non-Unix
5291 systems, the value can be set in config.h.
5292
5293 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5294 absolute limit. Changed the text of the error message to make this clear, and
5295 likewise updated the man page.
5296
5297 13. The limit of 99 on the number of capturing subpatterns has been removed.
5298 The new limit is 65535, which I hope will not be a "real" limit.
5299
5300
5301 Version 3.4 22-Aug-00
5302 ---------------------
5303
5304 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5305
5306 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5307
5308
5309 Version 3.3 01-Aug-00
5310 ---------------------
5311
5312 1. If an octal character was given, but the value was greater than \377, it
5313 was not getting masked to the least significant bits, as documented. This could
5314 lead to crashes in some systems.
5315
5316 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5317 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5318
5319 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5320 These just pass their arguments on to (pcre_free)(), but they are provided
5321 because some uses of PCRE bind it to non-C systems that can call its functions,
5322 but cannot call free() or pcre_free() directly.
5323
5324 4. Add "make test" as a synonym for "make check". Corrected some comments in
5325 the Makefile.
5326
5327 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5328 Makefile.
5329
5330 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5331 command called pgrep for grepping around the active processes.
5332
5333 7. Added the beginnings of support for UTF-8 character strings.
5334
5335 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5336 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5337 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5338 out for the ar command.)
5339
5340
5341 Version 3.2 12-May-00
5342 ---------------------
5343
5344 This is purely a bug fixing release.
5345
5346 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5347 of ZA. This was just one example of several cases that could provoke this bug,
5348 which was introduced by change 9 of version 2.00. The code for breaking
5349 infinite loops after an iteration that matches an empty string was't working
5350 correctly.
5351
5352 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5353 when matched against abbab (for example). After matching an empty string, it
5354 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5355 caused it to match further down the string than it should.
5356
5357 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5358 was there because it doesn't seem to be needed, and it causes trouble on some
5359 systems, as it is not a Standard C header. It has been removed.
5360
5361 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5362 were reported on the Macintosh. The changes were from
5363
5364 while ((c = *(++ptr)) != 0 && c != '\n');
5365 to
5366 while ((c = *(++ptr)) != 0 && c != '\n') ;
5367
5368 Totally extraordinary, but if that's what it takes...
5369
5370 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5371 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5372 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5373 assumes the way PCRE uses memmove() (always moving upwards).
5374
5375 6. PCRE is being used in one environment where strchr() is not available. There
5376 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5377 faster code anyway.
5378
5379
5380 Version 3.1 09-Feb-00
5381 ---------------------
5382
5383 The only change in this release is the fixing of some bugs in Makefile.in for
5384 the "install" target:
5385
5386 (1) It was failing to install pcreposix.h.
5387
5388 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5389
5390
5391 Version 3.0 01-Feb-00
5392 ---------------------
5393
5394 1. Add support for the /+ modifier to perltest (to output $` like it does in
5395 pcretest).
5396
5397 2. Add support for the /g modifier to perltest.
5398
5399 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5400 matches null strings.
5401
5402 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5403 pattern. Perl treats empty patterns specially - it reuses the most recent
5404 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5405 effect.
5406
5407 5. The POSIX interface was broken in that it was just handing over the POSIX
5408 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5409 required a bigger vector, with some working space on the end. This means that
5410 the POSIX wrapper now has to get and free some memory, and copy the results.
5411
5412 6. Added some simple autoconf support, placing the test data and the
5413 documentation in separate directories, re-organizing some of the
5414 information files, and making it build pcre-config (a GNU standard). Also added
5415 libtool support for building PCRE as a shared library, which is now the
5416 default.
5417
5418 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5419 09 are not valid octal constants. Single digits will be used for minor values
5420 less than 10.
5421
5422 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5423 existing programs that set these in the POSIX interface can use PCRE without
5424 modification.
5425
5426 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5427 return all that pcre_info() returns, plus additional data. The pcre_info()
5428 function is retained for compatibility, but is considered to be obsolete.
5429
5430 10. Added experimental recursion feature (?R) to handle one common case that
5431 Perl 5.6 will be able to do with (?p{...}).
5432
5433 11. Added support for POSIX character classes like [:alpha:], which Perl is
5434 adopting.
5435
5436
5437 Version 2.08 31-Aug-99
5438 ----------------------
5439
5440 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5441 trying to match at the startoffset position, but instead was moving forward to
5442 the next newline as if a previous match had failed.
5443
5444 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5445 and could get into a loop if a null string was matched other than at the start
5446 of the subject.
5447
5448 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5449 be distinguished at compile time, and for completeness also added PCRE_DATE.
5450
5451 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5452 in GnuWin32 environments.
5453
5454
5455 Version 2.07 29-Jul-99
5456 ----------------------
5457
5458 1. The documentation is now supplied in plain text form and HTML as well as in
5459 the form of man page sources.
5460
5461 2. C++ compilers don't like assigning (void *) values to other pointer types.
5462 In particular this affects malloc(). Although there is no problem in Standard
5463 C, I've put in casts to keep C++ compilers happy.
5464
5465 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5466 should be (const char *).
5467
5468 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5469 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5470 However, I haven't made this a standard facility. The documentation doesn't
5471 mention it, and the Makefile doesn't support it.
5472
5473 5. The Makefile now contains an "install" target, with editable destinations at
5474 the top of the file. The pcretest program is not installed.
5475
5476 6. pgrep -V now gives the PCRE version number and date.
5477
5478 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5479 causing the entire string to be ignored, instead of just the last character.
5480
5481 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5482 non-matching string, it can take a very, very long time, even for strings of
5483 quite modest length, because of the nested recursion. PCRE now does better in
5484 some of these cases. It does this by remembering the last required literal
5485 character in the pattern, and pre-searching the subject to ensure it is present
5486 before running the real match. In other words, it applies a heuristic to detect
5487 some types of certain failure quickly, and in the above example, if presented
5488 with a string that has no trailing " it gives "no match" very quickly.
5489
5490 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5491 other alternatives are tried instead.
5492
5493
5494 Version 2.06 09-Jun-99
5495 ----------------------
5496
5497 1. Change pcretest's output for amount of store used to show just the code
5498 space, because the remainder (the data block) varies in size between 32-bit and
5499 64-bit systems.
5500
5501 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5502 start matching at. This allows lookbehinds to work when searching for multiple
5503 occurrences in a string.
5504
5505 3. Added additional options to pcretest for testing multiple occurrences:
5506
5507 /+ outputs the rest of the string that follows a match
5508 /g loops for multiple occurrences, using the new startoffset argument
5509 /G loops for multiple occurrences by passing an incremented pointer
5510
5511 4. PCRE wasn't doing the "first character" optimization for patterns starting
5512 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5513 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5514 the letter 'x'. On long subject strings, this gives a significant speed-up.
5515
5516
5517 Version 2.05 21-Apr-99
5518 ----------------------
5519
5520 1. Changed the type of magic_number from int to long int so that it works
5521 properly on 16-bit systems.
5522
5523 2. Fixed a bug which caused patterns starting with .* not to work correctly
5524 when the subject string contained newline characters. PCRE was assuming
5525 anchoring for such patterns in all cases, which is not correct because .* will
5526 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5527 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5528 must be retried after every newline in the subject.
5529
5530
5531 Version 2.04 18-Feb-99
5532 ----------------------
5533
5534 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5535 computation of the store needed to hold the pattern was incorrect (too large).
5536 If such patterns were nested a few deep, this could multiply and become a real
5537 problem.
5538
5539 2. Added /M option to pcretest to show the memory requirement of a specific
5540 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5541
5542 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5543 compiled in such a way that the backtracking after subsequent failure was
5544 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5545 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5546
5547
5548 Version 2.03 02-Feb-99
5549 ----------------------
5550
5551 1. Fixed typo and small mistake in man page.
5552
5553 2. Added 4th condition (GPL supersedes if conflict) and created separate
5554 LICENCE file containing the conditions.
5555
5556 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5557 Perl, that is the internal \ allows the delimiter to be included in the
5558 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5559 the final delimiter, add \ to the end of the pattern (to test the error).
5560
5561 4. Added the convenience functions for extracting substrings after a successful
5562 match. Updated pcretest to make it able to test these functions.
5563
5564
5565 Version 2.02 14-Jan-99
5566 ----------------------
5567
5568 1. Initialized the working variables associated with each extraction so that
5569 their saving and restoring doesn't refer to uninitialized store.
5570
5571 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5572 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5573 fix the problem.
5574
5575 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5576 calls, and wasn't printing the correct value for compiling calls. Increased the
5577 default value of LOOPREPEAT, and the number of significant figures in the
5578 times.
5579
5580 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5581
5582 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5583 a building problem on Windows NT with a FAT file system.
5584
5585
5586 Version 2.01 21-Oct-98
5587 ----------------------
5588
5589 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5590 to character tables built by pcre_maketables() in the current locale. If NULL
5591 is passed, the default tables are used.
5592
5593
5594 Version 2.00 24-Sep-98
5595 ----------------------
5596
5597 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5598 it any more.
5599
5600 2. Allow quantification of (?>) groups, and make it work correctly.
5601
5602 3. The first character computation wasn't working for (?>) groups.
5603
5604 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5605 end of the subject) and add 5.005's \z, which really does match only at the
5606 very end of the subject.
5607
5608 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5609
5610 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5611 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5612 localized options. All options to pcre_study() were also removed.
5613
5614 7. Add other new features from 5.005:
5615
5616 $(?<= positive lookbehind
5617 $(?<! negative lookbehind
5618 (?imsx-imsx) added the unsetting capability
5619 such a setting is global if at outer level; local otherwise
5620 (?imsx-imsx:) non-capturing groups with option setting
5621 (?(cond)re|re) conditional pattern matching
5622
5623 A backreference to itself in a repeated group matches the previous
5624 captured string.
5625
5626 8. General tidying up of studying (both automatic and via "study")
5627 consequential on the addition of new assertions.
5628
5629 9. As in 5.005, unlimited repeated groups that could match an empty substring
5630 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5631 runtime if any iteration does actually match an empty substring.
5632
5633 10. Include the RunTest script in the distribution.
5634
5635 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5636 discrepancies, some of which were old and were also with respect to 5.004. They
5637 have now been fixed.
5638
5639
5640 Version 1.09 28-Apr-98
5641 ----------------------
5642
5643 1. A negated single character class followed by a quantifier with a minimum
5644 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5645 program crashes, or just wrong answers. This did not apply to negated classes
5646 containing more than one character, or to minima other than one.
5647
5648
5649 Version 1.08 27-Mar-98
5650 ----------------------
5651
5652 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5653
5654 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5655 latter must appear before anything that relies on it in the pattern.
5656
5657
5658 Version 1.07 16-Feb-98
5659 ----------------------
5660
5661 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5662 repeat of a potentially empty string).
5663
5664
5665 Version 1.06 23-Jan-98
5666 ----------------------
5667
5668 1. Added Markus Oberhumer's little patches for C++.
5669
5670 2. Literal strings longer than 255 characters were broken.
5671
5672
5673 Version 1.05 23-Dec-97
5674 ----------------------
5675
5676 1. Negated character classes containing more than one character were failing if
5677 PCRE_CASELESS was set at run time.
5678
5679
5680 Version 1.04 19-Dec-97
5681 ----------------------
5682
5683 1. Corrected the man page, where some "const" qualifiers had been omitted.
5684
5685 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5686 input syntax.
5687
5688 3. Fixed memory leak which occurred when a regex with back references was
5689 matched with an offsets vector that wasn't big enough. The temporary memory
5690 that is used in this case wasn't being freed if the match failed.
5691
5692 4. Tidied pcretest to ensure it frees memory that it gets.
5693
5694 5. Temporary memory was being obtained in the case where the passed offsets
5695 vector was exactly big enough.
5696
5697 6. Corrected definition of offsetof() from change 5 below.
5698
5699 7. I had screwed up change 6 below and broken the rules for the use of
5700 setjmp(). Now fixed.
5701
5702
5703 Version 1.03 18-Dec-97
5704 ----------------------
5705
5706 1. A erroneous regex with a missing opening parenthesis was correctly
5707 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5708 on some systems.
5709
5710 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5711 it was reported that one broken compiler failed on the former because "code" is
5712 also an independent variable.
5713
5714 3. The erroneous regex a[]b caused an array overrun reference.
5715
5716 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5717 fail on data ending with that character. (It was going on too far, and checking
5718 the next character, typically a binary zero.) This was specific to the
5719 optimized code for single-character negative classes.
5720
5721 5. Added a contributed patch from the TIN world which does the following:
5722
5723 + Add an undef for memmove, in case the the system defines a macro for it.
5724
5725 + Add a definition of offsetof(), in case there isn't one. (I don't know
5726 the reason behind this - offsetof() is part of the ANSI standard - but
5727 it does no harm).
5728
5729 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5730 most of the places where whitespace preceded '#'. I have given up and
5731 allowed the remaining 2 cases to be at the margin.
5732
5733 + Rename some variables in pcre to eliminate shadowing. This seems very
5734 pedantic, but does no harm, of course.
5735
5736 6. Moved the call to setjmp() into its own function, to get rid of warnings
5737 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5738
5739 7. Constructs such as \d{8,} were compiling into the equivalent of
5740 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5741 outcome, but in this particular case used more store than had been allocated,
5742 which caused the bug to be discovered because it threw up an internal error.
5743
5744 8. The debugging code in both pcre and pcretest for outputting the compiled
5745 form of a regex was going wrong in the case of back references followed by
5746 curly-bracketed repeats.
5747
5748
5749 Version 1.02 12-Dec-97
5750 ----------------------
5751
5752 1. Typos in pcre.3 and comments in the source fixed.
5753
5754 2. Applied a contributed patch to get rid of places where it used to remove
5755 'const' from variables, and fixed some signed/unsigned and uninitialized
5756 variable warnings.
5757
5758 3. Added the "runtest" target to Makefile.
5759
5760 4. Set default compiler flag to -O2 rather than just -O.
5761
5762
5763 Version 1.01 19-Nov-97
5764 ----------------------
5765
5766 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5767 like /([ab]*)*/, that is, for classes with more than one character in them.
5768
5769 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5770 as /((?>a*))*/ (a PCRE_EXTRA facility).
5771
5772
5773 Version 1.00 18-Nov-97
5774 ----------------------
5775
5776 1. Added compile-time macros to support systems such as SunOS4 which don't have
5777 memmove() or strerror() but have other things that can be used instead.
5778
5779 2. Arranged that "make clean" removes the executables.
5780
5781
5782 Version 0.99 27-Oct-97
5783 ----------------------
5784
5785 1. Fixed bug in code for optimizing classes with only one character. It was
5786 initializing a 32-byte map regardless, which could cause it to run off the end
5787 of the memory it had got.
5788
5789 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5790
5791
5792 Version 0.98 22-Oct-97
5793 ----------------------
5794
5795 1. Fixed bug in code for handling temporary memory usage when there are more
5796 back references than supplied space in the ovector. This could cause segfaults.
5797
5798
5799 Version 0.97 21-Oct-97
5800 ----------------------
5801
5802 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5803
5804 2. Optimized negated single characters not to use a bit map.
5805
5806 3. Brought error texts together as macro definitions; clarified some of them;
5807 fixed one that was wrong - it said "range out of order" when it meant "invalid
5808 escape sequence".
5809
5810 4. Changed some char * arguments to const char *.
5811
5812 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5813
5814 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5815 pcretest.
5816
5817
5818 Version 0.96 16-Oct-97
5819 ----------------------
5820
5821 1. Added a simple "pgrep" utility to the distribution.
5822
5823 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5824 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5825 where "ddd" means "one or more decimal digits".
5826
5827 3. Fixed serious bug. If a pattern had a back reference, but the call to
5828 pcre_exec() didn't supply a large enough ovector to record the related
5829 identifying subpattern, the match always failed. PCRE now remembers the number
5830 of the largest back reference, and gets some temporary memory in which to save
5831 the offsets during matching if necessary, in order to ensure that
5832 backreferences always work.
5833
5834 4. Increased the compatibility with Perl in a number of ways:
5835
5836 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5837 to request this handling. The option can be set at compile or exec time.
5838
5839 (b) $ matches before a terminating newline by default; an option
5840 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5841 mode). The option can be set at compile or exec time.
5842
5843 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5844 the same as Perl's. If the decimal number it represents is less than 10
5845 or there aren't that many previous left capturing parentheses, an octal
5846 escape is read. Inside a character class, it's always an octal escape,
5847 even if it is a single digit.
5848
5849 (d) An escaped but undefined alphabetic character is taken as a literal,
5850 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5851 escapes.
5852
5853 (e) {0} is now permitted. (The previous item is removed from the compiled
5854 pattern).
5855
5856 5. Changed all the names of code files so that the basic parts are no longer
5857 than 10 characters, and abolished the teeny "globals.c" file.
5858
5859 6. Changed the handling of character classes; they are now done with a 32-byte
5860 bit map always.
5861
5862 7. Added the -d and /D options to pcretest to make it possible to look at the
5863 internals of compilation without having to recompile pcre.
5864
5865
5866 Version 0.95 23-Sep-97
5867 ----------------------
5868
5869 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5870 \x20 at the start of a run of normal characters. These were being treated as
5871 real characters, instead of the source characters being re-checked.
5872
5873
5874 Version 0.94 18-Sep-97
5875 ----------------------
5876
5877 1. The functions are now thread-safe, with the caveat that the global variables
5878 containing pointers to malloc() and free() or alternative functions are the
5879 same for all threads.
5880
5881 2. Get pcre_study() to generate a bitmap of initial characters for non-
5882 anchored patterns when this is possible, and use it if passed to pcre_exec().
5883
5884
5885 Version 0.93 15-Sep-97
5886 ----------------------
5887
5888 1. /(b)|(:+)/ was computing an incorrect first character.
5889
5890 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5891 but not actually doing anything yet.
5892
5893 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5894 as Perl does (e.g. [-az] or [az-]).
5895
5896 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5897 all possible positions.
5898
5899 5. Split up into different modules to avoid including unneeded functions in a
5900 compiled binary. However, compile and exec are still in one module. The "study"
5901 function is split off.
5902
5903 6. The character tables are now in a separate module whose source is generated
5904 by an auxiliary program - but can then be edited by hand if required. There are
5905 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5906 toupper() in the code.
5907
5908 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5909 make them global. Abolish the function for setting them, as the caller can now
5910 set them directly.
5911
5912
5913 Version 0.92 11-Sep-97
5914 ----------------------
5915
5916 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5917 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5918
5919 2. Caseless matching was not working in character classes if the characters in
5920 the pattern were in upper case.
5921
5922 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5923
5924 4. Make PCRE_ANCHORED public and accept as a compile option.
5925
5926 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5927 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5928 pass them.
5929
5930 6. Give an error if bad option bits passed at compile or run time.
5931
5932 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5933 pcretest to cause it to pass that flag.
5934
5935 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5936 options, and the first character, if set.
5937
5938 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5939
5940
5941 Version 0.91 10-Sep-97
5942 ----------------------
5943
5944 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5945 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5946
5947 2. PCRE was looping on encountering an indefinitely repeated back reference to
5948 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5949 Perl does - treats the match as successful.
5950
5951 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5