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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1614 - (show annotations)
Sun Nov 29 17:38:25 2015 UTC (3 years, 10 months ago) by ph10
File size: 286186 byte(s)
Error occurred while calculating annotation data.
Give error for regexec with pmatch=NULL and REG_STARTEND set.
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
25 Version 8.38 23-November-2015
26 -----------------------------
27
28 1. If a group that contained a recursive back reference also contained a
29 forward reference subroutine call followed by a non-forward-reference
30 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
31 compile correct code, leading to undefined behaviour or an internally
32 detected error. This bug was discovered by the LLVM fuzzer.
33
34 2. Quantification of certain items (e.g. atomic back references) could cause
35 incorrect code to be compiled when recursive forward references were
36 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
37 This bug was discovered by the LLVM fuzzer.
38
39 3. A repeated conditional group whose condition was a reference by name caused
40 a buffer overflow if there was more than one group with the given name.
41 This bug was discovered by the LLVM fuzzer.
42
43 4. A recursive back reference by name within a group that had the same name as
44 another group caused a buffer overflow. For example:
45 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
46
47 5. A forward reference by name to a group whose number is the same as the
48 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
49 a buffer overflow at compile time. This bug was discovered by the LLVM
50 fuzzer.
51
52 6. A lookbehind assertion within a set of mutually recursive subpatterns could
53 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
54
55 7. Another buffer overflow bug involved duplicate named groups with a
56 reference between their definition, with a group that reset capture
57 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
58 fixed by always allowing for more memory, even if not needed. (A proper fix
59 is implemented in PCRE2, but it involves more refactoring.)
60
61 8. There was no check for integer overflow in subroutine calls such as (?123).
62
63 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
64 being treated as a literal 'l' instead of causing an error.
65
66 10. There was a buffer overflow if pcre_exec() was called with an ovector of
67 size 1. This bug was found by american fuzzy lop.
68
69 11. If a non-capturing group containing a conditional group that could match
70 an empty string was repeated, it was not identified as matching an empty
71 string itself. For example: /^(?:(?(1)x|)+)+$()/.
72
73 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
74 \a and \e in test subject lines.
75
76 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
77 instead of the EBCDIC value.
78
79 14. The handling of \c in an EBCDIC environment has been revised so that it is
80 now compatible with the specification in Perl's perlebcdic page.
81
82 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
83 ASCII/Unicode. This has now been added to the list of characters that are
84 recognized as white space in EBCDIC.
85
86 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
87 error (correctly) when used outside a class, but did not give an error
88 within a class.
89
90 17. \h within a class was incorrectly compiled in EBCDIC environments.
91
92 18. A pattern with an unmatched closing parenthesis that contained a backward
93 assertion which itself contained a forward reference caused buffer
94 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
95
96 19. JIT should return with error when the compiled pattern requires more stack
97 space than the maximum.
98
99 20. A possessively repeated conditional group that could match an empty string,
100 for example, /(?(R))*+/, was incorrectly compiled.
101
102 21. Fix infinite recursion in the JIT compiler when certain patterns such as
103 /(?:|a|){100}x/ are analysed.
104
105 22. Some patterns with character classes involving [: and \\ were incorrectly
106 compiled and could cause reading from uninitialized memory or an incorrect
107 error diagnosis.
108
109 23. Pathological patterns containing many nested occurrences of [: caused
110 pcre_compile() to run for a very long time.
111
112 24. A conditional group with only one branch has an implicit empty alternative
113 branch and must therefore be treated as potentially matching an empty
114 string.
115
116 25. If (?R was followed by - or + incorrect behaviour happened instead of a
117 diagnostic.
118
119 26. Arrange to give up on finding the minimum matching length for overly
120 complex patterns.
121
122 27. Similar to (4) above: in a pattern with duplicated named groups and an
123 occurrence of (?| it is possible for an apparently non-recursive back
124 reference to become recursive if a later named group with the relevant
125 number is encountered. This could lead to a buffer overflow. Wen Guanxing
126 from Venustech ADLAB discovered this bug.
127
128 28. If pcregrep was given the -q option with -c or -l, or when handling a
129 binary file, it incorrectly wrote output to stdout.
130
131 29. The JIT compiler did not restore the control verb head in case of *THEN
132 control verbs. This issue was found by Karl Skomski with a custom LLVM
133 fuzzer.
134
135 30. Error messages for syntax errors following \g and \k were giving inaccurate
136 offsets in the pattern.
137
138 31. Added a check for integer overflow in conditions (?(<digits>) and
139 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
140 fuzzer.
141
142 32. Handling recursive references such as (?2) when the reference is to a group
143 later in the pattern uses code that is very hacked about and error-prone.
144 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
145 give an internal error if it is obvious that compiling has gone wrong.
146
147 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
148 This issue was found by Karl Skomski with a custom LLVM fuzzer.
149
150 34. The JIT compiler should restore the control chain for empty possessive
151 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
152
153 35. Match limit check added to JIT recursion. This issue was found by Karl
154 Skomski with a custom LLVM fuzzer.
155
156 36. Yet another case similar to 27 above has been circumvented by an
157 unconditional allocation of extra memory. This issue is fixed "properly" in
158 PCRE2 by refactoring the way references are handled. Wen Guanxing
159 from Venustech ADLAB discovered this bug.
160
161 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
162 with a custom LLVM fuzzer.
163
164 38. Fixed a corner case of range optimization in JIT.
165
166 39. An incorrect error "overran compiling workspace" was given if there were
167 exactly enough group forward references such that the last one extended
168 into the workspace safety margin. The next one would have expanded the
169 workspace. The test for overflow was not including the safety margin.
170
171 40. A match limit issue is fixed in JIT which was found by Karl Skomski
172 with a custom LLVM fuzzer.
173
174 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
175 work under Windows. (Why has it taken so long for anyone to notice?)
176
177 42. In a character class such as [\W\p{Any}] where both a negative-type escape
178 ("not a word character") and a property escape were present, the property
179 escape was being ignored.
180
181 43. Fix crash caused by very long (*MARK) or (*THEN) names.
182
183 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
184 by a single ASCII character in a class item, was incorrectly compiled in
185 UCP mode. The POSIX class got lost, but only if the single character
186 followed it.
187
188 45. [:punct:] in UCP mode was matching some characters in the range 128-255
189 that should not have been matched.
190
191 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
192 class, all characters with code points greater than 255 are in the class.
193 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
194 such as \w are turned into Unicode properties), wide characters were not
195 correctly handled, and could fail to match.
196
197
198 Version 8.37 28-April-2015
199 --------------------------
200
201 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
202 for those parentheses to be closed with whatever has been captured so far.
203 However, it was failing to mark any other groups between the hightest
204 capture so far and the currrent group as "unset". Thus, the ovector for
205 those groups contained whatever was previously there. An example is the
206 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
207
208 2. If an assertion condition was quantified with a minimum of zero (an odd
209 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
210
211 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
212 unrecognized modifier, a crash could occur.
213
214 4. An attempt to do global matching in pcretest with a zero-length ovector
215 caused a crash.
216
217 5. Fixed a memory leak during matching that could occur for a subpattern
218 subroutine call (recursive or otherwise) if the number of captured groups
219 that had to be saved was greater than ten.
220
221 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
222 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
223 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
224
225 7. A UTF pattern containing a "not" match of a non-ASCII character and a
226 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
227
228 8. When a pattern is compiled, it remembers the highest back reference so that
229 when matching, if the ovector is too small, extra memory can be obtained to
230 use instead. A conditional subpattern whose condition is a check on a
231 capture having happened, such as, for example in the pattern
232 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
233 setting the highest backreference number. This mattered only if pcre_exec()
234 was called with an ovector that was too small to hold the capture, and there
235 was no other kind of back reference (a situation which is probably quite
236 rare). The effect of the bug was that the condition was always treated as
237 FALSE when the capture could not be consulted, leading to a incorrect
238 behaviour by pcre_exec(). This bug has been fixed.
239
240 9. A reference to a duplicated named group (either a back reference or a test
241 for being set in a conditional) that occurred in a part of the pattern where
242 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
243 to be incorrectly calculated, leading to overwriting.
244
245 10. A mutually recursive set of back references such as (\2)(\1) caused a
246 segfault at study time (while trying to find the minimum matching length).
247 The infinite loop is now broken (with the minimum length unset, that is,
248 zero).
249
250 11. If an assertion that was used as a condition was quantified with a minimum
251 of zero, matching went wrong. In particular, if the whole group had
252 unlimited repetition and could match an empty string, a segfault was
253 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
254 assertions to be quantified, but not if they are being used as conditions,
255 so the above pattern is faulted by Perl. PCRE has now been changed so that
256 it also rejects such patterns.
257
258 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
259 failed to allow the zero-repeat case if pcre2_exec() was called with an
260 ovector too small to capture the group.
261
262 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
263 Red Hat Product Security:
264
265 (a) A crash if /K and /F were both set with the option to save the compiled
266 pattern.
267
268 (b) Another crash if the option to print captured substrings in a callout
269 was combined with setting a null ovector, for example \O\C+ as a subject
270 string.
271
272 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
273 forward reference repeated a large (but limited) number of times within a
274 repeated outer group that has a zero minimum quantifier, caused incorrect
275 code to be compiled, leading to the error "internal error:
276 previously-checked referenced subpattern not found" when an incorrect
277 memory address was read. This bug was reported as "heap overflow",
278 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
279 CVE-2015-2325.
280
281 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
282 call within a group that also contained a recursive back reference caused
283 incorrect code to be compiled. This bug was reported as "heap overflow",
284 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
285 number CVE-2015-2326.
286
287 24. Computing the size of the JIT read-only data in advance has been a source
288 of various issues, and new ones are still appear unfortunately. To fix
289 existing and future issues, size computation is eliminated from the code,
290 and replaced by on-demand memory allocation.
291
292 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
293 adjacent to the end of the range, and the range contained characters with
294 more than one other case, caused incorrect behaviour when compiled in UTF
295 mode. In that example, the range a-j was left out of the class.
296
297 26. Fix JIT compilation of conditional blocks, which assertion
298 is converted to (*FAIL). E.g: /(?(?!))/.
299
300 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
301 discovered by the LLVM fuzzer.
302
303 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
304 when this assertion was used as a condition, for example (?(?!)a|b). In
305 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
306 error about an unsupported item.
307
308 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
309 possessification code could take exponential time to complete. A recursion
310 depth limit of 1000 has been imposed to limit the resources used by this
311 optimization.
312
313 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
314 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
315 because \S ensures they are all in the class. The code for doing this was
316 interacting badly with the code for computing the amount of space needed to
317 compile the pattern, leading to a buffer overflow. This bug was discovered
318 by the LLVM fuzzer.
319
320 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
321 other kinds of group caused stack overflow at compile time. This bug was
322 discovered by the LLVM fuzzer.
323
324 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
325 between a subroutine call and its quantifier was incorrectly compiled,
326 leading to buffer overflow or other errors. This bug was discovered by the
327 LLVM fuzzer.
328
329 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
330 assertion after (?(. The code was failing to check the character after
331 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
332 was discovered by the LLVM fuzzer.
333
334 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
335 a fixed maximum following a group that contains a subroutine reference was
336 incorrectly compiled and could trigger buffer overflow. This bug was
337 discovered by the LLVM fuzzer.
338
339 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
340 caused a stack overflow instead of the diagnosis of a non-fixed length
341 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
342
343 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
344 (e.g. /(?<=\Ka)/) could make pcregrep loop.
345
346 37. There was a similar problem to 36 in pcretest for global matches.
347
348 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
349 and a subsequent item in the pattern caused a non-match, backtracking over
350 the repeated \X did not stop, but carried on past the start of the subject,
351 causing reference to random memory and/or a segfault. There were also some
352 other cases where backtracking after \C could crash. This set of bugs was
353 discovered by the LLVM fuzzer.
354
355 39. The function for finding the minimum length of a matching string could take
356 a very long time if mutual recursion was present many times in a pattern,
357 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
358 method has been implemented. This infelicity was discovered by the LLVM
359 fuzzer.
360
361 40. Static linking against the PCRE library using the pkg-config module was
362 failing on missing pthread symbols.
363
364
365 Version 8.36 26-September-2014
366 ------------------------------
367
368 1. Got rid of some compiler warnings in the C++ modules that were shown up by
369 -Wmissing-field-initializers and -Wunused-parameter.
370
371 2. The tests for quantifiers being too big (greater than 65535) were being
372 applied after reading the number, and stupidly assuming that integer
373 overflow would give a negative number. The tests are now applied as the
374 numbers are read.
375
376 3. Tidy code in pcre_exec.c where two branches that used to be different are
377 now the same.
378
379 4. The JIT compiler did not generate match limit checks for certain
380 bracketed expressions with quantifiers. This may lead to exponential
381 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
382 issue should be resolved now.
383
384 5. Fixed an issue, which occures when nested alternatives are optimized
385 with table jumps.
386
387 6. Inserted two casts and changed some ints to size_t in the light of some
388 reported 64-bit compiler warnings (Bugzilla 1477).
389
390 7. Fixed a bug concerned with zero-minimum possessive groups that could match
391 an empty string, which sometimes were behaving incorrectly in the
392 interpreter (though correctly in the JIT matcher). This pcretest input is
393 an example:
394
395 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
396 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
397
398 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
399 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
400 for an empty string was breaking the inner loop and carrying on at a lower
401 level, when possessive repeated groups should always return to a higher
402 level as they have no backtrack points in them. The empty string test now
403 occurs at the outer level.
404
405 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
406 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
407
408 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
409 doesn't).
410
411 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
412 to hold a pattern to be incorrectly computed (too small) when there were
413 named back references to duplicated names. This could cause "internal
414 error: code overflow" or "double free or corruption" or other memory
415 handling errors.
416
417 11. When named subpatterns had the same prefixes, back references could be
418 confused. For example, in this pattern:
419
420 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
421
422 the reference to 'Name' was incorrectly treated as a reference to a
423 duplicate name.
424
425 12. A pattern such as /^s?c/mi8 where the optional character has more than
426 one "other case" was incorrectly compiled such that it would only try to
427 match starting at "c".
428
429 13. When a pattern starting with \s was studied, VT was not included in the
430 list of possible starting characters; this should have been part of the
431 8.34/18 patch.
432
433 14. If a character class started [\Qx]... where x is any character, the class
434 was incorrectly terminated at the ].
435
436 15. If a pattern that started with a caseless match for a character with more
437 than one "other case" was studied, PCRE did not set up the starting code
438 unit bit map for the list of possible characters. Now it does. This is an
439 optimization improvement, not a bug fix.
440
441 16. The Unicode data tables have been updated to Unicode 7.0.0.
442
443 17. Fixed a number of memory leaks in pcregrep.
444
445 18. Avoid a compiler warning (from some compilers) for a function call with
446 a cast that removes "const" from an lvalue by using an intermediate
447 variable (to which the compiler does not object).
448
449 19. Incorrect code was compiled if a group that contained an internal recursive
450 back reference was optional (had quantifier with a minimum of zero). This
451 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
452 caused segmentation faults because of stack overflows at compile time.
453
454 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
455 group that is quantified with an indefinite repeat, caused a compile-time
456 loop which used up all the system stack and provoked a segmentation fault.
457 This was not the same bug as 19 above.
458
459 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
460 Patch by Mike Frysinger.
461
462
463 Version 8.35 04-April-2014
464 --------------------------
465
466 1. A new flag is set, when property checks are present in an XCLASS.
467 When this flag is not set, PCRE can perform certain optimizations
468 such as studying these XCLASS-es.
469
470 2. The auto-possessification of character sets were improved: a normal
471 and an extended character set can be compared now. Furthermore
472 the JIT compiler optimizes more character set checks.
473
474 3. Got rid of some compiler warnings for potentially uninitialized variables
475 that show up only when compiled with -O2.
476
477 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
478 of a match later then the end of the match. The pcretest program was not
479 handling the case sensibly - it was outputting from the start to the next
480 binary zero. It now reports this situation in a message, and outputs the
481 text from the end to the start.
482
483 5. Fast forward search is improved in JIT. Instead of the first three
484 characters, any three characters with fixed position can be searched.
485 Search order: first, last, middle.
486
487 6. Improve character range checks in JIT. Characters are read by an inprecise
488 function now, which returns with an unknown value if the character code is
489 above a certain threshold (e.g: 256). The only limitation is that the value
490 must be bigger than the threshold as well. This function is useful when
491 the characters above the threshold are handled in the same way.
492
493 7. The macros whose names start with RAWUCHAR are placeholders for a future
494 mode in which only the bottom 21 bits of 32-bit data items are used. To
495 make this more memorable for those maintaining the code, the names have
496 been changed to start with UCHAR21, and an extensive comment has been added
497 to their definition.
498
499 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
500 to the export list in Makefile.am (they were accidentally omitted from the
501 8.34 tarball).
502
503 9. The informational output from pcretest used the phrase "starting byte set"
504 which is inappropriate for the 16-bit and 32-bit libraries. As the output
505 for "first char" and "need char" really means "non-UTF-char", I've changed
506 "byte" to "char", and slightly reworded the output. The documentation about
507 these values has also been (I hope) clarified.
508
509 10. Another JIT related optimization: use table jumps for selecting the correct
510 backtracking path, when more than four alternatives are present inside a
511 bracket.
512
513 11. Empty match is not possible, when the minimum length is greater than zero,
514 and there is no \K in the pattern. JIT should avoid empty match checks in
515 such cases.
516
517 12. In a caseless character class with UCP support, when a character with more
518 than one alternative case was not the first character of a range, not all
519 the alternative cases were added to the class. For example, s and \x{17f}
520 are both alternative cases for S: the class [RST] was handled correctly,
521 but [R-T] was not.
522
523 13. The configure.ac file always checked for pthread support when JIT was
524 enabled. This is not used in Windows, so I have put this test inside a
525 check for the presence of windows.h (which was already tested for).
526
527 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
528 The algorithm provides a way to skip certain starting offsets, and usually
529 faster than linear prefix searches.
530
531 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
532 as for 'fr_FR' and 'french'. For some reason, however, it then used the
533 Windows-specific input and output files, which have 'french' screwed in.
534 So this could never have worked. One of the problems with locales is that
535 they aren't always the same. I have now updated RunTest so that it checks
536 the output of the locale test (test 3) against three different output
537 files, and it allows the test to pass if any one of them matches. With luck
538 this should make the test pass on some versions of Solaris where it was
539 failing. Because of the uncertainty, the script did not used to stop if
540 test 3 failed; it now does. If further versions of a French locale ever
541 come to light, they can now easily be added.
542
543 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
544 there was a message during ./configure, but it did not stop. This now
545 provokes an error. The invalid example in README has been corrected.
546 If a value less than the minimum is given, the minimum value has always
547 been used, but now a warning is given.
548
549 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
550 was a bug in the test system, which is now fixed. Also, the list of various
551 configurations that are tested for each release did not have one with both
552 16/32 bits and --enable-bar-anycrlf. It now does.
553
554 18. pcretest was missing "-C bsr" for displaying the \R default setting.
555
556 19. Little endian PowerPC systems are supported now by the JIT compiler.
557
558 20. The fast forward newline mechanism could enter to an infinite loop on
559 certain invalid UTF-8 input. Although we don't support these cases
560 this issue can be fixed by a performance optimization.
561
562 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
563 not take account if existing stack usage. There is now a new global
564 variable called pcre_stack_guard that can be set to point to an external
565 function to check stack availability. It is called at the start of
566 processing every parenthesized group.
567
568 22. A typo in the code meant that in ungreedy mode the max/min qualifier
569 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
570 not match "ab".
571
572 23. When UTF was disabled, the JIT program reported some incorrect compile
573 errors. These messages are silenced now.
574
575 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
576 compiler.
577
578 25. Change all the temporary files used in RunGrepTest to be different to those
579 used by RunTest so that the tests can be run simultaneously, for example by
580 "make -j check".
581
582
583 Version 8.34 15-December-2013
584 -----------------------------
585
586 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
587 executable memory. Patch inspired by Carsten Klein.
588
589 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
590 this macro is never tested and has no effect, because the work to support
591 coverage involves only compiling and linking options and special targets in
592 the Makefile. The comment in config.h implied that defining the macro would
593 enable coverage support, which is totally false. There was also support for
594 setting this macro in the CMake files (my fault, I just copied it from
595 configure). SUPPORT_GCOV has now been removed.
596
597 3. Make a small performance improvement in strlen16() and strlen32() in
598 pcretest.
599
600 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
601 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
602 these cases). There was also one in pcretest.c.
603
604 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
605
606 6. In UTF mode, the code for checking whether a group could match an empty
607 string (which is used for indefinitely repeated groups to allow for
608 breaking an infinite loop) was broken when the group contained a repeated
609 negated single-character class with a character that occupied more than one
610 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
611 UTF-8 mode). The effect was undefined: the group might or might not be
612 deemed as matching an empty string, or the program might have crashed.
613
614 7. The code for checking whether a group could match an empty string was not
615 recognizing that \h, \H, \v, \V, and \R must match a character.
616
617 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
618 an empty string. If it can, pcretest shows this in its information output.
619
620 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
621 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
622 matched by pcre_exec() without using JIT:
623
624 (a) If the rest of the pattern did not match after a maximal run of
625 grapheme clusters, the code for backing up to try with fewer of them
626 did not always back up over a full grapheme when characters that do not
627 have the modifier quality were involved, e.g. Hangul syllables.
628
629 (b) If the match point in a subject started with modifier character, and
630 there was no match, the code could incorrectly back up beyond the match
631 point, and potentially beyond the first character in the subject,
632 leading to a segfault or an incorrect match result.
633
634 10. A conditional group with an assertion condition could lead to PCRE
635 recording an incorrect first data item for a match if no other first data
636 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
637 first data item, and therefore matched "ca" after "c" instead of at the
638 start.
639
640 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
641 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
642
643 12. The source of pcregrep now includes z/OS-specific code so that it can be
644 compiled for z/OS as part of the special z/OS distribution.
645
646 13. Added the -T and -TM options to pcretest.
647
648 14. The code in pcre_compile.c for creating the table of named capturing groups
649 has been refactored. Instead of creating the table dynamically during the
650 actual compiling pass, the information is remembered during the pre-compile
651 pass (on the stack unless there are more than 20 named groups, in which
652 case malloc() is used) and the whole table is created before the actual
653 compile happens. This has simplified the code (it is now nearly 150 lines
654 shorter) and prepared the way for better handling of references to groups
655 with duplicate names.
656
657 15. A back reference to a named subpattern when there is more than one of the
658 same name now checks them in the order in which they appear in the pattern.
659 The first one that is set is used for the reference. Previously only the
660 first one was inspected. This change makes PCRE more compatible with Perl.
661
662 16. Unicode character properties were updated from Unicode 6.3.0.
663
664 17. The compile-time code for auto-possessification has been refactored, based
665 on a patch by Zoltan Herczeg. It now happens after instead of during
666 compilation. The code is cleaner, and more cases are handled. The option
667 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
668 options in pcretest are provided to set it. It can also be set by
669 (*NO_AUTO_POSSESS) at the start of a pattern.
670
671 18. The character VT has been added to the default ("C" locale) set of
672 characters that match \s and are generally treated as white space,
673 following this same change in Perl 5.18. There is now no difference between
674 "Perl space" and "POSIX space". Whether VT is treated as white space in
675 other locales depends on the locale.
676
677 19. The code for checking named groups as conditions, either for being set or
678 for being recursed, has been refactored (this is related to 14 and 15
679 above). Processing unduplicated named groups should now be as fast at
680 numerical groups, and processing duplicated groups should be faster than
681 before.
682
683 20. Two patches to the CMake build system, by Alexander Barkov:
684
685 (1) Replace the "source" command by "." in CMakeLists.txt because
686 "source" is a bash-ism.
687
688 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
689 without these the CMake build does not work on Solaris.
690
691 21. Perl has changed its handling of \8 and \9. If there is no previously
692 encountered capturing group of those numbers, they are treated as the
693 literal characters 8 and 9 instead of a binary zero followed by the
694 literals. PCRE now does the same.
695
696 22. Following Perl, added \o{} to specify codepoints in octal, making it
697 possible to specify values greater than 0777 and also making them
698 unambiguous.
699
700 23. Perl now gives an error for missing closing braces after \x{... instead of
701 treating the string as literal. PCRE now does the same.
702
703 24. RunTest used to grumble if an inappropriate test was selected explicitly,
704 but just skip it when running all tests. This make it awkward to run ranges
705 of tests when one of them was inappropriate. Now it just skips any
706 inappropriate tests, as it always did when running all tests.
707
708 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
709 character types such as \d or \w, too many callouts were inserted, and the
710 data that they returned was rubbish.
711
712 26. In UCP mode, \s was not matching two of the characters that Perl matches,
713 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
714 were matched by \h. The code has now been refactored so that the lists of
715 the horizontal and vertical whitespace characters used for \h and \v (which
716 are defined only in one place) are now also used for \s.
717
718 27. Add JIT support for the 64 bit TileGX architecture.
719 Patch by Jiong Wang (Tilera Corporation).
720
721 28. Possessive quantifiers for classes (both explicit and automatically
722 generated) now use special opcodes instead of wrapping in ONCE brackets.
723
724 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
725 (because it's meaningless), this was not happening when PCRE_CASELESS was
726 set. Not wrong, but inefficient.
727
728 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
729 properties for \w, \d, etc) is present in a test regex. Otherwise if the
730 test contains no characters greater than 255, Perl doesn't realise it
731 should be using Unicode semantics.
732
733 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
734 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
735 does in Unicode mode.
736
737 32. Added the "forbid" facility to pcretest so that putting tests into the
738 wrong test files can sometimes be quickly detected.
739
740 33. There is now a limit (default 250) on the depth of nesting of parentheses.
741 This limit is imposed to control the amount of system stack used at compile
742 time. It can be changed at build time by --with-parens-nest-limit=xxx or
743 the equivalent in CMake.
744
745 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
746 errors. Perl warns for these when in warning mode, but PCRE has no facility
747 for giving warnings.
748
749 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
750 However, this was not working for (?!) because it is optimized to (*FAIL),
751 for which PCRE does not allow quantifiers. The optimization is now disabled
752 when a quantifier follows (?!). I can't see any use for this, but it makes
753 things uniform.
754
755 36. Perl no longer allows group names to start with digits, so I have made this
756 change also in PCRE. It simplifies the code a bit.
757
758 37. In extended mode, Perl ignores spaces before a + that indicates a
759 possessive quantifier. PCRE allowed a space before the quantifier, but not
760 before the possessive +. It now does.
761
762 38. The use of \K (reset reported match start) within a repeated possessive
763 group such as (a\Kb)*+ was not working.
764
765 40. Document that the same character tables must be used at compile time and
766 run time, and that the facility to pass tables to pcre_exec() and
767 pcre_dfa_exec() is for use only with saved/restored patterns.
768
769 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
770 features for Builds with MSVC:
771
772 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
773 builds). The .rc files can be used to set FileDescription and many other
774 attributes.
775
776 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
777 This allows higher-level build scripts which want .pdb files to avoid
778 hard-coding the exact files needed."
779
780 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
781 mean "start of word" and "end of word", respectively, as a transition aid.
782
783 43. A minimizing repeat of a class containing codepoints greater than 255 in
784 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
785 compiled to use the heap for recursion.
786
787 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
788 is configured.
789
790
791 Version 8.33 28-May-2013
792 ------------------------
793
794 1. Added 'U' to some constants that are compared to unsigned integers, to
795 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
796 variables that are added to signed variables, to ensure the result is
797 signed and can be negated.
798
799 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
800 CMake config files.
801
802 3. Revise the creation of config.h.generic so that all boolean macros are
803 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
804 overriding via -D on the command line possible.
805
806 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
807 to unsigned int is reported to make a quite noticeable speed difference in
808 a specific Windows environment. Testing on Linux did also appear to show
809 some benefit (and it is clearly not harmful). Also fixed the definition of
810 Xop which should be unsigned.
811
812 5. Related to (4), changing the definition of the intermediate variable cc
813 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
814 improvements.
815
816 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
817 unnecessary spaces.
818
819 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
820 and later.
821
822 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
823
824 9. Optimizing fast_forward_start_bits in JIT.
825
826 10. Adding support for callouts in JIT, and fixing some issues revealed
827 during this work. Namely:
828
829 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
830
831 (b) Minimum length was not checked before the matching is started.
832
833 11. The value of capture_last that is passed to callouts was incorrect in some
834 cases when there was a capture on one path that was subsequently abandoned
835 after a backtrack. Also, the capture_last value is now reset after a
836 recursion, since all captures are also reset in this case.
837
838 12. The interpreter no longer returns the "too many substrings" error in the
839 case when an overflowing capture is in a branch that is subsequently
840 abandoned after a backtrack.
841
842 13. In the pathological case when an offset vector of size 2 is used, pcretest
843 now prints out the matched string after a yield of 0 or 1.
844
845 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
846 Only supported by the JIT compiler at the moment.
847
848 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
849
850 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
851 offset of the starting point of the matching process, provided the offsets
852 vector is large enough.
853
854 17. The \A escape now records a lookbehind value of 1, though its execution
855 does not actually inspect the previous character. This is to ensure that,
856 in partial multi-segment matching, at least one character from the old
857 segment is retained when a new segment is processed. Otherwise, if there
858 are no lookbehinds in the pattern, \A might match incorrectly at the start
859 of a new segment.
860
861 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
862
863 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
864 gives some modest performance improvement in 8-bit mode.
865
866 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
867 be expressed in certain programming languages using Universal Character
868 Names.
869
870 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
871 which points out that "non characters" are not "characters that may not
872 appear in Unicode strings" but rather "characters that are reserved for
873 internal use and have only local meaning".
874
875 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
876 there was a conditional group that depended on an assertion, if the
877 assertion was false, the callout that immediately followed the alternation
878 in the condition was skipped when pcre_exec() was used for matching.
879
880 23. Allow an explicit callout to be inserted before an assertion that is the
881 condition for a conditional group, for compatibility with automatic
882 callouts, which always insert a callout at this point.
883
884 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
885 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
886
887 25. (*PRUNE) is now supported by the JIT compiler.
888
889 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
890
891 27. Fix the case where there are two or more SKIPs with arguments that may be
892 ignored.
893
894 28. (*SKIP) is now supported by the JIT compiler.
895
896 29. (*THEN) is now supported by the JIT compiler.
897
898 30. Update RunTest with additional test selector options.
899
900 31. The way PCRE handles backtracking verbs has been changed in two ways.
901
902 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
903 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
904 backtracking. In some cases this makes it more Perl-compatible, but Perl's
905 rather obscure rules do not always do the same thing.
906
907 (2) Previously, backtracking verbs were confined within assertions. This is
908 no longer the case for positive assertions, except for (*ACCEPT). Again,
909 this sometimes improves Perl compatibility, and sometimes does not.
910
911 32. A number of tests that were in test 2 because Perl did things differently
912 have been moved to test 1, because either Perl or PCRE has changed, and
913 these tests are now compatible.
914
915 32. Backtracking control verbs are now handled in the same way in JIT and
916 interpreter.
917
918 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
919 contained a forward subroutine reference caused a compile error.
920
921 34. Auto-detect and optimize limited repetitions in JIT.
922
923 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
924 blocking (*UTF) etc.
925
926 36. In the interpreter, maximizing pattern repetitions for characters and
927 character types now use tail recursion, which reduces stack usage.
928
929 37. The value of the max lookbehind was not correctly preserved if a compiled
930 and saved regex was reloaded on a host of different endianness.
931
932 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
933 of the compiled pattern block, expand the flags field from 16 to 32 bits
934 because it was almost full.
935
936 39. Try madvise first before posix_madvise.
937
938 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
939 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
940 original change locked out all matches of empty strings. This has been
941 changed so that one match of an empty string per line is recognized.
942 Subsequent searches on the same line (for colouring or for --only-matching,
943 for example) do not recognize empty strings.
944
945 41. Applied a user patch to fix a number of spelling mistakes in comments.
946
947 42. Data lines longer than 65536 caused pcretest to crash.
948
949 43. Clarified the data type for length and startoffset arguments for pcre_exec
950 and pcre_dfa_exec in the function-specific man pages, where they were
951 explicitly stated to be in bytes, never having been updated. I also added
952 some clarification to the pcreapi man page.
953
954 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
955 a segmentation fault.
956
957
958 Version 8.32 30-November-2012
959 -----------------------------
960
961 1. Improved JIT compiler optimizations for first character search and single
962 character iterators.
963
964 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
965 Patch by Daniel Richard G.
966
967 3. Single character iterator optimizations in the JIT compiler.
968
969 4. Improved JIT compiler optimizations for character ranges.
970
971 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
972 Reported by Giuseppe D'Angelo.
973
974 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
975 of a line, was being set incorrectly in cases where .* appeared inside
976 atomic brackets at the start of a pattern, or where there was a subsequent
977 *PRUNE or *SKIP.
978
979 7. Improved instruction cache flush for POWER/PowerPC.
980 Patch by Daniel Richard G.
981
982 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
983 grep:
984
985 (a) There is now no limit to the number of patterns to be matched.
986
987 (b) An error is given if a pattern is too long.
988
989 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
990 are now supported.
991
992 (d) --exclude-from and --include-from (multiple use) have been added.
993
994 (e) Exclusions and inclusions now apply to all files and directories, not
995 just to those obtained from scanning a directory recursively.
996
997 (f) Multiple uses of -f and --file-list are now supported.
998
999 (g) In a Windows environment, the default for -d has been changed from
1000 "read" (the GNU grep default) to "skip", because otherwise the presence
1001 of a directory in the file list provokes an error.
1002
1003 (h) The documentation has been revised and clarified in places.
1004
1005 9. Improve the matching speed of capturing brackets.
1006
1007 10. Changed the meaning of \X so that it now matches a Unicode extended
1008 grapheme cluster.
1009
1010 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1011 out POSIX threads when JIT support is configured.
1012
1013 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1014
1015 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1016 structure could go wrong in environments where size_t is not the same size
1017 as int.
1018
1019 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1020 set.
1021
1022 15. The EBCDIC support had decayed; later updates to the code had included
1023 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1024 general tidy up of EBCDIC-related issues, and the documentation was also
1025 not quite right. There is now a test that can be run on ASCII systems to
1026 check some of the EBCDIC-related things (but is it not a full test).
1027
1028 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1029 in a small tidy to the code.
1030
1031 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1032
1033 18. If the --only-matching (-o) option in pcregrep is specified multiple
1034 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1035 substrings matched by the 1st and 2nd capturing parentheses. A separating
1036 string can be specified by --om-separator (default empty).
1037
1038 19. Improving the first n character searches.
1039
1040 20. Turn case lists for horizontal and vertical white space into macros so that
1041 they are defined only once.
1042
1043 21. This set of changes together give more compatible Unicode case-folding
1044 behaviour for characters that have more than one other case when UCP
1045 support is available.
1046
1047 (a) The Unicode property table now has offsets into a new table of sets of
1048 three or more characters that are case-equivalent. The MultiStage2.py
1049 script that generates these tables (the pcre_ucd.c file) now scans
1050 CaseFolding.txt instead of UnicodeData.txt for character case
1051 information.
1052
1053 (b) The code for adding characters or ranges of characters to a character
1054 class has been abstracted into a generalized function that also handles
1055 case-independence. In UTF-mode with UCP support, this uses the new data
1056 to handle characters with more than one other case.
1057
1058 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1059 whose other case is greater than 256 are now correctly matched
1060 caselessly. Previously, the high codepoint matched the low one, but not
1061 vice versa.
1062
1063 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1064 of the new class addition function, using character lists defined as
1065 macros alongside the case definitions of 20 above.
1066
1067 (e) Caseless back references now work with characters that have more than
1068 one other case.
1069
1070 (f) General caseless matching of characters with more than one other case
1071 is supported.
1072
1073 22. Unicode character properties were updated from Unicode 6.2.0
1074
1075 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1076
1077 24. Add support for 32-bit character strings, and UTF-32
1078
1079 25. Major JIT compiler update (code refactoring and bugfixing).
1080 Experimental Sparc 32 support is added.
1081
1082 26. Applied a modified version of Daniel Richard G's patch to create
1083 pcre.h.generic and config.h.generic by "make" instead of in the
1084 PrepareRelease script.
1085
1086 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1087 pcre_compile.c when checking for a zero character.
1088
1089 28. Introducing a native interface for JIT. Through this interface, the compiled
1090 machine code can be directly executed. The purpose of this interface is to
1091 provide fast pattern matching, so several sanity checks are not performed.
1092 However, feature tests are still performed. The new interface provides
1093 1.4x speedup compared to the old one.
1094
1095 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1096 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1097 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1098
1099 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1100 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1101 the "old" RFC 2279). Instead, it ended up passing a negative length to
1102 pcre_exec().
1103
1104 31. Add support for GCC's visibility feature to hide internal functions.
1105
1106 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1107 "unknown -C option" after outputting 0 or 1.
1108
1109 33. There is now support for generating a code coverage report for the test
1110 suite in environments where gcc is the compiler and lcov is installed. This
1111 is mainly for the benefit of the developers.
1112
1113 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1114 unaddressable using valgrind annotations, allowing valgrind to detect
1115 invalid memory accesses. This is mainly for the benefit of the developers.
1116
1117 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1118
1119 26. Give configure error if --enable-cpp but no C++ compiler found.
1120
1121
1122 Version 8.31 06-July-2012
1123 -------------------------
1124
1125 1. Fixing a wrong JIT test case and some compiler warnings.
1126
1127 2. Removed a bashism from the RunTest script.
1128
1129 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1130 to unsigned type, result still unsigned" that was given by an MS compiler
1131 on encountering the code "-sizeof(xxx)".
1132
1133 4. Partial matching support is added to the JIT compiler.
1134
1135 5. Fixed several bugs concerned with partial matching of items that consist
1136 of more than one character:
1137
1138 (a) /^(..)\1/ did not partially match "aba" because checking references was
1139 done on an "all or nothing" basis. This also applied to repeated
1140 references.
1141
1142 (b) \R did not give a hard partial match if \r was found at the end of the
1143 subject.
1144
1145 (c) \X did not give a hard partial match after matching one or more
1146 characters at the end of the subject.
1147
1148 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1149 a partial match for the string "\r".
1150
1151 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1152 a partial match for a CR character at the end of the subject string.
1153
1154 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1155 running pcretest, the text "(JIT)" added to the output whenever JIT is
1156 actually used to run the match.
1157
1158 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1159 or /S+[+] with a digit between 1 and 7.
1160
1161 8. OP_NOT now supports any UTF character not just single-byte ones.
1162
1163 9. (*MARK) control verb is now supported by the JIT compiler.
1164
1165 10. The command "./RunTest list" lists the available tests without actually
1166 running any of them. (Because I keep forgetting what they all are.)
1167
1168 11. Add PCRE_INFO_MAXLOOKBEHIND.
1169
1170 12. Applied a (slightly modified) user-supplied patch that improves performance
1171 when the heap is used for recursion (compiled with --disable-stack-for-
1172 recursion). Instead of malloc and free for each heap frame each time a
1173 logical recursion happens, frames are retained on a chain and re-used where
1174 possible. This sometimes gives as much as 30% improvement.
1175
1176 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1177 call.
1178
1179 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1180
1181 15. It is now possible to link pcretest with libedit as an alternative to
1182 libreadline.
1183
1184 16. (*COMMIT) control verb is now supported by the JIT compiler.
1185
1186 17. The Unicode data tables have been updated to Unicode 6.1.0.
1187
1188 18. Added --file-list option to pcregrep.
1189
1190 19. Added binary file support to pcregrep, including the -a, --binary-files,
1191 -I, and --text options.
1192
1193 20. The madvise function is renamed for posix_madvise for QNX compatibility
1194 reasons. Fixed by Giuseppe D'Angelo.
1195
1196 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1197
1198 22. Changed the option for creating symbolic links for 16-bit man pages from
1199 -s to -sf so that re-installing does not cause issues.
1200
1201 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1202
1203 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1204 matches in certain environments (the workspace was not being correctly
1205 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1206 some of the workspace data at the beginning of a restart.
1207
1208 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1209 was not doing so when it should - probably a typo introduced by SVN 528
1210 (change 8.10/14).
1211
1212 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1213 \w+ when the character tables indicated that \x{c4} was a word character.
1214 There were several related cases, all because the tests for doing a table
1215 lookup were testing for characters less than 127 instead of 255.
1216
1217 27. If a pattern contains capturing parentheses that are not used in a match,
1218 their slots in the ovector are set to -1. For those that are higher than
1219 any matched groups, this happens at the end of processing. In the case when
1220 there were back references that the ovector was too small to contain
1221 (causing temporary malloc'd memory to be used during matching), and the
1222 highest capturing number was not used, memory off the end of the ovector
1223 was incorrectly being set to -1. (It was using the size of the temporary
1224 memory instead of the true size.)
1225
1226 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1227 ovector size, it uses memory at the end of the block that it has got.
1228
1229 29. Check for an overlong MARK name and give an error at compile time. The
1230 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1231
1232 30. JIT compiler update.
1233
1234 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1235 Rill for the patch.
1236
1237 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1238
1239 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1240
1241 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1242 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1243 was enabled.
1244
1245 35. Improve JIT code generation for greedy plus quantifier.
1246
1247 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1248 1 to "aa" instead of to an empty string. The bug affected repeated groups
1249 that could potentially match an empty string.
1250
1251 37. Optimizing single character iterators in JIT.
1252
1253 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1254 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1255 codepoints that are too big for the mode are faulted, and in a UTF mode,
1256 disallowed codepoints are also faulted.
1257
1258 39. If PCRE was compiled with UTF support, in three places in the DFA
1259 matcher there was code that should only have been obeyed in UTF mode, but
1260 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1261 processing when bytes with values greater than 127 were present. In 16-bit
1262 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1263 both cases the values are those that cannot be the first data item in a UTF
1264 character. The three items that might have provoked this were recursions,
1265 possessively repeated groups, and atomic groups.
1266
1267 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1268 and pcregrep, because some OS require shared objects to be explicitly
1269 passed to ld, causing the link step to fail if they are not.
1270
1271 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1272 mode, patterns that started with \h* or \R* might be incorrectly matched.
1273
1274
1275 Version 8.30 04-February-2012
1276 -----------------------------
1277
1278 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1279 name is defined in ctype.h.
1280
1281 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1282 only in quite long subpatterns.
1283
1284 3. Removed the function pcre_info(), which has been obsolete and deprecated
1285 since it was replaced by pcre_fullinfo() in February 2000.
1286
1287 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1288 match a (*MARK), and the match failed at the start of the subject, a
1289 reference to memory before the start of the subject could occur. This bug
1290 was introduced by fix 17 of release 8.21.
1291
1292 5. A reference to an unset group with zero minimum repetition was giving
1293 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1294 /(another)?(\1?)test/ matched against "hello world test". This bug was
1295 introduced in release 8.13.
1296
1297 6. Add support for 16-bit character strings (a large amount of work involving
1298 many changes and refactorings).
1299
1300 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1301 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1302 from a file.
1303
1304 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1305 rounding is not applied in this particular case).
1306
1307 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1308 if they appear, or are escaped, in patterns.
1309
1310 10. Get rid of a number of -Wunused-but-set-variable warnings.
1311
1312 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1313 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1314 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1315 also returns the mark "x". This bug applied to capturing parentheses,
1316 non-capturing parentheses, and atomic parentheses. It also applied to some
1317 assertions.
1318
1319 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1320 information out of configure.ac instead of relying on pcre.h.generic, which
1321 is not stored in the repository.
1322
1323 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1324 -lreadline.
1325
1326 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1327
1328 15. Applied Graycode's patch to put the top-level frame on the stack rather
1329 than the heap when not using the stack for recursion. This gives a
1330 performance improvement in many cases when recursion is not deep.
1331
1332 16. Experimental code added to "pcretest -C" to output the stack frame size.
1333
1334
1335 Version 8.21 12-Dec-2011
1336 ------------------------
1337
1338 1. Updating the JIT compiler.
1339
1340 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1341 are added as well.
1342
1343 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1344 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1345 calling _pcre_jit_exec. Some extra comments are added.
1346
1347 4. (*MARK) settings inside atomic groups that do not contain any capturing
1348 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1349 was introduced by change 18 for 8.20.
1350
1351 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1352 ECMA-262 standard.
1353
1354 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1355 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1356 This bug was probably introduced by change 9 of 8.13.
1357
1358 7. While fixing 6 above, I noticed that a number of other items were being
1359 incorrectly rejected as "not fixed length". This arose partly because newer
1360 opcodes had not been added to the fixed-length checking code. I have (a)
1361 corrected the bug and added tests for these items, and (b) arranged for an
1362 error to occur if an unknown opcode is encountered while checking for fixed
1363 length instead of just assuming "not fixed length". The items that were
1364 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1365 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1366 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1367
1368 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1369 being incorrectly compiled and would have given unpredicatble results.
1370
1371 9. A possessively repeated subpattern with minimum repeat count greater than
1372 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1373 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1374 the first (A) could occur when it should not.
1375
1376 10. Add a cast and remove a redundant test from the code.
1377
1378 11. JIT should use pcre_malloc/pcre_free for allocation.
1379
1380 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1381 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1382 is anything other than /usr, -L is still shown).
1383
1384 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1385
1386 14. Perl does not support \N without a following name in a [] class; PCRE now
1387 also gives an error.
1388
1389 15. If a forward reference was repeated with an upper limit of around 2000,
1390 it caused the error "internal error: overran compiling workspace". The
1391 maximum number of forward references (including repeats) was limited by the
1392 internal workspace, and dependent on the LINK_SIZE. The code has been
1393 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1394 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1395 of around 200,000 forward references. While doing this, I also speeded up
1396 the filling in of repeated forward references.
1397
1398 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1399 incorrectly expecting the subject to contain another "a" after the start.
1400
1401 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1402 in the match, the SKIP should be ignored. This was not happening; instead
1403 the SKIP was being treated as NOMATCH. For patterns such as
1404 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1405 tested.
1406
1407 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1408 now much more compatible with Perl, in particular in cases where the result
1409 is a non-match for a non-anchored pattern. For example, if
1410 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1411 "m", where previously it did not return a name. A side effect of this
1412 change is that for partial matches, the last encountered mark name is
1413 returned, as for non matches. A number of tests that were previously not
1414 Perl-compatible have been moved into the Perl-compatible test files. The
1415 refactoring has had the pleasing side effect of removing one argument from
1416 the match() function, thus reducing its stack requirements.
1417
1418 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1419 subsequent uses of /S (without +) incorrectly behaved like /S+.
1420
1421 21. Retrieve executable code size support for the JIT compiler and fixing
1422 some warnings.
1423
1424 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1425 not work when the shorter character appeared right at the end of the
1426 subject string.
1427
1428 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1429 systems.
1430
1431 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1432 output it when the /M option is used in pcretest.
1433
1434 25. The CheckMan script was not being included in the distribution. Also, added
1435 an explicit "perl" to run Perl scripts from the PrepareRelease script
1436 because this is reportedly needed in Windows.
1437
1438 26. If study data was being save in a file and studying had not found a set of
1439 "starts with" bytes for the pattern, the data written to the file (though
1440 never used) was taken from uninitialized memory and so caused valgrind to
1441 complain.
1442
1443 27. Updated RunTest.bat as provided by Sheri Pierce.
1444
1445 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1446
1447 29. Computation of memory usage for the table of capturing group names was
1448 giving an unnecessarily large value.
1449
1450
1451 Version 8.20 21-Oct-2011
1452 ------------------------
1453
1454 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1455 a POSIX class. After further experiments with Perl, which convinced me that
1456 Perl has bugs and confusions, a closing square bracket is no longer allowed
1457 in a POSIX name. This bug also affected patterns with classes that started
1458 with full stops.
1459
1460 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1461 captured substring, but while checking the failing first alternative,
1462 substring 1 is temporarily captured. If the output vector supplied to
1463 pcre_exec() was not big enough for this capture, the yield of the function
1464 was still zero ("insufficient space for captured substrings"). This cannot
1465 be totally fixed without adding another stack variable, which seems a lot
1466 of expense for a edge case. However, I have improved the situation in cases
1467 such as /(a)(b)x|abc/ matched against "abc", where the return code
1468 indicates that fewer than the maximum number of slots in the ovector have
1469 been set.
1470
1471 3. Related to (2) above: when there are more back references in a pattern than
1472 slots in the output vector, pcre_exec() uses temporary memory during
1473 matching, and copies in the captures as far as possible afterwards. It was
1474 using the entire output vector, but this conflicts with the specification
1475 that only 2/3 is used for passing back captured substrings. Now it uses
1476 only the first 2/3, for compatibility. This is, of course, another edge
1477 case.
1478
1479 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1480 main code base, and can be used by building with --enable-jit. When this is
1481 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1482 runtime --no-jit option is given.
1483
1484 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1485 ovector, the return from the function was zero, implying that there were
1486 other matches that did not fit. The correct "exactly full" value is now
1487 returned.
1488
1489 6. If a subpattern that was called recursively or as a subroutine contained
1490 (*PRUNE) or any other control that caused it to give a non-standard return,
1491 invalid errors such as "Error -26 (nested recursion at the same subject
1492 position)" or even infinite loops could occur.
1493
1494 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1495 computing the minimum length on reaching *ACCEPT, and so ended up with the
1496 wrong value of 1 rather than 0. Further investigation indicates that
1497 computing a minimum subject length in the presence of *ACCEPT is difficult
1498 (think back references, subroutine calls), and so I have changed the code
1499 so that no minimum is registered for a pattern that contains *ACCEPT.
1500
1501 8. If (*THEN) was present in the first (true) branch of a conditional group,
1502 it was not handled as intended. [But see 16 below.]
1503
1504 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1505 Sheri Pierce.
1506
1507 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1508 the first byte in a match must be "a".
1509
1510 11. Change 17 for 8.13 increased the recursion depth for patterns like
1511 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1512 pattern contains any instances of (*THEN). If it does not, the old
1513 optimizations are restored. It would be nice to do this on a per-group
1514 basis, but at the moment that is not feasible.
1515
1516 12. In some environments, the output of pcretest -C is CRLF terminated. This
1517 broke RunTest's code that checks for the link size. A single white space
1518 character after the value is now allowed for.
1519
1520 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1521 For "fr", it uses the Windows-specific input and output files.
1522
1523 14. If (*THEN) appeared in a group that was called recursively or as a
1524 subroutine, it did not work as intended. [But see next item.]
1525
1526 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1527 pattern fragments (but not containing any | characters). If A and B are
1528 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1529 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1530 D. In other words, Perl considers parentheses that do not contain any |
1531 characters to be part of a surrounding alternative, whereas PCRE was
1532 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1533 differently. PCRE now behaves in the same way as Perl, except in the case
1534 of subroutine/recursion calls such as (?1) which have in any case always
1535 been different (but PCRE had them first :-).
1536
1537 16. Related to 15 above: Perl does not treat the | in a conditional group as
1538 creating alternatives. Such a group is treated in the same way as an
1539 ordinary group without any | characters when processing (*THEN). PCRE has
1540 been changed to match Perl's behaviour.
1541
1542 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1543 RunGrepTest script failed.
1544
1545 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1546 inevitable for groups that contain captures, but it can lead to a lot of
1547 stack use in large patterns. The old behaviour has been restored for atomic
1548 groups that do not contain any capturing parentheses.
1549
1550 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1551 suppress the check for a minimum subject length at run time. (If it was
1552 given to pcre_exec() or pcre_dfa_exec() it did work.)
1553
1554 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1555 fail to work when decoding hex characters in data strings in EBCDIC
1556 environments.
1557
1558 21. It appears that in at least one Mac OS environment, the isxdigit() function
1559 is implemented as a macro that evaluates to its argument more than once,
1560 contravening the C 90 Standard (I haven't checked a later standard). There
1561 was an instance in pcretest which caused it to go wrong when processing
1562 \x{...} escapes in subject strings. The has been rewritten to avoid using
1563 things like p++ in the argument of isxdigit().
1564
1565
1566 Version 8.13 16-Aug-2011
1567 ------------------------
1568
1569 1. The Unicode data tables have been updated to Unicode 6.0.0.
1570
1571 2. Two minor typos in pcre_internal.h have been fixed.
1572
1573 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1574 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1575 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1576
1577 4. There were a number of related bugs in the code for matching backrefences
1578 caselessly in UTF-8 mode when codes for the characters concerned were
1579 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1580 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1581 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1582 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1583 2-byte code at the end of the subject (it thought there wasn't enough data
1584 left).
1585
1586 5. Comprehensive information about what went wrong is now returned by
1587 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1588 as the output vector has at least 2 elements. The offset of the start of
1589 the failing character and a reason code are placed in the vector.
1590
1591 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1592 now returned is for the first byte of the failing character, instead of the
1593 last byte inspected. This is an incompatible change, but I hope it is small
1594 enough not to be a problem. It makes the returned offset consistent with
1595 pcre_exec() and pcre_dfa_exec().
1596
1597 7. pcretest now gives a text phrase as well as the error number when
1598 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1599 failure, the offset and reason code are output.
1600
1601 8. When \R was used with a maximizing quantifier it failed to skip backwards
1602 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1603 back over a single character (\n). This seems wrong (because it treated the
1604 two characters as a single entity when going forwards), conflicts with the
1605 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1606 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1607 has been changed.
1608
1609 9. Some internal refactoring has changed the processing so that the handling
1610 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1611 time (the PCRE_DOTALL option was changed this way some time ago: version
1612 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1613 which was always a bit of a fudge. It also means that there is one less
1614 argument for the match() function, which reduces its stack requirements
1615 slightly. This change also fixes an incompatibility with Perl: the pattern
1616 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1617
1618 10. More internal refactoring has drastically reduced the number of recursive
1619 calls to match() for possessively repeated groups such as (abc)++ when
1620 using pcre_exec().
1621
1622 11. While implementing 10, a number of bugs in the handling of groups were
1623 discovered and fixed:
1624
1625 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1626 (a|)*(?1) gave a compile-time internal error.
1627 ((a|)+)+ did not notice that the outer group could match an empty string.
1628 (^a|^)+ was not marked as anchored.
1629 (.*a|.*)+ was not marked as matching at start or after a newline.
1630
1631 12. Yet more internal refactoring has removed another argument from the match()
1632 function. Special calls to this function are now indicated by setting a
1633 value in a variable in the "match data" data block.
1634
1635 13. Be more explicit in pcre_study() instead of relying on "default" for
1636 opcodes that mean there is no starting character; this means that when new
1637 ones are added and accidentally left out of pcre_study(), testing should
1638 pick them up.
1639
1640 14. The -s option of pcretest has been documented for ages as being an old
1641 synonym of -m (show memory usage). I have changed it to mean "force study
1642 for every regex", that is, assume /S for every regex. This is similar to -i
1643 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1644 using it. It makes it easier to run collections of tests with and without
1645 study enabled, and thereby test pcre_study() more easily. All the standard
1646 tests are now run with and without -s (but some patterns can be marked as
1647 "never study" - see 20 below).
1648
1649 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1650 restoration of the capturing data to the outer values was not happening
1651 correctly.
1652
1653 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1654 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1655 pattern had matched an empty string, and so incorrectly returned a no
1656 match.
1657
1658 17. There was optimizing code for the last branch of non-capturing parentheses,
1659 and also for the obeyed branch of a conditional subexpression, which used
1660 tail recursion to cut down on stack usage. Unfortunately, now that there is
1661 the possibility of (*THEN) occurring in these branches, tail recursion is
1662 no longer possible because the return has to be checked for (*THEN). These
1663 two optimizations have therefore been removed. [But see 8.20/11 above.]
1664
1665 18. If a pattern containing \R was studied, it was assumed that \R always
1666 matched two bytes, thus causing the minimum subject length to be
1667 incorrectly computed because \R can also match just one byte.
1668
1669 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1670 was incorrectly computed.
1671
1672 20. If /S is present twice on a test pattern in pcretest input, it now
1673 *disables* studying, thereby overriding the use of -s on the command line
1674 (see 14 above). This is necessary for one or two tests to keep the output
1675 identical in both cases.
1676
1677 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1678 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1679
1680 22. When an atomic group that contained a capturing parenthesis was
1681 successfully matched, but the branch in which it appeared failed, the
1682 capturing was not being forgotten if a higher numbered group was later
1683 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1684 group 1 to "a", when in fact it should be unset. This applied to multi-
1685 branched capturing and non-capturing groups, repeated or not, and also to
1686 positive assertions (capturing in negative assertions does not happen
1687 in PCRE) and also to nested atomic groups.
1688
1689 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1690 subject after a captured substring, to make it easier to tell which of a
1691 number of identical substrings has been captured.
1692
1693 24. The way atomic groups are processed by pcre_exec() has been changed so that
1694 if they are repeated, backtracking one repetition now resets captured
1695 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1696 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1697 "aaa". Previously, it would have been "a". As part of this code
1698 refactoring, the way recursive calls are handled has also been changed.
1699
1700 25. If an assertion condition captured any substrings, they were not passed
1701 back unless some other capturing happened later. For example, if
1702 (?(?=(a))a) was matched against "a", no capturing was returned.
1703
1704 26. When studying a pattern that contained subroutine calls or assertions,
1705 the code for finding the minimum length of a possible match was handling
1706 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1707 group 1 called group 2 while simultaneously a separate group 2 called group
1708 1). A stack overflow occurred in this case. I have fixed this by limiting
1709 the recursion depth to 10.
1710
1711 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1712 Fortmann. This supports explicit test numbers on the command line, and has
1713 argument validation and error reporting.
1714
1715 28. An instance of \X with an unlimited repeat could fail if at any point the
1716 first character it looked at was a mark character.
1717
1718 29. Some minor code refactoring concerning Unicode properties and scripts
1719 should reduce the stack requirement of match() slightly.
1720
1721 30. Added the '=' option to pcretest to check the setting of unused capturing
1722 slots at the end of the pattern, which are documented as being -1, but are
1723 not included in the return count.
1724
1725 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1726 compiled something random. Now it gives a compile-time error (as does
1727 Perl).
1728
1729 32. A *MARK encountered during the processing of a positive assertion is now
1730 recorded and passed back (compatible with Perl).
1731
1732 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1733 had alternative anchored branches, the search for a second match in a line
1734 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1735 matched the line "0102" twice. The same bug affected patterns that started
1736 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1737 twice.
1738
1739 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1740 does, and because of capturing effects, quantifying parenthesized
1741 assertions may at times be useful. Quantifiers are now allowed for
1742 parenthesized assertions.
1743
1744 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1745
1746 36. \g was being checked for fancy things in a character class, when it should
1747 just be a literal "g".
1748
1749 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1750 appearance of a nested POSIX class supersedes an apparent external class.
1751 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1752 unescaped square brackets may also appear as part of class names. For
1753 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1754 more like Perl. (But see 8.20/1 above.)
1755
1756 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1757 was because it thought it was \N{name}, which is not supported).
1758
1759 39. Add minix to OS list not supporting the -S option in pcretest.
1760
1761 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1762 cannot analyze patterns in sufficient detail to catch mutual recursions
1763 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1764 subgroup is called recursively as a subpattern for a second time at the
1765 same position in the subject string. In previous releases this might have
1766 been caught by the recursion limit, or it might have run out of stack.
1767
1768 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1769 happen only once. PCRE was, however incorrectly giving a compile time error
1770 "recursive call could loop indefinitely" because it cannot analyze the
1771 pattern in sufficient detail. The compile time test no longer happens when
1772 PCRE is compiling a conditional subpattern, but actual runaway loops are
1773 now caught at runtime (see 40 above).
1774
1775 42. It seems that Perl allows any characters other than a closing parenthesis
1776 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1777 has been changed to be the same.
1778
1779 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1780 as not to get warnings when autogen.sh is called. Also changed
1781 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1782
1783 44. To help people who use pcregrep to scan files containing exceedingly long
1784 lines, the following changes have been made:
1785
1786 (a) The default value of the buffer size parameter has been increased from
1787 8K to 20K. (The actual buffer used is three times this size.)
1788
1789 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1790 PCRE is built.
1791
1792 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1793 to be set at run time.
1794
1795 (d) Numerical values in pcregrep options can be followed by K or M, for
1796 example --buffer-size=50K.
1797
1798 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1799 given and the return code is set to 2.
1800
1801 45. Add a pointer to the latest mark to the callout data block.
1802
1803 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1804 partial match of an empty string instead of no match. This was specific to
1805 the use of ".".
1806
1807 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1808 complete match instead of a partial match. This bug was dependent on both
1809 the PCRE_UTF8 and PCRE_DOTALL options being set.
1810
1811 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1812 starting byte set, because \b was not being ignored.
1813
1814
1815 Version 8.12 15-Jan-2011
1816 ------------------------
1817
1818 1. Fixed some typos in the markup of the man pages, and wrote a script that
1819 checks for such things as part of the documentation building process.
1820
1821 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1822 --match-limit and --recursion-limit options (added for 8.11). In
1823 particular, this made one of the standard tests fail. (The integer value
1824 went into the wrong half of a long int.)
1825
1826 3. If the --colour option was given to pcregrep with -v (invert match), it
1827 did strange things, either producing crazy output, or crashing. It should,
1828 of course, ignore a request for colour when reporting lines that do not
1829 match.
1830
1831 4. Another pcregrep bug caused similar problems if --colour was specified with
1832 -M (multiline) and the pattern match finished with a line ending.
1833
1834 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1835 matched in multiline mode, the following line was shown as part of the
1836 match. This seems wrong, so I have changed it.
1837
1838 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1839 the check for further matches in the same line (so they could be coloured)
1840 to overrun the end of the current line. If another match was found, it was
1841 incorrectly shown (and then shown again when found in the next line).
1842
1843 7. If pcregrep was compiled under Windows, there was a reference to the
1844 function pcregrep_exit() before it was defined. I am assuming this was
1845 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1846 reported by a user. I've moved the definition above the reference.
1847
1848
1849 Version 8.11 10-Dec-2010
1850 ------------------------
1851
1852 1. (*THEN) was not working properly if there were untried alternatives prior
1853 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1854 backtracked to try for "b" instead of moving to the next alternative branch
1855 at the same level (in this case, to look for "c"). The Perl documentation
1856 is clear that when (*THEN) is backtracked onto, it goes to the "next
1857 alternative in the innermost enclosing group".
1858
1859 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1860 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1861 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1862 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1863 (*THEN).
1864
1865 3. If \s appeared in a character class, it removed the VT character from
1866 the class, even if it had been included by some previous item, for example
1867 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1868 of \s, but is part of the POSIX "space" class.)
1869
1870 4. A partial match never returns an empty string (because you can always
1871 match an empty string at the end of the subject); however the checking for
1872 an empty string was starting at the "start of match" point. This has been
1873 changed to the "earliest inspected character" point, because the returned
1874 data for a partial match starts at this character. This means that, for
1875 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1876 (previously it gave "no match").
1877
1878 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1879 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1880 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1881 has an implication that the given string is incomplete (because a partial
1882 match is preferred over a full match). For this reason, these items now
1883 give a partial match in this situation. [Aside: previously, the one case
1884 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1885 match rather than a full match, which was wrong by the old rules, but is
1886 now correct.]
1887
1888 6. There was a bug in the handling of #-introduced comments, recognized when
1889 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1890 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1891 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1892 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1893 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1894 places in pcre_compile().
1895
1896 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1897 comments when looking ahead for named forward references to subpatterns,
1898 the only newline sequence it recognized was NL. It now handles newlines
1899 according to the set newline convention.
1900
1901 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1902 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1903 cater for a lack of strerror(). These oversights have been fixed.
1904
1905 9. Added --match-limit and --recursion-limit to pcregrep.
1906
1907 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1908
1909 11. When the -o option was used, pcregrep was setting a return code of 1, even
1910 when matches were found, and --line-buffered was not being honoured.
1911
1912 12. Added an optional parentheses number to the -o and --only-matching options
1913 of pcregrep.
1914
1915 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1916 can match an empty string. The code to do it in pcretest and pcredemo
1917 needed fixing:
1918
1919 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1920 only one byte after an empty string match just before CRLF (this case
1921 just got forgotten; "any" and "anycrlf" were OK).
1922
1923 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1924 mode when an empty string match preceded an ASCII character followed by
1925 a non-ASCII character. (The code for advancing by one character rather
1926 than one byte was nonsense.)
1927
1928 (c) The pcredemo.c sample program did not have any code at all to handle
1929 the cases when CRLF is a valid newline sequence.
1930
1931 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1932 as a starting offset was within the subject string. There is now a new
1933 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1934 negative or greater than the length of the string. In order to test this,
1935 pcretest is extended to allow the setting of negative starting offsets.
1936
1937 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1938 starting offset points to the beginning of a UTF-8 character was
1939 unnecessarily clumsy. I tidied it up.
1940
1941 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1942 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1943
1944 17. Nobody had reported that the --include_dir option, which was added in
1945 release 7.7 should have been called --include-dir (hyphen, not underscore)
1946 for compatibility with GNU grep. I have changed it to --include-dir, but
1947 left --include_dir as an undocumented synonym, and the same for
1948 --exclude-dir, though that is not available in GNU grep, at least as of
1949 release 2.5.4.
1950
1951 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1952 characters from a string of bytes) have been redefined so as not to use
1953 loops, in order to improve performance in some environments. At the same
1954 time, I abstracted some of the common code into auxiliary macros to save
1955 repetition (this should not affect the compiled code).
1956
1957 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1958 compile-time error is now given if \c is not followed by an ASCII
1959 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1960 different, and any byte value is allowed.)
1961
1962 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1963 START_OPTIMIZE option, which is now allowed at compile time - but just
1964 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1965 to pcregrep and other applications that have no direct access to PCRE
1966 options. The new /Y option in pcretest sets this option when calling
1967 pcre_compile().
1968
1969 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1970 back references. Groups containing recursive back references were forced to
1971 be atomic by that change, but in the case of named groups, the amount of
1972 memory required was incorrectly computed, leading to "Failed: internal
1973 error: code overflow". This has been fixed.
1974
1975 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1976 pcretest.c, to avoid build problems in some Borland environments.
1977
1978
1979 Version 8.10 25-Jun-2010
1980 ------------------------
1981
1982 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1983 THEN.
1984
1985 2. (*ACCEPT) was not working when inside an atomic group.
1986
1987 3. Inside a character class, \B is treated as a literal by default, but
1988 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1989 causes the error). The code is unchanged, but I tidied the documentation.
1990
1991 4. Inside a character class, PCRE always treated \R and \X as literals,
1992 whereas Perl faults them if its -w option is set. I have changed PCRE so
1993 that it faults them when PCRE_EXTRA is set.
1994
1995 5. Added support for \N, which always matches any character other than
1996 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1997
1998 6. When compiling pcregrep with newer versions of gcc which may have
1999 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2000 declared with attribute warn_unused_result" were given. Just casting the
2001 result to (void) does not stop the warnings; a more elaborate fudge is
2002 needed. I've used a macro to implement this.
2003
2004 7. Minor change to pcretest.c to avoid a compiler warning.
2005
2006 8. Added four artifical Unicode properties to help with an option to make
2007 \s etc use properties (see next item). The new properties are: Xan
2008 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2009
2010 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2011 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2012 this option. Modified pcretest to add /W to test this facility. Added
2013 REG_UCP to make it available via the POSIX interface.
2014
2015 10. Added --line-buffered to pcregrep.
2016
2017 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2018 studied, and the match started with a letter with a code point greater than
2019 127 whose first byte was different to the first byte of the other case of
2020 the letter, the other case of this starting letter was not recognized
2021 (#976).
2022
2023 12. If a pattern that was studied started with a repeated Unicode property
2024 test, for example, \p{Nd}+, there was the theoretical possibility of
2025 setting up an incorrect bitmap of starting bytes, but fortunately it could
2026 not have actually happened in practice until change 8 above was made (it
2027 added property types that matched character-matching opcodes).
2028
2029 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2030 possible starting bytes for non-anchored patterns.
2031
2032 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2033 \R, and also a number of cases that involve Unicode properties, both
2034 explicit and implicit when PCRE_UCP is set.
2035
2036 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2037 input, it could crash or give wrong results if characters with values
2038 greater than 0xc0 were present in the subject string. (Detail: it assumed
2039 UTF-8 input when processing these items.)
2040
2041 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2042 size_t is 64-bit (#991).
2043
2044 17. Added a check for running out of memory when PCRE is compiled with
2045 --disable-stack-for-recursion (#990).
2046
2047 18. If the last data line in a file for pcretest does not have a newline on
2048 the end, a newline was missing in the output.
2049
2050 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2051 less than 128) in its various bitmaps. However, there is a facility for
2052 generating tables according to the current locale when PCRE is compiled. It
2053 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2054 characters, are recognized by isspace() and therefore were getting set in
2055 these tables, and indeed these tables seem to approximate to ISO 8859. This
2056 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2057 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2058 which of course cannot start UTF-8 characters. I have changed the code so
2059 that only real ASCII characters (less than 128) and the correct starting
2060 bytes for UTF-8 encodings are set for characters greater than 127 when in
2061 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2062 altogether.)
2063
2064 20. Added the /T option to pcretest so as to be able to run tests with non-
2065 standard character tables, thus making it possible to include the tests
2066 used for 19 above in the standard set of tests.
2067
2068 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2069 reference to a subpattern the other side of a comment that contains an
2070 opening parenthesis caused either an internal compiling error, or a
2071 reference to the wrong subpattern.
2072
2073
2074 Version 8.02 19-Mar-2010
2075 ------------------------
2076
2077 1. The Unicode data tables have been updated to Unicode 5.2.0.
2078
2079 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2080 configured.
2081
2082 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2083 original author of that file, following a query about its status.
2084
2085 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2086 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2087
2088 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2089 quantifier applied to a forward-referencing subroutine call, could compile
2090 incorrect code or give the error "internal error: previously-checked
2091 referenced subpattern not found".
2092
2093 6. Both MS Visual Studio and Symbian OS have problems with initializing
2094 variables to point to external functions. For these systems, therefore,
2095 pcre_malloc etc. are now initialized to local functions that call the
2096 relevant global functions.
2097
2098 7. There were two entries missing in the vectors called coptable and poptable
2099 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2100 I've fixed the data, and added a kludgy way of testing at compile time that
2101 the lengths are correct (equal to the number of opcodes).
2102
2103 8. Following on from 7, I added a similar kludge to check the length of the
2104 eint vector in pcreposix.c.
2105
2106 9. Error texts for pcre_compile() are held as one long string to avoid too
2107 much relocation at load time. To find a text, the string is searched,
2108 counting zeros. There was no check for running off the end of the string,
2109 which could happen if a new error number was added without updating the
2110 string.
2111
2112 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2113
2114 11. \K was not working if it appeared in an atomic group or in a group that
2115 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2116 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2117 the assertion is positive, but not if it is negative.
2118
2119 12. Change 11 fortuitously reduced the size of the stack frame used in the
2120 "match()" function of pcre_exec.c by one pointer. Forthcoming
2121 implementation of support for (*MARK) will need an extra pointer on the
2122 stack; I have reserved it now, so that the stack frame size does not
2123 decrease.
2124
2125 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2126 item in branch that calls a recursion is a subroutine call - as in the
2127 second branch in the above example - was incorrectly given the compile-
2128 time error "recursive call could loop indefinitely" because pcre_compile()
2129 was not correctly checking the subroutine for matching a non-empty string.
2130
2131 14. The checks for overrunning compiling workspace could trigger after an
2132 overrun had occurred. This is a "should never occur" error, but it can be
2133 triggered by pathological patterns such as hundreds of nested parentheses.
2134 The checks now trigger 100 bytes before the end of the workspace.
2135
2136 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2137
2138
2139 Version 8.01 19-Jan-2010
2140 ------------------------
2141
2142 1. If a pattern contained a conditional subpattern with only one branch (in
2143 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2144 computed the wrong minimum data length (which is of course zero for such
2145 subpatterns). This could cause incorrect "no match" results.
2146
2147 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2148 the pattern is reset in the first branch, pcre_compile() failed with
2149 "internal error: code overflow at offset...". This happened only when
2150 the reset was to the original external option setting. (An optimization
2151 abstracts leading options settings into an external setting, which was the
2152 cause of this.)
2153
2154 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2155 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2156 assertion pattern did not match (meaning that the assertion was true), it
2157 was incorrectly treated as false if the SKIP had been reached during the
2158 matching. This also applied to assertions used as conditions.
2159
2160 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2161 assertion subpattern, including such a pattern used as a condition,
2162 unpredictable results occurred, instead of the error return
2163 PCRE_ERROR_DFA_UITEM.
2164
2165 5. The C++ GlobalReplace function was not working like Perl for the special
2166 situation when an empty string is matched. It now does the fancy magic
2167 stuff that is necessary.
2168
2169 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2170 removed. (These were left over from very, very early versions of PCRE.)
2171
2172 7. Some cosmetic changes to the code to make life easier when compiling it
2173 as part of something else:
2174
2175 (a) Change DEBUG to PCRE_DEBUG.
2176
2177 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2178 called "current" as "current_branch", to prevent a collision with the
2179 Linux macro when compiled as a kernel module.
2180
2181 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2182 prevent a collision with the Linux macro when compiled as a kernel
2183 module.
2184
2185 8. In pcre_compile() there are some checks for integer overflows that used to
2186 cast potentially large values to (double). This has been changed to that
2187 when building, a check for int64_t is made, and if it is found, it is used
2188 instead, thus avoiding the use of floating point arithmetic. (There is no
2189 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2190 double.
2191
2192 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2193 2005 (difference between two addresses compared to an unsigned value).
2194
2195 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2196 custom one, because of the following reported problem in Windows:
2197
2198 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2199 under Win32.
2200 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2201 therefore missing the function definition.
2202 - The compiler thus generates a "C" signature for the test function.
2203 - The linker fails to find the "C" function.
2204 - PCRE fails to configure if asked to do so against libbz2.
2205
2206 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2207 messages were output:
2208
2209 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2210 rerunning libtoolize, to keep the correct libtool macros in-tree.
2211 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2212
2213 I have done both of these things.
2214
2215 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2216 most of the time, it *can* run out if it is given a pattern that contains a
2217 runaway infinite recursion. I updated the discussion in the pcrestack man
2218 page.
2219
2220 13. Now that we have gone to the x.xx style of version numbers, the minor
2221 version may start with zero. Using 08 or 09 is a bad idea because users
2222 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2223 interpreted as invalid octal numbers. I've updated the previous comment in
2224 configure.ac, and also added a check that gives an error if 08 or 09 are
2225 used.
2226
2227 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2228 causing partial matching to fail when the end of the subject matched \W
2229 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2230
2231 15. There were some discrepancies between the declarations in pcre_internal.h
2232 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2233 their definitions. The declarations used "const uschar *" and the
2234 definitions used USPTR. Even though USPTR is normally defined as "const
2235 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2236 reported that: "This difference in casting confuses some C++ compilers, for
2237 example, SunCC recognizes above declarations as different functions and
2238 generates broken code for hbpcre." I have changed the declarations to use
2239 USPTR.
2240
2241 16. GNU libtool is named differently on some systems. The autogen.sh script now
2242 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2243 (FreeBSD).
2244
2245 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2246 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2247 comment: "Figure out how to create a longlong from a string: strtoll and
2248 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2249 instance, but it only takes 2 args instead of 3!"
2250
2251 18. A subtle bug concerned with back references has been fixed by a change of
2252 specification, with a corresponding code fix. A pattern such as
2253 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2254 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2255 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2256 same bug. Such groups have to be quantified to be useful, or contained
2257 inside another quantified group. (If there's no repetition, the reference
2258 can never match.) The problem arises because, having left the group and
2259 moved on to the rest of the pattern, a later failure that backtracks into
2260 the group uses the captured value from the final iteration of the group
2261 rather than the correct earlier one. I have fixed this in PCRE by forcing
2262 any group that contains a reference to itself to be an atomic group; that
2263 is, there cannot be any backtracking into it once it has completed. This is
2264 similar to recursive and subroutine calls.
2265
2266
2267 Version 8.00 19-Oct-09
2268 ----------------------
2269
2270 1. The table for translating pcre_compile() error codes into POSIX error codes
2271 was out-of-date, and there was no check on the pcre_compile() error code
2272 being within the table. This could lead to an OK return being given in
2273 error.
2274
2275 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2276 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2277 in a Windows environment.
2278
2279 3. The pcregrep --count option prints the count for each file even when it is
2280 zero, as does GNU grep. However, pcregrep was also printing all files when
2281 --files-with-matches was added. Now, when both options are given, it prints
2282 counts only for those files that have at least one match. (GNU grep just
2283 prints the file name in this circumstance, but including the count seems
2284 more useful - otherwise, why use --count?) Also ensured that the
2285 combination -clh just lists non-zero counts, with no names.
2286
2287 4. The long form of the pcregrep -F option was incorrectly implemented as
2288 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2289 but it seems right to fix it, and I didn't think it was worth preserving
2290 the old behaviour.
2291
2292 5. The command line items --regex=pattern and --regexp=pattern were not
2293 recognized by pcregrep, which required --regex pattern or --regexp pattern
2294 (with a space rather than an '='). The man page documented the '=' forms,
2295 which are compatible with GNU grep; these now work.
2296
2297 6. No libpcreposix.pc file was created for pkg-config; there was just
2298 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2299
2300 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2301 when UCP support is not needed, by modifying the Python script that
2302 generates it from Unicode data files. This should not matter if the module
2303 is correctly used as a library, but I received one complaint about 50K of
2304 unwanted data. My guess is that the person linked everything into his
2305 program rather than using a library. Anyway, it does no harm.
2306
2307 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2308 was a minimum greater than 1 for a wide character in a possessive
2309 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2310 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2311 character. Chaos in the form of incorrect output or a compiling loop could
2312 result.
2313
2314 9. The restrictions on what a pattern can contain when partial matching is
2315 requested for pcre_exec() have been removed. All patterns can now be
2316 partially matched by this function. In addition, if there are at least two
2317 slots in the offset vector, the offset of the earliest inspected character
2318 for the match and the offset of the end of the subject are set in them when
2319 PCRE_ERROR_PARTIAL is returned.
2320
2321 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2322 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2323 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2324 and may be more useful for multi-segment matching.
2325
2326 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2327 used to be given if ever the end of the subject was reached; now it is
2328 given only if matching could not proceed because another character was
2329 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2330 string "Z", which now yields "no match" instead of "partial match". In the
2331 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2332 final character ended with (*FAIL).
2333
2334 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2335 if the pattern had a "must contain" character that was already found in the
2336 earlier partial match, unless partial matching was again requested. For
2337 example, with the pattern /dog.(body)?/, the "must contain" character is
2338 "g". If the first part-match was for the string "dog", restarting with
2339 "sbody" failed. This bug has been fixed.
2340
2341 13. The string returned by pcre_dfa_exec() after a partial match has been
2342 changed so that it starts at the first inspected character rather than the
2343 first character of the match. This makes a difference only if the pattern
2344 starts with a lookbehind assertion or \b or \B (\K is not supported by
2345 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2346 matching functions compatible, and I think it's the right thing to do.
2347
2348 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2349 so that the demonstration program is easily available in environments where
2350 PCRE has not been installed from source.
2351
2352 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2353 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2354 library.
2355
2356 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2357 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2358 is not the first non-POSIX option to be added. Clearly some people find
2359 these options useful.
2360
2361 17. If a caller to the POSIX matching function regexec() passes a non-zero
2362 value for nmatch with a NULL value for pmatch, the value of
2363 nmatch is forced to zero.
2364
2365 18. RunGrepTest did not have a test for the availability of the -u option of
2366 the diff command, as RunTest does. It now checks in the same way as
2367 RunTest, and also checks for the -b option.
2368
2369 19. If an odd number of negated classes containing just a single character
2370 interposed, within parentheses, between a forward reference to a named
2371 subpattern and the definition of the subpattern, compilation crashed with
2372 an internal error, complaining that it could not find the referenced
2373 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2374 [The bug was that it was starting one character too far in when skipping
2375 over the character class, thus treating the ] as data rather than
2376 terminating the class. This meant it could skip too much.]
2377
2378 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2379 /g option in pcretest when the pattern contains \K, which makes it possible
2380 to have an empty string match not at the start, even when the pattern is
2381 anchored. Updated pcretest and pcredemo to use this option.
2382
2383 21. If the maximum number of capturing subpatterns in a recursion was greater
2384 than the maximum at the outer level, the higher number was returned, but
2385 with unset values at the outer level. The correct (outer level) value is
2386 now given.
2387
2388 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2389 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2390 make it do so. The string so far is captured, making this feature
2391 compatible with Perl.
2392
2393 23. The tests have been re-organized, adding tests 11 and 12, to make it
2394 possible to check the Perl 5.10 features against Perl 5.10.
2395
2396 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2397 pattern matches a fixed length string. PCRE did not allow this; now it
2398 does. Neither allows recursion.
2399
2400 25. I finally figured out how to implement a request to provide the minimum
2401 length of subject string that was needed in order to match a given pattern.
2402 (It was back references and recursion that I had previously got hung up
2403 on.) This code has now been added to pcre_study(); it finds a lower bound
2404 to the length of subject needed. It is not necessarily the greatest lower
2405 bound, but using it to avoid searching strings that are too short does give
2406 some useful speed-ups. The value is available to calling programs via
2407 pcre_fullinfo().
2408
2409 26. While implementing 25, I discovered to my embarrassment that pcretest had
2410 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2411 study optimizations had never been tested with that matching function.
2412 Oops. What is worse, even when it was passed study data, there was a bug in
2413 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2414 were also very few tests of studied patterns with pcre_dfa_exec().
2415
2416 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2417 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2418 on the other side of the coin, they are no longer allowed to have different
2419 names, because these cannot be distinguished in PCRE, and this has caused
2420 confusion. (This is a difference from Perl.)
2421
2422 28. When duplicate subpattern names are present (necessarily with different
2423 numbers, as required by 27 above), and a test is made by name in a
2424 conditional pattern, either for a subpattern having been matched, or for
2425 recursion in such a pattern, all the associated numbered subpatterns are
2426 tested, and the overall condition is true if the condition is true for any
2427 one of them. This is the way Perl works, and is also more like the way
2428 testing by number works.
2429
2430
2431 Version 7.9 11-Apr-09
2432 ---------------------
2433
2434 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2435 (pcretest), all targets were linked against these libraries. This included
2436 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2437 libraries. This caused unwanted dependencies to be created. This problem
2438 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2439 pcretest is linked with readline.
2440
2441 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2442 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2443 moved outside it again, because FALSE and TRUE are already defined in AIX,
2444 but BOOL is not.
2445
2446 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2447 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2448
2449 4. The pcregrep documentation said spaces were inserted as well as colons (or
2450 hyphens) following file names and line numbers when outputting matching
2451 lines. This is not true; no spaces are inserted. I have also clarified the
2452 wording for the --colour (or --color) option.
2453
2454 5. In pcregrep, when --colour was used with -o, the list of matching strings
2455 was not coloured; this is different to GNU grep, so I have changed it to be
2456 the same.
2457
2458 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2459 each matching line was coloured. Now it goes on to look for further matches
2460 of any of the test patterns, which is the same behaviour as GNU grep.
2461
2462 7. A pattern that could match an empty string could cause pcregrep to loop; it
2463 doesn't make sense to accept an empty string match in pcregrep, so I have
2464 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2465 seems to be how GNU grep behaves. [But see later change 40 for release
2466 8.33.]
2467
2468 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2469 start or after a newline", because the conditional assertion was not being
2470 correctly handled. The rule now is that both the assertion and what follows
2471 in the first alternative must satisfy the test.
2472
2473 9. If auto-callout was enabled in a pattern with a conditional group whose
2474 condition was an assertion, PCRE could crash during matching, both with
2475 pcre_exec() and pcre_dfa_exec().
2476
2477 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2478 used for matching.
2479
2480 11. Unicode property support in character classes was not working for
2481 characters (bytes) greater than 127 when not in UTF-8 mode.
2482
2483 12. Added the -M command line option to pcretest.
2484
2485 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2486
2487 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2488
2489 16. Added comments and documentation about mis-use of no_arg in the C++
2490 wrapper.
2491
2492 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2493 from Martin Jerabek that uses macro names for all relevant character and
2494 string constants.
2495
2496 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2497 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2498 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2499 these, but not everybody uses configure.
2500
2501 19. A conditional group that had only one branch was not being correctly
2502 recognized as an item that could match an empty string. This meant that an
2503 enclosing group might also not be so recognized, causing infinite looping
2504 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2505 with the subject "ab", where knowledge that the repeated group can match
2506 nothing is needed in order to break the loop.
2507
2508 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2509 exec(), but without supplying a callout function, matching went wrong.
2510
2511 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2512 leak if the size of the offset vector was greater than 30. When the vector
2513 is smaller, the saved offsets during recursion go onto a local stack
2514 vector, but for larger vectors malloc() is used. It was failing to free
2515 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2516 error, in fact).
2517
2518 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2519 heapframe that is used only when UTF-8 support is enabled. This caused no
2520 problem, but was untidy.
2521
2522 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2523 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2524 included within another project.
2525
2526 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2527 slightly modified by me:
2528
2529 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2530 not building pcregrep.
2531
2532 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2533 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2534
2535 25. Forward references, both numeric and by name, in patterns that made use of
2536 duplicate group numbers, could behave incorrectly or give incorrect errors,
2537 because when scanning forward to find the reference group, PCRE was not
2538 taking into account the duplicate group numbers. A pattern such as
2539 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2540
2541 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2542 the feature of a custom pointer more persuasive (as requested by a user).
2543
2544 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2545 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2546 pre-defined.
2547
2548 28. Added support for (*UTF8) at the start of a pattern.
2549
2550 29. Arrange for flags added by the "release type" setting in CMake to be shown
2551 in the configuration summary.
2552
2553
2554 Version 7.8 05-Sep-08
2555 ---------------------
2556
2557 1. Replaced UCP searching code with optimized version as implemented for Ad
2558 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2559 stage table and inline lookup instead of a function, giving speed ups of 2
2560 to 5 times on some simple patterns that I tested. Permission was given to
2561 distribute the MultiStage2.py script that generates the tables (it's not in
2562 the tarball, but is in the Subversion repository).
2563
2564 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2565 scripts.
2566
2567 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2568 a group with a zero qualifier. The result of the study could be incorrect,
2569 or the function might crash, depending on the pattern.
2570
2571 4. Caseless matching was not working for non-ASCII characters in back
2572 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2573 It now works when Unicode Property Support is available.
2574
2575 5. In pcretest, an escape such as \x{de} in the data was always generating
2576 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2577 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2578 truncation.
2579
2580 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2581
2582 7. Added two (int) casts to pcregrep when printing the difference of two
2583 pointers, in case they are 64-bit values.
2584
2585 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2586 test 2 if it fails.
2587
2588 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2589 and a #define of that name to empty if it is not externally set. This is to
2590 allow users of MSVC to set it if necessary.
2591
2592 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2593 the convenience functions in the pcre_get.c source file.
2594
2595 11. An option change at the start of a pattern that had top-level alternatives
2596 could cause overwriting and/or a crash. This command provoked a crash in
2597 some environments:
2598
2599 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2600
2601 This potential security problem was recorded as CVE-2008-2371.
2602
2603 12. For a pattern where the match had to start at the beginning or immediately
2604 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2605 pcre_dfa_exec() could read past the end of the passed subject if there was
2606 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2607 pcretest so that it places the subject at the end of its malloc-ed buffer.
2608
2609 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2610 exec() might read past the end of the data buffer in UTF-8 mode.
2611
2612 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2613 the data contained the byte 0x85 as part of a UTF-8 character within its
2614 first line. This applied both to normal and DFA matching.
2615
2616 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2617 /^[^d]*?$/8 failed to match "abc".
2618
2619 16. Added a missing copyright notice to pcrecpp_internal.h.
2620
2621 17. Make it more clear in the documentation that values returned from
2622 pcre_exec() in ovector are byte offsets, not character counts.
2623
2624 18. Tidied a few places to stop certain compilers from issuing warnings.
2625
2626 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2627 supplied by Stefan Weber. I made a further small update for 7.8 because
2628 there is a change of source arrangements: the pcre_searchfuncs.c module is
2629 replaced by pcre_ucd.c.
2630
2631
2632 Version 7.7 07-May-08
2633 ---------------------
2634
2635 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2636 a string to a long long, pretend we don't even have a long long." This is
2637 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2638
2639 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2640 pre-7.6 versions, which defined a global no_arg variable instead of putting
2641 it in the RE class. (See also #8 below.)
2642
2643 3. Remove a line of dead code, identified by coverity and reported by Nuno
2644 Lopes.
2645
2646 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2647
2648 (1) The include/exclude patterns were being applied to the whole pathnames
2649 of files, instead of just to the final components.
2650
2651 (2) If there was more than one level of directory, the subdirectories were
2652 skipped unless they satisfied the include/exclude conditions. This is
2653 inconsistent with GNU grep (and could even be seen as contrary to the
2654 pcregrep specification - which I improved to make it absolutely clear).
2655 The action now is always to scan all levels of directory, and just
2656 apply the include/exclude patterns to regular files.
2657
2658 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2659 --exclude_dir in the tests to avoid scanning .svn directories.
2660
2661 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2662 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2663 doesn't support NULs in patterns.
2664
2665 7. Added some missing "const"s to declarations of static tables in
2666 pcre_compile.c and pcre_dfa_exec.c.
2667
2668 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2669 caused by fix #2 above. (Subsequently also a second patch to fix the
2670 first patch. And a third patch - this was a messy problem.)
2671
2672 9. Applied Craig's patch to remove the use of push_back().
2673
2674 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2675 matching function regexec().
2676
2677 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2678 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2679 references. PCRE supports relative numbers with this syntax (I don't think
2680 Oniguruma does).
2681
2682 12. Previously, a group with a zero repeat such as (...){0} was completely
2683 omitted from the compiled regex. However, this means that if the group
2684 was called as a subroutine from elsewhere in the pattern, things went wrong
2685 (an internal error was given). Such groups are now left in the compiled
2686 pattern, with a new opcode that causes them to be skipped at execution
2687 time.
2688
2689 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2690 to the way PCRE behaves:
2691
2692 (a) A lone ] character is dis-allowed (Perl treats it as data).
2693
2694 (b) A back reference to an unmatched subpattern matches an empty string
2695 (Perl fails the current match path).
2696
2697 (c) A data ] in a character class must be notated as \] because if the
2698 first data character in a class is ], it defines an empty class. (In
2699 Perl it is not possible to have an empty class.) The empty class []
2700 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2701 The negative empty class [^] matches any one character, independently
2702 of the DOTALL setting.
2703
2704 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2705 non-existent subpattern following a character class starting with ']' and
2706 containing () gave an internal compiling error instead of "reference to
2707 non-existent subpattern". Fortunately, when the pattern did exist, the
2708 compiled code was correct. (When scanning forwards to check for the
2709 existence of the subpattern, it was treating the data ']' as terminating
2710 the class, so got the count wrong. When actually compiling, the reference
2711 was subsequently set up correctly.)
2712
2713 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2714 it was being rejected as not supported by pcre_dfa_exec(), even though
2715 other assertions are supported. I have made pcre_dfa_exec() support
2716 (*FAIL).
2717
2718 16. The implementation of 13c above involved the invention of a new opcode,
2719 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2720 cannot be changed at match time, I realized I could make a small
2721 improvement to matching performance by compiling OP_ALLANY instead of
2722 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2723 on the OP_ANY path.
2724
2725 17. Compiling pcretest on Windows with readline support failed without the
2726 following two fixes: (1) Make the unistd.h include conditional on
2727 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2728
2729 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2730 ncurses library to be included for pcretest when ReadLine support is
2731 requested, but also to allow for it to be overridden. This patch came from
2732 Daniel Bergström.
2733
2734 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2735 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2736 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2737 spotting this.
2738
2739
2740 Version 7.6 28-Jan-08
2741 ---------------------
2742
2743 1. A character class containing a very large number of characters with
2744 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2745 overflow.
2746
2747 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2748 HAVE_LONG_LONG is not defined.
2749
2750 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2751 bring them up to date and include new features. This patch includes:
2752
2753 - Fixed PH's badly added libz and libbz2 support.
2754 - Fixed a problem with static linking.
2755 - Added pcredemo. [But later removed - see 7 below.]
2756 - Fixed dftables problem and added an option.
2757 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2758 HAVE_LONG_LONG.
2759 - Added readline support for pcretest.
2760 - Added an listing of the option settings after cmake has run.
2761
2762 4. A user submitted a patch to Makefile that makes it easy to create
2763 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2764 Makefile.am that cause it to include this special target, without
2765 affecting anything else. Note that the same mingw target plus all
2766 the other distribution libraries and programs are now supported
2767 when configuring with CMake (see 6 below) instead of with
2768 Configure/Make.
2769
2770 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2771 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2772 exported in the Windows port". It has not yet been confirmed that the patch
2773 solves the problem, but it does no harm.
2774
2775 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2776 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2777 with CMake, and also correct the comment about stack recursion.
2778
2779 7. Remove the automatic building of pcredemo from the ./configure system and
2780 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2781 of a program that users should build themselves after PCRE is installed, so
2782 building it automatically is not really right. What is more, it gave
2783 trouble in some build environments.
2784
2785 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2786
2787
2788 Version 7.5 10-Jan-08
2789 ---------------------
2790
2791 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2792 values in parens when parsing an RE using the C++ wrapper."
2793
2794 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2795 Characters greater than 255 were excluded from the class instead of being
2796 included.
2797
2798 3. The same bug as (2) above applied to negated POSIX classes such as
2799 [:^space:].
2800
2801 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2802 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2803 I have changed it.
2804
2805 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2806 first named subpattern) and a construct such as (?&a) would reference the
2807 first named subpattern whose name started with "a" (in other words, the
2808 length check was missing). Both these problems are fixed. "Subpattern name
2809 expected" is now given for (?&) (a zero-length name), and this patch also
2810 makes it give the same error for \k'' (previously it complained that that
2811 was a reference to a non-existent subpattern).
2812
2813 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2814 this is right because (?- can be followed by option settings as well as by
2815 digits. I have, however, made the messages clearer.
2816
2817 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2818 than the number used in the conditional) now cause a compile-time error.
2819 This is actually not compatible with Perl, which accepts such patterns, but
2820 treats the conditional as always being FALSE (as PCRE used to), but it
2821 seems to me that giving a diagnostic is better.
2822
2823 8. Change "alphameric" to the more common word "alphanumeric" in comments
2824 and messages.
2825
2826 9. Fix two occurrences of "backslash" in comments that should have been
2827 "backspace".
2828
2829 10. Remove two redundant lines of code that can never be obeyed (their function
2830 was moved elsewhere).
2831
2832 11. The program that makes PCRE's Unicode character property table had a bug
2833 which caused it to generate incorrect table entries for sequences of
2834 characters that have the same character type, but are in different scripts.
2835 It amalgamated them into a single range, with the script of the first of
2836 them. In other words, some characters were in the wrong script. There were
2837 thirteen such cases, affecting characters in the following ranges:
2838
2839 U+002b0 - U+002c1
2840 U+0060c - U+0060d
2841 U+0061e - U+00612
2842 U+0064b - U+0065e
2843 U+0074d - U+0076d
2844 U+01800 - U+01805
2845 U+01d00 - U+01d77
2846 U+01d9b - U+01dbf
2847 U+0200b - U+0200f
2848 U+030fc - U+030fe
2849 U+03260 - U+0327f
2850 U+0fb46 - U+0fbb1
2851 U+10450 - U+1049d
2852
2853 12. The -o option (show only the matching part of a line) for pcregrep was not
2854 compatible with GNU grep in that, if there was more than one match in a
2855 line, it showed only the first of them. It now behaves in the same way as
2856 GNU grep.
2857
2858 13. If the -o and -v options were combined for pcregrep, it printed a blank
2859 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2860 does the same. The return code can be used to tell if there were any
2861 non-matching lines.
2862
2863 14. Added --file-offsets and --line-offsets to pcregrep.
2864
2865 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2866 infinitely looping recursion. The bug was that positive lookaheads were not
2867 being skipped when checking for a possible empty match (negative lookaheads
2868 and both kinds of lookbehind were skipped).
2869
2870 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2871 inclusion of <windows.h> to before rather than after the definition of
2872 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2873
2874 17. Specifying a possessive quantifier with a specific limit for a Unicode
2875 character property caused pcre_compile() to compile bad code, which led at
2876 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2877 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2878 caused the error; without that there was no problem.
2879
2880 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2881
2882 19. Added --enable-pcretest-libreadline.
2883
2884 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2885 RE::GlobalReplace(). As a result, the number of replacements returned was
2886 double what it should be. I removed one of the increments, but Craig sent a
2887 later patch that removed the other one (the right fix) and added unit tests
2888 that check the return values (which was not done before).
2889
2890 21. Several CMake things:
2891
2892 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2893 the names libpcre and libpcreposix, not just pcre and pcreposix.
2894
2895 (2) The above change means that pcretest and pcregrep are now correctly
2896 linked with the newly-built libraries, not previously installed ones.
2897
2898 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2899
2900 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2901 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2902 UTF-8 newline character). The key issue is that the pattern starts .*;
2903 this means that the match must be either at the beginning, or after a
2904 newline. The bug was in the code for advancing after a failed match and
2905 checking that the new position followed a newline. It was not taking
2906 account of UTF-8 characters correctly.
2907
2908 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2909 character classes. PCRE was not treating the sequence [:...:] as a
2910 character class unless the ... were all letters. Perl, however, seems to
2911 allow any characters between [: and :], though of course it rejects as
2912 unknown any "names" that contain non-letters, because all the known class
2913 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2914 for example, whereas PCRE did not - it did not recognize a POSIX character
2915 class. This seemed a bit dangerous, so the code has been changed to be
2916 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2917 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2918 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2919 Perl does, and where it didn't before.
2920
2921 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2922 Windows environments %n is disabled by default.
2923
2924
2925 Version 7.4 21-Sep-07
2926 ---------------------
2927
2928 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2929 means that a class such as [\s] counted as "explicit reference to CR or
2930 LF". That isn't really right - the whole point of the change was to try to
2931 help when there was an actual mention of one of the two characters. So now
2932 the change happens only if \r or \n (or a literal CR or LF) character is
2933 encountered.
2934
2935 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2936 of both had grown to the point where there were only 3 bits left.
2937 Fortunately, there was spare space in the data structure, and so I have
2938 moved the internal flags into a new 16-bit field to free up more option
2939 bits.
2940
2941 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2942 but did not set the internal JCHANGED flag - either of these is enough to
2943 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2944 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2945 start sets both bits.
2946
2947 4. Added options (at build time, compile time, exec time) to change \R from
2948 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2949
2950 5. doc/pcresyntax.html was missing from the distribution.
2951
2952 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2953 compatibility, even though it is no longer used.
2954
2955 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2956 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2957 windows.h file is present (where different names are used). [This was
2958 reversed later after testing - see 16 below.]
2959
2960 8. Changed all #include <config.h> to #include "config.h". There were also
2961 some further <pcre.h> cases that I changed to "pcre.h".
2962
2963 9. When pcregrep was used with the --colour option, it missed the line ending
2964 sequence off the lines that it output.
2965
2966 10. It was pointed out to me that arrays of string pointers cause lots of
2967 relocations when a shared library is dynamically loaded. A technique of
2968 using a single long string with a table of offsets can drastically reduce
2969 these. I have refactored PCRE in four places to do this. The result is
2970 dramatic:
2971
2972 Originally: 290
2973 After changing UCP table: 187
2974 After changing error message table: 43
2975 After changing table of "verbs" 36
2976 After changing table of Posix names 22
2977
2978 Thanks to the folks working on Gregex for glib for this insight.
2979
2980 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2981 unicode-properties was also set.
2982
2983 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2984
2985 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2986 checked only for CRLF.
2987
2988 14. Added casts to pcretest.c to avoid compiler warnings.
2989
2990 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2991
2992 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2993 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2994 entirely. This removes changes made in 7 above.
2995
2996 17. The CMake files have been updated, and there is now more information about
2997 building with CMake in the NON-UNIX-USE document.
2998
2999
3000 Version 7.3 28-Aug-07
3001 ---------------------
3002
3003 1. In the rejigging of the build system that eventually resulted in 7.1, the
3004 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3005 brackets there is not right, since it causes compilers to look for an
3006 installed pcre.h, not the version that is in the source that is being
3007 compiled (which of course may be different). I have changed it back to:
3008
3009 #include "pcre.h"
3010
3011 I have a vague recollection that the change was concerned with compiling in
3012 different directories, but in the new build system, that is taken care of
3013 by the VPATH setting the Makefile.
3014
3015 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3016 when the subject happened to end in the byte 0x85 (e.g. if the last
3017 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3018 characters but of course it shouldn't be taken as a newline when it is part
3019 of another character. The bug was that, for an unlimited repeat of . in
3020 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3021 characters when looking for a newline.
3022
3023 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3024
3025 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3026 in debug output.
3027
3028 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3029 long printing in the pcrecpp unittest when running under MinGW.
3030
3031 6. ESC_K was left out of the EBCDIC table.
3032
3033 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3034 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3035 limit also applies to "virtual nesting" when a pattern is recursive, and in
3036 this case 1000 isn't so big. I have been able to remove this limit at the
3037 expense of backing off one optimization in certain circumstances. Normally,
3038 when pcre_exec() would call its internal match() function recursively and
3039 immediately return the result unconditionally, it uses a "tail recursion"
3040 feature to save stack. However, when a subpattern that can match an empty
3041 string has an unlimited repetition quantifier, it no longer makes this
3042 optimization. That gives it a stack frame in which to save the data for
3043 checking that an empty string has been matched. Previously this was taken
3044 from the 1000-entry workspace that had been reserved. So now there is no
3045 explicit limit, but more stack is used.
3046
3047 8. Applied Daniel's patches to solve problems with the import/export magic
3048 syntax that is required for Windows, and which was going wrong for the
3049 pcreposix and pcrecpp parts of the library. These were overlooked when this
3050 problem was solved for the main library.
3051
3052 9. There were some crude static tests to avoid integer overflow when computing
3053 the size of patterns that contain repeated groups with explicit upper
3054 limits. As the maximum quantifier is 65535, the maximum group length was
3055 set at 30,000 so that the product of these two numbers did not overflow a
3056 32-bit integer. However, it turns out that people want to use groups that
3057 are longer than 30,000 bytes (though not repeat them that many times).
3058 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3059 made it possible to implement the integer overflow checks in a much more
3060 dynamic way, which I have now done. The artificial limitation on group
3061 length has been removed - we now have only the limit on the total length of
3062 the compiled pattern, which depends on the LINK_SIZE setting.
3063
3064 10. Fixed a bug in the documentation for get/copy named substring when
3065 duplicate names are permitted. If none of the named substrings are set, the
3066 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3067 empty string.
3068
3069 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3070 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3071 because the ] is interpreted as the first data character and the
3072 terminating ] is not found. PCRE has been made compatible with Perl in this
3073 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3074 cause memory overwriting.
3075
3076 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3077 string has been matched (to stop an infinite loop). It was not recognizing
3078 a conditional subpattern that could match an empty string if that
3079 subpattern was within another subpattern. For example, it looped when
3080 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3081 condition was not nested. This bug has been fixed.
3082
3083 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3084 past the start of the subject in the presence of bytes with the top bit
3085 set, for example "\x8aBCD".
3086
3087 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3088 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3089
3090 14. Optimized (?!) to (*FAIL).
3091
3092 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3093 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3094 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3095 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3096 does: it's just the validity check that is more restrictive.
3097
3098 16. Inserted checks for integer overflows during escape sequence (backslash)
3099 processing, and also fixed erroneous offset values for syntax errors during
3100 backslash processing.
3101
3102 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3103 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3104
3105 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3106 caused an overrun.
3107
3108 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3109 something other than just ASCII characters) inside a group that had an
3110 unlimited repeat caused a loop at compile time (while checking to see
3111 whether the group could match an empty string).
3112
3113 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3114 [\P{Any}] did so. (Error in the code for printing property names.)
3115
3116 21. An orphan \E inside a character class could cause a crash.
3117
3118 22. A repeated capturing bracket such as (A)? could cause a wild memory
3119 reference during compilation.
3120
3121 23. There are several functions in pcre_compile() that scan along a compiled
3122 expression for various reasons (e.g. to see if it's fixed length for look
3123 behind). There were bugs in these functions when a repeated \p or \P was
3124 present in the pattern. These operators have additional parameters compared
3125 with \d, etc, and these were not being taken into account when moving along
3126 the compiled data. Specifically:
3127
3128 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3129 length.
3130
3131 (b) An item such as \pL+ within a repeated group could cause crashes or
3132 loops.
3133
3134 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3135 "reference to non-existent subpattern" error.
3136
3137 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3138
3139 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3140 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3141
3142 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3143
3144 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3145 character were causing crashes (broken optimization).
3146
3147 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3148 \p or \P) caused a compile-time loop.
3149
3150 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3151 break. For example, the unstudied pattern [\r\n]A does not match the string
3152 "\r\nA" because change 7.0/46 below moves the current point on by two
3153 characters after failing to match at the start. However, the pattern \nA
3154 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3155 the same is true. There doesn't seem any very clean way out of this, but
3156 what I have chosen to do makes the common cases work: PCRE now takes note
3157 of whether there can be an explicit match for \r or \n anywhere in the
3158 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3159 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3160 pattern has explicit CR or LF references.
3161
3162 29. Added (*CR) etc for changing newline setting at start of pattern.
3163
3164
3165 Version 7.2 19-Jun-07
3166 ---------------------
3167
3168 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3169 which is apparently normally available under Windows.
3170
3171 2. Re-jig the pcregrep tests with different newline settings in an attempt
3172 to make them independent of the local environment's newline setting.
3173
3174 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3175
3176 4. Some of the "internals" tests were previously cut out when the link size
3177 was not 2, because the output contained actual offsets. The recent new
3178 "Z" feature of pcretest means that these can be cut out, making the tests
3179 usable with all link sizes.
3180
3181 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3182 stack recursion. This gives a massive performance boost under BSD, but just
3183 a small improvement under Linux. However, it saves one field in the frame
3184 in all cases.
3185
3186 6. Added more features from the forthcoming Perl 5.10:
3187
3188 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3189 recursion call. It refers to the nth most recently opened parentheses.
3190
3191 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3192 to be opened parentheses.
3193
3194 (c) Conditions that refer to capturing parentheses can be specified
3195 relatively, for example, (?(-2)... or (?(+3)...
3196
3197 (d) \K resets the start of the current match so that everything before
3198 is not part of it.
3199
3200 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3201
3202 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3203 reference syntax.
3204
3205 (g) (?| introduces a group in which the numbering of parentheses in each
3206 alternative starts with the same number.
3207
3208 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3209
3210 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3211 PCRE_INFO_JCHANGED.
3212
3213 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3214 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3215 for detecting groups that can match an empty string.
3216
3217 9. A pattern with a very large number of alternatives (more than several
3218 hundred) was running out of internal workspace during the pre-compile
3219 phase, where pcre_compile() figures out how much memory will be needed. A
3220 bit of new cunning has reduced the workspace needed for groups with
3221 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3222 workspace instead of running out of the 4096 that are available.
3223
3224 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3225
3226 11. Applied patch from Google to remove an optimization that didn't quite work.
3227 The report of the bug said:
3228
3229 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3230 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3231 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3232
3233 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3234 it matched the wrong number of bytes.
3235
3236
3237 Version 7.1 24-Apr-07
3238 ---------------------
3239
3240 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3241 that is more "standard", making use of automake and other Autotools. There
3242 is some re-arrangement of the files and adjustment of comments consequent
3243 on this.
3244
3245 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3246 for recursive directory scanning broke on some systems because the files
3247 are not scanned in any specific order and on different systems the order
3248 was different. A call to "sort" has been inserted into RunGrepTest for the
3249 approprate test as a short-term fix. In the longer term there may be an
3250 alternative.
3251
3252 3. I had an email from Eric Raymond about problems translating some of PCRE's
3253 man pages to HTML (despite the fact that I distribute HTML pages, some
3254 people do their own conversions for various reasons). The problems
3255 concerned the use of low-level troff macros .br and .in. I have therefore
3256 removed all such uses from the man pages (some were redundant, some could
3257 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3258 HTML has been updated to handle .nf/.fi and to complain if it encounters
3259 .br or .in.
3260
3261 4. Updated comments in configure.ac that get placed in config.h.in and also
3262 arranged for config.h to be included in the distribution, with the name
3263 config.h.generic, for the benefit of those who have to compile without
3264 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3265
3266 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3267 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3268 makevp.bat for the current PCRE, using the additional files
3269 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3270
3271 6. A Windows user reported a minor discrepancy with test 2, which turned out
3272 to be caused by a trailing space on an input line that had got lost in his
3273 copy. The trailing space was an accident, so I've just removed it.
3274
3275 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3276 that is needed.
3277
3278 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3279 as "const" (a) because they are and (b) because it helps the PHP
3280 maintainers who have recently made a script to detect big data structures
3281 in the php code that should be moved to the .rodata section. I remembered
3282 to update Builducptable as well, so it won't revert if ucptable.h is ever
3283 re-created.
3284
3285 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3286 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3287 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3288 support is not required. This saves 1.5-2K of code, which is important in
3289 some applications.
3290
3291 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3292 so as not to refer to the tables, even though these functions will never be
3293 called when UTF-8 support is disabled. Otherwise there are problems with a
3294 shared library.
3295
3296 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3297
3298 (a) It was defining its arguments as char * instead of void *.
3299
3300 (b) It was assuming that all moves were upwards in memory; this was true
3301 a long time ago when I wrote it, but is no longer the case.
3302
3303 The emulated memove() is provided for those environments that have neither
3304 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3305 is clearly not the case, as these two bugs were recently reported.
3306
3307 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3308 and Detrail to create the HTML documentation, the .txt form of the man
3309 pages, and it removes trailing spaces from listed files. It also creates
3310 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3311 case, it wraps all the #defines with #ifndefs. This script should be run
3312 before "make dist".
3313
3314 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3315 with Unicode property support.
3316
3317 (a) For a maximizing quantifier, if the two different cases of the
3318 character were of different lengths in their UTF-8 codings (there are
3319 some cases like this - I found 11), and the matching function had to
3320 back up over a mixture of the two cases, it incorrectly assumed they
3321 were both the same length.
3322
3323 (b) When PCRE was configured to use the heap rather than the stack for
3324 recursion during matching, it was not correctly preserving the data for
3325 the other case of a UTF-8 character when checking ahead for a match
3326 while processing a minimizing repeat. If the check also involved
3327 matching a wide character, but failed, corruption could cause an
3328 erroneous result when trying to check for a repeat of the original
3329 character.
3330
3331 13. Some tidying changes to the testing mechanism:
3332
3333 (a) The RunTest script now detects the internal link size and whether there
3334 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3335 values substituted by "configure". (The RunGrepTest script already did
3336 this for UTF-8.) The configure.ac script no longer substitutes the
3337 relevant variables.
3338
3339 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3340 with length and offset values. This means that the output is different
3341 for different internal link sizes. Test 2 is skipped for link sizes
3342 other than 2 because of this, bypassing the problem. Unfortunately,
3343 there was also a test in test 3 (the locale tests) that used /B and
3344 failed for link sizes other than 2. Rather than cut the whole test out,
3345 I have added a new /Z option to pcretest that replaces the length and
3346 offset values with spaces. This is now used to make test 3 independent
3347 of link size. (Test 2 will be tidied up later.)
3348
3349 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3350 segmentation fault instead of returning the appropriate error message.
3351
3352 15. In multiline mode when the newline sequence was set to "any", the pattern
3353 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3354 This doesn't seem right; it now treats the CRLF combination as the line
3355 ending, and so does not match in that case. It's only a pattern such as ^$
3356 that would hit this one: something like ^ABC$ would have failed after \r
3357 and then tried again after \r\n.
3358
3359 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3360 in an attempt to make files that differ only in their line terminators
3361 compare equal. This works on Linux.
3362
3363 17. Under certain error circumstances pcregrep might try to free random memory
3364 as it exited. This is now fixed, thanks to valgrind.
3365
3366 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3367 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3368 was because its rules for how to advance for /g after matching an empty
3369 string at the end of a line did not allow for this case. They now check for
3370 it specially.
3371
3372 20. pcretest is supposed to handle patterns and data of any length, by
3373 extending its buffers when necessary. It was getting this wrong when the
3374 buffer for a data line had to be extended.
3375
3376 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3377 CRLF as a newline sequence.
3378
3379 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3380 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3381 I have nevertheless tidied it up.
3382
3383 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3384
3385 24. Added a man page for pcre-config.
3386
3387
3388 Version 7.0 19-Dec-06
3389 ---------------------
3390
3391 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3392 moving to gcc 4.1.1.
3393
3394 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3395 sys/time.h, which is documented as needed for this function. It doesn't
3396 seem to matter on Linux, but it showed up on some releases of OS X.
3397
3398 3. It seems that there are systems where bytes whose values are greater than
3399 127 match isprint() in the "C" locale. The "C" locale should be the
3400 default when a C program starts up. In most systems, only ASCII printing
3401 characters match isprint(). This difference caused the output from pcretest
3402 to vary, making some of the tests fail. I have changed pcretest so that:
3403
3404 (a) When it is outputting text in the compiled version of a pattern, bytes
3405 other than 32-126 are always shown as hex escapes.
3406
3407 (b) When it is outputting text that is a matched part of a subject string,
3408 it does the same, unless a different locale has been set for the match
3409 (using the /L modifier). In this case, it uses isprint() to decide.
3410
3411 4. Fixed a major bug that caused incorrect computation of the amount of memory
3412 required for a compiled pattern when options that changed within the
3413 pattern affected the logic of the preliminary scan that determines the
3414 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3415 that the computed length was too small. The symptoms of this bug were
3416 either the PCRE error "internal error: code overflow" from pcre_compile(),
3417 or a glibc crash with a message such as "pcretest: free(): invalid next
3418 size (fast)". Examples of patterns that provoked this bug (shown in
3419 pcretest format) are:
3420
3421 /(?-x: )/x
3422 /(?x)(?-x: \s*#\s*)/
3423 /((?i)[\x{c0}])/8
3424 /(?i:[\x{c0}])/8
3425
3426 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3427 is now done differently.
3428
3429 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3430 wrapper classes; (b) implement a new function in the C++ scanner that is
3431 more efficient than the old way of doing things because it avoids levels of
3432 recursion in the regex matching; (c) add a paragraph to the documentation
3433 for the FullMatch() function.
3434
3435 6. The escape sequence \n was being treated as whatever was defined as
3436 "newline". Not only was this contrary to the documentation, which states
3437 that \n is character 10 (hex 0A), but it also went horribly wrong when
3438 "newline" was defined as CRLF. This has been fixed.
3439
3440 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3441 was being set to -1 for the "end of line" case (supposedly a value that no
3442 character can have). Though this value is never used (the check for end of
3443 line is "zero bytes in current character"), it caused compiler complaints.
3444 I've changed it to 0xffffffff.
3445
3446 8. In pcre_version.c, the version string was being built by a sequence of
3447 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3448 string (as it is for production releases) called a macro with an empty
3449 argument. The C standard says the result of this is undefined. The gcc
3450 compiler treats it as an empty string (which was what was wanted) but it is
3451 reported that Visual C gives an error. The source has been hacked around to
3452 avoid this problem.
3453
3454 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3455 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3456 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3457 of them did).
3458
3459 10. Originally, pcretest opened its input and output without "b"; then I was
3460 told that "b" was needed in some environments, so it was added for release
3461 5.0 to both the input and output. (It makes no difference on Unix-like
3462 systems.) Later I was told that it is wrong for the input on Windows. I've
3463 now abstracted the modes into two macros, to make it easier to fiddle with
3464 them, and removed "b" from the input mode under Windows.
3465
3466 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3467
3468 12. Added -help and --help to pcretest as an official way of being reminded
3469 of the options.
3470
3471 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3472 and pcrecpp.cc because they annoy compilers at high warning levels.
3473
3474 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3475
3476 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3477 scripts are not C programs :-) and which was not noticed because it works
3478 on Linux.
3479
3480 16. pcretest is supposed to handle any length of pattern and data line (as one
3481 line or as a continued sequence of lines) by extending its input buffer if
3482 necessary. This feature was broken for very long pattern lines, leading to
3483 a string of junk being passed to pcre_compile() if the pattern was longer
3484 than about 50K.
3485
3486 17. I have done a major re-factoring of the way pcre_compile() computes the
3487 amount of memory needed for a compiled pattern. Previously, there was code
3488 that made a preliminary scan of the pattern in order to do this. That was
3489 OK when PCRE was new, but as the facilities have expanded, it has become
3490 harder and harder to keep it in step with the real compile phase, and there
3491 have been a number of bugs (see for example, 4 above). I have now found a
3492 cunning way of running the real compile function in a "fake" mode that
3493 enables it to compute how much memory it would need, while actually only
3494 ever using a few hundred bytes of working memory and without too many
3495 tests of the mode. This should make future maintenance and development
3496 easier. A side effect of this work is that the limit of 200 on the nesting
3497 depth of parentheses has been removed (though this was never a serious
3498 limitation, I suspect). However, there is a downside: pcre_compile() now
3499 runs more slowly than before (30% or more, depending on the pattern). I
3500 hope this isn't a big issue. There is no effect on runtime performance.
3501
3502 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3503 newline (only possible for the last line of a file) and it was a
3504 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3505
3506 19. Added additional timing features to pcretest. (1) The -tm option now times
3507 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3508 separate command line item, by a number that specifies the number of
3509 repeats to use when timing. The default is 50000; this gives better
3510 precision, but takes uncomfortably long for very large patterns.
3511
3512 20. Extended pcre_study() to be more clever in cases where a branch of a
3513 subpattern has no definite first character. For example, (a*|b*)[cd] would
3514 previously give no result from pcre_study(). Now it recognizes that the
3515 first character must be a, b, c, or d.
3516
3517 21. There was an incorrect error "recursive call could loop indefinitely" if
3518 a subpattern (or the entire pattern) that was being tested for matching an
3519 empty string contained only one non-empty item after a nested subpattern.
3520 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3521 incorrectly, because the \d was being skipped in the check.
3522
3523 22. The pcretest program now has a new pattern option /B and a command line
3524 option -b, which is equivalent to adding /B to every pattern. This causes
3525 it to show the compiled bytecode, without the additional information that
3526 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3527 is the same as /B/I).
3528
3529 23. A new optimization is now able automatically to treat some sequences such
3530 as a*b as a*+b. More specifically, if something simple (such as a character
3531 or a simple class like \d) has an unlimited quantifier, and is followed by
3532 something that cannot possibly match the quantified thing, the quantifier
3533 is automatically "possessified".
3534
3535 24. A recursive reference to a subpattern whose number was greater than 39
3536 went wrong under certain circumstances in UTF-8 mode. This bug could also
3537 have affected the operation of pcre_study().
3538
3539 25. Realized that a little bit of performance could be had by replacing
3540 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3541
3542 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3543
3544 27. Possessive quantifiers such as a++ were previously implemented by turning
3545 them into atomic groups such as ($>a+). Now they have their own opcodes,
3546 which improves performance. This includes the automatically created ones
3547 from 23 above.
3548
3549 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3550 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3551 the first matched character to be a colon. This applied both to named and
3552 numbered groups.
3553
3554 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3555
3556 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3557 building PCRE on Windows easier, so I have included it in the distribution.
3558
3559 31. There is now a check in pcretest against a ridiculously large number being
3560 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3561 loop, the loop is abandoned.
3562
3563 32. Forward references to subpatterns in conditions such as (?(2)...) where
3564 subpattern 2 is defined later cause pcre_compile() to search forwards in
3565 the pattern for the relevant set of parentheses. This search went wrong
3566 when there were unescaped parentheses in a character class, parentheses
3567 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3568
3569 33. "Subroutine" calls and backreferences were previously restricted to
3570 referencing subpatterns earlier in the regex. This restriction has now
3571 been removed.
3572
3573 34. Added a number of extra features that are going to be in Perl 5.10. On the
3574 whole, these are just syntactic alternatives for features that PCRE had
3575 previously implemented using the Python syntax or my own invention. The
3576 other formats are all retained for compatibility.
3577
3578 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3579 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3580 also .NET compatible.
3581
3582 (b) A recursion or subroutine call to a named group can now be defined as
3583 (?&name) as well as (?P>name).
3584
3585 (c) A backreference to a named group can now be defined as \k<name> or
3586 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3587 5.10, are also .NET compatible.
3588
3589 (d) A conditional reference to a named group can now use the syntax
3590 (?(<name>) or (?('name') as well as (?(name).
3591
3592 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3593 groups (named and numbered) that are never evaluated inline, but can be
3594 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3595 is always false. There may be only one alternative in such a group.
3596
3597 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3598 as the simple (?(R). The condition is true only if the most recent
3599 recursion is that of the given number or name. It does not search out
3600 through the entire recursion stack.
3601
3602 (g) The escape \gN or \g{N} has been added, where N is a positive or
3603 negative number, specifying an absolute or relative reference.
3604
3605 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3606 some "unreachable code" warnings.
3607
3608 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3609 things, this adds five new scripts.
3610
3611 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3612 There were also incompatibilities regarding the handling of \Q..\E inside
3613 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3614 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3615
3616 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3617 matches an empty string, and forcibly breaks the loop. There were bugs in
3618 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3619 against aaaa the result was just "a" rather than "aaaa", for example. Two
3620 separate and independent bugs (that affected different cases) have been
3621 fixed.
3622
3623 39. Refactored the code to abolish the use of different opcodes for small
3624 capturing bracket numbers. This is a tidy that I avoided doing when I
3625 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3626 The new approach is not only tidier, it makes it possible to reduce the
3627 memory needed to fix the previous bug (38).
3628
3629 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3630 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3631 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3632 mode.
3633
3634 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3635 report.
3636
3637 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3638 copy construction and assignment in the C++ wrapper.
3639
3640 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3641 couple of bugs that could have given wrong results in the "--newline=crlf"
3642 case.
3643
3644 44. Added a number of casts and did some reorganization of signed/unsigned int
3645 variables following suggestions from Dair Grant. Also renamed the variable
3646 "this" as "item" because it is a C++ keyword.
3647
3648 45. Arranged for dftables to add
3649
3650 #include "pcre_internal.h"
3651
3652 to pcre_chartables.c because without it, gcc 4.x may remove the array
3653 definition from the final binary if PCRE is built into a static library and
3654 dead code stripping is activated.
3655
3656 46. For an unanchored pattern, if a match attempt fails at the start of a
3657 newline sequence, and the newline setting is CRLF or ANY, and the next two
3658 characters are CRLF, advance by two characters instead of one.
3659
3660
3661 Version 6.7 04-Jul-06
3662 ---------------------
3663
3664 1. In order to handle tests when input lines are enormously long, pcretest has
3665 been re-factored so that it automatically extends its buffers when
3666 necessary. The code is crude, but this _is_ just a test program. The
3667 default size has been increased from 32K to 50K.
3668
3669 2. The code in pcre_study() was using the value of the re argument before
3670 testing it for NULL. (Of course, in any sensible call of the function, it
3671 won't be NULL.)
3672
3673 3. The memmove() emulation function in pcre_internal.h, which is used on
3674 systems that lack both memmove() and bcopy() - that is, hardly ever -
3675 was missing a "static" storage class specifier.
3676
3677 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3678 containing an extended class (one that cannot be represented by a bitmap
3679 because it contains high-valued characters or Unicode property items, e.g.
3680 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3681 pattern, but PCRE should not loop if you do not (it no longer does).
3682 [Detail: two cases were found: (a) a repeated subpattern containing an
3683 extended class; (b) a recursive reference to a subpattern that followed a
3684 previous extended class. It wasn't skipping over the extended class
3685 correctly when UTF-8 mode was not set.]
3686
3687 5. A negated single-character class was not being recognized as fixed-length
3688 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3689 compile error "lookbehind assertion is not fixed length".
3690
3691 6. The RunPerlTest auxiliary script was showing an unexpected difference
3692 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3693 write a Perl script that can interpret lines of an input file either as
3694 byte characters or as UTF-8, which is what "perltest" was being required to
3695 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3696 can't do is switch easily at run time between having the "use utf8;" pragma
3697 or not. In the end, I fudged it by using the RunPerlTest script to insert
3698 "use utf8;" explicitly for the UTF-8 tests.
3699
3700 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3701 the end of the subject string, contrary to the documentation and to what
3702 Perl does. This was true of both matching functions. Now it matches only at
3703 the start of the subject and immediately after *internal* newlines.
3704
3705 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3706 a pointer to an int instead of a pointer to an unsigned long int. This
3707 caused problems on 64-bit systems.
3708
3709 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3710 instance of the 'standard' template library not being so standard".
3711
3712 10. There was no check on the number of named subpatterns nor the maximum
3713 length of a subpattern name. The product of these values is used to compute
3714 the size of the memory block for a compiled pattern. By supplying a very
3715 long subpattern name and a large number of named subpatterns, the size
3716 computation could be caused to overflow. This is now prevented by limiting
3717 the length of names to 32 characters, and the number of named subpatterns
3718 to 10,000.
3719
3720 11. Subpatterns that are repeated with specific counts have to be replicated in
3721 the compiled pattern. The size of memory for this was computed from the
3722 length of the subpattern and the repeat count. The latter is limited to
3723 65535, but there was no limit on the former, meaning that integer overflow
3724 could in principle occur. The compiled length of a repeated subpattern is
3725 now limited to 30,000 bytes in order to prevent this.
3726
3727 12. Added the optional facility to have named substrings with the same name.
3728
3729 13. Added the ability to use a named substring as a condition, using the
3730 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3731 are numbers (not recommended). Forward references are permitted.
3732
3733 14. Added forward references in named backreferences (if you see what I mean).
3734
3735 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3736 pattern could run off the end of the subject. For example, the pattern
3737 "(?s)(.{1,5})"8 did this with the subject "ab".
3738
3739 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3740 PCRE_CASELESS was set when matching characters that were quantified with ?
3741 or *.
3742
3743 17. A character class other than a single negated character that had a minimum
3744 but no maximum quantifier - for example [ab]{6,} - was not handled
3745 correctly by pce_dfa_exec(). It would match only one character.
3746
3747 18. A valid (though odd) pattern that looked like a POSIX character
3748 class but used an invalid character after [ (for example [[,abc,]]) caused
3749 pcre_compile() to give the error "Failed: internal error: code overflow" or
3750 in some cases to crash with a glibc free() error. This could even happen if
3751 the pattern terminated after [[ but there just happened to be a sequence of
3752 letters, a binary zero, and a closing ] in the memory that followed.
3753
3754 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3755 over the years. Originally (before any Unicode support), just the bottom 8
3756 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3757 output from "man perlunicode" includes this:
3758
3759 The regular expression compiler produces polymorphic opcodes. That
3760 is, the pattern adapts to the data and automatically switches to
3761 the Unicode character scheme when presented with Unicode data--or
3762 instead uses a traditional byte scheme when presented with byte
3763 data.
3764
3765 Sadly, a wide octal escape does not cause a switch, and in a string with
3766 no other multibyte characters, these octal escapes are treated as before.
3767 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3768 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3769 Unicode string.
3770
3771 I have not perpetrated such confusion in PCRE. Up till now, it took just
3772 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3773 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3774 translate to the appropriate multibyte character.
3775
3776 29. Applied some refactoring to reduce the number of warnings from Microsoft
3777 and Borland compilers. This has included removing the fudge introduced
3778 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3779 a warning about an unused variable.
3780
3781 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3782 characters since release 4.0, because Perl (from release 5.004) does not.
3783 [Or at least, is documented not to: some releases seem to be in conflict
3784 with the documentation.] However, when a pattern was studied with
3785 pcre_study() and all its branches started with \s, PCRE still included VT
3786 as a possible starting character. Of course, this did no harm; it just
3787 caused an unnecessary match attempt.
3788
3789 22. Removed a now-redundant internal flag bit that recorded the fact that case
3790 dependency changed within the pattern. This was once needed for "required
3791 byte" processing, but is no longer used. This recovers a now-scarce options
3792 bit. Also moved the least significant internal flag bit to the most-
3793 significant bit of the word, which was not previously used (hangover from
3794 the days when it was an int rather than a uint) to free up another bit for
3795 the future.
3796
3797 23. Added support for CRLF line endings as well as CR and LF. As well as the
3798 default being selectable at build time, it can now be changed at runtime
3799 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3800 specify that it is scanning data with non-default line endings.
3801
3802 24. Changed the definition of CXXLINK to make it agree with the definition of
3803 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3804
3805 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3806 recursions. This makes a big different to stack usage for some patterns.
3807
3808 26. If a subpattern containing a named recursion or subroutine reference such
3809 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3810 the space required for the compiled pattern went wrong and gave too small a
3811 value. Depending on the environment, this could lead to "Failed: internal
3812 error: code overflow at offset 49" or "glibc detected double free or
3813 corruption" errors.
3814
3815 27. Applied patches from Google (a) to support the new newline modes and (b) to
3816 advance over multibyte UTF-8 characters in GlobalReplace.
3817
3818 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3819 difference for some implementation of PCRE in some Windows version.
3820
3821 29. Added some extra testing facilities to pcretest:
3822
3823 \q<number> in a data line sets the "match limit" value
3824 \Q<number> in a data line sets the "match recursion limt" value
3825 -S <number> sets the stack size, where <number> is in megabytes
3826
3827 The -S option isn't available for Windows.
3828
3829
3830 Version 6.6 06-Feb-06
3831 ---------------------
3832
3833 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3834 in pcreposix.h. I have copied the definition from pcre.h.
3835
3836 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3837 because pcre.h is no longer a built file.
3838
3839 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3840 not normally included in the compiled code.
3841
3842
3843 Version 6.5 01-Feb-06
3844 ---------------------
3845
3846 1. When using the partial match feature with pcre_dfa_exec(), it was not
3847 anchoring the second and subsequent partial matches at the new starting
3848 point. This could lead to incorrect results. For example, with the pattern
3849 /1234/, partially matching against "123" and then "a4" gave a match.
3850
3851 2. Changes to pcregrep:
3852
3853 (a) All non-match returns from pcre_exec() were being treated as failures
3854 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3855 error message is output. Some extra information is given for the
3856 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3857 probably the only errors that are likely to be caused by users (by
3858 specifying a regex that has nested indefinite repeats, for instance).
3859 If there are more than 20 of these errors, pcregrep is abandoned.
3860
3861 (b) A binary zero was treated as data while matching, but terminated the
3862 output line if it was written out. This has been fixed: binary zeroes
3863 are now no different to any other data bytes.
3864
3865 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3866 used to set a locale for matching. The --locale=xxxx long option has
3867 been added (no short equivalent) to specify a locale explicitly on the
3868 pcregrep command, overriding the environment variables.
3869
3870 (d) When -B was used with -n, some line numbers in the output were one less
3871 than they should have been.
3872
3873 (e) Added the -o (--only-matching) option.
3874
3875 (f) If -A or -C was used with -c (count only), some lines of context were
3876 accidentally printed for the final match.
3877
3878 (g) Added the -H (--with-filename) option.
3879
3880 (h) The combination of options -rh failed to suppress file names for files
3881 that were found from directory arguments.
3882
3883 (i) Added the -D (--devices) and -d (--directories) options.
3884
3885 (j) Added the -F (--fixed-strings) option.
3886
3887 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3888
3889 (l) Added the --colo(u)r option.
3890
3891 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3892 is not present by default.
3893
3894 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3895 items such as (?R) or (?1), when the recursion could match a number of
3896 alternatives. If it matched one of the alternatives, but subsequently,
3897 outside the recursion, there was a failure, the code tried to back up into
3898 the recursion. However, because of the way PCRE is implemented, this is not
3899 possible, and the result was an incorrect result from the match.
3900
3901 In order to prevent this happening, the specification of recursion has
3902 been changed so that all such subpatterns are automatically treated as
3903 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3904
3905 4. I had overlooked the fact that, in some locales, there are characters for
3906 which isalpha() is true but neither isupper() nor islower() are true. In
3907 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3908 and ordfeminine) are like this. This affected the treatment of \w and \W
3909 when they appeared in character classes, but not when they appeared outside
3910 a character class. The bit map for "word" characters is now created
3911 separately from the results of isalnum() instead of just taking it from the
3912 upper, lower, and digit maps. (Plus the underscore character, of course.)
3913
3914 5. The above bug also affected the handling of POSIX character classes such as
3915 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3916 permanent tables. Instead, the bit maps for such a class were previously
3917 created as the appropriate unions of the upper, lower, and digit bitmaps.
3918 Now they are created by subtraction from the [[:word:]] class, which has
3919 its own bitmap.
3920
3921 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3922 It is created by subtracting the vertical space characters (\x09, \x0a,
3923 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3924 subtraction was done in the overall bitmap for a character class, meaning
3925 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3926 be recognized. This bug has been fixed.
3927
3928 7. Patches from the folks at Google:
3929
3930 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3931 real life, but is still worth protecting against".
3932
3933 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3934 regular expressions".
3935
3936 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3937 have it.
3938
3939 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3940 "configure" and the latter not, in order to fix a problem somebody had
3941 with compiling the Arg class on HP-UX.
3942
3943 (e) Improve the error-handling of the C++ wrapper a little bit.
3944
3945 (f) New tests for checking recursion limiting.
3946
3947 8. The pcre_memmove() function, which is used only if the environment does not
3948 have a standard memmove() function (and is therefore rarely compiled),
3949 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3950 returning a result (though PCRE never actually uses the result).
3951
3952 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3953 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3954 returned instead of calling malloc() with an overflowing number that would
3955 most likely cause subsequent chaos.
3956
3957 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3958
3959 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3960 with this option is matched, the nmatch and pmatch options of regexec() are
3961 ignored.
3962
3963 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3964 provided in case anyone wants to the the POSIX interface with UTF-8
3965 strings.
3966
3967 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3968 C++ linking (needed for some HP-UX environments).
3969
3970 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3971 (unused parameter) and in the pcre_printint() function (omitted "default"
3972 switch label when the default is to do nothing).
3973
3974 15. Added some code to make it possible, when PCRE is compiled as a C++
3975 library, to replace subject pointers for pcre_exec() with a smart pointer
3976 class, thus making it possible to process discontinuous strings.
3977
3978 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3979 much the same function. They were added by different people who were trying
3980 to make PCRE easy to compile on non-Unix systems. It has been suggested
3981 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3982 for compiling on Windows systems. I have therefore replaced it with
3983 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3984 defaults to "extern" for C or "extern C" for C++, which works fine on
3985 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3986 SCOPE with something explicit in config.h. In addition:
3987
3988 (a) pcreposix.h still had just "extern" instead of either of these macros;
3989 I have replaced it with PCRE_DATA_SCOPE.
3990
3991 (b) Functions such as _pcre_xclass(), which are internal to the library,
3992 but external in the C sense, all had PCRE_EXPORT in their definitions.
3993 This is apparently wrong for the Windows case, so I have removed it.
3994 (It makes no difference on Unix-like systems.)
3995
3996 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3997 of recursive calls to match(). This is different to MATCH_LIMIT because
3998 that limits the total number of calls to match(), not all of which increase
3999 the depth of recursion. Limiting the recursion depth limits the amount of
4000 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4001 when PCRE is compiled, and changed at run time. A patch from Google adds
4002 this functionality to the C++ interface.
4003
4004 18. Changes to the handling of Unicode character properties:
4005
4006 (a) Updated the table to Unicode 4.1.0.
4007
4008 (b) Recognize characters that are not in the table as "Cn" (undefined).
4009
4010 (c) I revised the way the table is implemented to a much improved format
4011 which includes recognition of ranges. It now supports the ranges that
4012 are defined in UnicodeData.txt, and it also amalgamates other
4013 characters into ranges. This has reduced the number of entries in the
4014 table from around 16,000 to around 3,000, thus reducing its size
4015 considerably. I realized I did not need to use a tree structure after
4016 all - a binary chop search is just as efficient. Having reduced the
4017 number of entries, I extended their size from 6 bytes to 8 bytes to
4018 allow for more data.
4019
4020 (d) Added support for Unicode script names via properties such as \p{Han}.
4021
4022 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4023 matching that character.
4024
4025 20. When matching a repeated Unicode property with a minimum greater than zero,
4026 (for example \pL{2,}), PCRE could look past the end of the subject if it
4027 reached it while seeking the minimum number of characters. This could
4028 happen only if some of the characters were more than one byte long, because
4029 there is a check for at least the minimum number of bytes.
4030
4031 21. Refactored the implementation of \p and \P so as to be more general, to
4032 allow for more different types of property in future. This has changed the
4033 compiled form incompatibly. Anybody with saved compiled patterns that use
4034 \p or \P will have to recompile them.
4035
4036 22. Added "Any" and "L&" to the supported property types.
4037
4038 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4039 but give a compile time error if the value is greater than 0xff.
4040
4041 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4042 accidentally not being installed or uninstalled.
4043
4044 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4045 made were to insert the current release number. This seemed silly, because
4046 it made things harder for people building PCRE on systems that don't run
4047 "configure". I have turned pcre.h into a distributed file, no longer built
4048 by "configure", with the version identification directly included. There is
4049 no longer a pcre.h.in file.
4050
4051 However, this change necessitated a change to the pcre-config script as
4052 well. It is built from pcre-config.in, and one of the substitutions was the
4053 release number. I have updated configure.ac so that ./configure now finds
4054 the release number by grepping pcre.h.
4055
4056 26. Added the ability to run the tests under valgrind.
4057
4058
4059 Version 6.4 05-Sep-05
4060 ---------------------
4061
4062 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4063 "--" to be printed when multiple files were scanned, even when none of the
4064 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4065 consider it to be a bug, and have restored the previous behaviour.
4066
4067 2. A couple of code tidies to get rid of compiler warnings.
4068
4069 3. The pcretest program used to cheat by referring to symbols in the library
4070 whose names begin with _pcre_. These are internal symbols that are not
4071 really supposed to be visible externally, and in some environments it is
4072 possible to suppress them. The cheating is now confined to including
4073 certain files from the library's source, which is a bit cleaner.
4074
4075 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4076 file's purpose clearer.
4077
4078 5. Reorganized pcre_ucp_findchar().
4079
4080
4081 Version 6.3 15-Aug-05
4082 ---------------------
4083
4084 1. The file libpcre.pc.in did not have general read permission in the tarball.
4085
4086 2. There were some problems when building without C++ support:
4087
4088 (a) If C++ support was not built, "make install" and "make test" still
4089 tried to test it.
4090
4091 (b) There were problems when the value of CXX was explicitly set. Some
4092 changes have been made to try to fix these, and ...
4093
4094 (c) --disable-cpp can now be used to explicitly disable C++ support.
4095
4096 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4097 backslash in a target when C++ was disabled. This confuses some
4098 versions of "make", apparently. Using an intermediate variable solves
4099 this. (Same for CPP_LOBJ.)
4100
4101 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4102 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4103 necessary on certain architectures.
4104
4105 4. Added a setting of -export-symbols-regex to the link command to remove
4106 those symbols that are exported in the C sense, but actually are local
4107 within the library, and not documented. Their names all begin with
4108 "_pcre_". This is not a perfect job, because (a) we have to except some
4109 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4110 available (and never for static libraries). I have made a note to try to
4111 find a way round (a) in the future.
4112
4113
4114 Version 6.2 01-Aug-05
4115 ---------------------
4116
4117 1. There was no test for integer overflow of quantifier values. A construction
4118 such as {1111111111111111} would give undefined results. What is worse, if
4119 a minimum quantifier for a parenthesized subpattern overflowed and became
4120 negative, the calculation of the memory size went wrong. This could have
4121 led to memory overwriting.
4122
4123 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4124
4125 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4126 operating environments where this matters.
4127
4128 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4129 PCRE options from within the C++ wrapper.
4130
4131 5. Named capturing subpatterns were not being correctly counted when a pattern
4132 was compiled. This caused two problems: (a) If there were more than 100
4133 such subpatterns, the calculation of the memory needed for the whole
4134 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4135 back references of the form \12, where the number was greater than 9, were
4136 not recognized as back references, even though there were sufficient
4137 previous subpatterns.
4138
4139 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4140 versions of gcc, e.g. 2.95.4.
4141
4142
4143 Version 6.1 21-Jun-05
4144 ---------------------
4145
4146 1. There was one reference to the variable "posix" in pcretest.c that was not
4147 surrounded by "#if !defined NOPOSIX".
4148
4149 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4150 the cross-check on the old pcre_info() function, for the benefit of the
4151 cut-down version of PCRE that is currently imported into Exim.
4152
4153 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4154 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4155 patterns that start (?i) but I don't think that matters. The use of (?i) is
4156 just an example; this all applies to the other options as well.
4157
4158 4. Since libtool seems to echo the compile commands it is issuing, the output
4159 from "make" can be reduced a bit by putting "@" in front of each libtool
4160 compile command.
4161
4162 5. Patch from the folks at Google for configure.in to be a bit more thorough
4163 in checking for a suitable C++ installation before trying to compile the
4164 C++ stuff. This should fix a reported problem when a compiler was present,
4165 but no suitable headers.
4166
4167 6. The man pages all had just "PCRE" as their title. I have changed them to
4168 be the relevant file name. I have also arranged that these names are
4169 retained in the file doc/pcre.txt, which is a concatenation in text format
4170 of all the man pages except the little individual ones for each function.
4171
4172 7. The NON-UNIX-USE file had not been updated for the different set of source
4173 files that come with release 6. I also added a few comments about the C++
4174 wrapper.
4175
4176
4177 Version 6.0 07-Jun-05
4178 ---------------------
4179
4180 1. Some minor internal re-organization to help with my DFA experiments.
4181
4182 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4183 didn't matter for the library itself when fully configured, but did matter
4184 when compiling without UCP support, or within Exim, where the ucp files are
4185 not imported.
4186
4187 3. Refactoring of the library code to split up the various functions into
4188 different source modules. The addition of the new DFA matching code (see
4189 below) to a single monolithic source would have made it really too
4190 unwieldy, quite apart from causing all the code to be include in a
4191 statically linked application, when only some functions are used. This is
4192 relevant even without the DFA addition now that patterns can be compiled in
4193 one application and matched in another.
4194
4195 The downside of splitting up is that there have to be some external
4196 functions and data tables that are used internally in different modules of
4197 the library but which are not part of the API. These have all had their
4198 names changed to start with "_pcre_" so that they are unlikely to clash
4199 with other external names.
4200
4201 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4202 a different (DFA) algorithm. Although it is slower than the original
4203 function, it does have some advantages for certain types of matching
4204 problem.
4205
4206 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4207 including restarting after a partial match.
4208
4209 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4210 defined when compiling for Windows was sent to me. I have put it into the
4211 code, though I have no means of testing or verifying it.
4212
4213 7. Added the pcre_refcount() auxiliary function.
4214
4215 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4216 match before or at the first newline in the subject string. In pcretest,
4217 the /f option on a pattern can be used to set this.
4218
4219 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4220 would behave wrongly. This has been present in PCRE since release 4.0.
4221
4222 10. A number of changes to the pcregrep command:
4223
4224 (a) Refactored how -x works; insert ^(...)$ instead of setting
4225 PCRE_ANCHORED and checking the length, in preparation for adding
4226 something similar for -w.
4227
4228 (b) Added the -w (match as a word) option.
4229
4230 (c) Refactored the way lines are read and buffered so as to have more
4231 than one at a time available.
4232
4233 (d) Implemented a pcregrep test script.
4234
4235 (e) Added the -M (multiline match) option. This allows patterns to match
4236 over several lines of the subject. The buffering ensures that at least
4237 8K, or the rest of the document (whichever is the shorter) is available
4238 for matching (and similarly the previous 8K for lookbehind assertions).
4239
4240 (f) Changed the --help output so that it now says
4241
4242 -w, --word-regex(p)
4243
4244 instead of two lines, one with "regex" and the other with "regexp"
4245 because that confused at least one person since the short forms are the
4246 same. (This required a bit of code, as the output is generated
4247 automatically from a table. It wasn't just a text change.)
4248
4249 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4250 option but starts with a hyphen. Could be a pattern or a path name
4251 starting with a hyphen, for instance.
4252
4253 (h) "-" can be given as a file name to represent stdin.
4254
4255 (i) When file names are being printed, "(standard input)" is used for
4256 the standard input, for compatibility with GNU grep. Previously
4257 "<stdin>" was used.
4258
4259 (j) The option --label=xxx can be used to supply a name to be used for
4260 stdin when file names are being printed. There is no short form.
4261
4262 (k) Re-factored the options decoding logic because we are going to add
4263 two more options that take data. Such options can now be given in four
4264 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4265
4266 (l) Added the -A, -B, and -C options for requesting that lines of context
4267 around matches be printed.
4268
4269 (m) Added the -L option to print the names of files that do not contain
4270 any matching lines, that is, the complement of -l.
4271
4272 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4273 continue to scan other files.
4274
4275 (o) The -s option was incorrectly implemented. For compatibility with other
4276 greps, it now suppresses the error message for a non-existent or non-
4277 accessible file (but not the return code). There is a new option called
4278 -q that suppresses the output of matching lines, which was what -s was
4279 previously doing.
4280
4281 (p) Added --include and --exclude options to specify files for inclusion
4282 and exclusion when recursing.
4283
4284 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4285 Hopefully, it now does.
4286
4287 12. Missing cast in pcre_study().
4288
4289 13. Added an "uninstall" target to the makefile.
4290
4291 14. Replaced "extern" in the function prototypes in Makefile.in with
4292 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4293 world, but is set differently for Windows.
4294
4295 15. Added a second compiling function called pcre_compile2(). The only
4296 difference is that it has an extra argument, which is a pointer to an
4297 integer error code. When there is a compile-time failure, this is set
4298 non-zero, in addition to the error test pointer being set to point to an
4299 error message. The new argument may be NULL if no error number is required
4300 (but then you may as well call pcre_compile(), which is now just a
4301 wrapper). This facility is provided because some applications need a
4302 numeric error indication, but it has also enabled me to tidy up the way
4303 compile-time errors are handled in the POSIX wrapper.
4304
4305 16. Added VPATH=.libs to the makefile; this should help when building with one
4306 prefix path and installing with another. (Or so I'm told by someone who
4307 knows more about this stuff than I do.)
4308
4309 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4310 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4311 match everything, including newlines. This is not POSIX-compatible, but
4312 somebody wanted the feature. From pcretest it can be activated by using
4313 both the P and the s flags.
4314
4315 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4316
4317 19. libpcre.pc was being incorrectly installed as executable.
4318
4319 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4320 it now also looks for '\r' so that it will work unmodified on Windows.
4321
4322 21. Added Google's contributed C++ wrapper to the distribution.
4323
4324 22. Added some untidy missing memory free() calls in pcretest, to keep
4325 Electric Fence happy when testing.
4326
4327
4328
4329 Version 5.0 13-Sep-04
4330 ---------------------
4331
4332 1. Internal change: literal characters are no longer packed up into items
4333 containing multiple characters in a single byte-string. Each character
4334 is now matched using a separate opcode. However, there may be more than one
4335 byte in the character in UTF-8 mode.
4336
4337 2. The pcre_callout_block structure has two new fields: pattern_position and
4338 next_item_length. These contain the offset in the pattern to the next match
4339 item, and its length, respectively.
4340
4341 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4342 insertion of callouts before each pattern item. Added the /C option to
4343 pcretest to make use of this.
4344
4345 4. On the advice of a Windows user, the lines
4346
4347 #if defined(_WIN32) || defined(WIN32)
4348 _setmode( _fileno( stdout ), 0x8000 );
4349 #endif /* defined(_WIN32) || defined(WIN32) */
4350
4351 have been added to the source of pcretest. This apparently does useful
4352 magic in relation to line terminators.
4353
4354 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4355 for the benefit of those environments where the "b" makes a difference.
4356
4357 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4358 to know about it. I have put a hack into configure.in that adds in code
4359 to set GCC=yes if CC=icc. This seems to end up at a point in the
4360 generated configure script that is early enough to affect the setting of
4361 compiler options, which is what is needed, but I have no means of testing
4362 whether it really works. (The user who reported this had patched the
4363 generated configure script, which of course I cannot do.)
4364
4365 LATER: After change 22 below (new libtool files), the configure script
4366 seems to know about icc (and also ecc). Therefore, I have commented out
4367 this hack in configure.in.
4368
4369 7. Added support for pkg-config (2 patches were sent in).
4370
4371 8. Negated POSIX character classes that used a combination of internal tables
4372 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4373 [[:^ascii]]. Typically, they would match almost any characters. The other
4374 POSIX classes were not broken in this way.
4375
4376 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4377 to find the match, as PCRE was deluded into thinking that the match had to
4378 start at the start point or following a newline. The same bug applied to
4379 patterns with negative forward assertions or any backward assertions
4380 preceding ".*" at the start, unless the pattern required a fixed first
4381 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4382
4383 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4384 starting at the last subject character, bytes beyond the end of the subject
4385 string were read.
4386
4387 11. Renamed the variable "class" as "classbits" to make life easier for C++
4388 users. (Previously there was a macro definition, but it apparently wasn't
4389 enough.)
4390
4391 12. Added the new field "tables" to the extra data so that tables can be passed
4392 in at exec time, or the internal tables can be re-selected. This allows
4393 a compiled regex to be saved and re-used at a later time by a different
4394 program that might have everything at different addresses.
4395
4396 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4397 -R library as well as a -L library.
4398
4399 14. The debugging options of pcretest (-d on the command line or D on a
4400 pattern) showed incorrect output for anything following an extended class
4401 that contained multibyte characters and which was followed by a quantifier.
4402
4403 15. Added optional support for general category Unicode character properties
4404 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4405 support. It adds about 90K to the size of the library. The meanings of the
4406 inbuilt class escapes such as \d and \s have NOT been changed.
4407
4408 16. Updated pcredemo.c to include calls to free() to release the memory for the
4409 compiled pattern.
4410
4411 17. The generated file chartables.c was being created in the source directory
4412 instead of in the building directory. This caused the build to fail if the
4413 source directory was different from the building directory, and was
4414 read-only.
4415
4416 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4417 file. No doubt somebody will tell me if they don't make sense... Also added
4418 Dan Mooney's comments about building on OpenVMS.
4419
4420 19. Added support for partial matching via the PCRE_PARTIAL option for
4421 pcre_exec() and the \P data escape in pcretest.
4422
4423 20. Extended pcretest with 3 new pattern features:
4424
4425 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4426 write the compiled pattern to the file whose name is "rest-of-line".
4427 This is a straight binary dump of the data, with the saved pointer to
4428 the character tables forced to be NULL. The study data, if any, is
4429 written too. After writing, pcretest reads a new pattern.
4430
4431 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4432 compiled pattern from the given file. There must not be any
4433 occurrences of "<" in the file name (pretty unlikely); if there are,
4434 pcretest will instead treat the initial "<" as a pattern delimiter.
4435 After reading in the pattern, pcretest goes on to read data lines as
4436 usual.
4437
4438 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4439 and 16-bit fields in a compiled pattern, to simulate a pattern that
4440 was compiled on a host of opposite endianness.
4441
4442 21. The pcre-exec() function can now cope with patterns that were compiled on
4443 hosts of opposite endianness, with this restriction:
4444
4445 As for any compiled expression that is saved and used later, the tables
4446 pointer field cannot be preserved; the extra_data field in the arguments
4447 to pcre_exec() should be used to pass in a tables address if a value
4448 other than the default internal tables were used at compile time.
4449
4450 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4451 now diagnosed as an error. Previously, most of the time, a negative number
4452 would have been treated as zero, but if in addition "ovector" was passed as
4453 NULL, a crash could occur.
4454
4455 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4456 new versions from the libtool 1.5 distribution (the last one is a copy of
4457 a file called libtool.m4). This seems to have fixed the need to patch
4458 "configure" to support Darwin 1.3 (which I used to do). However, I still
4459 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4460 workstation).
4461
4462 24. Changed the PCRE licence to be the more standard "BSD" licence.
4463
4464
4465 Version 4.5 01-Dec-03
4466 ---------------------
4467
4468 1. There has been some re-arrangement of the code for the match() function so
4469 that it can be compiled in a version that does not call itself recursively.
4470 Instead, it keeps those local variables that need separate instances for
4471 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4472 needs to "recurse". Keeping track of where control must go is done by means
4473 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4474 hide most of the details from the main code, and operates only if
4475 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4476 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4477 operating.
4478
4479 To make it easier for callers to provide specially tailored get/free
4480 functions for this usage, two new functions, pcre_stack_malloc, and
4481 pcre_stack_free, are used. They are always called in strict stacking order,
4482 and the size of block requested is always the same.
4483
4484 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4485 PCRE has been compiled to use the stack or the heap for recursion. The
4486 -C option of pcretest uses this to show which version is compiled.
4487
4488 A new data escape \S, is added to pcretest; it causes the amounts of store
4489 obtained and freed by both kinds of malloc/free at match time to be added
4490 to the output.
4491
4492 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4493 what's available on my current Linux desktop machine.
4494
4495 3. When matching a UTF-8 string, the test for a valid string at the start has
4496 been extended. If start_offset is not zero, PCRE now checks that it points
4497 to a byte that is the start of a UTF-8 character. If not, it returns
4498 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4499 this is necessary because there may be backward assertions in the pattern.
4500 When matching the same subject several times, it may save resources to use
4501 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4502
4503 4. The code for checking the validity of UTF-8 strings has been tightened so
4504 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4505 containing "overlong sequences".
4506
4507 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4508 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4509 should have been "&", but it just so happened that all the cases this let
4510 through by mistake were picked up later in the function.
4511
4512 6. I had used a variable called "isblank" - this is a C99 function, causing
4513 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4514
4515 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4516 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4517 so the version is shown; (c) stop "make test" echoing "./RunTest".
4518
4519 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4520
4521 9. The prototype for memmove() for systems that don't have it was using
4522 size_t, but the inclusion of the header that defines size_t was later. I've
4523 moved the #includes for the C headers earlier to avoid this.
4524
4525 10. Added some adjustments to the code to make it easier to compiler on certain
4526 special systems:
4527
4528 (a) Some "const" qualifiers were missing.
4529 (b) Added the macro EXPORT before all exported functions; by default this
4530 is defined to be empty.
4531 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4532 that it reads its output file name as an argument instead of writing
4533 to the standard output and assuming this can be redirected.
4534
4535 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4536 class containing characters with values greater than 255, PCRE compilation
4537 went into a loop.
4538
4539 12. A recursive reference to a subpattern that was within another subpattern
4540 that had a minimum quantifier of zero caused PCRE to crash. For example,
4541 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4542 recursion. If the recursively-called subpattern itself had a zero repeat,
4543 that was OK.
4544
4545 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4546 buffer into which it was copied (for escape processing) was still set at
4547 1024, so long lines caused crashes.
4548
4549 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4550 "internal error: code overflow...". This applied to any character class
4551 that was followed by a possessive quantifier.
4552
4553 15. Modified the Makefile to add libpcre.la as a prerequisite for
4554 libpcreposix.la because I was told this is needed for a parallel build to
4555 work.
4556
4557 16. If a pattern that contained .* following optional items at the start was
4558 studied, the wrong optimizing data was generated, leading to matching
4559 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4560 matching string must start with a or b or c. The correct conclusion for
4561 this pattern is that a match can start with any character.
4562
4563
4564 Version 4.4 13-Aug-03
4565 ---------------------
4566
4567 1. In UTF-8 mode, a character class containing characters with values between
4568 127 and 255 was not handled correctly if the compiled pattern was studied.
4569 In fixing this, I have also improved the studying algorithm for such
4570 classes (slightly).
4571
4572 2. Three internal functions had redundant arguments passed to them. Removal
4573 might give a very teeny performance improvement.
4574
4575 3. Documentation bug: the value of the capture_top field in a callout is *one
4576 more than* the number of the hightest numbered captured substring.
4577
4578 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4579 in incorrectly linking with a previously installed version. They now link
4580 explicitly with libpcre.la.
4581
4582 5. configure.in no longer needs to recognize Cygwin specially.
4583
4584 6. A problem in pcre.in for Windows platforms is fixed.
4585
4586 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4587 pcretest, it used to include the size of the study block as part of its
4588 output. Unfortunately, the structure contains a field that has a different
4589 size on different hardware architectures. This meant that the tests that
4590 showed this size failed. As the block is currently always of a fixed size,
4591 this information isn't actually particularly useful in pcretest output, so
4592 I have just removed it.
4593
4594 8. Three pre-processor statements accidentally did not start in column 1.
4595 Sadly, there are *still* compilers around that complain, even though
4596 standard C has not required this for well over a decade. Sigh.
4597
4598 9. In pcretest, the code for checking callouts passed small integers in the
4599 callout_data field, which is a void * field. However, some picky compilers
4600 complained about the casts involved for this on 64-bit systems. Now
4601 pcretest passes the address of the small integer instead, which should get
4602 rid of the warnings.
4603
4604 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4605 both compile and run time, and gives an error if an invalid UTF-8 sequence
4606 is found. There is a option for disabling this check in cases where the
4607 string is known to be correct and/or the maximum performance is wanted.
4608
4609 11. In response to a bug report, I changed one line in Makefile.in from
4610
4611 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4612 to
4613 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4614
4615 to look similar to other lines, but I have no way of telling whether this
4616 is the right thing to do, as I do not use Windows. No doubt I'll get told
4617 if it's wrong...
4618
4619
4620 Version 4.3 21-May-03
4621 ---------------------
4622
4623 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4624 Makefile.
4625
4626 2. Some refactoring to improve the quality of the code:
4627
4628 (i) The utf8_table... variables are now declared "const".
4629
4630 (ii) The code for \cx, which used the "case flipping" table to upper case
4631 lower case letters, now just substracts 32. This is ASCII-specific,
4632 but the whole concept of \cx is ASCII-specific, so it seems
4633 reasonable.
4634
4635 (iii) PCRE was using its character types table to recognize decimal and
4636 hexadecimal digits in the pattern. This is silly, because it handles
4637 only 0-9, a-f, and A-F, but the character types table is locale-
4638 specific, which means strange things might happen. A private
4639 table is now used for this - though it costs 256 bytes, a table is
4640 much faster than multiple explicit tests. Of course, the standard
4641 character types table is still used for matching digits in subject
4642 strings against \d.
4643
4644 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4645 ending in _t are). So I've renamed it as ESC_tee.
4646
4647 3. The first argument for regexec() in the POSIX wrapper should have been
4648 defined as "const".
4649
4650 4. Changed pcretest to use malloc() for its buffers so that they can be
4651 Electric Fenced for debugging.
4652
4653 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4654 to read one or more bytes before the start of the subject string. Often this
4655 had no effect on PCRE's behaviour, but in some circumstances it could
4656 provoke a segmentation fault.
4657
4658 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4659 to try to read one or more bytes before the start of the subject string.
4660
4661 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4662 UTF-8 support could misbehave in various ways if the subject string
4663 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4664 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4665 back over UTF-8 characters.)
4666
4667
4668 Version 4.2 14-Apr-03
4669 ---------------------
4670
4671 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4672
4673 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4674 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4675 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4676 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4677 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4678 and BUILD_EXEEXT
4679 Note: automatic setting of the BUILD variables is not yet working
4680 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4681 compile-time but not at link-time
4682 [LINK]: use for linking executables only
4683 make different versions for Windows and non-Windows
4684 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4685 libraries
4686 [LINK_FOR_BUILD]: new variable
4687 [OBJEXT]: use throughout
4688 [EXEEXT]: use throughout
4689 <winshared>: new target
4690 <wininstall>: new target
4691 <dftables.o>: use native compiler
4692 <dftables>: use native linker
4693 <install>: handle Windows platform correctly
4694 <clean>: ditto
4695 <check>: ditto
4696 copy DLL to top builddir before testing
4697
4698 As part of these changes, -no-undefined was removed again. This was reported
4699 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4700 in any case.
4701
4702 3. Some tidies to get rid of compiler warnings:
4703
4704 . In the match_data structure, match_limit was an unsigned long int, whereas
4705 match_call_count was an int. I've made them both unsigned long ints.
4706
4707 . In pcretest the fact that a const uschar * doesn't automatically cast to
4708 a void * provoked a warning.
4709
4710 . Turning on some more compiler warnings threw up some "shadow" variables
4711 and a few more missing casts.
4712
4713 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4714 option, a class that contained a single character with a value between 128
4715 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4716
4717 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4718 option, a class that contained several characters, but with at least one
4719 whose value was between 128 and 255 caused PCRE to crash.
4720
4721
4722 Version 4.1 12-Mar-03
4723 ---------------------
4724
4725 1. Compiling with gcc -pedantic found a couple of places where casts were
4726 needed, and a string in dftables.c that was longer than standard compilers are
4727 required to support.
4728
4729 2. Compiling with Sun's compiler found a few more places where the code could
4730 be tidied up in order to avoid warnings.
4731
4732 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4733 first of these names is deprecated in the latest Autoconf in favour of the name
4734 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4735 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4736 analogy I have changed it to CFLAGS_FOR_BUILD.
4737
4738 4. Added -no-undefined to the linking command in the Makefile, because this is
4739 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4740 linking step for the pcreposix library.
4741
4742 5. PCRE was failing to diagnose the case of two named groups with the same
4743 name.
4744
4745 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4746 literal character that is needed in the subject for a match, and scans along to
4747 ensure that it is present before embarking on the full matching process. This
4748 saves time in cases of nested unlimited repeats that are never going to match.
4749 Problem: the scan can take a lot of time if the subject is very long (e.g.
4750 megabytes), thus penalizing straightforward matches. It is now done only if the
4751 amount of subject to be scanned is less than 1000 bytes.
4752
4753 7. A lesser problem with the same optimization is that it was recording the
4754 first character of an anchored pattern as "needed", thus provoking a search
4755 right along the subject, even when the first match of the pattern was going to
4756 fail. The "needed" character is now not set for anchored patterns, unless it
4757 follows something in the pattern that is of non-fixed length. Thus, it still
4758 fulfils its original purpose of finding quick non-matches in cases of nested
4759 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4760
4761
4762 Version 4.0 17-Feb-03
4763 ---------------------
4764
4765 1. If a comment in an extended regex that started immediately after a meta-item
4766 extended to the end of string, PCRE compiled incorrect data. This could lead to
4767 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4768
4769 2. Moved to autoconf 2.53 and libtool 1.4.2.
4770
4771 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4772 the special perltest8 script is no longer needed - all the tests can be run
4773 from a single perltest script.
4774
4775 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4776 by \s. It has now been removed in PCRE. This means it isn't recognized as
4777 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4778 class [:space:] *does* include VT, thereby creating a mess.
4779
4780 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4781 space and tab.
4782
4783 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4784 its new features into the main test script, reducing the number of scripts.
4785
4786 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4787 were backward compatible, and made the (?i) apply to the whole pattern, as if
4788 /i were given. Now it behaves more logically, and applies the option setting
4789 only to what follows. PCRE has been changed to follow suit. However, if it
4790 finds options settings right at the start of the pattern, it extracts them into
4791 the global options, as before. Thus, they show up in the info data.
4792
4793 8. Added support for the \Q...\E escape sequence. Characters in between are
4794 treated as literals. This is slightly different from Perl in that $ and @ are
4795 also handled as literals inside the quotes. In Perl, they will cause variable
4796 interpolation. Note the following examples:
4797
4798 Pattern PCRE matches Perl matches
4799
4800 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4801 \Qabc\$xyz\E abc\$xyz abc\$xyz
4802 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4803
4804 For compatibility with Perl, \Q...\E sequences are recognized inside character
4805 classes as well as outside them.
4806
4807 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4808 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4809 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4810 signed/unsigned warnings.
4811
4812 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4813 option for pcretest, so I've replaced it by a simple function that does just
4814 that job.
4815
4816 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4817 "pcregrep -".
4818
4819 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4820 Java package. This provides some syntactic sugar for simple cases of what my
4821 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4822 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4823 item, you can use this simplified notation. Note that only makes sense with
4824 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4825 greediness, whatever the setting of the PCRE_UNGREEDY option.
4826
4827 13. A change of greediness default within a pattern was not taking effect at
4828 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4829 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4830 was abstracted outside.
4831
4832 14. PCRE now supports the \G assertion. It is true when the current matching
4833 position is at the start point of the match. This differs from \A when the
4834 starting offset is non-zero. Used with the /g option of pcretest (or similar
4835 code), it works in the same way as it does for Perl's /g option. If all
4836 alternatives of a regex begin with \G, the expression is anchored to the start
4837 match position, and the "anchored" flag is set in the compiled expression.
4838
4839 15. Some bugs concerning the handling of certain option changes within patterns
4840 have been fixed. These applied to options other than (?ims). For example,
4841 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4842 been the other way round. Some of this was related to change 7 above.
4843
4844 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4845 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4846 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4847 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4848
4849 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4850 mode. Unlike ".", it always matches newline, whatever the setting of
4851 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4852 assertions. Perl allows it, but it doesn't (in general) work because it can't
4853 calculate the length of the lookbehind. At least, that's the case for Perl
4854 5.8.0 - I've been told they are going to document that it doesn't work in
4855 future.
4856
4857 18. Added an error diagnosis for escapes that PCRE does not support: these are
4858 \L, \l, \N, \P, \p, \U, \u, and \X.
4859
4860 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4861 reading past the end of the pattern in cases such as /[abcd/.
4862
4863 20. PCRE was getting more memory than necessary for patterns with classes that
4864 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4865
4866 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4867 compiling PCRE for use with Virtual Pascal.
4868
4869 22. Small fix to the Makefile to make it work properly if the build is done
4870 outside the source tree.
4871
4872 23. Added a new extension: a condition to go with recursion. If a conditional
4873 subpattern starts with (?(R) the "true" branch is used if recursion has
4874 happened, whereas the "false" branch is used only at the top level.
4875
4876 24. When there was a very long string of literal characters (over 255 bytes
4877 without UTF support, over 250 bytes with UTF support), the computation of how
4878 much memory was required could be incorrect, leading to segfaults or other
4879 strange effects.
4880
4881 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4882 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4883 there was a subsequent back reference to those brackets. This meant that, for
4884 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4885 possible to check for precisely this case. All we can do is abandon the
4886 optimization if .* occurs inside capturing brackets when there are any back
4887 references whatsoever. (See below for a better fix that came later.)
4888
4889 26. The handling of the optimization for finding the first character of a
4890 non-anchored pattern, and for finding a character that is required later in the
4891 match were failing in some cases. This didn't break the matching; it just
4892 failed to optimize when it could. The way this is done has been re-implemented.
4893
4894 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4895
4896 28. Added a new feature that provides some of the functionality that Perl
4897 provides with (?{...}). The facility is termed a "callout". The way it is done
4898 in PCRE is for the caller to provide an optional function, by setting
4899 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4900 global variable. By default it is unset, which disables all calling out. To get
4901 the function called, the regex must include (?C) at appropriate points. This
4902 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4903 This provides a means of identifying different callout points. When PCRE
4904 reaches such a point in the regex, if pcre_callout has been set, the external
4905 function is called. It is provided with data in a structure called
4906 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4907 matching continues; if it returns a non-zero value, the match at the current
4908 point fails. However, backtracking will occur if possible. [This was changed
4909 later and other features added - see item 49 below.]
4910
4911 29. pcretest is upgraded to test the callout functionality. It provides a
4912 callout function that displays information. By default, it shows the start of
4913 the match and the current position in the text. There are some new data escapes
4914 to vary what happens:
4915
4916 \C+ in addition, show current contents of captured substrings
4917 \C- do not supply a callout function
4918 \C!n return 1 when callout number n is reached
4919 \C!n!m return 1 when callout number n is reached for the mth time
4920
4921 30. If pcregrep was called with the -l option and just a single file name, it
4922 output "<stdin>" if a match was found, instead of the file name.
4923
4924 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4925 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4926 pcre_exec(). This saves a malloc/free per call. The default value of
4927 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4928 when configuring.
4929
4930 32. The default maximum size of a compiled pattern is 64K. There have been a
4931 few cases of people hitting this limit. The code now uses macros to handle the
4932 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4933 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4934 configuring. Tests 2 and 5 work only with 2-byte links because they output
4935 debugging information about compiled patterns.
4936
4937 33. Internal code re-arrangements:
4938
4939 (a) Moved the debugging function for printing out a compiled regex into
4940 its own source file (printint.c) and used #include to pull it into
4941 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4942 separate copies.
4943
4944 (b) Defined the list of op-code names for debugging as a macro in
4945 internal.h so that it is next to the definition of the opcodes.
4946
4947 (c) Defined a table of op-code lengths for simpler skipping along compiled
4948 code. This is again a macro in internal.h so that it is next to the
4949 definition of the opcodes.
4950
4951 34. Added support for recursive calls to individual subpatterns, along the
4952 lines of Robin Houston's patch (but implemented somewhat differently).
4953
4954 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4955 allow it to read and process whole directories in Win32. This code was
4956 contributed by Lionel Fourquaux; it has not been tested by me.
4957
4958 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4959 used to name a group. Names consist of alphanumerics and underscores, and must
4960 be unique. Back references use the syntax (?P=name) and recursive calls use
4961 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4962 numbers. The function pcre_fullinfo() can be used after compilation to extract
4963 a name/number map. There are three relevant calls:
4964
4965 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4966 PCRE_INFO_NAMECOUNT yields the number of entries
4967 PCRE_INFO_NAMETABLE yields a pointer to the map.
4968
4969 The map is a vector of fixed-size entries. The size of each entry depends on
4970 the length of the longest name used. The first two bytes of each entry are the
4971 group number, most significant byte first. There follows the corresponding
4972 name, zero terminated. The names are in alphabetical order.
4973
4974 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4975 case instead of 255. Making it the same both with and without UTF-8 support
4976 means that the same test output works with both.
4977
4978 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4979 calling malloc() with a zero argument.
4980
4981 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4982 optimization. I've improved things by keeping a bitmap of backreferences with
4983 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4984 fact referenced, the optimization can be applied. It is unlikely that a
4985 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4986 the match to follow \n) will appear inside brackets with a number greater than
4987 31, but if it does, any back reference > 31 suppresses the optimization.
4988
4989 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4990 of disabling numbered capturing parentheses. Any opening parenthesis that is
4991 not followed by ? behaves as if it were followed by ?: but named parentheses
4992 can still be used for capturing (and they will acquire numbers in the usual
4993 way).
4994
4995 41. Redesigned the return codes from the match() function into yes/no/error so
4996 that errors can be passed back from deep inside the nested calls. A malloc
4997 failure while inside a recursive subpattern call now causes the
4998 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4999
5000 42. It is now possible to set a limit on the number of times the match()
5001 function is called in a call to pcre_exec(). This facility makes it possible to
5002 limit the amount of recursion and backtracking, though not in a directly
5003 obvious way, because the match() function is used in a number of different
5004 circumstances. The count starts from zero for each position in the subject
5005 string (for non-anchored patterns). The default limit is, for compatibility, a
5006 large number, namely 10 000 000. You can change this in two ways:
5007
5008 (a) When configuring PCRE before making, you can use --with-match-limit=n
5009 to set a default value for the compiled library.
5010
5011 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5012 a different value is set. See 45 below.
5013
5014 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5015
5016 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5017 of things that can be changed at compile time. The first argument specifies
5018 what is wanted and the second points to where the information is to be placed.
5019 The current list of available information is:
5020
5021 PCRE_CONFIG_UTF8
5022
5023 The output is an integer that is set to one if UTF-8 support is available;
5024 otherwise it is set to zero.
5025
5026 PCRE_CONFIG_NEWLINE
5027
5028 The output is an integer that it set to the value of the code that is used for
5029 newline. It is either LF (10) or CR (13).
5030
5031 PCRE_CONFIG_LINK_SIZE
5032
5033 The output is an integer that contains the number of bytes used for internal
5034 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5035
5036 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
5037
5038 The output is an integer that contains the threshold above which the POSIX
5039 interface uses malloc() for output vectors. See item 31 above.
5040
5041 PCRE_CONFIG_MATCH_LIMIT
5042
5043 The output is an unsigned integer that contains the default limit of the number
5044 of match() calls in a pcre_exec() execution. See 42 above.
5045
5046 44. pcretest has been upgraded by the addition of the -C option. This causes it
5047 to extract all the available output from the new pcre_config() function, and to
5048 output it. The program then exits immediately.
5049
5050 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5051 order to support additional features. One way would have been to define
5052 pcre_exec2() (for example) with extra arguments, but this would not have been
5053 extensible, and would also have required all calls to the original function to
5054 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5055 is used for passing in "extra" data from pcre_study().
5056
5057 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5058 contains the following fields:
5059
5060 flags a bitmap indicating which of the following fields are set
5061 study_data opaque data from pcre_study()
5062 match_limit a way of specifying a limit on match() calls for a specific
5063 call to pcre_exec()
5064 callout_data data for callouts (see 49 below)
5065
5066 The flag bits are also defined in pcre.h, and are
5067
5068 PCRE_EXTRA_STUDY_DATA
5069 PCRE_EXTRA_MATCH_LIMIT
5070 PCRE_EXTRA_CALLOUT_DATA
5071
5072 The pcre_study() function now returns one of these new pcre_extra blocks, with
5073 the actual study data pointed to by the study_data field, and the
5074 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5075 before. That is, this change is entirely upwards-compatible and requires no
5076 change to existing code.
5077
5078 If you want to pass in additional data to pcre_exec(), you can either place it
5079 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5080 block.
5081
5082 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5083 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5084 times with different match limits, until it finds the minimum value needed for
5085 pcre_exec() to complete. The value is then output. This can be instructive; for
5086 most simple matches the number is quite small, but for pathological cases it
5087 gets very large very quickly.
5088
5089 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5090 returns the size of the data block pointed to by the study_data field in a
5091 pcre_extra block, that is, the value that was passed as the argument to
5092 pcre_malloc() when PCRE was getting memory in which to place the information
5093 created by pcre_study(). The fourth argument should point to a size_t variable.
5094 pcretest has been extended so that this information is shown after a successful
5095 pcre_study() call when information about the compiled regex is being displayed.
5096
5097 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5098 because what follows is always an absolute path. (Later: it turns out that this
5099 is more than cosmetic for MinGW, because it doesn't like empty path
5100 components.)
5101
5102 49. Some changes have been made to the callout feature (see 28 above):
5103
5104 (i) A callout function now has three choices for what it returns:
5105
5106 0 => success, carry on matching
5107 > 0 => failure at this point, but backtrack if possible
5108 < 0 => serious error, return this value from pcre_exec()
5109
5110 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5111 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5112 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5113 use by callout functions. It will never be used by PCRE itself.
5114
5115 (ii) The pcre_extra structure (see 45 above) has a void * field called
5116 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5117 pcre_callout_block structure has a field of the same name. The contents of
5118 the field passed in the pcre_extra structure are passed to the callout
5119 function in the corresponding field in the callout block. This makes it
5120 easier to use the same callout-containing regex from multiple threads. For
5121 testing, the pcretest program has a new data escape
5122
5123 \C*n pass the number n (may be negative) as callout_data
5124
5125 If the callout function in pcretest receives a non-zero value as
5126 callout_data, it returns that value.
5127
5128 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5129 there were some redundant $(CFLAGS) in commands that are now specified as
5130 $(LINK), which already includes $(CFLAGS).
5131
5132 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5133 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5134 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5135 one-byte characters throughout. Note that case-insensitive matching applies
5136 only to characters whose values are less than 256. PCRE doesn't support the
5137 notion of cases for higher-valued characters.
5138
5139 (i) A character class whose characters are all within 0-255 is handled as
5140 a bit map, and the map is inverted for negative classes. Previously, a
5141 character > 255 always failed to match such a class; however it should
5142 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5143
5144 (ii) A negated character class with a single character < 255 is coded as
5145 "not this character" (OP_NOT). This wasn't working properly when the test
5146 character was multibyte, either singly or repeated.
5147
5148 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5149 mode, for example: \x{100}{2,3}.
5150
5151 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5152 singly or repeated) now correctly test multibyte characters. However,
5153 PCRE doesn't recognize any characters with values greater than 255 as
5154 digits, spaces, or word characters. Such characters always match \D, \S,
5155 and \W, and never match \d, \s, or \w.
5156
5157 (v) Classes may now contain characters and character ranges with values
5158 greater than 255. For example: [ab\x{100}-\x{400}].
5159
5160 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5161 PCRE in UTF-8 mode.
5162
5163 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5164 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5165 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5166 value.)
5167
5168 53. The single man page has become too large. I have therefore split it up into
5169 a number of separate man pages. These also give rise to individual HTML pages;
5170 these are now put in a separate directory, and there is an index.html page that
5171 lists them all. Some hyperlinking between the pages has been installed.
5172
5173 54. Added convenience functions for handling named capturing parentheses.
5174
5175 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5176 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5177 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5178 are faulted.
5179
5180 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5181 calling configure. These values are used when compiling the dftables.c program
5182 which is run to generate the source of the default character tables. They
5183 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5184 you will need to set these values.
5185
5186 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5187
5188
5189 Version 3.9 02-Jan-02
5190 ---------------------
5191
5192 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5193
5194 2. If --disable-static was given, the building process failed when trying to
5195 build pcretest and pcregrep. (For some reason it was using libtool to compile
5196 them, which is not right, as they aren't part of the library.)
5197
5198
5199 Version 3.8 18-Dec-01
5200 ---------------------
5201
5202 1. The experimental UTF-8 code was completely screwed up. It was packing the
5203 bytes in the wrong order. How dumb can you get?
5204
5205
5206 Version 3.7 29-Oct-01
5207 ---------------------
5208
5209 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5210 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5211 this didn't happen under Solaris 8, where I normally test things.
5212
5213 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5214 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5215 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5216 this fix an hour or so after the initial 3.7 release.)
5217
5218
5219 Version 3.6 23-Oct-01
5220 ---------------------
5221
5222 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5223 offsets passed as NULL with zero offset count.
5224
5225 2. The config.guess and config.sub files had not been updated when I moved to
5226 the latest autoconf.
5227
5228
5229 Version 3.5 15-Aug-01
5230 ---------------------
5231
5232 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5233 had been forgotten.
5234
5235 2. By using declared but undefined structures, we can avoid using "void"
5236 definitions in pcre.h while keeping the internal definitions of the structures
5237 private.
5238
5239 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5240 user point of view, this means that both static and shared libraries are built
5241 by default, but this can be individually controlled. More of the work of
5242 handling this static/shared cases is now inside libtool instead of PCRE's make
5243 file.
5244
5245 4. The pcretest utility is now installed along with pcregrep because it is
5246 useful for users (to test regexs) and by doing this, it automatically gets
5247 relinked by libtool. The documentation has been turned into a man page, so
5248 there are now .1, .txt, and .html versions in /doc.
5249
5250 5. Upgrades to pcregrep:
5251 (i) Added long-form option names like gnu grep.
5252 (ii) Added --help to list all options with an explanatory phrase.
5253 (iii) Added -r, --recursive to recurse into sub-directories.
5254 (iv) Added -f, --file to read patterns from a file.
5255
5256 6. pcre_exec() was referring to its "code" argument before testing that
5257 argument for NULL (and giving an error if it was NULL).
5258
5259 7. Upgraded Makefile.in to allow for compiling in a different directory from
5260 the source directory.
5261
5262 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5263 options bits, the pointer it was passed was to an int instead of to an unsigned
5264 long int. This mattered only on 64-bit systems.
5265
5266 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5267 generated) instead of pcre.in, which it its source. Also made the same change
5268 in several of the .c files.
5269
5270 10. A new release of gcc defines printf() as a macro, which broke pcretest
5271 because it had an ifdef in the middle of a string argument for printf(). Fixed
5272 by using separate calls to printf().
5273
5274 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5275 script, to force use of CR or LF instead of \n in the source. On non-Unix
5276 systems, the value can be set in config.h.
5277
5278 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5279 absolute limit. Changed the text of the error message to make this clear, and
5280 likewise updated the man page.
5281
5282 13. The limit of 99 on the number of capturing subpatterns has been removed.
5283 The new limit is 65535, which I hope will not be a "real" limit.
5284
5285
5286 Version 3.4 22-Aug-00
5287 ---------------------
5288
5289 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5290
5291 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5292
5293
5294 Version 3.3 01-Aug-00
5295 ---------------------
5296
5297 1. If an octal character was given, but the value was greater than \377, it
5298 was not getting masked to the least significant bits, as documented. This could
5299 lead to crashes in some systems.
5300
5301 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5302 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5303
5304 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5305 These just pass their arguments on to (pcre_free)(), but they are provided
5306 because some uses of PCRE bind it to non-C systems that can call its functions,
5307 but cannot call free() or pcre_free() directly.
5308
5309 4. Add "make test" as a synonym for "make check". Corrected some comments in
5310 the Makefile.
5311
5312 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5313 Makefile.
5314
5315 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5316 command called pgrep for grepping around the active processes.
5317
5318 7. Added the beginnings of support for UTF-8 character strings.
5319
5320 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5321 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5322 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5323 out for the ar command.)
5324
5325
5326 Version 3.2 12-May-00
5327 ---------------------
5328
5329 This is purely a bug fixing release.
5330
5331 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5332 of ZA. This was just one example of several cases that could provoke this bug,
5333 which was introduced by change 9 of version 2.00. The code for breaking
5334 infinite loops after an iteration that matches an empty string was't working
5335 correctly.
5336
5337 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5338 when matched against abbab (for example). After matching an empty string, it
5339 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5340 caused it to match further down the string than it should.
5341
5342 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5343 was there because it doesn't seem to be needed, and it causes trouble on some
5344 systems, as it is not a Standard C header. It has been removed.
5345
5346 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5347 were reported on the Macintosh. The changes were from
5348
5349 while ((c = *(++ptr)) != 0 && c != '\n');
5350 to
5351 while ((c = *(++ptr)) != 0 && c != '\n') ;
5352
5353 Totally extraordinary, but if that's what it takes...
5354
5355 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5356 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5357 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5358 assumes the way PCRE uses memmove() (always moving upwards).
5359
5360 6. PCRE is being used in one environment where strchr() is not available. There
5361 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5362 faster code anyway.
5363
5364
5365 Version 3.1 09-Feb-00
5366 ---------------------
5367
5368 The only change in this release is the fixing of some bugs in Makefile.in for
5369 the "install" target:
5370
5371 (1) It was failing to install pcreposix.h.
5372
5373 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5374
5375
5376 Version 3.0 01-Feb-00
5377 ---------------------
5378
5379 1. Add support for the /+ modifier to perltest (to output $` like it does in
5380 pcretest).
5381
5382 2. Add support for the /g modifier to perltest.
5383
5384 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5385 matches null strings.
5386
5387 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5388 pattern. Perl treats empty patterns specially - it reuses the most recent
5389 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5390 effect.
5391
5392 5. The POSIX interface was broken in that it was just handing over the POSIX
5393 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5394 required a bigger vector, with some working space on the end. This means that
5395 the POSIX wrapper now has to get and free some memory, and copy the results.
5396
5397 6. Added some simple autoconf support, placing the test data and the
5398 documentation in separate directories, re-organizing some of the
5399 information files, and making it build pcre-config (a GNU standard). Also added
5400 libtool support for building PCRE as a shared library, which is now the
5401 default.
5402
5403 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5404 09 are not valid octal constants. Single digits will be used for minor values
5405 less than 10.
5406
5407 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5408 existing programs that set these in the POSIX interface can use PCRE without
5409 modification.
5410
5411 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5412 return all that pcre_info() returns, plus additional data. The pcre_info()
5413 function is retained for compatibility, but is considered to be obsolete.
5414
5415 10. Added experimental recursion feature (?R) to handle one common case that
5416 Perl 5.6 will be able to do with (?p{...}).
5417
5418 11. Added support for POSIX character classes like [:alpha:], which Perl is
5419 adopting.
5420
5421
5422 Version 2.08 31-Aug-99
5423 ----------------------
5424
5425 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5426 trying to match at the startoffset position, but instead was moving forward to
5427 the next newline as if a previous match had failed.
5428
5429 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5430 and could get into a loop if a null string was matched other than at the start
5431 of the subject.
5432
5433 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5434 be distinguished at compile time, and for completeness also added PCRE_DATE.
5435
5436 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5437 in GnuWin32 environments.
5438
5439
5440 Version 2.07 29-Jul-99
5441 ----------------------
5442
5443 1. The documentation is now supplied in plain text form and HTML as well as in
5444 the form of man page sources.
5445
5446 2. C++ compilers don't like assigning (void *) values to other pointer types.
5447 In particular this affects malloc(). Although there is no problem in Standard
5448 C, I've put in casts to keep C++ compilers happy.
5449
5450 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5451 should be (const char *).
5452
5453 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5454 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5455 However, I haven't made this a standard facility. The documentation doesn't
5456 mention it, and the Makefile doesn't support it.
5457
5458 5. The Makefile now contains an "install" target, with editable destinations at
5459 the top of the file. The pcretest program is not installed.
5460
5461 6. pgrep -V now gives the PCRE version number and date.
5462
5463 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5464 causing the entire string to be ignored, instead of just the last character.
5465
5466 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5467 non-matching string, it can take a very, very long time, even for strings of
5468 quite modest length, because of the nested recursion. PCRE now does better in
5469 some of these cases. It does this by remembering the last required literal
5470 character in the pattern, and pre-searching the subject to ensure it is present
5471 before running the real match. In other words, it applies a heuristic to detect
5472 some types of certain failure quickly, and in the above example, if presented
5473 with a string that has no trailing " it gives "no match" very quickly.
5474
5475 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5476 other alternatives are tried instead.
5477
5478
5479 Version 2.06 09-Jun-99
5480 ----------------------
5481
5482 1. Change pcretest's output for amount of store used to show just the code
5483 space, because the remainder (the data block) varies in size between 32-bit and
5484 64-bit systems.
5485
5486 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5487 start matching at. This allows lookbehinds to work when searching for multiple
5488 occurrences in a string.
5489
5490 3. Added additional options to pcretest for testing multiple occurrences:
5491
5492 /+ outputs the rest of the string that follows a match
5493 /g loops for multiple occurrences, using the new startoffset argument
5494 /G loops for multiple occurrences by passing an incremented pointer
5495
5496 4. PCRE wasn't doing the "first character" optimization for patterns starting
5497 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5498 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5499 the letter 'x'. On long subject strings, this gives a significant speed-up.
5500
5501
5502 Version 2.05 21-Apr-99
5503 ----------------------
5504
5505 1. Changed the type of magic_number from int to long int so that it works
5506 properly on 16-bit systems.
5507
5508 2. Fixed a bug which caused patterns starting with .* not to work correctly
5509 when the subject string contained newline characters. PCRE was assuming
5510 anchoring for such patterns in all cases, which is not correct because .* will
5511 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5512 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5513 must be retried after every newline in the subject.
5514
5515
5516 Version 2.04 18-Feb-99
5517 ----------------------
5518
5519 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5520 computation of the store needed to hold the pattern was incorrect (too large).
5521 If such patterns were nested a few deep, this could multiply and become a real
5522 problem.
5523
5524 2. Added /M option to pcretest to show the memory requirement of a specific
5525 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5526
5527 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5528 compiled in such a way that the backtracking after subsequent failure was
5529 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5530 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5531
5532
5533 Version 2.03 02-Feb-99
5534 ----------------------
5535
5536 1. Fixed typo and small mistake in man page.
5537
5538 2. Added 4th condition (GPL supersedes if conflict) and created separate
5539 LICENCE file containing the conditions.
5540
5541 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5542 Perl, that is the internal \ allows the delimiter to be included in the
5543 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5544 the final delimiter, add \ to the end of the pattern (to test the error).
5545
5546 4. Added the convenience functions for extracting substrings after a successful
5547 match. Updated pcretest to make it able to test these functions.
5548
5549
5550 Version 2.02 14-Jan-99
5551 ----------------------
5552
5553 1. Initialized the working variables associated with each extraction so that
5554 their saving and restoring doesn't refer to uninitialized store.
5555
5556 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5557 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5558 fix the problem.
5559
5560 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5561 calls, and wasn't printing the correct value for compiling calls. Increased the
5562 default value of LOOPREPEAT, and the number of significant figures in the
5563 times.
5564
5565 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5566
5567 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5568 a building problem on Windows NT with a FAT file system.
5569
5570
5571 Version 2.01 21-Oct-98
5572 ----------------------
5573
5574 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5575 to character tables built by pcre_maketables() in the current locale. If NULL
5576 is passed, the default tables are used.
5577
5578
5579 Version 2.00 24-Sep-98
5580 ----------------------
5581
5582 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5583 it any more.
5584
5585 2. Allow quantification of (?>) groups, and make it work correctly.
5586
5587 3. The first character computation wasn't working for (?>) groups.
5588
5589 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5590 end of the subject) and add 5.005's \z, which really does match only at the
5591 very end of the subject.
5592
5593 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5594
5595 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5596 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5597 localized options. All options to pcre_study() were also removed.
5598
5599 7. Add other new features from 5.005:
5600
5601 $(?<= positive lookbehind
5602 $(?<! negative lookbehind
5603 (?imsx-imsx) added the unsetting capability
5604 such a setting is global if at outer level; local otherwise
5605 (?imsx-imsx:) non-capturing groups with option setting
5606 (?(cond)re|re) conditional pattern matching
5607
5608 A backreference to itself in a repeated group matches the previous
5609 captured string.
5610
5611 8. General tidying up of studying (both automatic and via "study")
5612 consequential on the addition of new assertions.
5613
5614 9. As in 5.005, unlimited repeated groups that could match an empty substring
5615 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5616 runtime if any iteration does actually match an empty substring.
5617
5618 10. Include the RunTest script in the distribution.
5619
5620 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5621 discrepancies, some of which were old and were also with respect to 5.004. They
5622 have now been fixed.
5623
5624
5625 Version 1.09 28-Apr-98
5626 ----------------------
5627
5628 1. A negated single character class followed by a quantifier with a minimum
5629 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5630 program crashes, or just wrong answers. This did not apply to negated classes
5631 containing more than one character, or to minima other than one.
5632
5633
5634 Version 1.08 27-Mar-98
5635 ----------------------
5636
5637 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5638
5639 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5640 latter must appear before anything that relies on it in the pattern.
5641
5642
5643 Version 1.07 16-Feb-98
5644 ----------------------
5645
5646 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5647 repeat of a potentially empty string).
5648
5649
5650 Version 1.06 23-Jan-98
5651 ----------------------
5652
5653 1. Added Markus Oberhumer's little patches for C++.
5654
5655 2. Literal strings longer than 255 characters were broken.
5656
5657
5658 Version 1.05 23-Dec-97
5659 ----------------------
5660
5661 1. Negated character classes containing more than one character were failing if
5662 PCRE_CASELESS was set at run time.
5663
5664
5665 Version 1.04 19-Dec-97
5666 ----------------------
5667
5668 1. Corrected the man page, where some "const" qualifiers had been omitted.
5669
5670 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5671 input syntax.
5672
5673 3. Fixed memory leak which occurred when a regex with back references was
5674 matched with an offsets vector that wasn't big enough. The temporary memory
5675 that is used in this case wasn't being freed if the match failed.
5676
5677 4. Tidied pcretest to ensure it frees memory that it gets.
5678
5679 5. Temporary memory was being obtained in the case where the passed offsets
5680 vector was exactly big enough.
5681
5682 6. Corrected definition of offsetof() from change 5 below.
5683
5684 7. I had screwed up change 6 below and broken the rules for the use of
5685 setjmp(). Now fixed.
5686
5687
5688 Version 1.03 18-Dec-97
5689 ----------------------
5690
5691 1. A erroneous regex with a missing opening parenthesis was correctly
5692 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5693 on some systems.
5694
5695 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5696 it was reported that one broken compiler failed on the former because "code" is
5697 also an independent variable.
5698
5699 3. The erroneous regex a[]b caused an array overrun reference.
5700
5701 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5702 fail on data ending with that character. (It was going on too far, and checking
5703 the next character, typically a binary zero.) This was specific to the
5704 optimized code for single-character negative classes.
5705
5706 5. Added a contributed patch from the TIN world which does the following:
5707
5708 + Add an undef for memmove, in case the the system defines a macro for it.
5709
5710 + Add a definition of offsetof(), in case there isn't one. (I don't know
5711 the reason behind this - offsetof() is part of the ANSI standard - but
5712 it does no harm).
5713
5714 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5715 most of the places where whitespace preceded '#'. I have given up and
5716 allowed the remaining 2 cases to be at the margin.
5717
5718 + Rename some variables in pcre to eliminate shadowing. This seems very
5719 pedantic, but does no harm, of course.
5720
5721 6. Moved the call to setjmp() into its own function, to get rid of warnings
5722 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5723
5724 7. Constructs such as \d{8,} were compiling into the equivalent of
5725 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5726 outcome, but in this particular case used more store than had been allocated,
5727 which caused the bug to be discovered because it threw up an internal error.
5728
5729 8. The debugging code in both pcre and pcretest for outputting the compiled
5730 form of a regex was going wrong in the case of back references followed by
5731 curly-bracketed repeats.
5732
5733
5734 Version 1.02 12-Dec-97
5735 ----------------------
5736
5737 1. Typos in pcre.3 and comments in the source fixed.
5738
5739 2. Applied a contributed patch to get rid of places where it used to remove
5740 'const' from variables, and fixed some signed/unsigned and uninitialized
5741 variable warnings.
5742
5743 3. Added the "runtest" target to Makefile.
5744
5745 4. Set default compiler flag to -O2 rather than just -O.
5746
5747
5748 Version 1.01 19-Nov-97
5749 ----------------------
5750
5751 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5752 like /([ab]*)*/, that is, for classes with more than one character in them.
5753
5754 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5755 as /((?>a*))*/ (a PCRE_EXTRA facility).
5756
5757
5758 Version 1.00 18-Nov-97
5759 ----------------------
5760
5761 1. Added compile-time macros to support systems such as SunOS4 which don't have
5762 memmove() or strerror() but have other things that can be used instead.
5763
5764 2. Arranged that "make clean" removes the executables.
5765
5766
5767 Version 0.99 27-Oct-97
5768 ----------------------
5769
5770 1. Fixed bug in code for optimizing classes with only one character. It was
5771 initializing a 32-byte map regardless, which could cause it to run off the end
5772 of the memory it had got.
5773
5774 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5775
5776
5777 Version 0.98 22-Oct-97
5778 ----------------------
5779
5780 1. Fixed bug in code for handling temporary memory usage when there are more
5781 back references than supplied space in the ovector. This could cause segfaults.
5782
5783
5784 Version 0.97 21-Oct-97
5785 ----------------------
5786
5787 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5788
5789 2. Optimized negated single characters not to use a bit map.
5790
5791 3. Brought error texts together as macro definitions; clarified some of them;
5792 fixed one that was wrong - it said "range out of order" when it meant "invalid
5793 escape sequence".
5794
5795 4. Changed some char * arguments to const char *.
5796
5797 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5798
5799 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5800 pcretest.
5801
5802
5803 Version 0.96 16-Oct-97
5804 ----------------------
5805
5806 1. Added a simple "pgrep" utility to the distribution.
5807
5808 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5809 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5810 where "ddd" means "one or more decimal digits".
5811
5812 3. Fixed serious bug. If a pattern had a back reference, but the call to
5813 pcre_exec() didn't supply a large enough ovector to record the related
5814 identifying subpattern, the match always failed. PCRE now remembers the number
5815 of the largest back reference, and gets some temporary memory in which to save
5816 the offsets during matching if necessary, in order to ensure that
5817 backreferences always work.
5818
5819 4. Increased the compatibility with Perl in a number of ways:
5820
5821 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5822 to request this handling. The option can be set at compile or exec time.
5823
5824 (b) $ matches before a terminating newline by default; an option
5825 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5826 mode). The option can be set at compile or exec time.
5827
5828 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5829 the same as Perl's. If the decimal number it represents is less than 10
5830 or there aren't that many previous left capturing parentheses, an octal
5831 escape is read. Inside a character class, it's always an octal escape,
5832 even if it is a single digit.
5833
5834 (d) An escaped but undefined alphabetic character is taken as a literal,
5835 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5836 escapes.
5837
5838 (e) {0} is now permitted. (The previous item is removed from the compiled
5839 pattern).
5840
5841 5. Changed all the names of code files so that the basic parts are no longer
5842 than 10 characters, and abolished the teeny "globals.c" file.
5843
5844 6. Changed the handling of character classes; they are now done with a 32-byte
5845 bit map always.
5846
5847 7. Added the -d and /D options to pcretest to make it possible to look at the
5848 internals of compilation without having to recompile pcre.
5849
5850
5851 Version 0.95 23-Sep-97
5852 ----------------------
5853
5854 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5855 \x20 at the start of a run of normal characters. These were being treated as
5856 real characters, instead of the source characters being re-checked.
5857
5858
5859 Version 0.94 18-Sep-97
5860 ----------------------
5861
5862 1. The functions are now thread-safe, with the caveat that the global variables
5863 containing pointers to malloc() and free() or alternative functions are the
5864 same for all threads.
5865
5866 2. Get pcre_study() to generate a bitmap of initial characters for non-
5867 anchored patterns when this is possible, and use it if passed to pcre_exec().
5868
5869
5870 Version 0.93 15-Sep-97
5871 ----------------------
5872
5873 1. /(b)|(:+)/ was computing an incorrect first character.
5874
5875 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5876 but not actually doing anything yet.
5877
5878 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5879 as Perl does (e.g. [-az] or [az-]).
5880
5881 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5882 all possible positions.
5883
5884 5. Split up into different modules to avoid including unneeded functions in a
5885 compiled binary. However, compile and exec are still in one module. The "study"
5886 function is split off.
5887
5888 6. The character tables are now in a separate module whose source is generated
5889 by an auxiliary program - but can then be edited by hand if required. There are
5890 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5891 toupper() in the code.
5892
5893 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5894 make them global. Abolish the function for setting them, as the caller can now
5895 set them directly.
5896
5897
5898 Version 0.92 11-Sep-97
5899 ----------------------
5900
5901 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5902 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5903
5904 2. Caseless matching was not working in character classes if the characters in
5905 the pattern were in upper case.
5906
5907 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5908
5909 4. Make PCRE_ANCHORED public and accept as a compile option.
5910
5911 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5912 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5913 pass them.
5914
5915 6. Give an error if bad option bits passed at compile or run time.
5916
5917 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5918 pcretest to cause it to pass that flag.
5919
5920 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5921 options, and the first character, if set.
5922
5923 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5924
5925
5926 Version 0.91 10-Sep-97
5927 ----------------------
5928
5929 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5930 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5931
5932 2. PCRE was looping on encountering an indefinitely repeated back reference to
5933 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5934 Perl does - treats the match as successful.
5935
5936 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5