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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1647 - (show annotations)
Thu May 19 18:02:51 2016 UTC (3 years, 6 months ago) by zherczeg
File size: 289086 byte(s)
A racing condition is fixed in JIT reported by Mozilla.
1 ChangeLog for PCRE
2 ------------------
3
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
6
7 Version 8.39 xx-xxxxxx-201x
8 ---------------------------
9
10 1. If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
11 an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
12 misbehaved. This bug was found by the LLVM fuzzer.
13
14 2. Similar to the above, if an isolated \E was present between an item and its
15 qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
16 bug was found by the LLVM fuzzer.
17
18 3. Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
19 working correctly in UCP mode.
20
21 4. The POSIX wrapper function regexec() crashed if the option REG_STARTEND
22 was set when the pmatch argument was NULL. It now returns REG_INVARG.
23
24 5. Allow for up to 32-bit numbers in the ordin() function in pcregrep.
25
26 6. An empty \Q\E sequence between an item and its qualifier caused
27 pcre_compile() to misbehave when auto callouts were enabled. This bug was
28 found by the LLVM fuzzer.
29
30 7. If a pattern that was compiled with PCRE_EXTENDED started with white
31 space or a #-type comment that was followed by (?-x), which turns off
32 PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
33 pcre_compile() assumed that (?-x) applied to the whole pattern and
34 consequently mis-compiled it. This bug was found by the LLVM fuzzer.
35
36 8. An call of pcre_copy_named_substring() for a named substring whose number
37 was greater than the space in the ovector could cause a crash.
38
39 9. Yet another buffer overflow bug involved duplicate named groups with a
40 group that reset capture numbers (compare 8.38/7 below). Once again, I have
41 just allowed for more memory, even if not needed. (A proper fix is
42 implemented in PCRE2, but it involves a lot of refactoring.)
43
44 10. pcre_get_substring_list() crashed if the use of \K in a match caused the
45 start of the match to be earlier than the end.
46
47 11. Migrating appropriate PCRE2 JIT improvements to PCRE.
48
49 12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
50 assertion, caused pcretest to generate incorrect output, and also to read
51 uninitialized memory (detected by ASAN or valgrind).
52
53 13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
54 nested set of parentheses of sufficient size caused an overflow of the
55 compiling workspace (which was diagnosed, but of course is not desirable).
56
57 14. And yet another buffer overflow bug involving duplicate named groups, this
58 time nested, with a nested back reference. Yet again, I have just allowed
59 for more memory, because anything more needs all the refactoring that has
60 been done for PCRE2. An example pattern that provoked this bug is:
61 /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
62 registered as CVE-2016-1283.
63
64 15. pcretest went into a loop if global matching was requested with an ovector
65 size less than 2. It now gives an error message. This bug was found by
66 afl-fuzz.
67
68 16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
69 ("assertion expected") when (?(?C) was not followed by an opening
70 parenthesis.
71
72 17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
73 actually affect anything, by sheer luck.
74
75 18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
76 static compilation.
77
78 19. Modified the RunTest script to incorporate a valgrind suppressions file so
79 that certain errors, provoked by the SSE2 instruction set when JIT is used,
80 are ignored.
81
82 20. A racing condition is fixed in JIT reported by Mozilla.
83
84
85 Version 8.38 23-November-2015
86 -----------------------------
87
88 1. If a group that contained a recursive back reference also contained a
89 forward reference subroutine call followed by a non-forward-reference
90 subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
91 compile correct code, leading to undefined behaviour or an internally
92 detected error. This bug was discovered by the LLVM fuzzer.
93
94 2. Quantification of certain items (e.g. atomic back references) could cause
95 incorrect code to be compiled when recursive forward references were
96 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
97 This bug was discovered by the LLVM fuzzer.
98
99 3. A repeated conditional group whose condition was a reference by name caused
100 a buffer overflow if there was more than one group with the given name.
101 This bug was discovered by the LLVM fuzzer.
102
103 4. A recursive back reference by name within a group that had the same name as
104 another group caused a buffer overflow. For example:
105 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
106
107 5. A forward reference by name to a group whose number is the same as the
108 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
109 a buffer overflow at compile time. This bug was discovered by the LLVM
110 fuzzer.
111
112 6. A lookbehind assertion within a set of mutually recursive subpatterns could
113 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
114
115 7. Another buffer overflow bug involved duplicate named groups with a
116 reference between their definition, with a group that reset capture
117 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
118 fixed by always allowing for more memory, even if not needed. (A proper fix
119 is implemented in PCRE2, but it involves more refactoring.)
120
121 8. There was no check for integer overflow in subroutine calls such as (?123).
122
123 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
124 being treated as a literal 'l' instead of causing an error.
125
126 10. There was a buffer overflow if pcre_exec() was called with an ovector of
127 size 1. This bug was found by american fuzzy lop.
128
129 11. If a non-capturing group containing a conditional group that could match
130 an empty string was repeated, it was not identified as matching an empty
131 string itself. For example: /^(?:(?(1)x|)+)+$()/.
132
133 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
134 \a and \e in test subject lines.
135
136 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
137 instead of the EBCDIC value.
138
139 14. The handling of \c in an EBCDIC environment has been revised so that it is
140 now compatible with the specification in Perl's perlebcdic page.
141
142 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
143 ASCII/Unicode. This has now been added to the list of characters that are
144 recognized as white space in EBCDIC.
145
146 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
147 error (correctly) when used outside a class, but did not give an error
148 within a class.
149
150 17. \h within a class was incorrectly compiled in EBCDIC environments.
151
152 18. A pattern with an unmatched closing parenthesis that contained a backward
153 assertion which itself contained a forward reference caused buffer
154 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
155
156 19. JIT should return with error when the compiled pattern requires more stack
157 space than the maximum.
158
159 20. A possessively repeated conditional group that could match an empty string,
160 for example, /(?(R))*+/, was incorrectly compiled.
161
162 21. Fix infinite recursion in the JIT compiler when certain patterns such as
163 /(?:|a|){100}x/ are analysed.
164
165 22. Some patterns with character classes involving [: and \\ were incorrectly
166 compiled and could cause reading from uninitialized memory or an incorrect
167 error diagnosis.
168
169 23. Pathological patterns containing many nested occurrences of [: caused
170 pcre_compile() to run for a very long time.
171
172 24. A conditional group with only one branch has an implicit empty alternative
173 branch and must therefore be treated as potentially matching an empty
174 string.
175
176 25. If (?R was followed by - or + incorrect behaviour happened instead of a
177 diagnostic.
178
179 26. Arrange to give up on finding the minimum matching length for overly
180 complex patterns.
181
182 27. Similar to (4) above: in a pattern with duplicated named groups and an
183 occurrence of (?| it is possible for an apparently non-recursive back
184 reference to become recursive if a later named group with the relevant
185 number is encountered. This could lead to a buffer overflow. Wen Guanxing
186 from Venustech ADLAB discovered this bug.
187
188 28. If pcregrep was given the -q option with -c or -l, or when handling a
189 binary file, it incorrectly wrote output to stdout.
190
191 29. The JIT compiler did not restore the control verb head in case of *THEN
192 control verbs. This issue was found by Karl Skomski with a custom LLVM
193 fuzzer.
194
195 30. Error messages for syntax errors following \g and \k were giving inaccurate
196 offsets in the pattern.
197
198 31. Added a check for integer overflow in conditions (?(<digits>) and
199 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
200 fuzzer.
201
202 32. Handling recursive references such as (?2) when the reference is to a group
203 later in the pattern uses code that is very hacked about and error-prone.
204 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
205 give an internal error if it is obvious that compiling has gone wrong.
206
207 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
208 This issue was found by Karl Skomski with a custom LLVM fuzzer.
209
210 34. The JIT compiler should restore the control chain for empty possessive
211 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
212
213 35. Match limit check added to JIT recursion. This issue was found by Karl
214 Skomski with a custom LLVM fuzzer.
215
216 36. Yet another case similar to 27 above has been circumvented by an
217 unconditional allocation of extra memory. This issue is fixed "properly" in
218 PCRE2 by refactoring the way references are handled. Wen Guanxing
219 from Venustech ADLAB discovered this bug.
220
221 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
222 with a custom LLVM fuzzer.
223
224 38. Fixed a corner case of range optimization in JIT.
225
226 39. An incorrect error "overran compiling workspace" was given if there were
227 exactly enough group forward references such that the last one extended
228 into the workspace safety margin. The next one would have expanded the
229 workspace. The test for overflow was not including the safety margin.
230
231 40. A match limit issue is fixed in JIT which was found by Karl Skomski
232 with a custom LLVM fuzzer.
233
234 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
235 work under Windows. (Why has it taken so long for anyone to notice?)
236
237 42. In a character class such as [\W\p{Any}] where both a negative-type escape
238 ("not a word character") and a property escape were present, the property
239 escape was being ignored.
240
241 43. Fix crash caused by very long (*MARK) or (*THEN) names.
242
243 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
244 by a single ASCII character in a class item, was incorrectly compiled in
245 UCP mode. The POSIX class got lost, but only if the single character
246 followed it.
247
248 45. [:punct:] in UCP mode was matching some characters in the range 128-255
249 that should not have been matched.
250
251 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
252 class, all characters with code points greater than 255 are in the class.
253 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
254 such as \w are turned into Unicode properties), wide characters were not
255 correctly handled, and could fail to match.
256
257
258 Version 8.37 28-April-2015
259 --------------------------
260
261 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
262 for those parentheses to be closed with whatever has been captured so far.
263 However, it was failing to mark any other groups between the hightest
264 capture so far and the currrent group as "unset". Thus, the ovector for
265 those groups contained whatever was previously there. An example is the
266 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
267
268 2. If an assertion condition was quantified with a minimum of zero (an odd
269 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
270
271 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
272 unrecognized modifier, a crash could occur.
273
274 4. An attempt to do global matching in pcretest with a zero-length ovector
275 caused a crash.
276
277 5. Fixed a memory leak during matching that could occur for a subpattern
278 subroutine call (recursive or otherwise) if the number of captured groups
279 that had to be saved was greater than ten.
280
281 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
282 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
283 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
284
285 7. A UTF pattern containing a "not" match of a non-ASCII character and a
286 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
287
288 8. When a pattern is compiled, it remembers the highest back reference so that
289 when matching, if the ovector is too small, extra memory can be obtained to
290 use instead. A conditional subpattern whose condition is a check on a
291 capture having happened, such as, for example in the pattern
292 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
293 setting the highest backreference number. This mattered only if pcre_exec()
294 was called with an ovector that was too small to hold the capture, and there
295 was no other kind of back reference (a situation which is probably quite
296 rare). The effect of the bug was that the condition was always treated as
297 FALSE when the capture could not be consulted, leading to a incorrect
298 behaviour by pcre_exec(). This bug has been fixed.
299
300 9. A reference to a duplicated named group (either a back reference or a test
301 for being set in a conditional) that occurred in a part of the pattern where
302 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
303 to be incorrectly calculated, leading to overwriting.
304
305 10. A mutually recursive set of back references such as (\2)(\1) caused a
306 segfault at study time (while trying to find the minimum matching length).
307 The infinite loop is now broken (with the minimum length unset, that is,
308 zero).
309
310 11. If an assertion that was used as a condition was quantified with a minimum
311 of zero, matching went wrong. In particular, if the whole group had
312 unlimited repetition and could match an empty string, a segfault was
313 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
314 assertions to be quantified, but not if they are being used as conditions,
315 so the above pattern is faulted by Perl. PCRE has now been changed so that
316 it also rejects such patterns.
317
318 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
319 failed to allow the zero-repeat case if pcre2_exec() was called with an
320 ovector too small to capture the group.
321
322 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
323 Red Hat Product Security:
324
325 (a) A crash if /K and /F were both set with the option to save the compiled
326 pattern.
327
328 (b) Another crash if the option to print captured substrings in a callout
329 was combined with setting a null ovector, for example \O\C+ as a subject
330 string.
331
332 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
333 forward reference repeated a large (but limited) number of times within a
334 repeated outer group that has a zero minimum quantifier, caused incorrect
335 code to be compiled, leading to the error "internal error:
336 previously-checked referenced subpattern not found" when an incorrect
337 memory address was read. This bug was reported as "heap overflow",
338 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
339 CVE-2015-2325.
340
341 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
342 call within a group that also contained a recursive back reference caused
343 incorrect code to be compiled. This bug was reported as "heap overflow",
344 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
345 number CVE-2015-2326.
346
347 24. Computing the size of the JIT read-only data in advance has been a source
348 of various issues, and new ones are still appear unfortunately. To fix
349 existing and future issues, size computation is eliminated from the code,
350 and replaced by on-demand memory allocation.
351
352 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
353 adjacent to the end of the range, and the range contained characters with
354 more than one other case, caused incorrect behaviour when compiled in UTF
355 mode. In that example, the range a-j was left out of the class.
356
357 26. Fix JIT compilation of conditional blocks, which assertion
358 is converted to (*FAIL). E.g: /(?(?!))/.
359
360 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
361 discovered by the LLVM fuzzer.
362
363 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
364 when this assertion was used as a condition, for example (?(?!)a|b). In
365 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
366 error about an unsupported item.
367
368 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
369 possessification code could take exponential time to complete. A recursion
370 depth limit of 1000 has been imposed to limit the resources used by this
371 optimization.
372
373 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
374 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
375 because \S ensures they are all in the class. The code for doing this was
376 interacting badly with the code for computing the amount of space needed to
377 compile the pattern, leading to a buffer overflow. This bug was discovered
378 by the LLVM fuzzer.
379
380 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
381 other kinds of group caused stack overflow at compile time. This bug was
382 discovered by the LLVM fuzzer.
383
384 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
385 between a subroutine call and its quantifier was incorrectly compiled,
386 leading to buffer overflow or other errors. This bug was discovered by the
387 LLVM fuzzer.
388
389 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
390 assertion after (?(. The code was failing to check the character after
391 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
392 was discovered by the LLVM fuzzer.
393
394 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
395 a fixed maximum following a group that contains a subroutine reference was
396 incorrectly compiled and could trigger buffer overflow. This bug was
397 discovered by the LLVM fuzzer.
398
399 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
400 caused a stack overflow instead of the diagnosis of a non-fixed length
401 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
402
403 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
404 (e.g. /(?<=\Ka)/) could make pcregrep loop.
405
406 37. There was a similar problem to 36 in pcretest for global matches.
407
408 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
409 and a subsequent item in the pattern caused a non-match, backtracking over
410 the repeated \X did not stop, but carried on past the start of the subject,
411 causing reference to random memory and/or a segfault. There were also some
412 other cases where backtracking after \C could crash. This set of bugs was
413 discovered by the LLVM fuzzer.
414
415 39. The function for finding the minimum length of a matching string could take
416 a very long time if mutual recursion was present many times in a pattern,
417 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
418 method has been implemented. This infelicity was discovered by the LLVM
419 fuzzer.
420
421 40. Static linking against the PCRE library using the pkg-config module was
422 failing on missing pthread symbols.
423
424
425 Version 8.36 26-September-2014
426 ------------------------------
427
428 1. Got rid of some compiler warnings in the C++ modules that were shown up by
429 -Wmissing-field-initializers and -Wunused-parameter.
430
431 2. The tests for quantifiers being too big (greater than 65535) were being
432 applied after reading the number, and stupidly assuming that integer
433 overflow would give a negative number. The tests are now applied as the
434 numbers are read.
435
436 3. Tidy code in pcre_exec.c where two branches that used to be different are
437 now the same.
438
439 4. The JIT compiler did not generate match limit checks for certain
440 bracketed expressions with quantifiers. This may lead to exponential
441 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
442 issue should be resolved now.
443
444 5. Fixed an issue, which occures when nested alternatives are optimized
445 with table jumps.
446
447 6. Inserted two casts and changed some ints to size_t in the light of some
448 reported 64-bit compiler warnings (Bugzilla 1477).
449
450 7. Fixed a bug concerned with zero-minimum possessive groups that could match
451 an empty string, which sometimes were behaving incorrectly in the
452 interpreter (though correctly in the JIT matcher). This pcretest input is
453 an example:
454
455 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
456 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
457
458 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
459 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
460 for an empty string was breaking the inner loop and carrying on at a lower
461 level, when possessive repeated groups should always return to a higher
462 level as they have no backtrack points in them. The empty string test now
463 occurs at the outer level.
464
465 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
466 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
467
468 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
469 doesn't).
470
471 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
472 to hold a pattern to be incorrectly computed (too small) when there were
473 named back references to duplicated names. This could cause "internal
474 error: code overflow" or "double free or corruption" or other memory
475 handling errors.
476
477 11. When named subpatterns had the same prefixes, back references could be
478 confused. For example, in this pattern:
479
480 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
481
482 the reference to 'Name' was incorrectly treated as a reference to a
483 duplicate name.
484
485 12. A pattern such as /^s?c/mi8 where the optional character has more than
486 one "other case" was incorrectly compiled such that it would only try to
487 match starting at "c".
488
489 13. When a pattern starting with \s was studied, VT was not included in the
490 list of possible starting characters; this should have been part of the
491 8.34/18 patch.
492
493 14. If a character class started [\Qx]... where x is any character, the class
494 was incorrectly terminated at the ].
495
496 15. If a pattern that started with a caseless match for a character with more
497 than one "other case" was studied, PCRE did not set up the starting code
498 unit bit map for the list of possible characters. Now it does. This is an
499 optimization improvement, not a bug fix.
500
501 16. The Unicode data tables have been updated to Unicode 7.0.0.
502
503 17. Fixed a number of memory leaks in pcregrep.
504
505 18. Avoid a compiler warning (from some compilers) for a function call with
506 a cast that removes "const" from an lvalue by using an intermediate
507 variable (to which the compiler does not object).
508
509 19. Incorrect code was compiled if a group that contained an internal recursive
510 back reference was optional (had quantifier with a minimum of zero). This
511 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
512 caused segmentation faults because of stack overflows at compile time.
513
514 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
515 group that is quantified with an indefinite repeat, caused a compile-time
516 loop which used up all the system stack and provoked a segmentation fault.
517 This was not the same bug as 19 above.
518
519 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
520 Patch by Mike Frysinger.
521
522
523 Version 8.35 04-April-2014
524 --------------------------
525
526 1. A new flag is set, when property checks are present in an XCLASS.
527 When this flag is not set, PCRE can perform certain optimizations
528 such as studying these XCLASS-es.
529
530 2. The auto-possessification of character sets were improved: a normal
531 and an extended character set can be compared now. Furthermore
532 the JIT compiler optimizes more character set checks.
533
534 3. Got rid of some compiler warnings for potentially uninitialized variables
535 that show up only when compiled with -O2.
536
537 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
538 of a match later then the end of the match. The pcretest program was not
539 handling the case sensibly - it was outputting from the start to the next
540 binary zero. It now reports this situation in a message, and outputs the
541 text from the end to the start.
542
543 5. Fast forward search is improved in JIT. Instead of the first three
544 characters, any three characters with fixed position can be searched.
545 Search order: first, last, middle.
546
547 6. Improve character range checks in JIT. Characters are read by an inprecise
548 function now, which returns with an unknown value if the character code is
549 above a certain threshold (e.g: 256). The only limitation is that the value
550 must be bigger than the threshold as well. This function is useful when
551 the characters above the threshold are handled in the same way.
552
553 7. The macros whose names start with RAWUCHAR are placeholders for a future
554 mode in which only the bottom 21 bits of 32-bit data items are used. To
555 make this more memorable for those maintaining the code, the names have
556 been changed to start with UCHAR21, and an extensive comment has been added
557 to their definition.
558
559 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
560 to the export list in Makefile.am (they were accidentally omitted from the
561 8.34 tarball).
562
563 9. The informational output from pcretest used the phrase "starting byte set"
564 which is inappropriate for the 16-bit and 32-bit libraries. As the output
565 for "first char" and "need char" really means "non-UTF-char", I've changed
566 "byte" to "char", and slightly reworded the output. The documentation about
567 these values has also been (I hope) clarified.
568
569 10. Another JIT related optimization: use table jumps for selecting the correct
570 backtracking path, when more than four alternatives are present inside a
571 bracket.
572
573 11. Empty match is not possible, when the minimum length is greater than zero,
574 and there is no \K in the pattern. JIT should avoid empty match checks in
575 such cases.
576
577 12. In a caseless character class with UCP support, when a character with more
578 than one alternative case was not the first character of a range, not all
579 the alternative cases were added to the class. For example, s and \x{17f}
580 are both alternative cases for S: the class [RST] was handled correctly,
581 but [R-T] was not.
582
583 13. The configure.ac file always checked for pthread support when JIT was
584 enabled. This is not used in Windows, so I have put this test inside a
585 check for the presence of windows.h (which was already tested for).
586
587 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
588 The algorithm provides a way to skip certain starting offsets, and usually
589 faster than linear prefix searches.
590
591 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
592 as for 'fr_FR' and 'french'. For some reason, however, it then used the
593 Windows-specific input and output files, which have 'french' screwed in.
594 So this could never have worked. One of the problems with locales is that
595 they aren't always the same. I have now updated RunTest so that it checks
596 the output of the locale test (test 3) against three different output
597 files, and it allows the test to pass if any one of them matches. With luck
598 this should make the test pass on some versions of Solaris where it was
599 failing. Because of the uncertainty, the script did not used to stop if
600 test 3 failed; it now does. If further versions of a French locale ever
601 come to light, they can now easily be added.
602
603 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
604 there was a message during ./configure, but it did not stop. This now
605 provokes an error. The invalid example in README has been corrected.
606 If a value less than the minimum is given, the minimum value has always
607 been used, but now a warning is given.
608
609 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
610 was a bug in the test system, which is now fixed. Also, the list of various
611 configurations that are tested for each release did not have one with both
612 16/32 bits and --enable-bar-anycrlf. It now does.
613
614 18. pcretest was missing "-C bsr" for displaying the \R default setting.
615
616 19. Little endian PowerPC systems are supported now by the JIT compiler.
617
618 20. The fast forward newline mechanism could enter to an infinite loop on
619 certain invalid UTF-8 input. Although we don't support these cases
620 this issue can be fixed by a performance optimization.
621
622 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
623 not take account if existing stack usage. There is now a new global
624 variable called pcre_stack_guard that can be set to point to an external
625 function to check stack availability. It is called at the start of
626 processing every parenthesized group.
627
628 22. A typo in the code meant that in ungreedy mode the max/min qualifier
629 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
630 not match "ab".
631
632 23. When UTF was disabled, the JIT program reported some incorrect compile
633 errors. These messages are silenced now.
634
635 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
636 compiler.
637
638 25. Change all the temporary files used in RunGrepTest to be different to those
639 used by RunTest so that the tests can be run simultaneously, for example by
640 "make -j check".
641
642
643 Version 8.34 15-December-2013
644 -----------------------------
645
646 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
647 executable memory. Patch inspired by Carsten Klein.
648
649 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
650 this macro is never tested and has no effect, because the work to support
651 coverage involves only compiling and linking options and special targets in
652 the Makefile. The comment in config.h implied that defining the macro would
653 enable coverage support, which is totally false. There was also support for
654 setting this macro in the CMake files (my fault, I just copied it from
655 configure). SUPPORT_GCOV has now been removed.
656
657 3. Make a small performance improvement in strlen16() and strlen32() in
658 pcretest.
659
660 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
661 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
662 these cases). There was also one in pcretest.c.
663
664 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
665
666 6. In UTF mode, the code for checking whether a group could match an empty
667 string (which is used for indefinitely repeated groups to allow for
668 breaking an infinite loop) was broken when the group contained a repeated
669 negated single-character class with a character that occupied more than one
670 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
671 UTF-8 mode). The effect was undefined: the group might or might not be
672 deemed as matching an empty string, or the program might have crashed.
673
674 7. The code for checking whether a group could match an empty string was not
675 recognizing that \h, \H, \v, \V, and \R must match a character.
676
677 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
678 an empty string. If it can, pcretest shows this in its information output.
679
680 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
681 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
682 matched by pcre_exec() without using JIT:
683
684 (a) If the rest of the pattern did not match after a maximal run of
685 grapheme clusters, the code for backing up to try with fewer of them
686 did not always back up over a full grapheme when characters that do not
687 have the modifier quality were involved, e.g. Hangul syllables.
688
689 (b) If the match point in a subject started with modifier character, and
690 there was no match, the code could incorrectly back up beyond the match
691 point, and potentially beyond the first character in the subject,
692 leading to a segfault or an incorrect match result.
693
694 10. A conditional group with an assertion condition could lead to PCRE
695 recording an incorrect first data item for a match if no other first data
696 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
697 first data item, and therefore matched "ca" after "c" instead of at the
698 start.
699
700 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
701 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
702
703 12. The source of pcregrep now includes z/OS-specific code so that it can be
704 compiled for z/OS as part of the special z/OS distribution.
705
706 13. Added the -T and -TM options to pcretest.
707
708 14. The code in pcre_compile.c for creating the table of named capturing groups
709 has been refactored. Instead of creating the table dynamically during the
710 actual compiling pass, the information is remembered during the pre-compile
711 pass (on the stack unless there are more than 20 named groups, in which
712 case malloc() is used) and the whole table is created before the actual
713 compile happens. This has simplified the code (it is now nearly 150 lines
714 shorter) and prepared the way for better handling of references to groups
715 with duplicate names.
716
717 15. A back reference to a named subpattern when there is more than one of the
718 same name now checks them in the order in which they appear in the pattern.
719 The first one that is set is used for the reference. Previously only the
720 first one was inspected. This change makes PCRE more compatible with Perl.
721
722 16. Unicode character properties were updated from Unicode 6.3.0.
723
724 17. The compile-time code for auto-possessification has been refactored, based
725 on a patch by Zoltan Herczeg. It now happens after instead of during
726 compilation. The code is cleaner, and more cases are handled. The option
727 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
728 options in pcretest are provided to set it. It can also be set by
729 (*NO_AUTO_POSSESS) at the start of a pattern.
730
731 18. The character VT has been added to the default ("C" locale) set of
732 characters that match \s and are generally treated as white space,
733 following this same change in Perl 5.18. There is now no difference between
734 "Perl space" and "POSIX space". Whether VT is treated as white space in
735 other locales depends on the locale.
736
737 19. The code for checking named groups as conditions, either for being set or
738 for being recursed, has been refactored (this is related to 14 and 15
739 above). Processing unduplicated named groups should now be as fast at
740 numerical groups, and processing duplicated groups should be faster than
741 before.
742
743 20. Two patches to the CMake build system, by Alexander Barkov:
744
745 (1) Replace the "source" command by "." in CMakeLists.txt because
746 "source" is a bash-ism.
747
748 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
749 without these the CMake build does not work on Solaris.
750
751 21. Perl has changed its handling of \8 and \9. If there is no previously
752 encountered capturing group of those numbers, they are treated as the
753 literal characters 8 and 9 instead of a binary zero followed by the
754 literals. PCRE now does the same.
755
756 22. Following Perl, added \o{} to specify codepoints in octal, making it
757 possible to specify values greater than 0777 and also making them
758 unambiguous.
759
760 23. Perl now gives an error for missing closing braces after \x{... instead of
761 treating the string as literal. PCRE now does the same.
762
763 24. RunTest used to grumble if an inappropriate test was selected explicitly,
764 but just skip it when running all tests. This make it awkward to run ranges
765 of tests when one of them was inappropriate. Now it just skips any
766 inappropriate tests, as it always did when running all tests.
767
768 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
769 character types such as \d or \w, too many callouts were inserted, and the
770 data that they returned was rubbish.
771
772 26. In UCP mode, \s was not matching two of the characters that Perl matches,
773 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
774 were matched by \h. The code has now been refactored so that the lists of
775 the horizontal and vertical whitespace characters used for \h and \v (which
776 are defined only in one place) are now also used for \s.
777
778 27. Add JIT support for the 64 bit TileGX architecture.
779 Patch by Jiong Wang (Tilera Corporation).
780
781 28. Possessive quantifiers for classes (both explicit and automatically
782 generated) now use special opcodes instead of wrapping in ONCE brackets.
783
784 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
785 (because it's meaningless), this was not happening when PCRE_CASELESS was
786 set. Not wrong, but inefficient.
787
788 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
789 properties for \w, \d, etc) is present in a test regex. Otherwise if the
790 test contains no characters greater than 255, Perl doesn't realise it
791 should be using Unicode semantics.
792
793 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
794 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
795 does in Unicode mode.
796
797 32. Added the "forbid" facility to pcretest so that putting tests into the
798 wrong test files can sometimes be quickly detected.
799
800 33. There is now a limit (default 250) on the depth of nesting of parentheses.
801 This limit is imposed to control the amount of system stack used at compile
802 time. It can be changed at build time by --with-parens-nest-limit=xxx or
803 the equivalent in CMake.
804
805 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
806 errors. Perl warns for these when in warning mode, but PCRE has no facility
807 for giving warnings.
808
809 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
810 However, this was not working for (?!) because it is optimized to (*FAIL),
811 for which PCRE does not allow quantifiers. The optimization is now disabled
812 when a quantifier follows (?!). I can't see any use for this, but it makes
813 things uniform.
814
815 36. Perl no longer allows group names to start with digits, so I have made this
816 change also in PCRE. It simplifies the code a bit.
817
818 37. In extended mode, Perl ignores spaces before a + that indicates a
819 possessive quantifier. PCRE allowed a space before the quantifier, but not
820 before the possessive +. It now does.
821
822 38. The use of \K (reset reported match start) within a repeated possessive
823 group such as (a\Kb)*+ was not working.
824
825 40. Document that the same character tables must be used at compile time and
826 run time, and that the facility to pass tables to pcre_exec() and
827 pcre_dfa_exec() is for use only with saved/restored patterns.
828
829 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
830 features for Builds with MSVC:
831
832 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
833 builds). The .rc files can be used to set FileDescription and many other
834 attributes.
835
836 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
837 This allows higher-level build scripts which want .pdb files to avoid
838 hard-coding the exact files needed."
839
840 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
841 mean "start of word" and "end of word", respectively, as a transition aid.
842
843 43. A minimizing repeat of a class containing codepoints greater than 255 in
844 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
845 compiled to use the heap for recursion.
846
847 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
848 is configured.
849
850
851 Version 8.33 28-May-2013
852 ------------------------
853
854 1. Added 'U' to some constants that are compared to unsigned integers, to
855 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
856 variables that are added to signed variables, to ensure the result is
857 signed and can be negated.
858
859 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
860 CMake config files.
861
862 3. Revise the creation of config.h.generic so that all boolean macros are
863 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
864 overriding via -D on the command line possible.
865
866 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
867 to unsigned int is reported to make a quite noticeable speed difference in
868 a specific Windows environment. Testing on Linux did also appear to show
869 some benefit (and it is clearly not harmful). Also fixed the definition of
870 Xop which should be unsigned.
871
872 5. Related to (4), changing the definition of the intermediate variable cc
873 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
874 improvements.
875
876 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
877 unnecessary spaces.
878
879 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
880 and later.
881
882 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
883
884 9. Optimizing fast_forward_start_bits in JIT.
885
886 10. Adding support for callouts in JIT, and fixing some issues revealed
887 during this work. Namely:
888
889 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
890
891 (b) Minimum length was not checked before the matching is started.
892
893 11. The value of capture_last that is passed to callouts was incorrect in some
894 cases when there was a capture on one path that was subsequently abandoned
895 after a backtrack. Also, the capture_last value is now reset after a
896 recursion, since all captures are also reset in this case.
897
898 12. The interpreter no longer returns the "too many substrings" error in the
899 case when an overflowing capture is in a branch that is subsequently
900 abandoned after a backtrack.
901
902 13. In the pathological case when an offset vector of size 2 is used, pcretest
903 now prints out the matched string after a yield of 0 or 1.
904
905 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
906 Only supported by the JIT compiler at the moment.
907
908 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
909
910 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
911 offset of the starting point of the matching process, provided the offsets
912 vector is large enough.
913
914 17. The \A escape now records a lookbehind value of 1, though its execution
915 does not actually inspect the previous character. This is to ensure that,
916 in partial multi-segment matching, at least one character from the old
917 segment is retained when a new segment is processed. Otherwise, if there
918 are no lookbehinds in the pattern, \A might match incorrectly at the start
919 of a new segment.
920
921 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
922
923 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
924 gives some modest performance improvement in 8-bit mode.
925
926 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
927 be expressed in certain programming languages using Universal Character
928 Names.
929
930 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
931 which points out that "non characters" are not "characters that may not
932 appear in Unicode strings" but rather "characters that are reserved for
933 internal use and have only local meaning".
934
935 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
936 there was a conditional group that depended on an assertion, if the
937 assertion was false, the callout that immediately followed the alternation
938 in the condition was skipped when pcre_exec() was used for matching.
939
940 23. Allow an explicit callout to be inserted before an assertion that is the
941 condition for a conditional group, for compatibility with automatic
942 callouts, which always insert a callout at this point.
943
944 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
945 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
946
947 25. (*PRUNE) is now supported by the JIT compiler.
948
949 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
950
951 27. Fix the case where there are two or more SKIPs with arguments that may be
952 ignored.
953
954 28. (*SKIP) is now supported by the JIT compiler.
955
956 29. (*THEN) is now supported by the JIT compiler.
957
958 30. Update RunTest with additional test selector options.
959
960 31. The way PCRE handles backtracking verbs has been changed in two ways.
961
962 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
963 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
964 backtracking. In some cases this makes it more Perl-compatible, but Perl's
965 rather obscure rules do not always do the same thing.
966
967 (2) Previously, backtracking verbs were confined within assertions. This is
968 no longer the case for positive assertions, except for (*ACCEPT). Again,
969 this sometimes improves Perl compatibility, and sometimes does not.
970
971 32. A number of tests that were in test 2 because Perl did things differently
972 have been moved to test 1, because either Perl or PCRE has changed, and
973 these tests are now compatible.
974
975 32. Backtracking control verbs are now handled in the same way in JIT and
976 interpreter.
977
978 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
979 contained a forward subroutine reference caused a compile error.
980
981 34. Auto-detect and optimize limited repetitions in JIT.
982
983 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
984 blocking (*UTF) etc.
985
986 36. In the interpreter, maximizing pattern repetitions for characters and
987 character types now use tail recursion, which reduces stack usage.
988
989 37. The value of the max lookbehind was not correctly preserved if a compiled
990 and saved regex was reloaded on a host of different endianness.
991
992 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
993 of the compiled pattern block, expand the flags field from 16 to 32 bits
994 because it was almost full.
995
996 39. Try madvise first before posix_madvise.
997
998 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
999 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1000 original change locked out all matches of empty strings. This has been
1001 changed so that one match of an empty string per line is recognized.
1002 Subsequent searches on the same line (for colouring or for --only-matching,
1003 for example) do not recognize empty strings.
1004
1005 41. Applied a user patch to fix a number of spelling mistakes in comments.
1006
1007 42. Data lines longer than 65536 caused pcretest to crash.
1008
1009 43. Clarified the data type for length and startoffset arguments for pcre_exec
1010 and pcre_dfa_exec in the function-specific man pages, where they were
1011 explicitly stated to be in bytes, never having been updated. I also added
1012 some clarification to the pcreapi man page.
1013
1014 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1015 a segmentation fault.
1016
1017
1018 Version 8.32 30-November-2012
1019 -----------------------------
1020
1021 1. Improved JIT compiler optimizations for first character search and single
1022 character iterators.
1023
1024 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1025 Patch by Daniel Richard G.
1026
1027 3. Single character iterator optimizations in the JIT compiler.
1028
1029 4. Improved JIT compiler optimizations for character ranges.
1030
1031 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1032 Reported by Giuseppe D'Angelo.
1033
1034 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1035 of a line, was being set incorrectly in cases where .* appeared inside
1036 atomic brackets at the start of a pattern, or where there was a subsequent
1037 *PRUNE or *SKIP.
1038
1039 7. Improved instruction cache flush for POWER/PowerPC.
1040 Patch by Daniel Richard G.
1041
1042 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
1043 grep:
1044
1045 (a) There is now no limit to the number of patterns to be matched.
1046
1047 (b) An error is given if a pattern is too long.
1048
1049 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1050 are now supported.
1051
1052 (d) --exclude-from and --include-from (multiple use) have been added.
1053
1054 (e) Exclusions and inclusions now apply to all files and directories, not
1055 just to those obtained from scanning a directory recursively.
1056
1057 (f) Multiple uses of -f and --file-list are now supported.
1058
1059 (g) In a Windows environment, the default for -d has been changed from
1060 "read" (the GNU grep default) to "skip", because otherwise the presence
1061 of a directory in the file list provokes an error.
1062
1063 (h) The documentation has been revised and clarified in places.
1064
1065 9. Improve the matching speed of capturing brackets.
1066
1067 10. Changed the meaning of \X so that it now matches a Unicode extended
1068 grapheme cluster.
1069
1070 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1071 out POSIX threads when JIT support is configured.
1072
1073 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1074
1075 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1076 structure could go wrong in environments where size_t is not the same size
1077 as int.
1078
1079 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1080 set.
1081
1082 15. The EBCDIC support had decayed; later updates to the code had included
1083 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1084 general tidy up of EBCDIC-related issues, and the documentation was also
1085 not quite right. There is now a test that can be run on ASCII systems to
1086 check some of the EBCDIC-related things (but is it not a full test).
1087
1088 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1089 in a small tidy to the code.
1090
1091 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1092
1093 18. If the --only-matching (-o) option in pcregrep is specified multiple
1094 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1095 substrings matched by the 1st and 2nd capturing parentheses. A separating
1096 string can be specified by --om-separator (default empty).
1097
1098 19. Improving the first n character searches.
1099
1100 20. Turn case lists for horizontal and vertical white space into macros so that
1101 they are defined only once.
1102
1103 21. This set of changes together give more compatible Unicode case-folding
1104 behaviour for characters that have more than one other case when UCP
1105 support is available.
1106
1107 (a) The Unicode property table now has offsets into a new table of sets of
1108 three or more characters that are case-equivalent. The MultiStage2.py
1109 script that generates these tables (the pcre_ucd.c file) now scans
1110 CaseFolding.txt instead of UnicodeData.txt for character case
1111 information.
1112
1113 (b) The code for adding characters or ranges of characters to a character
1114 class has been abstracted into a generalized function that also handles
1115 case-independence. In UTF-mode with UCP support, this uses the new data
1116 to handle characters with more than one other case.
1117
1118 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1119 whose other case is greater than 256 are now correctly matched
1120 caselessly. Previously, the high codepoint matched the low one, but not
1121 vice versa.
1122
1123 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1124 of the new class addition function, using character lists defined as
1125 macros alongside the case definitions of 20 above.
1126
1127 (e) Caseless back references now work with characters that have more than
1128 one other case.
1129
1130 (f) General caseless matching of characters with more than one other case
1131 is supported.
1132
1133 22. Unicode character properties were updated from Unicode 6.2.0
1134
1135 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1136
1137 24. Add support for 32-bit character strings, and UTF-32
1138
1139 25. Major JIT compiler update (code refactoring and bugfixing).
1140 Experimental Sparc 32 support is added.
1141
1142 26. Applied a modified version of Daniel Richard G's patch to create
1143 pcre.h.generic and config.h.generic by "make" instead of in the
1144 PrepareRelease script.
1145
1146 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1147 pcre_compile.c when checking for a zero character.
1148
1149 28. Introducing a native interface for JIT. Through this interface, the compiled
1150 machine code can be directly executed. The purpose of this interface is to
1151 provide fast pattern matching, so several sanity checks are not performed.
1152 However, feature tests are still performed. The new interface provides
1153 1.4x speedup compared to the old one.
1154
1155 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1156 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1157 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1158
1159 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1160 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1161 the "old" RFC 2279). Instead, it ended up passing a negative length to
1162 pcre_exec().
1163
1164 31. Add support for GCC's visibility feature to hide internal functions.
1165
1166 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1167 "unknown -C option" after outputting 0 or 1.
1168
1169 33. There is now support for generating a code coverage report for the test
1170 suite in environments where gcc is the compiler and lcov is installed. This
1171 is mainly for the benefit of the developers.
1172
1173 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1174 unaddressable using valgrind annotations, allowing valgrind to detect
1175 invalid memory accesses. This is mainly for the benefit of the developers.
1176
1177 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1178
1179 26. Give configure error if --enable-cpp but no C++ compiler found.
1180
1181
1182 Version 8.31 06-July-2012
1183 -------------------------
1184
1185 1. Fixing a wrong JIT test case and some compiler warnings.
1186
1187 2. Removed a bashism from the RunTest script.
1188
1189 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1190 to unsigned type, result still unsigned" that was given by an MS compiler
1191 on encountering the code "-sizeof(xxx)".
1192
1193 4. Partial matching support is added to the JIT compiler.
1194
1195 5. Fixed several bugs concerned with partial matching of items that consist
1196 of more than one character:
1197
1198 (a) /^(..)\1/ did not partially match "aba" because checking references was
1199 done on an "all or nothing" basis. This also applied to repeated
1200 references.
1201
1202 (b) \R did not give a hard partial match if \r was found at the end of the
1203 subject.
1204
1205 (c) \X did not give a hard partial match after matching one or more
1206 characters at the end of the subject.
1207
1208 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1209 a partial match for the string "\r".
1210
1211 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1212 a partial match for a CR character at the end of the subject string.
1213
1214 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1215 running pcretest, the text "(JIT)" added to the output whenever JIT is
1216 actually used to run the match.
1217
1218 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1219 or /S+[+] with a digit between 1 and 7.
1220
1221 8. OP_NOT now supports any UTF character not just single-byte ones.
1222
1223 9. (*MARK) control verb is now supported by the JIT compiler.
1224
1225 10. The command "./RunTest list" lists the available tests without actually
1226 running any of them. (Because I keep forgetting what they all are.)
1227
1228 11. Add PCRE_INFO_MAXLOOKBEHIND.
1229
1230 12. Applied a (slightly modified) user-supplied patch that improves performance
1231 when the heap is used for recursion (compiled with --disable-stack-for-
1232 recursion). Instead of malloc and free for each heap frame each time a
1233 logical recursion happens, frames are retained on a chain and re-used where
1234 possible. This sometimes gives as much as 30% improvement.
1235
1236 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1237 call.
1238
1239 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1240
1241 15. It is now possible to link pcretest with libedit as an alternative to
1242 libreadline.
1243
1244 16. (*COMMIT) control verb is now supported by the JIT compiler.
1245
1246 17. The Unicode data tables have been updated to Unicode 6.1.0.
1247
1248 18. Added --file-list option to pcregrep.
1249
1250 19. Added binary file support to pcregrep, including the -a, --binary-files,
1251 -I, and --text options.
1252
1253 20. The madvise function is renamed for posix_madvise for QNX compatibility
1254 reasons. Fixed by Giuseppe D'Angelo.
1255
1256 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1257
1258 22. Changed the option for creating symbolic links for 16-bit man pages from
1259 -s to -sf so that re-installing does not cause issues.
1260
1261 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1262
1263 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1264 matches in certain environments (the workspace was not being correctly
1265 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1266 some of the workspace data at the beginning of a restart.
1267
1268 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1269 was not doing so when it should - probably a typo introduced by SVN 528
1270 (change 8.10/14).
1271
1272 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1273 \w+ when the character tables indicated that \x{c4} was a word character.
1274 There were several related cases, all because the tests for doing a table
1275 lookup were testing for characters less than 127 instead of 255.
1276
1277 27. If a pattern contains capturing parentheses that are not used in a match,
1278 their slots in the ovector are set to -1. For those that are higher than
1279 any matched groups, this happens at the end of processing. In the case when
1280 there were back references that the ovector was too small to contain
1281 (causing temporary malloc'd memory to be used during matching), and the
1282 highest capturing number was not used, memory off the end of the ovector
1283 was incorrectly being set to -1. (It was using the size of the temporary
1284 memory instead of the true size.)
1285
1286 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1287 ovector size, it uses memory at the end of the block that it has got.
1288
1289 29. Check for an overlong MARK name and give an error at compile time. The
1290 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1291
1292 30. JIT compiler update.
1293
1294 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1295 Rill for the patch.
1296
1297 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1298
1299 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1300
1301 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1302 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1303 was enabled.
1304
1305 35. Improve JIT code generation for greedy plus quantifier.
1306
1307 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1308 1 to "aa" instead of to an empty string. The bug affected repeated groups
1309 that could potentially match an empty string.
1310
1311 37. Optimizing single character iterators in JIT.
1312
1313 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1314 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1315 codepoints that are too big for the mode are faulted, and in a UTF mode,
1316 disallowed codepoints are also faulted.
1317
1318 39. If PCRE was compiled with UTF support, in three places in the DFA
1319 matcher there was code that should only have been obeyed in UTF mode, but
1320 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1321 processing when bytes with values greater than 127 were present. In 16-bit
1322 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1323 both cases the values are those that cannot be the first data item in a UTF
1324 character. The three items that might have provoked this were recursions,
1325 possessively repeated groups, and atomic groups.
1326
1327 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1328 and pcregrep, because some OS require shared objects to be explicitly
1329 passed to ld, causing the link step to fail if they are not.
1330
1331 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1332 mode, patterns that started with \h* or \R* might be incorrectly matched.
1333
1334
1335 Version 8.30 04-February-2012
1336 -----------------------------
1337
1338 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1339 name is defined in ctype.h.
1340
1341 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1342 only in quite long subpatterns.
1343
1344 3. Removed the function pcre_info(), which has been obsolete and deprecated
1345 since it was replaced by pcre_fullinfo() in February 2000.
1346
1347 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1348 match a (*MARK), and the match failed at the start of the subject, a
1349 reference to memory before the start of the subject could occur. This bug
1350 was introduced by fix 17 of release 8.21.
1351
1352 5. A reference to an unset group with zero minimum repetition was giving
1353 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1354 /(another)?(\1?)test/ matched against "hello world test". This bug was
1355 introduced in release 8.13.
1356
1357 6. Add support for 16-bit character strings (a large amount of work involving
1358 many changes and refactorings).
1359
1360 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1361 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1362 from a file.
1363
1364 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1365 rounding is not applied in this particular case).
1366
1367 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1368 if they appear, or are escaped, in patterns.
1369
1370 10. Get rid of a number of -Wunused-but-set-variable warnings.
1371
1372 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1373 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1374 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1375 also returns the mark "x". This bug applied to capturing parentheses,
1376 non-capturing parentheses, and atomic parentheses. It also applied to some
1377 assertions.
1378
1379 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1380 information out of configure.ac instead of relying on pcre.h.generic, which
1381 is not stored in the repository.
1382
1383 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1384 -lreadline.
1385
1386 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1387
1388 15. Applied Graycode's patch to put the top-level frame on the stack rather
1389 than the heap when not using the stack for recursion. This gives a
1390 performance improvement in many cases when recursion is not deep.
1391
1392 16. Experimental code added to "pcretest -C" to output the stack frame size.
1393
1394
1395 Version 8.21 12-Dec-2011
1396 ------------------------
1397
1398 1. Updating the JIT compiler.
1399
1400 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1401 are added as well.
1402
1403 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1404 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1405 calling _pcre_jit_exec. Some extra comments are added.
1406
1407 4. (*MARK) settings inside atomic groups that do not contain any capturing
1408 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1409 was introduced by change 18 for 8.20.
1410
1411 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1412 ECMA-262 standard.
1413
1414 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1415 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1416 This bug was probably introduced by change 9 of 8.13.
1417
1418 7. While fixing 6 above, I noticed that a number of other items were being
1419 incorrectly rejected as "not fixed length". This arose partly because newer
1420 opcodes had not been added to the fixed-length checking code. I have (a)
1421 corrected the bug and added tests for these items, and (b) arranged for an
1422 error to occur if an unknown opcode is encountered while checking for fixed
1423 length instead of just assuming "not fixed length". The items that were
1424 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1425 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1426 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1427
1428 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1429 being incorrectly compiled and would have given unpredicatble results.
1430
1431 9. A possessively repeated subpattern with minimum repeat count greater than
1432 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1433 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1434 the first (A) could occur when it should not.
1435
1436 10. Add a cast and remove a redundant test from the code.
1437
1438 11. JIT should use pcre_malloc/pcre_free for allocation.
1439
1440 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1441 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1442 is anything other than /usr, -L is still shown).
1443
1444 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1445
1446 14. Perl does not support \N without a following name in a [] class; PCRE now
1447 also gives an error.
1448
1449 15. If a forward reference was repeated with an upper limit of around 2000,
1450 it caused the error "internal error: overran compiling workspace". The
1451 maximum number of forward references (including repeats) was limited by the
1452 internal workspace, and dependent on the LINK_SIZE. The code has been
1453 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1454 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1455 of around 200,000 forward references. While doing this, I also speeded up
1456 the filling in of repeated forward references.
1457
1458 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1459 incorrectly expecting the subject to contain another "a" after the start.
1460
1461 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1462 in the match, the SKIP should be ignored. This was not happening; instead
1463 the SKIP was being treated as NOMATCH. For patterns such as
1464 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1465 tested.
1466
1467 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1468 now much more compatible with Perl, in particular in cases where the result
1469 is a non-match for a non-anchored pattern. For example, if
1470 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1471 "m", where previously it did not return a name. A side effect of this
1472 change is that for partial matches, the last encountered mark name is
1473 returned, as for non matches. A number of tests that were previously not
1474 Perl-compatible have been moved into the Perl-compatible test files. The
1475 refactoring has had the pleasing side effect of removing one argument from
1476 the match() function, thus reducing its stack requirements.
1477
1478 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1479 subsequent uses of /S (without +) incorrectly behaved like /S+.
1480
1481 21. Retrieve executable code size support for the JIT compiler and fixing
1482 some warnings.
1483
1484 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1485 not work when the shorter character appeared right at the end of the
1486 subject string.
1487
1488 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1489 systems.
1490
1491 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1492 output it when the /M option is used in pcretest.
1493
1494 25. The CheckMan script was not being included in the distribution. Also, added
1495 an explicit "perl" to run Perl scripts from the PrepareRelease script
1496 because this is reportedly needed in Windows.
1497
1498 26. If study data was being save in a file and studying had not found a set of
1499 "starts with" bytes for the pattern, the data written to the file (though
1500 never used) was taken from uninitialized memory and so caused valgrind to
1501 complain.
1502
1503 27. Updated RunTest.bat as provided by Sheri Pierce.
1504
1505 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1506
1507 29. Computation of memory usage for the table of capturing group names was
1508 giving an unnecessarily large value.
1509
1510
1511 Version 8.20 21-Oct-2011
1512 ------------------------
1513
1514 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1515 a POSIX class. After further experiments with Perl, which convinced me that
1516 Perl has bugs and confusions, a closing square bracket is no longer allowed
1517 in a POSIX name. This bug also affected patterns with classes that started
1518 with full stops.
1519
1520 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1521 captured substring, but while checking the failing first alternative,
1522 substring 1 is temporarily captured. If the output vector supplied to
1523 pcre_exec() was not big enough for this capture, the yield of the function
1524 was still zero ("insufficient space for captured substrings"). This cannot
1525 be totally fixed without adding another stack variable, which seems a lot
1526 of expense for a edge case. However, I have improved the situation in cases
1527 such as /(a)(b)x|abc/ matched against "abc", where the return code
1528 indicates that fewer than the maximum number of slots in the ovector have
1529 been set.
1530
1531 3. Related to (2) above: when there are more back references in a pattern than
1532 slots in the output vector, pcre_exec() uses temporary memory during
1533 matching, and copies in the captures as far as possible afterwards. It was
1534 using the entire output vector, but this conflicts with the specification
1535 that only 2/3 is used for passing back captured substrings. Now it uses
1536 only the first 2/3, for compatibility. This is, of course, another edge
1537 case.
1538
1539 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1540 main code base, and can be used by building with --enable-jit. When this is
1541 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1542 runtime --no-jit option is given.
1543
1544 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1545 ovector, the return from the function was zero, implying that there were
1546 other matches that did not fit. The correct "exactly full" value is now
1547 returned.
1548
1549 6. If a subpattern that was called recursively or as a subroutine contained
1550 (*PRUNE) or any other control that caused it to give a non-standard return,
1551 invalid errors such as "Error -26 (nested recursion at the same subject
1552 position)" or even infinite loops could occur.
1553
1554 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1555 computing the minimum length on reaching *ACCEPT, and so ended up with the
1556 wrong value of 1 rather than 0. Further investigation indicates that
1557 computing a minimum subject length in the presence of *ACCEPT is difficult
1558 (think back references, subroutine calls), and so I have changed the code
1559 so that no minimum is registered for a pattern that contains *ACCEPT.
1560
1561 8. If (*THEN) was present in the first (true) branch of a conditional group,
1562 it was not handled as intended. [But see 16 below.]
1563
1564 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1565 Sheri Pierce.
1566
1567 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1568 the first byte in a match must be "a".
1569
1570 11. Change 17 for 8.13 increased the recursion depth for patterns like
1571 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1572 pattern contains any instances of (*THEN). If it does not, the old
1573 optimizations are restored. It would be nice to do this on a per-group
1574 basis, but at the moment that is not feasible.
1575
1576 12. In some environments, the output of pcretest -C is CRLF terminated. This
1577 broke RunTest's code that checks for the link size. A single white space
1578 character after the value is now allowed for.
1579
1580 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1581 For "fr", it uses the Windows-specific input and output files.
1582
1583 14. If (*THEN) appeared in a group that was called recursively or as a
1584 subroutine, it did not work as intended. [But see next item.]
1585
1586 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1587 pattern fragments (but not containing any | characters). If A and B are
1588 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1589 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1590 D. In other words, Perl considers parentheses that do not contain any |
1591 characters to be part of a surrounding alternative, whereas PCRE was
1592 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1593 differently. PCRE now behaves in the same way as Perl, except in the case
1594 of subroutine/recursion calls such as (?1) which have in any case always
1595 been different (but PCRE had them first :-).
1596
1597 16. Related to 15 above: Perl does not treat the | in a conditional group as
1598 creating alternatives. Such a group is treated in the same way as an
1599 ordinary group without any | characters when processing (*THEN). PCRE has
1600 been changed to match Perl's behaviour.
1601
1602 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1603 RunGrepTest script failed.
1604
1605 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1606 inevitable for groups that contain captures, but it can lead to a lot of
1607 stack use in large patterns. The old behaviour has been restored for atomic
1608 groups that do not contain any capturing parentheses.
1609
1610 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1611 suppress the check for a minimum subject length at run time. (If it was
1612 given to pcre_exec() or pcre_dfa_exec() it did work.)
1613
1614 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1615 fail to work when decoding hex characters in data strings in EBCDIC
1616 environments.
1617
1618 21. It appears that in at least one Mac OS environment, the isxdigit() function
1619 is implemented as a macro that evaluates to its argument more than once,
1620 contravening the C 90 Standard (I haven't checked a later standard). There
1621 was an instance in pcretest which caused it to go wrong when processing
1622 \x{...} escapes in subject strings. The has been rewritten to avoid using
1623 things like p++ in the argument of isxdigit().
1624
1625
1626 Version 8.13 16-Aug-2011
1627 ------------------------
1628
1629 1. The Unicode data tables have been updated to Unicode 6.0.0.
1630
1631 2. Two minor typos in pcre_internal.h have been fixed.
1632
1633 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1634 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1635 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1636
1637 4. There were a number of related bugs in the code for matching backrefences
1638 caselessly in UTF-8 mode when codes for the characters concerned were
1639 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1640 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1641 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1642 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1643 2-byte code at the end of the subject (it thought there wasn't enough data
1644 left).
1645
1646 5. Comprehensive information about what went wrong is now returned by
1647 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1648 as the output vector has at least 2 elements. The offset of the start of
1649 the failing character and a reason code are placed in the vector.
1650
1651 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1652 now returned is for the first byte of the failing character, instead of the
1653 last byte inspected. This is an incompatible change, but I hope it is small
1654 enough not to be a problem. It makes the returned offset consistent with
1655 pcre_exec() and pcre_dfa_exec().
1656
1657 7. pcretest now gives a text phrase as well as the error number when
1658 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1659 failure, the offset and reason code are output.
1660
1661 8. When \R was used with a maximizing quantifier it failed to skip backwards
1662 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1663 back over a single character (\n). This seems wrong (because it treated the
1664 two characters as a single entity when going forwards), conflicts with the
1665 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1666 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1667 has been changed.
1668
1669 9. Some internal refactoring has changed the processing so that the handling
1670 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1671 time (the PCRE_DOTALL option was changed this way some time ago: version
1672 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1673 which was always a bit of a fudge. It also means that there is one less
1674 argument for the match() function, which reduces its stack requirements
1675 slightly. This change also fixes an incompatibility with Perl: the pattern
1676 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1677
1678 10. More internal refactoring has drastically reduced the number of recursive
1679 calls to match() for possessively repeated groups such as (abc)++ when
1680 using pcre_exec().
1681
1682 11. While implementing 10, a number of bugs in the handling of groups were
1683 discovered and fixed:
1684
1685 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1686 (a|)*(?1) gave a compile-time internal error.
1687 ((a|)+)+ did not notice that the outer group could match an empty string.
1688 (^a|^)+ was not marked as anchored.
1689 (.*a|.*)+ was not marked as matching at start or after a newline.
1690
1691 12. Yet more internal refactoring has removed another argument from the match()
1692 function. Special calls to this function are now indicated by setting a
1693 value in a variable in the "match data" data block.
1694
1695 13. Be more explicit in pcre_study() instead of relying on "default" for
1696 opcodes that mean there is no starting character; this means that when new
1697 ones are added and accidentally left out of pcre_study(), testing should
1698 pick them up.
1699
1700 14. The -s option of pcretest has been documented for ages as being an old
1701 synonym of -m (show memory usage). I have changed it to mean "force study
1702 for every regex", that is, assume /S for every regex. This is similar to -i
1703 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1704 using it. It makes it easier to run collections of tests with and without
1705 study enabled, and thereby test pcre_study() more easily. All the standard
1706 tests are now run with and without -s (but some patterns can be marked as
1707 "never study" - see 20 below).
1708
1709 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1710 restoration of the capturing data to the outer values was not happening
1711 correctly.
1712
1713 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1714 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1715 pattern had matched an empty string, and so incorrectly returned a no
1716 match.
1717
1718 17. There was optimizing code for the last branch of non-capturing parentheses,
1719 and also for the obeyed branch of a conditional subexpression, which used
1720 tail recursion to cut down on stack usage. Unfortunately, now that there is
1721 the possibility of (*THEN) occurring in these branches, tail recursion is
1722 no longer possible because the return has to be checked for (*THEN). These
1723 two optimizations have therefore been removed. [But see 8.20/11 above.]
1724
1725 18. If a pattern containing \R was studied, it was assumed that \R always
1726 matched two bytes, thus causing the minimum subject length to be
1727 incorrectly computed because \R can also match just one byte.
1728
1729 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1730 was incorrectly computed.
1731
1732 20. If /S is present twice on a test pattern in pcretest input, it now
1733 *disables* studying, thereby overriding the use of -s on the command line
1734 (see 14 above). This is necessary for one or two tests to keep the output
1735 identical in both cases.
1736
1737 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1738 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1739
1740 22. When an atomic group that contained a capturing parenthesis was
1741 successfully matched, but the branch in which it appeared failed, the
1742 capturing was not being forgotten if a higher numbered group was later
1743 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1744 group 1 to "a", when in fact it should be unset. This applied to multi-
1745 branched capturing and non-capturing groups, repeated or not, and also to
1746 positive assertions (capturing in negative assertions does not happen
1747 in PCRE) and also to nested atomic groups.
1748
1749 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1750 subject after a captured substring, to make it easier to tell which of a
1751 number of identical substrings has been captured.
1752
1753 24. The way atomic groups are processed by pcre_exec() has been changed so that
1754 if they are repeated, backtracking one repetition now resets captured
1755 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1756 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1757 "aaa". Previously, it would have been "a". As part of this code
1758 refactoring, the way recursive calls are handled has also been changed.
1759
1760 25. If an assertion condition captured any substrings, they were not passed
1761 back unless some other capturing happened later. For example, if
1762 (?(?=(a))a) was matched against "a", no capturing was returned.
1763
1764 26. When studying a pattern that contained subroutine calls or assertions,
1765 the code for finding the minimum length of a possible match was handling
1766 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1767 group 1 called group 2 while simultaneously a separate group 2 called group
1768 1). A stack overflow occurred in this case. I have fixed this by limiting
1769 the recursion depth to 10.
1770
1771 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1772 Fortmann. This supports explicit test numbers on the command line, and has
1773 argument validation and error reporting.
1774
1775 28. An instance of \X with an unlimited repeat could fail if at any point the
1776 first character it looked at was a mark character.
1777
1778 29. Some minor code refactoring concerning Unicode properties and scripts
1779 should reduce the stack requirement of match() slightly.
1780
1781 30. Added the '=' option to pcretest to check the setting of unused capturing
1782 slots at the end of the pattern, which are documented as being -1, but are
1783 not included in the return count.
1784
1785 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1786 compiled something random. Now it gives a compile-time error (as does
1787 Perl).
1788
1789 32. A *MARK encountered during the processing of a positive assertion is now
1790 recorded and passed back (compatible with Perl).
1791
1792 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1793 had alternative anchored branches, the search for a second match in a line
1794 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1795 matched the line "0102" twice. The same bug affected patterns that started
1796 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1797 twice.
1798
1799 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1800 does, and because of capturing effects, quantifying parenthesized
1801 assertions may at times be useful. Quantifiers are now allowed for
1802 parenthesized assertions.
1803
1804 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1805
1806 36. \g was being checked for fancy things in a character class, when it should
1807 just be a literal "g".
1808
1809 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1810 appearance of a nested POSIX class supersedes an apparent external class.
1811 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1812 unescaped square brackets may also appear as part of class names. For
1813 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1814 more like Perl. (But see 8.20/1 above.)
1815
1816 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1817 was because it thought it was \N{name}, which is not supported).
1818
1819 39. Add minix to OS list not supporting the -S option in pcretest.
1820
1821 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1822 cannot analyze patterns in sufficient detail to catch mutual recursions
1823 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1824 subgroup is called recursively as a subpattern for a second time at the
1825 same position in the subject string. In previous releases this might have
1826 been caught by the recursion limit, or it might have run out of stack.
1827
1828 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1829 happen only once. PCRE was, however incorrectly giving a compile time error
1830 "recursive call could loop indefinitely" because it cannot analyze the
1831 pattern in sufficient detail. The compile time test no longer happens when
1832 PCRE is compiling a conditional subpattern, but actual runaway loops are
1833 now caught at runtime (see 40 above).
1834
1835 42. It seems that Perl allows any characters other than a closing parenthesis
1836 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1837 has been changed to be the same.
1838
1839 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1840 as not to get warnings when autogen.sh is called. Also changed
1841 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1842
1843 44. To help people who use pcregrep to scan files containing exceedingly long
1844 lines, the following changes have been made:
1845
1846 (a) The default value of the buffer size parameter has been increased from
1847 8K to 20K. (The actual buffer used is three times this size.)
1848
1849 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1850 PCRE is built.
1851
1852 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1853 to be set at run time.
1854
1855 (d) Numerical values in pcregrep options can be followed by K or M, for
1856 example --buffer-size=50K.
1857
1858 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1859 given and the return code is set to 2.
1860
1861 45. Add a pointer to the latest mark to the callout data block.
1862
1863 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1864 partial match of an empty string instead of no match. This was specific to
1865 the use of ".".
1866
1867 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1868 complete match instead of a partial match. This bug was dependent on both
1869 the PCRE_UTF8 and PCRE_DOTALL options being set.
1870
1871 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1872 starting byte set, because \b was not being ignored.
1873
1874
1875 Version 8.12 15-Jan-2011
1876 ------------------------
1877
1878 1. Fixed some typos in the markup of the man pages, and wrote a script that
1879 checks for such things as part of the documentation building process.
1880
1881 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1882 --match-limit and --recursion-limit options (added for 8.11). In
1883 particular, this made one of the standard tests fail. (The integer value
1884 went into the wrong half of a long int.)
1885
1886 3. If the --colour option was given to pcregrep with -v (invert match), it
1887 did strange things, either producing crazy output, or crashing. It should,
1888 of course, ignore a request for colour when reporting lines that do not
1889 match.
1890
1891 4. Another pcregrep bug caused similar problems if --colour was specified with
1892 -M (multiline) and the pattern match finished with a line ending.
1893
1894 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1895 matched in multiline mode, the following line was shown as part of the
1896 match. This seems wrong, so I have changed it.
1897
1898 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1899 the check for further matches in the same line (so they could be coloured)
1900 to overrun the end of the current line. If another match was found, it was
1901 incorrectly shown (and then shown again when found in the next line).
1902
1903 7. If pcregrep was compiled under Windows, there was a reference to the
1904 function pcregrep_exit() before it was defined. I am assuming this was
1905 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1906 reported by a user. I've moved the definition above the reference.
1907
1908
1909 Version 8.11 10-Dec-2010
1910 ------------------------
1911
1912 1. (*THEN) was not working properly if there were untried alternatives prior
1913 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1914 backtracked to try for "b" instead of moving to the next alternative branch
1915 at the same level (in this case, to look for "c"). The Perl documentation
1916 is clear that when (*THEN) is backtracked onto, it goes to the "next
1917 alternative in the innermost enclosing group".
1918
1919 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1920 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1921 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1922 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1923 (*THEN).
1924
1925 3. If \s appeared in a character class, it removed the VT character from
1926 the class, even if it had been included by some previous item, for example
1927 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1928 of \s, but is part of the POSIX "space" class.)
1929
1930 4. A partial match never returns an empty string (because you can always
1931 match an empty string at the end of the subject); however the checking for
1932 an empty string was starting at the "start of match" point. This has been
1933 changed to the "earliest inspected character" point, because the returned
1934 data for a partial match starts at this character. This means that, for
1935 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1936 (previously it gave "no match").
1937
1938 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1939 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1940 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1941 has an implication that the given string is incomplete (because a partial
1942 match is preferred over a full match). For this reason, these items now
1943 give a partial match in this situation. [Aside: previously, the one case
1944 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1945 match rather than a full match, which was wrong by the old rules, but is
1946 now correct.]
1947
1948 6. There was a bug in the handling of #-introduced comments, recognized when
1949 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1950 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1951 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1952 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1953 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1954 places in pcre_compile().
1955
1956 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1957 comments when looking ahead for named forward references to subpatterns,
1958 the only newline sequence it recognized was NL. It now handles newlines
1959 according to the set newline convention.
1960
1961 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1962 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1963 cater for a lack of strerror(). These oversights have been fixed.
1964
1965 9. Added --match-limit and --recursion-limit to pcregrep.
1966
1967 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1968
1969 11. When the -o option was used, pcregrep was setting a return code of 1, even
1970 when matches were found, and --line-buffered was not being honoured.
1971
1972 12. Added an optional parentheses number to the -o and --only-matching options
1973 of pcregrep.
1974
1975 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1976 can match an empty string. The code to do it in pcretest and pcredemo
1977 needed fixing:
1978
1979 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1980 only one byte after an empty string match just before CRLF (this case
1981 just got forgotten; "any" and "anycrlf" were OK).
1982
1983 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1984 mode when an empty string match preceded an ASCII character followed by
1985 a non-ASCII character. (The code for advancing by one character rather
1986 than one byte was nonsense.)
1987
1988 (c) The pcredemo.c sample program did not have any code at all to handle
1989 the cases when CRLF is a valid newline sequence.
1990
1991 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1992 as a starting offset was within the subject string. There is now a new
1993 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1994 negative or greater than the length of the string. In order to test this,
1995 pcretest is extended to allow the setting of negative starting offsets.
1996
1997 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1998 starting offset points to the beginning of a UTF-8 character was
1999 unnecessarily clumsy. I tidied it up.
2000
2001 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2002 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2003
2004 17. Nobody had reported that the --include_dir option, which was added in
2005 release 7.7 should have been called --include-dir (hyphen, not underscore)
2006 for compatibility with GNU grep. I have changed it to --include-dir, but
2007 left --include_dir as an undocumented synonym, and the same for
2008 --exclude-dir, though that is not available in GNU grep, at least as of
2009 release 2.5.4.
2010
2011 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2012 characters from a string of bytes) have been redefined so as not to use
2013 loops, in order to improve performance in some environments. At the same
2014 time, I abstracted some of the common code into auxiliary macros to save
2015 repetition (this should not affect the compiled code).
2016
2017 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2018 compile-time error is now given if \c is not followed by an ASCII
2019 character, that is, a byte less than 128. (In EBCDIC mode, the code is
2020 different, and any byte value is allowed.)
2021
2022 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2023 START_OPTIMIZE option, which is now allowed at compile time - but just
2024 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2025 to pcregrep and other applications that have no direct access to PCRE
2026 options. The new /Y option in pcretest sets this option when calling
2027 pcre_compile().
2028
2029 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2030 back references. Groups containing recursive back references were forced to
2031 be atomic by that change, but in the case of named groups, the amount of
2032 memory required was incorrectly computed, leading to "Failed: internal
2033 error: code overflow". This has been fixed.
2034
2035 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2036 pcretest.c, to avoid build problems in some Borland environments.
2037
2038
2039 Version 8.10 25-Jun-2010
2040 ------------------------
2041
2042 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2043 THEN.
2044
2045 2. (*ACCEPT) was not working when inside an atomic group.
2046
2047 3. Inside a character class, \B is treated as a literal by default, but
2048 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2049 causes the error). The code is unchanged, but I tidied the documentation.
2050
2051 4. Inside a character class, PCRE always treated \R and \X as literals,
2052 whereas Perl faults them if its -w option is set. I have changed PCRE so
2053 that it faults them when PCRE_EXTRA is set.
2054
2055 5. Added support for \N, which always matches any character other than
2056 newline. (It is the same as "." when PCRE_DOTALL is not set.)
2057
2058 6. When compiling pcregrep with newer versions of gcc which may have
2059 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2060 declared with attribute warn_unused_result" were given. Just casting the
2061 result to (void) does not stop the warnings; a more elaborate fudge is
2062 needed. I've used a macro to implement this.
2063
2064 7. Minor change to pcretest.c to avoid a compiler warning.
2065
2066 8. Added four artifical Unicode properties to help with an option to make
2067 \s etc use properties (see next item). The new properties are: Xan
2068 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2069
2070 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2071 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2072 this option. Modified pcretest to add /W to test this facility. Added
2073 REG_UCP to make it available via the POSIX interface.
2074
2075 10. Added --line-buffered to pcregrep.
2076
2077 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2078 studied, and the match started with a letter with a code point greater than
2079 127 whose first byte was different to the first byte of the other case of
2080 the letter, the other case of this starting letter was not recognized
2081 (#976).
2082
2083 12. If a pattern that was studied started with a repeated Unicode property
2084 test, for example, \p{Nd}+, there was the theoretical possibility of
2085 setting up an incorrect bitmap of starting bytes, but fortunately it could
2086 not have actually happened in practice until change 8 above was made (it
2087 added property types that matched character-matching opcodes).
2088
2089 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2090 possible starting bytes for non-anchored patterns.
2091
2092 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2093 \R, and also a number of cases that involve Unicode properties, both
2094 explicit and implicit when PCRE_UCP is set.
2095
2096 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2097 input, it could crash or give wrong results if characters with values
2098 greater than 0xc0 were present in the subject string. (Detail: it assumed
2099 UTF-8 input when processing these items.)
2100
2101 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2102 size_t is 64-bit (#991).
2103
2104 17. Added a check for running out of memory when PCRE is compiled with
2105 --disable-stack-for-recursion (#990).
2106
2107 18. If the last data line in a file for pcretest does not have a newline on
2108 the end, a newline was missing in the output.
2109
2110 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2111 less than 128) in its various bitmaps. However, there is a facility for
2112 generating tables according to the current locale when PCRE is compiled. It
2113 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2114 characters, are recognized by isspace() and therefore were getting set in
2115 these tables, and indeed these tables seem to approximate to ISO 8859. This
2116 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2117 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2118 which of course cannot start UTF-8 characters. I have changed the code so
2119 that only real ASCII characters (less than 128) and the correct starting
2120 bytes for UTF-8 encodings are set for characters greater than 127 when in
2121 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2122 altogether.)
2123
2124 20. Added the /T option to pcretest so as to be able to run tests with non-
2125 standard character tables, thus making it possible to include the tests
2126 used for 19 above in the standard set of tests.
2127
2128 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2129 reference to a subpattern the other side of a comment that contains an
2130 opening parenthesis caused either an internal compiling error, or a
2131 reference to the wrong subpattern.
2132
2133
2134 Version 8.02 19-Mar-2010
2135 ------------------------
2136
2137 1. The Unicode data tables have been updated to Unicode 5.2.0.
2138
2139 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2140 configured.
2141
2142 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2143 original author of that file, following a query about its status.
2144
2145 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2146 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2147
2148 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2149 quantifier applied to a forward-referencing subroutine call, could compile
2150 incorrect code or give the error "internal error: previously-checked
2151 referenced subpattern not found".
2152
2153 6. Both MS Visual Studio and Symbian OS have problems with initializing
2154 variables to point to external functions. For these systems, therefore,
2155 pcre_malloc etc. are now initialized to local functions that call the
2156 relevant global functions.
2157
2158 7. There were two entries missing in the vectors called coptable and poptable
2159 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2160 I've fixed the data, and added a kludgy way of testing at compile time that
2161 the lengths are correct (equal to the number of opcodes).
2162
2163 8. Following on from 7, I added a similar kludge to check the length of the
2164 eint vector in pcreposix.c.
2165
2166 9. Error texts for pcre_compile() are held as one long string to avoid too
2167 much relocation at load time. To find a text, the string is searched,
2168 counting zeros. There was no check for running off the end of the string,
2169 which could happen if a new error number was added without updating the
2170 string.
2171
2172 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2173
2174 11. \K was not working if it appeared in an atomic group or in a group that
2175 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2176 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2177 the assertion is positive, but not if it is negative.
2178
2179 12. Change 11 fortuitously reduced the size of the stack frame used in the
2180 "match()" function of pcre_exec.c by one pointer. Forthcoming
2181 implementation of support for (*MARK) will need an extra pointer on the
2182 stack; I have reserved it now, so that the stack frame size does not
2183 decrease.
2184
2185 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2186 item in branch that calls a recursion is a subroutine call - as in the
2187 second branch in the above example - was incorrectly given the compile-
2188 time error "recursive call could loop indefinitely" because pcre_compile()
2189 was not correctly checking the subroutine for matching a non-empty string.
2190
2191 14. The checks for overrunning compiling workspace could trigger after an
2192 overrun had occurred. This is a "should never occur" error, but it can be
2193 triggered by pathological patterns such as hundreds of nested parentheses.
2194 The checks now trigger 100 bytes before the end of the workspace.
2195
2196 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2197
2198
2199 Version 8.01 19-Jan-2010
2200 ------------------------
2201
2202 1. If a pattern contained a conditional subpattern with only one branch (in
2203 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2204 computed the wrong minimum data length (which is of course zero for such
2205 subpatterns). This could cause incorrect "no match" results.
2206
2207 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2208 the pattern is reset in the first branch, pcre_compile() failed with
2209 "internal error: code overflow at offset...". This happened only when
2210 the reset was to the original external option setting. (An optimization
2211 abstracts leading options settings into an external setting, which was the
2212 cause of this.)
2213
2214 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2215 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2216 assertion pattern did not match (meaning that the assertion was true), it
2217 was incorrectly treated as false if the SKIP had been reached during the
2218 matching. This also applied to assertions used as conditions.
2219
2220 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2221 assertion subpattern, including such a pattern used as a condition,
2222 unpredictable results occurred, instead of the error return
2223 PCRE_ERROR_DFA_UITEM.
2224
2225 5. The C++ GlobalReplace function was not working like Perl for the special
2226 situation when an empty string is matched. It now does the fancy magic
2227 stuff that is necessary.
2228
2229 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2230 removed. (These were left over from very, very early versions of PCRE.)
2231
2232 7. Some cosmetic changes to the code to make life easier when compiling it
2233 as part of something else:
2234
2235 (a) Change DEBUG to PCRE_DEBUG.
2236
2237 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2238 called "current" as "current_branch", to prevent a collision with the
2239 Linux macro when compiled as a kernel module.
2240
2241 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2242 prevent a collision with the Linux macro when compiled as a kernel
2243 module.
2244
2245 8. In pcre_compile() there are some checks for integer overflows that used to
2246 cast potentially large values to (double). This has been changed to that
2247 when building, a check for int64_t is made, and if it is found, it is used
2248 instead, thus avoiding the use of floating point arithmetic. (There is no
2249 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2250 double.
2251
2252 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2253 2005 (difference between two addresses compared to an unsigned value).
2254
2255 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2256 custom one, because of the following reported problem in Windows:
2257
2258 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2259 under Win32.
2260 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2261 therefore missing the function definition.
2262 - The compiler thus generates a "C" signature for the test function.
2263 - The linker fails to find the "C" function.
2264 - PCRE fails to configure if asked to do so against libbz2.
2265
2266 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2267 messages were output:
2268
2269 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2270 rerunning libtoolize, to keep the correct libtool macros in-tree.
2271 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2272
2273 I have done both of these things.
2274
2275 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2276 most of the time, it *can* run out if it is given a pattern that contains a
2277 runaway infinite recursion. I updated the discussion in the pcrestack man
2278 page.
2279
2280 13. Now that we have gone to the x.xx style of version numbers, the minor
2281 version may start with zero. Using 08 or 09 is a bad idea because users
2282 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2283 interpreted as invalid octal numbers. I've updated the previous comment in
2284 configure.ac, and also added a check that gives an error if 08 or 09 are
2285 used.
2286
2287 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2288 causing partial matching to fail when the end of the subject matched \W
2289 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2290
2291 15. There were some discrepancies between the declarations in pcre_internal.h
2292 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2293 their definitions. The declarations used "const uschar *" and the
2294 definitions used USPTR. Even though USPTR is normally defined as "const
2295 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2296 reported that: "This difference in casting confuses some C++ compilers, for
2297 example, SunCC recognizes above declarations as different functions and
2298 generates broken code for hbpcre." I have changed the declarations to use
2299 USPTR.
2300
2301 16. GNU libtool is named differently on some systems. The autogen.sh script now
2302 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2303 (FreeBSD).
2304
2305 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2306 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2307 comment: "Figure out how to create a longlong from a string: strtoll and
2308 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2309 instance, but it only takes 2 args instead of 3!"
2310
2311 18. A subtle bug concerned with back references has been fixed by a change of
2312 specification, with a corresponding code fix. A pattern such as
2313 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2314 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2315 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2316 same bug. Such groups have to be quantified to be useful, or contained
2317 inside another quantified group. (If there's no repetition, the reference
2318 can never match.) The problem arises because, having left the group and
2319 moved on to the rest of the pattern, a later failure that backtracks into
2320 the group uses the captured value from the final iteration of the group
2321 rather than the correct earlier one. I have fixed this in PCRE by forcing
2322 any group that contains a reference to itself to be an atomic group; that
2323 is, there cannot be any backtracking into it once it has completed. This is
2324 similar to recursive and subroutine calls.
2325
2326
2327 Version 8.00 19-Oct-09
2328 ----------------------
2329
2330 1. The table for translating pcre_compile() error codes into POSIX error codes
2331 was out-of-date, and there was no check on the pcre_compile() error code
2332 being within the table. This could lead to an OK return being given in
2333 error.
2334
2335 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2336 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2337 in a Windows environment.
2338
2339 3. The pcregrep --count option prints the count for each file even when it is
2340 zero, as does GNU grep. However, pcregrep was also printing all files when
2341 --files-with-matches was added. Now, when both options are given, it prints
2342 counts only for those files that have at least one match. (GNU grep just
2343 prints the file name in this circumstance, but including the count seems
2344 more useful - otherwise, why use --count?) Also ensured that the
2345 combination -clh just lists non-zero counts, with no names.
2346
2347 4. The long form of the pcregrep -F option was incorrectly implemented as
2348 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2349 but it seems right to fix it, and I didn't think it was worth preserving
2350 the old behaviour.
2351
2352 5. The command line items --regex=pattern and --regexp=pattern were not
2353 recognized by pcregrep, which required --regex pattern or --regexp pattern
2354 (with a space rather than an '='). The man page documented the '=' forms,
2355 which are compatible with GNU grep; these now work.
2356
2357 6. No libpcreposix.pc file was created for pkg-config; there was just
2358 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2359
2360 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2361 when UCP support is not needed, by modifying the Python script that
2362 generates it from Unicode data files. This should not matter if the module
2363 is correctly used as a library, but I received one complaint about 50K of
2364 unwanted data. My guess is that the person linked everything into his
2365 program rather than using a library. Anyway, it does no harm.
2366
2367 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2368 was a minimum greater than 1 for a wide character in a possessive
2369 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2370 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2371 character. Chaos in the form of incorrect output or a compiling loop could
2372 result.
2373
2374 9. The restrictions on what a pattern can contain when partial matching is
2375 requested for pcre_exec() have been removed. All patterns can now be
2376 partially matched by this function. In addition, if there are at least two
2377 slots in the offset vector, the offset of the earliest inspected character
2378 for the match and the offset of the end of the subject are set in them when
2379 PCRE_ERROR_PARTIAL is returned.
2380
2381 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2382 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2383 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2384 and may be more useful for multi-segment matching.
2385
2386 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2387 used to be given if ever the end of the subject was reached; now it is
2388 given only if matching could not proceed because another character was
2389 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2390 string "Z", which now yields "no match" instead of "partial match". In the
2391 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2392 final character ended with (*FAIL).
2393
2394 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2395 if the pattern had a "must contain" character that was already found in the
2396 earlier partial match, unless partial matching was again requested. For
2397 example, with the pattern /dog.(body)?/, the "must contain" character is
2398 "g". If the first part-match was for the string "dog", restarting with
2399 "sbody" failed. This bug has been fixed.
2400
2401 13. The string returned by pcre_dfa_exec() after a partial match has been
2402 changed so that it starts at the first inspected character rather than the
2403 first character of the match. This makes a difference only if the pattern
2404 starts with a lookbehind assertion or \b or \B (\K is not supported by
2405 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2406 matching functions compatible, and I think it's the right thing to do.
2407
2408 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2409 so that the demonstration program is easily available in environments where
2410 PCRE has not been installed from source.
2411
2412 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2413 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2414 library.
2415
2416 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2417 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2418 is not the first non-POSIX option to be added. Clearly some people find
2419 these options useful.
2420
2421 17. If a caller to the POSIX matching function regexec() passes a non-zero
2422 value for nmatch with a NULL value for pmatch, the value of
2423 nmatch is forced to zero.
2424
2425 18. RunGrepTest did not have a test for the availability of the -u option of
2426 the diff command, as RunTest does. It now checks in the same way as
2427 RunTest, and also checks for the -b option.
2428
2429 19. If an odd number of negated classes containing just a single character
2430 interposed, within parentheses, between a forward reference to a named
2431 subpattern and the definition of the subpattern, compilation crashed with
2432 an internal error, complaining that it could not find the referenced
2433 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2434 [The bug was that it was starting one character too far in when skipping
2435 over the character class, thus treating the ] as data rather than
2436 terminating the class. This meant it could skip too much.]
2437
2438 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2439 /g option in pcretest when the pattern contains \K, which makes it possible
2440 to have an empty string match not at the start, even when the pattern is
2441 anchored. Updated pcretest and pcredemo to use this option.
2442
2443 21. If the maximum number of capturing subpatterns in a recursion was greater
2444 than the maximum at the outer level, the higher number was returned, but
2445 with unset values at the outer level. The correct (outer level) value is
2446 now given.
2447
2448 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2449 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2450 make it do so. The string so far is captured, making this feature
2451 compatible with Perl.
2452
2453 23. The tests have been re-organized, adding tests 11 and 12, to make it
2454 possible to check the Perl 5.10 features against Perl 5.10.
2455
2456 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2457 pattern matches a fixed length string. PCRE did not allow this; now it
2458 does. Neither allows recursion.
2459
2460 25. I finally figured out how to implement a request to provide the minimum
2461 length of subject string that was needed in order to match a given pattern.
2462 (It was back references and recursion that I had previously got hung up
2463 on.) This code has now been added to pcre_study(); it finds a lower bound
2464 to the length of subject needed. It is not necessarily the greatest lower
2465 bound, but using it to avoid searching strings that are too short does give
2466 some useful speed-ups. The value is available to calling programs via
2467 pcre_fullinfo().
2468
2469 26. While implementing 25, I discovered to my embarrassment that pcretest had
2470 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2471 study optimizations had never been tested with that matching function.
2472 Oops. What is worse, even when it was passed study data, there was a bug in
2473 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2474 were also very few tests of studied patterns with pcre_dfa_exec().
2475
2476 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2477 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2478 on the other side of the coin, they are no longer allowed to have different
2479 names, because these cannot be distinguished in PCRE, and this has caused
2480 confusion. (This is a difference from Perl.)
2481
2482 28. When duplicate subpattern names are present (necessarily with different
2483 numbers, as required by 27 above), and a test is made by name in a
2484 conditional pattern, either for a subpattern having been matched, or for
2485 recursion in such a pattern, all the associated numbered subpatterns are
2486 tested, and the overall condition is true if the condition is true for any
2487 one of them. This is the way Perl works, and is also more like the way
2488 testing by number works.
2489
2490
2491 Version 7.9 11-Apr-09
2492 ---------------------
2493
2494 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2495 (pcretest), all targets were linked against these libraries. This included
2496 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2497 libraries. This caused unwanted dependencies to be created. This problem
2498 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2499 pcretest is linked with readline.
2500
2501 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2502 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2503 moved outside it again, because FALSE and TRUE are already defined in AIX,
2504 but BOOL is not.
2505
2506 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2507 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2508
2509 4. The pcregrep documentation said spaces were inserted as well as colons (or
2510 hyphens) following file names and line numbers when outputting matching
2511 lines. This is not true; no spaces are inserted. I have also clarified the
2512 wording for the --colour (or --color) option.
2513
2514 5. In pcregrep, when --colour was used with -o, the list of matching strings
2515 was not coloured; this is different to GNU grep, so I have changed it to be
2516 the same.
2517
2518 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2519 each matching line was coloured. Now it goes on to look for further matches
2520 of any of the test patterns, which is the same behaviour as GNU grep.
2521
2522 7. A pattern that could match an empty string could cause pcregrep to loop; it
2523 doesn't make sense to accept an empty string match in pcregrep, so I have
2524 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2525 seems to be how GNU grep behaves. [But see later change 40 for release
2526 8.33.]
2527
2528 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2529 start or after a newline", because the conditional assertion was not being
2530 correctly handled. The rule now is that both the assertion and what follows
2531 in the first alternative must satisfy the test.
2532
2533 9. If auto-callout was enabled in a pattern with a conditional group whose
2534 condition was an assertion, PCRE could crash during matching, both with
2535 pcre_exec() and pcre_dfa_exec().
2536
2537 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2538 used for matching.
2539
2540 11. Unicode property support in character classes was not working for
2541 characters (bytes) greater than 127 when not in UTF-8 mode.
2542
2543 12. Added the -M command line option to pcretest.
2544
2545 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2546
2547 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2548
2549 16. Added comments and documentation about mis-use of no_arg in the C++
2550 wrapper.
2551
2552 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2553 from Martin Jerabek that uses macro names for all relevant character and
2554 string constants.
2555
2556 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2557 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2558 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2559 these, but not everybody uses configure.
2560
2561 19. A conditional group that had only one branch was not being correctly
2562 recognized as an item that could match an empty string. This meant that an
2563 enclosing group might also not be so recognized, causing infinite looping
2564 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2565 with the subject "ab", where knowledge that the repeated group can match
2566 nothing is needed in order to break the loop.
2567
2568 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2569 exec(), but without supplying a callout function, matching went wrong.
2570
2571 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2572 leak if the size of the offset vector was greater than 30. When the vector
2573 is smaller, the saved offsets during recursion go onto a local stack
2574 vector, but for larger vectors malloc() is used. It was failing to free
2575 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2576 error, in fact).
2577
2578 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2579 heapframe that is used only when UTF-8 support is enabled. This caused no
2580 problem, but was untidy.
2581
2582 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2583 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2584 included within another project.
2585
2586 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2587 slightly modified by me:
2588
2589 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2590 not building pcregrep.
2591
2592 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2593 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2594
2595 25. Forward references, both numeric and by name, in patterns that made use of
2596 duplicate group numbers, could behave incorrectly or give incorrect errors,
2597 because when scanning forward to find the reference group, PCRE was not
2598 taking into account the duplicate group numbers. A pattern such as
2599 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2600
2601 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2602 the feature of a custom pointer more persuasive (as requested by a user).
2603
2604 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2605 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2606 pre-defined.
2607
2608 28. Added support for (*UTF8) at the start of a pattern.
2609
2610 29. Arrange for flags added by the "release type" setting in CMake to be shown
2611 in the configuration summary.
2612
2613
2614 Version 7.8 05-Sep-08
2615 ---------------------
2616
2617 1. Replaced UCP searching code with optimized version as implemented for Ad
2618 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2619 stage table and inline lookup instead of a function, giving speed ups of 2
2620 to 5 times on some simple patterns that I tested. Permission was given to
2621 distribute the MultiStage2.py script that generates the tables (it's not in
2622 the tarball, but is in the Subversion repository).
2623
2624 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2625 scripts.
2626
2627 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2628 a group with a zero qualifier. The result of the study could be incorrect,
2629 or the function might crash, depending on the pattern.
2630
2631 4. Caseless matching was not working for non-ASCII characters in back
2632 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2633 It now works when Unicode Property Support is available.
2634
2635 5. In pcretest, an escape such as \x{de} in the data was always generating
2636 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2637 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2638 truncation.
2639
2640 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2641
2642 7. Added two (int) casts to pcregrep when printing the difference of two
2643 pointers, in case they are 64-bit values.
2644
2645 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2646 test 2 if it fails.
2647
2648 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2649 and a #define of that name to empty if it is not externally set. This is to
2650 allow users of MSVC to set it if necessary.
2651
2652 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2653 the convenience functions in the pcre_get.c source file.
2654
2655 11. An option change at the start of a pattern that had top-level alternatives
2656 could cause overwriting and/or a crash. This command provoked a crash in
2657 some environments:
2658
2659 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2660
2661 This potential security problem was recorded as CVE-2008-2371.
2662
2663 12. For a pattern where the match had to start at the beginning or immediately
2664 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2665 pcre_dfa_exec() could read past the end of the passed subject if there was
2666 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2667 pcretest so that it places the subject at the end of its malloc-ed buffer.
2668
2669 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2670 exec() might read past the end of the data buffer in UTF-8 mode.
2671
2672 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2673 the data contained the byte 0x85 as part of a UTF-8 character within its
2674 first line. This applied both to normal and DFA matching.
2675
2676 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2677 /^[^d]*?$/8 failed to match "abc".
2678
2679 16. Added a missing copyright notice to pcrecpp_internal.h.
2680
2681 17. Make it more clear in the documentation that values returned from
2682 pcre_exec() in ovector are byte offsets, not character counts.
2683
2684 18. Tidied a few places to stop certain compilers from issuing warnings.
2685
2686 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2687 supplied by Stefan Weber. I made a further small update for 7.8 because
2688 there is a change of source arrangements: the pcre_searchfuncs.c module is
2689 replaced by pcre_ucd.c.
2690
2691
2692 Version 7.7 07-May-08
2693 ---------------------
2694
2695 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2696 a string to a long long, pretend we don't even have a long long." This is
2697 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2698
2699 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2700 pre-7.6 versions, which defined a global no_arg variable instead of putting
2701 it in the RE class. (See also #8 below.)
2702
2703 3. Remove a line of dead code, identified by coverity and reported by Nuno
2704 Lopes.
2705
2706 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2707
2708 (1) The include/exclude patterns were being applied to the whole pathnames
2709 of files, instead of just to the final components.
2710
2711 (2) If there was more than one level of directory, the subdirectories were
2712 skipped unless they satisfied the include/exclude conditions. This is
2713 inconsistent with GNU grep (and could even be seen as contrary to the
2714 pcregrep specification - which I improved to make it absolutely clear).
2715 The action now is always to scan all levels of directory, and just
2716 apply the include/exclude patterns to regular files.
2717
2718 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2719 --exclude_dir in the tests to avoid scanning .svn directories.
2720
2721 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2722 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2723 doesn't support NULs in patterns.
2724
2725 7. Added some missing "const"s to declarations of static tables in
2726 pcre_compile.c and pcre_dfa_exec.c.
2727
2728 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2729 caused by fix #2 above. (Subsequently also a second patch to fix the
2730 first patch. And a third patch - this was a messy problem.)
2731
2732 9. Applied Craig's patch to remove the use of push_back().
2733
2734 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2735 matching function regexec().
2736
2737 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2738 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2739 references. PCRE supports relative numbers with this syntax (I don't think
2740 Oniguruma does).
2741
2742 12. Previously, a group with a zero repeat such as (...){0} was completely
2743 omitted from the compiled regex. However, this means that if the group
2744 was called as a subroutine from elsewhere in the pattern, things went wrong
2745 (an internal error was given). Such groups are now left in the compiled
2746 pattern, with a new opcode that causes them to be skipped at execution
2747 time.
2748
2749 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2750 to the way PCRE behaves:
2751
2752 (a) A lone ] character is dis-allowed (Perl treats it as data).
2753
2754 (b) A back reference to an unmatched subpattern matches an empty string
2755 (Perl fails the current match path).
2756
2757 (c) A data ] in a character class must be notated as \] because if the
2758 first data character in a class is ], it defines an empty class. (In
2759 Perl it is not possible to have an empty class.) The empty class []
2760 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2761 The negative empty class [^] matches any one character, independently
2762 of the DOTALL setting.
2763
2764 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2765 non-existent subpattern following a character class starting with ']' and
2766 containing () gave an internal compiling error instead of "reference to
2767 non-existent subpattern". Fortunately, when the pattern did exist, the
2768 compiled code was correct. (When scanning forwards to check for the
2769 existence of the subpattern, it was treating the data ']' as terminating
2770 the class, so got the count wrong. When actually compiling, the reference
2771 was subsequently set up correctly.)
2772
2773 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2774 it was being rejected as not supported by pcre_dfa_exec(), even though
2775 other assertions are supported. I have made pcre_dfa_exec() support
2776 (*FAIL).
2777
2778 16. The implementation of 13c above involved the invention of a new opcode,
2779 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2780 cannot be changed at match time, I realized I could make a small
2781 improvement to matching performance by compiling OP_ALLANY instead of
2782 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2783 on the OP_ANY path.
2784
2785 17. Compiling pcretest on Windows with readline support failed without the
2786 following two fixes: (1) Make the unistd.h include conditional on
2787 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2788
2789 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2790 ncurses library to be included for pcretest when ReadLine support is
2791 requested, but also to allow for it to be overridden. This patch came from
2792 Daniel Bergström.
2793
2794 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2795 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2796 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2797 spotting this.
2798
2799
2800 Version 7.6 28-Jan-08
2801 ---------------------
2802
2803 1. A character class containing a very large number of characters with
2804 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2805 overflow.
2806
2807 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2808 HAVE_LONG_LONG is not defined.
2809
2810 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2811 bring them up to date and include new features. This patch includes:
2812
2813 - Fixed PH's badly added libz and libbz2 support.
2814 - Fixed a problem with static linking.
2815 - Added pcredemo. [But later removed - see 7 below.]
2816 - Fixed dftables problem and added an option.
2817 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2818 HAVE_LONG_LONG.
2819 - Added readline support for pcretest.
2820 - Added an listing of the option settings after cmake has run.
2821
2822 4. A user submitted a patch to Makefile that makes it easy to create
2823 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2824 Makefile.am that cause it to include this special target, without
2825 affecting anything else. Note that the same mingw target plus all
2826 the other distribution libraries and programs are now supported
2827 when configuring with CMake (see 6 below) instead of with
2828 Configure/Make.
2829
2830 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2831 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2832 exported in the Windows port". It has not yet been confirmed that the patch
2833 solves the problem, but it does no harm.
2834
2835 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2836 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2837 with CMake, and also correct the comment about stack recursion.
2838
2839 7. Remove the automatic building of pcredemo from the ./configure system and
2840 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2841 of a program that users should build themselves after PCRE is installed, so
2842 building it automatically is not really right. What is more, it gave
2843 trouble in some build environments.
2844
2845 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2846
2847
2848 Version 7.5 10-Jan-08
2849 ---------------------
2850
2851 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2852 values in parens when parsing an RE using the C++ wrapper."
2853
2854 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2855 Characters greater than 255 were excluded from the class instead of being
2856 included.
2857
2858 3. The same bug as (2) above applied to negated POSIX classes such as
2859 [:^space:].
2860
2861 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2862 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2863 I have changed it.
2864
2865 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2866 first named subpattern) and a construct such as (?&a) would reference the
2867 first named subpattern whose name started with "a" (in other words, the
2868 length check was missing). Both these problems are fixed. "Subpattern name
2869 expected" is now given for (?&) (a zero-length name), and this patch also
2870 makes it give the same error for \k'' (previously it complained that that
2871 was a reference to a non-existent subpattern).
2872
2873 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2874 this is right because (?- can be followed by option settings as well as by
2875 digits. I have, however, made the messages clearer.
2876
2877 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2878 than the number used in the conditional) now cause a compile-time error.
2879 This is actually not compatible with Perl, which accepts such patterns, but
2880 treats the conditional as always being FALSE (as PCRE used to), but it
2881 seems to me that giving a diagnostic is better.
2882
2883 8. Change "alphameric" to the more common word "alphanumeric" in comments
2884 and messages.
2885
2886 9. Fix two occurrences of "backslash" in comments that should have been
2887 "backspace".
2888
2889 10. Remove two redundant lines of code that can never be obeyed (their function
2890 was moved elsewhere).
2891
2892 11. The program that makes PCRE's Unicode character property table had a bug
2893 which caused it to generate incorrect table entries for sequences of
2894 characters that have the same character type, but are in different scripts.
2895 It amalgamated them into a single range, with the script of the first of
2896 them. In other words, some characters were in the wrong script. There were
2897 thirteen such cases, affecting characters in the following ranges:
2898
2899 U+002b0 - U+002c1
2900 U+0060c - U+0060d
2901 U+0061e - U+00612
2902 U+0064b - U+0065e
2903 U+0074d - U+0076d
2904 U+01800 - U+01805
2905 U+01d00 - U+01d77
2906 U+01d9b - U+01dbf
2907 U+0200b - U+0200f
2908 U+030fc - U+030fe
2909 U+03260 - U+0327f
2910 U+0fb46 - U+0fbb1
2911 U+10450 - U+1049d
2912
2913 12. The -o option (show only the matching part of a line) for pcregrep was not
2914 compatible with GNU grep in that, if there was more than one match in a
2915 line, it showed only the first of them. It now behaves in the same way as
2916 GNU grep.
2917
2918 13. If the -o and -v options were combined for pcregrep, it printed a blank
2919 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2920 does the same. The return code can be used to tell if there were any
2921 non-matching lines.
2922
2923 14. Added --file-offsets and --line-offsets to pcregrep.
2924
2925 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2926 infinitely looping recursion. The bug was that positive lookaheads were not
2927 being skipped when checking for a possible empty match (negative lookaheads
2928 and both kinds of lookbehind were skipped).
2929
2930 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2931 inclusion of <windows.h> to before rather than after the definition of
2932 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2933
2934 17. Specifying a possessive quantifier with a specific limit for a Unicode
2935 character property caused pcre_compile() to compile bad code, which led at
2936 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2937 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2938 caused the error; without that there was no problem.
2939
2940 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2941
2942 19. Added --enable-pcretest-libreadline.
2943
2944 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2945 RE::GlobalReplace(). As a result, the number of replacements returned was
2946 double what it should be. I removed one of the increments, but Craig sent a
2947 later patch that removed the other one (the right fix) and added unit tests
2948 that check the return values (which was not done before).
2949
2950 21. Several CMake things:
2951
2952 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2953 the names libpcre and libpcreposix, not just pcre and pcreposix.
2954
2955 (2) The above change means that pcretest and pcregrep are now correctly
2956 linked with the newly-built libraries, not previously installed ones.
2957
2958 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2959
2960 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2961 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2962 UTF-8 newline character). The key issue is that the pattern starts .*;
2963 this means that the match must be either at the beginning, or after a
2964 newline. The bug was in the code for advancing after a failed match and
2965 checking that the new position followed a newline. It was not taking
2966 account of UTF-8 characters correctly.
2967
2968 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2969 character classes. PCRE was not treating the sequence [:...:] as a
2970 character class unless the ... were all letters. Perl, however, seems to
2971 allow any characters between [: and :], though of course it rejects as
2972 unknown any "names" that contain non-letters, because all the known class
2973 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2974 for example, whereas PCRE did not - it did not recognize a POSIX character
2975 class. This seemed a bit dangerous, so the code has been changed to be
2976 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2977 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2978 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2979 Perl does, and where it didn't before.
2980
2981 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2982 Windows environments %n is disabled by default.
2983
2984
2985 Version 7.4 21-Sep-07
2986 ---------------------
2987
2988 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2989 means that a class such as [\s] counted as "explicit reference to CR or
2990 LF". That isn't really right - the whole point of the change was to try to
2991 help when there was an actual mention of one of the two characters. So now
2992 the change happens only if \r or \n (or a literal CR or LF) character is
2993 encountered.
2994
2995 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2996 of both had grown to the point where there were only 3 bits left.
2997 Fortunately, there was spare space in the data structure, and so I have
2998 moved the internal flags into a new 16-bit field to free up more option
2999 bits.
3000
3001 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3002 but did not set the internal JCHANGED flag - either of these is enough to
3003 control the way the "get" function works - but the PCRE_INFO_JCHANGED
3004 facility is supposed to tell if (?J) was ever used, so now (?J) at the
3005 start sets both bits.
3006
3007 4. Added options (at build time, compile time, exec time) to change \R from
3008 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3009
3010 5. doc/pcresyntax.html was missing from the distribution.
3011
3012 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3013 compatibility, even though it is no longer used.
3014
3015 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3016 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3017 windows.h file is present (where different names are used). [This was
3018 reversed later after testing - see 16 below.]
3019
3020 8. Changed all #include <config.h> to #include "config.h". There were also
3021 some further <pcre.h> cases that I changed to "pcre.h".
3022
3023 9. When pcregrep was used with the --colour option, it missed the line ending
3024 sequence off the lines that it output.
3025
3026 10. It was pointed out to me that arrays of string pointers cause lots of
3027 relocations when a shared library is dynamically loaded. A technique of
3028 using a single long string with a table of offsets can drastically reduce
3029 these. I have refactored PCRE in four places to do this. The result is
3030 dramatic:
3031
3032 Originally: 290
3033 After changing UCP table: 187
3034 After changing error message table: 43
3035 After changing table of "verbs" 36
3036 After changing table of Posix names 22
3037
3038 Thanks to the folks working on Gregex for glib for this insight.
3039
3040 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3041 unicode-properties was also set.
3042
3043 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3044
3045 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3046 checked only for CRLF.
3047
3048 14. Added casts to pcretest.c to avoid compiler warnings.
3049
3050 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3051
3052 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3053 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3054 entirely. This removes changes made in 7 above.
3055
3056 17. The CMake files have been updated, and there is now more information about
3057 building with CMake in the NON-UNIX-USE document.
3058
3059
3060 Version 7.3 28-Aug-07
3061 ---------------------
3062
3063 1. In the rejigging of the build system that eventually resulted in 7.1, the
3064 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3065 brackets there is not right, since it causes compilers to look for an
3066 installed pcre.h, not the version that is in the source that is being
3067 compiled (which of course may be different). I have changed it back to:
3068
3069 #include "pcre.h"
3070
3071 I have a vague recollection that the change was concerned with compiling in
3072 different directories, but in the new build system, that is taken care of
3073 by the VPATH setting the Makefile.
3074
3075 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3076 when the subject happened to end in the byte 0x85 (e.g. if the last
3077 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3078 characters but of course it shouldn't be taken as a newline when it is part
3079 of another character. The bug was that, for an unlimited repeat of . in
3080 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3081 characters when looking for a newline.
3082
3083 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3084
3085 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3086 in debug output.
3087
3088 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3089 long printing in the pcrecpp unittest when running under MinGW.
3090
3091 6. ESC_K was left out of the EBCDIC table.
3092
3093 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3094 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3095 limit also applies to "virtual nesting" when a pattern is recursive, and in
3096 this case 1000 isn't so big. I have been able to remove this limit at the
3097 expense of backing off one optimization in certain circumstances. Normally,
3098 when pcre_exec() would call its internal match() function recursively and
3099 immediately return the result unconditionally, it uses a "tail recursion"
3100 feature to save stack. However, when a subpattern that can match an empty
3101 string has an unlimited repetition quantifier, it no longer makes this
3102 optimization. That gives it a stack frame in which to save the data for
3103 checking that an empty string has been matched. Previously this was taken
3104 from the 1000-entry workspace that had been reserved. So now there is no
3105 explicit limit, but more stack is used.
3106
3107 8. Applied Daniel's patches to solve problems with the import/export magic
3108 syntax that is required for Windows, and which was going wrong for the
3109 pcreposix and pcrecpp parts of the library. These were overlooked when this
3110 problem was solved for the main library.
3111
3112 9. There were some crude static tests to avoid integer overflow when computing
3113 the size of patterns that contain repeated groups with explicit upper
3114 limits. As the maximum quantifier is 65535, the maximum group length was
3115 set at 30,000 so that the product of these two numbers did not overflow a
3116 32-bit integer. However, it turns out that people want to use groups that
3117 are longer than 30,000 bytes (though not repeat them that many times).
3118 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3119 made it possible to implement the integer overflow checks in a much more
3120 dynamic way, which I have now done. The artificial limitation on group
3121 length has been removed - we now have only the limit on the total length of
3122 the compiled pattern, which depends on the LINK_SIZE setting.
3123
3124 10. Fixed a bug in the documentation for get/copy named substring when
3125 duplicate names are permitted. If none of the named substrings are set, the
3126 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3127 empty string.
3128
3129 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3130 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3131 because the ] is interpreted as the first data character and the
3132 terminating ] is not found. PCRE has been made compatible with Perl in this
3133 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3134 cause memory overwriting.
3135
3136 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3137 string has been matched (to stop an infinite loop). It was not recognizing
3138 a conditional subpattern that could match an empty string if that
3139 subpattern was within another subpattern. For example, it looped when
3140 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3141 condition was not nested. This bug has been fixed.
3142
3143 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3144 past the start of the subject in the presence of bytes with the top bit
3145 set, for example "\x8aBCD".
3146
3147 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3148 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3149
3150 14. Optimized (?!) to (*FAIL).
3151
3152 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3153 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3154 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3155 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3156 does: it's just the validity check that is more restrictive.
3157
3158 16. Inserted checks for integer overflows during escape sequence (backslash)
3159 processing, and also fixed erroneous offset values for syntax errors during
3160 backslash processing.
3161
3162 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3163 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3164
3165 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3166 caused an overrun.
3167
3168 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3169 something other than just ASCII characters) inside a group that had an
3170 unlimited repeat caused a loop at compile time (while checking to see
3171 whether the group could match an empty string).
3172
3173 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3174 [\P{Any}] did so. (Error in the code for printing property names.)
3175
3176 21. An orphan \E inside a character class could cause a crash.
3177
3178 22. A repeated capturing bracket such as (A)? could cause a wild memory
3179 reference during compilation.
3180
3181 23. There are several functions in pcre_compile() that scan along a compiled
3182 expression for various reasons (e.g. to see if it's fixed length for look
3183 behind). There were bugs in these functions when a repeated \p or \P was
3184 present in the pattern. These operators have additional parameters compared
3185 with \d, etc, and these were not being taken into account when moving along
3186 the compiled data. Specifically:
3187
3188 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3189 length.
3190
3191 (b) An item such as \pL+ within a repeated group could cause crashes or
3192 loops.
3193
3194 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3195 "reference to non-existent subpattern" error.
3196
3197 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3198
3199 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3200 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3201
3202 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3203
3204 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3205 character were causing crashes (broken optimization).
3206
3207 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3208 \p or \P) caused a compile-time loop.
3209
3210 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3211 break. For example, the unstudied pattern [\r\n]A does not match the string
3212 "\r\nA" because change 7.0/46 below moves the current point on by two
3213 characters after failing to match at the start. However, the pattern \nA
3214 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3215 the same is true. There doesn't seem any very clean way out of this, but
3216 what I have chosen to do makes the common cases work: PCRE now takes note
3217 of whether there can be an explicit match for \r or \n anywhere in the
3218 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3219 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3220 pattern has explicit CR or LF references.
3221
3222 29. Added (*CR) etc for changing newline setting at start of pattern.
3223
3224
3225 Version 7.2 19-Jun-07
3226 ---------------------
3227
3228 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3229 which is apparently normally available under Windows.
3230
3231 2. Re-jig the pcregrep tests with different newline settings in an attempt
3232 to make them independent of the local environment's newline setting.
3233
3234 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3235
3236 4. Some of the "internals" tests were previously cut out when the link size
3237 was not 2, because the output contained actual offsets. The recent new
3238 "Z" feature of pcretest means that these can be cut out, making the tests
3239 usable with all link sizes.
3240
3241 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3242 stack recursion. This gives a massive performance boost under BSD, but just
3243 a small improvement under Linux. However, it saves one field in the frame
3244 in all cases.
3245
3246 6. Added more features from the forthcoming Perl 5.10:
3247
3248 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3249 recursion call. It refers to the nth most recently opened parentheses.
3250
3251 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3252 to be opened parentheses.
3253
3254 (c) Conditions that refer to capturing parentheses can be specified
3255 relatively, for example, (?(-2)... or (?(+3)...
3256
3257 (d) \K resets the start of the current match so that everything before
3258 is not part of it.
3259
3260 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3261
3262 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3263 reference syntax.
3264
3265 (g) (?| introduces a group in which the numbering of parentheses in each
3266 alternative starts with the same number.
3267
3268 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3269
3270 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3271 PCRE_INFO_JCHANGED.
3272
3273 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3274 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3275 for detecting groups that can match an empty string.
3276
3277 9. A pattern with a very large number of alternatives (more than several
3278 hundred) was running out of internal workspace during the pre-compile
3279 phase, where pcre_compile() figures out how much memory will be needed. A
3280 bit of new cunning has reduced the workspace needed for groups with
3281 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3282 workspace instead of running out of the 4096 that are available.
3283
3284 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3285
3286 11. Applied patch from Google to remove an optimization that didn't quite work.
3287 The report of the bug said:
3288
3289 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3290 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3291 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3292
3293 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3294 it matched the wrong number of bytes.
3295
3296
3297 Version 7.1 24-Apr-07
3298 ---------------------
3299
3300 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3301 that is more "standard", making use of automake and other Autotools. There
3302 is some re-arrangement of the files and adjustment of comments consequent
3303 on this.
3304
3305 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3306 for recursive directory scanning broke on some systems because the files
3307 are not scanned in any specific order and on different systems the order
3308 was different. A call to "sort" has been inserted into RunGrepTest for the
3309 approprate test as a short-term fix. In the longer term there may be an
3310 alternative.
3311
3312 3. I had an email from Eric Raymond about problems translating some of PCRE's
3313 man pages to HTML (despite the fact that I distribute HTML pages, some
3314 people do their own conversions for various reasons). The problems
3315 concerned the use of low-level troff macros .br and .in. I have therefore
3316 removed all such uses from the man pages (some were redundant, some could
3317 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3318 HTML has been updated to handle .nf/.fi and to complain if it encounters
3319 .br or .in.
3320
3321 4. Updated comments in configure.ac that get placed in config.h.in and also
3322 arranged for config.h to be included in the distribution, with the name
3323 config.h.generic, for the benefit of those who have to compile without
3324 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3325
3326 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3327 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3328 makevp.bat for the current PCRE, using the additional files
3329 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3330
3331 6. A Windows user reported a minor discrepancy with test 2, which turned out
3332 to be caused by a trailing space on an input line that had got lost in his
3333 copy. The trailing space was an accident, so I've just removed it.
3334
3335 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3336 that is needed.
3337
3338 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3339 as "const" (a) because they are and (b) because it helps the PHP
3340 maintainers who have recently made a script to detect big data structures
3341 in the php code that should be moved to the .rodata section. I remembered
3342 to update Builducptable as well, so it won't revert if ucptable.h is ever
3343 re-created.
3344
3345 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3346 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3347 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3348 support is not required. This saves 1.5-2K of code, which is important in
3349 some applications.
3350
3351 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3352 so as not to refer to the tables, even though these functions will never be
3353 called when UTF-8 support is disabled. Otherwise there are problems with a
3354 shared library.
3355
3356 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3357
3358 (a) It was defining its arguments as char * instead of void *.
3359
3360 (b) It was assuming that all moves were upwards in memory; this was true
3361 a long time ago when I wrote it, but is no longer the case.
3362
3363 The emulated memove() is provided for those environments that have neither
3364 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3365 is clearly not the case, as these two bugs were recently reported.
3366
3367 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3368 and Detrail to create the HTML documentation, the .txt form of the man
3369 pages, and it removes trailing spaces from listed files. It also creates
3370 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3371 case, it wraps all the #defines with #ifndefs. This script should be run
3372 before "make dist".
3373
3374 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3375 with Unicode property support.
3376
3377 (a) For a maximizing quantifier, if the two different cases of the
3378 character were of different lengths in their UTF-8 codings (there are
3379 some cases like this - I found 11), and the matching function had to
3380 back up over a mixture of the two cases, it incorrectly assumed they
3381 were both the same length.
3382
3383 (b) When PCRE was configured to use the heap rather than the stack for
3384 recursion during matching, it was not correctly preserving the data for
3385 the other case of a UTF-8 character when checking ahead for a match
3386 while processing a minimizing repeat. If the check also involved
3387 matching a wide character, but failed, corruption could cause an
3388 erroneous result when trying to check for a repeat of the original
3389 character.
3390
3391 13. Some tidying changes to the testing mechanism:
3392
3393 (a) The RunTest script now detects the internal link size and whether there
3394 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3395 values substituted by "configure". (The RunGrepTest script already did
3396 this for UTF-8.) The configure.ac script no longer substitutes the
3397 relevant variables.
3398
3399 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3400 with length and offset values. This means that the output is different
3401 for different internal link sizes. Test 2 is skipped for link sizes
3402 other than 2 because of this, bypassing the problem. Unfortunately,
3403 there was also a test in test 3 (the locale tests) that used /B and
3404 failed for link sizes other than 2. Rather than cut the whole test out,
3405 I have added a new /Z option to pcretest that replaces the length and
3406 offset values with spaces. This is now used to make test 3 independent
3407 of link size. (Test 2 will be tidied up later.)
3408
3409 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3410 segmentation fault instead of returning the appropriate error message.
3411
3412 15. In multiline mode when the newline sequence was set to "any", the pattern
3413 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3414 This doesn't seem right; it now treats the CRLF combination as the line
3415 ending, and so does not match in that case. It's only a pattern such as ^$
3416 that would hit this one: something like ^ABC$ would have failed after \r
3417 and then tried again after \r\n.
3418
3419 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3420 in an attempt to make files that differ only in their line terminators
3421 compare equal. This works on Linux.
3422
3423 17. Under certain error circumstances pcregrep might try to free random memory
3424 as it exited. This is now fixed, thanks to valgrind.
3425
3426 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3427 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3428 was because its rules for how to advance for /g after matching an empty
3429 string at the end of a line did not allow for this case. They now check for
3430 it specially.
3431
3432 20. pcretest is supposed to handle patterns and data of any length, by
3433 extending its buffers when necessary. It was getting this wrong when the
3434 buffer for a data line had to be extended.
3435
3436 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3437 CRLF as a newline sequence.
3438
3439 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3440 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3441 I have nevertheless tidied it up.
3442
3443 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3444
3445 24. Added a man page for pcre-config.
3446
3447
3448 Version 7.0 19-Dec-06
3449 ---------------------
3450
3451 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3452 moving to gcc 4.1.1.
3453
3454 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3455 sys/time.h, which is documented as needed for this function. It doesn't
3456 seem to matter on Linux, but it showed up on some releases of OS X.
3457
3458 3. It seems that there are systems where bytes whose values are greater than
3459 127 match isprint() in the "C" locale. The "C" locale should be the
3460 default when a C program starts up. In most systems, only ASCII printing
3461 characters match isprint(). This difference caused the output from pcretest
3462 to vary, making some of the tests fail. I have changed pcretest so that:
3463
3464 (a) When it is outputting text in the compiled version of a pattern, bytes
3465 other than 32-126 are always shown as hex escapes.
3466
3467 (b) When it is outputting text that is a matched part of a subject string,
3468 it does the same, unless a different locale has been set for the match
3469 (using the /L modifier). In this case, it uses isprint() to decide.
3470
3471 4. Fixed a major bug that caused incorrect computation of the amount of memory
3472 required for a compiled pattern when options that changed within the
3473 pattern affected the logic of the preliminary scan that determines the
3474 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3475 that the computed length was too small. The symptoms of this bug were
3476 either the PCRE error "internal error: code overflow" from pcre_compile(),
3477 or a glibc crash with a message such as "pcretest: free(): invalid next
3478 size (fast)". Examples of patterns that provoked this bug (shown in
3479 pcretest format) are:
3480
3481 /(?-x: )/x
3482 /(?x)(?-x: \s*#\s*)/
3483 /((?i)[\x{c0}])/8
3484 /(?i:[\x{c0}])/8
3485
3486 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3487 is now done differently.
3488
3489 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3490 wrapper classes; (b) implement a new function in the C++ scanner that is
3491 more efficient than the old way of doing things because it avoids levels of
3492 recursion in the regex matching; (c) add a paragraph to the documentation
3493 for the FullMatch() function.
3494
3495 6. The escape sequence \n was being treated as whatever was defined as
3496 "newline". Not only was this contrary to the documentation, which states
3497 that \n is character 10 (hex 0A), but it also went horribly wrong when
3498 "newline" was defined as CRLF. This has been fixed.
3499
3500 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3501 was being set to -1 for the "end of line" case (supposedly a value that no
3502 character can have). Though this value is never used (the check for end of
3503 line is "zero bytes in current character"), it caused compiler complaints.
3504 I've changed it to 0xffffffff.
3505
3506 8. In pcre_version.c, the version string was being built by a sequence of
3507 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3508 string (as it is for production releases) called a macro with an empty
3509 argument. The C standard says the result of this is undefined. The gcc
3510 compiler treats it as an empty string (which was what was wanted) but it is
3511 reported that Visual C gives an error. The source has been hacked around to
3512 avoid this problem.
3513
3514 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3515 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3516 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3517 of them did).
3518
3519 10. Originally, pcretest opened its input and output without "b"; then I was
3520 told that "b" was needed in some environments, so it was added for release
3521 5.0 to both the input and output. (It makes no difference on Unix-like
3522 systems.) Later I was told that it is wrong for the input on Windows. I've
3523 now abstracted the modes into two macros, to make it easier to fiddle with
3524 them, and removed "b" from the input mode under Windows.
3525
3526 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3527
3528 12. Added -help and --help to pcretest as an official way of being reminded
3529 of the options.
3530
3531 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3532 and pcrecpp.cc because they annoy compilers at high warning levels.
3533
3534 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3535
3536 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3537 scripts are not C programs :-) and which was not noticed because it works
3538 on Linux.
3539
3540 16. pcretest is supposed to handle any length of pattern and data line (as one
3541 line or as a continued sequence of lines) by extending its input buffer if
3542 necessary. This feature was broken for very long pattern lines, leading to
3543 a string of junk being passed to pcre_compile() if the pattern was longer
3544 than about 50K.
3545
3546 17. I have done a major re-factoring of the way pcre_compile() computes the
3547 amount of memory needed for a compiled pattern. Previously, there was code
3548 that made a preliminary scan of the pattern in order to do this. That was
3549 OK when PCRE was new, but as the facilities have expanded, it has become
3550 harder and harder to keep it in step with the real compile phase, and there
3551 have been a number of bugs (see for example, 4 above). I have now found a
3552 cunning way of running the real compile function in a "fake" mode that
3553 enables it to compute how much memory it would need, while actually only
3554 ever using a few hundred bytes of working memory and without too many
3555 tests of the mode. This should make future maintenance and development
3556 easier. A side effect of this work is that the limit of 200 on the nesting
3557 depth of parentheses has been removed (though this was never a serious
3558 limitation, I suspect). However, there is a downside: pcre_compile() now
3559 runs more slowly than before (30% or more, depending on the pattern). I
3560 hope this isn't a big issue. There is no effect on runtime performance.
3561
3562 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3563 newline (only possible for the last line of a file) and it was a
3564 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3565
3566 19. Added additional timing features to pcretest. (1) The -tm option now times
3567 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3568 separate command line item, by a number that specifies the number of
3569 repeats to use when timing. The default is 50000; this gives better
3570 precision, but takes uncomfortably long for very large patterns.
3571
3572 20. Extended pcre_study() to be more clever in cases where a branch of a
3573 subpattern has no definite first character. For example, (a*|b*)[cd] would
3574 previously give no result from pcre_study(). Now it recognizes that the
3575 first character must be a, b, c, or d.
3576
3577 21. There was an incorrect error "recursive call could loop indefinitely" if
3578 a subpattern (or the entire pattern) that was being tested for matching an
3579 empty string contained only one non-empty item after a nested subpattern.
3580 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3581 incorrectly, because the \d was being skipped in the check.
3582
3583 22. The pcretest program now has a new pattern option /B and a command line
3584 option -b, which is equivalent to adding /B to every pattern. This causes
3585 it to show the compiled bytecode, without the additional information that
3586 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3587 is the same as /B/I).
3588
3589 23. A new optimization is now able automatically to treat some sequences such
3590 as a*b as a*+b. More specifically, if something simple (such as a character
3591 or a simple class like \d) has an unlimited quantifier, and is followed by
3592 something that cannot possibly match the quantified thing, the quantifier
3593 is automatically "possessified".
3594
3595 24. A recursive reference to a subpattern whose number was greater than 39
3596 went wrong under certain circumstances in UTF-8 mode. This bug could also
3597 have affected the operation of pcre_study().
3598
3599 25. Realized that a little bit of performance could be had by replacing
3600 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3601
3602 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3603
3604 27. Possessive quantifiers such as a++ were previously implemented by turning
3605 them into atomic groups such as ($>a+). Now they have their own opcodes,
3606 which improves performance. This includes the automatically created ones
3607 from 23 above.
3608
3609 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3610 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3611 the first matched character to be a colon. This applied both to named and
3612 numbered groups.
3613
3614 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3615
3616 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3617 building PCRE on Windows easier, so I have included it in the distribution.
3618
3619 31. There is now a check in pcretest against a ridiculously large number being
3620 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3621 loop, the loop is abandoned.
3622
3623 32. Forward references to subpatterns in conditions such as (?(2)...) where
3624 subpattern 2 is defined later cause pcre_compile() to search forwards in
3625 the pattern for the relevant set of parentheses. This search went wrong
3626 when there were unescaped parentheses in a character class, parentheses
3627 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3628
3629 33. "Subroutine" calls and backreferences were previously restricted to
3630 referencing subpatterns earlier in the regex. This restriction has now
3631 been removed.
3632
3633 34. Added a number of extra features that are going to be in Perl 5.10. On the
3634 whole, these are just syntactic alternatives for features that PCRE had
3635 previously implemented using the Python syntax or my own invention. The
3636 other formats are all retained for compatibility.
3637
3638 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3639 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3640 also .NET compatible.
3641
3642 (b) A recursion or subroutine call to a named group can now be defined as
3643 (?&name) as well as (?P>name).
3644
3645 (c) A backreference to a named group can now be defined as \k<name> or
3646 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3647 5.10, are also .NET compatible.
3648
3649 (d) A conditional reference to a named group can now use the syntax
3650 (?(<name>) or (?('name') as well as (?(name).
3651
3652 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3653 groups (named and numbered) that are never evaluated inline, but can be
3654 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3655 is always false. There may be only one alternative in such a group.
3656
3657 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3658 as the simple (?(R). The condition is true only if the most recent
3659 recursion is that of the given number or name. It does not search out
3660 through the entire recursion stack.
3661
3662 (g) The escape \gN or \g{N} has been added, where N is a positive or
3663 negative number, specifying an absolute or relative reference.
3664
3665 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3666 some "unreachable code" warnings.
3667
3668 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3669 things, this adds five new scripts.
3670
3671 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3672 There were also incompatibilities regarding the handling of \Q..\E inside
3673 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3674 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3675
3676 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3677 matches an empty string, and forcibly breaks the loop. There were bugs in
3678 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3679 against aaaa the result was just "a" rather than "aaaa", for example. Two
3680 separate and independent bugs (that affected different cases) have been
3681 fixed.
3682
3683 39. Refactored the code to abolish the use of different opcodes for small
3684 capturing bracket numbers. This is a tidy that I avoided doing when I
3685 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3686 The new approach is not only tidier, it makes it possible to reduce the
3687 memory needed to fix the previous bug (38).
3688
3689 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3690 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3691 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3692 mode.
3693
3694 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3695 report.
3696
3697 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3698 copy construction and assignment in the C++ wrapper.
3699
3700 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3701 couple of bugs that could have given wrong results in the "--newline=crlf"
3702 case.
3703
3704 44. Added a number of casts and did some reorganization of signed/unsigned int
3705 variables following suggestions from Dair Grant. Also renamed the variable
3706 "this" as "item" because it is a C++ keyword.
3707
3708 45. Arranged for dftables to add
3709
3710 #include "pcre_internal.h"
3711
3712 to pcre_chartables.c because without it, gcc 4.x may remove the array
3713 definition from the final binary if PCRE is built into a static library and
3714 dead code stripping is activated.
3715
3716 46. For an unanchored pattern, if a match attempt fails at the start of a
3717 newline sequence, and the newline setting is CRLF or ANY, and the next two
3718 characters are CRLF, advance by two characters instead of one.
3719
3720
3721 Version 6.7 04-Jul-06
3722 ---------------------
3723
3724 1. In order to handle tests when input lines are enormously long, pcretest has
3725 been re-factored so that it automatically extends its buffers when
3726 necessary. The code is crude, but this _is_ just a test program. The
3727 default size has been increased from 32K to 50K.
3728
3729 2. The code in pcre_study() was using the value of the re argument before
3730 testing it for NULL. (Of course, in any sensible call of the function, it
3731 won't be NULL.)
3732
3733 3. The memmove() emulation function in pcre_internal.h, which is used on
3734 systems that lack both memmove() and bcopy() - that is, hardly ever -
3735 was missing a "static" storage class specifier.
3736
3737 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3738 containing an extended class (one that cannot be represented by a bitmap
3739 because it contains high-valued characters or Unicode property items, e.g.
3740 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3741 pattern, but PCRE should not loop if you do not (it no longer does).
3742 [Detail: two cases were found: (a) a repeated subpattern containing an
3743 extended class; (b) a recursive reference to a subpattern that followed a
3744 previous extended class. It wasn't skipping over the extended class
3745 correctly when UTF-8 mode was not set.]
3746
3747 5. A negated single-character class was not being recognized as fixed-length
3748 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3749 compile error "lookbehind assertion is not fixed length".
3750
3751 6. The RunPerlTest auxiliary script was showing an unexpected difference
3752 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3753 write a Perl script that can interpret lines of an input file either as
3754 byte characters or as UTF-8, which is what "perltest" was being required to
3755 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3756 can't do is switch easily at run time between having the "use utf8;" pragma
3757 or not. In the end, I fudged it by using the RunPerlTest script to insert
3758 "use utf8;" explicitly for the UTF-8 tests.
3759
3760 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3761 the end of the subject string, contrary to the documentation and to what
3762 Perl does. This was true of both matching functions. Now it matches only at
3763 the start of the subject and immediately after *internal* newlines.
3764
3765 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3766 a pointer to an int instead of a pointer to an unsigned long int. This
3767 caused problems on 64-bit systems.
3768
3769 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3770 instance of the 'standard' template library not being so standard".
3771
3772 10. There was no check on the number of named subpatterns nor the maximum
3773 length of a subpattern name. The product of these values is used to compute
3774 the size of the memory block for a compiled pattern. By supplying a very
3775 long subpattern name and a large number of named subpatterns, the size
3776 computation could be caused to overflow. This is now prevented by limiting
3777 the length of names to 32 characters, and the number of named subpatterns
3778 to 10,000.
3779
3780 11. Subpatterns that are repeated with specific counts have to be replicated in
3781 the compiled pattern. The size of memory for this was computed from the
3782 length of the subpattern and the repeat count. The latter is limited to
3783 65535, but there was no limit on the former, meaning that integer overflow
3784 could in principle occur. The compiled length of a repeated subpattern is
3785 now limited to 30,000 bytes in order to prevent this.
3786
3787 12. Added the optional facility to have named substrings with the same name.
3788
3789 13. Added the ability to use a named substring as a condition, using the
3790 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3791 are numbers (not recommended). Forward references are permitted.
3792
3793 14. Added forward references in named backreferences (if you see what I mean).
3794
3795 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3796 pattern could run off the end of the subject. For example, the pattern
3797 "(?s)(.{1,5})"8 did this with the subject "ab".
3798
3799 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3800 PCRE_CASELESS was set when matching characters that were quantified with ?
3801 or *.
3802
3803 17. A character class other than a single negated character that had a minimum
3804 but no maximum quantifier - for example [ab]{6,} - was not handled
3805 correctly by pce_dfa_exec(). It would match only one character.
3806
3807 18. A valid (though odd) pattern that looked like a POSIX character
3808 class but used an invalid character after [ (for example [[,abc,]]) caused
3809 pcre_compile() to give the error "Failed: internal error: code overflow" or
3810 in some cases to crash with a glibc free() error. This could even happen if
3811 the pattern terminated after [[ but there just happened to be a sequence of
3812 letters, a binary zero, and a closing ] in the memory that followed.
3813
3814 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3815 over the years. Originally (before any Unicode support), just the bottom 8
3816 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3817 output from "man perlunicode" includes this:
3818
3819 The regular expression compiler produces polymorphic opcodes. That
3820 is, the pattern adapts to the data and automatically switches to
3821 the Unicode character scheme when presented with Unicode data--or
3822 instead uses a traditional byte scheme when presented with byte
3823 data.
3824
3825 Sadly, a wide octal escape does not cause a switch, and in a string with
3826 no other multibyte characters, these octal escapes are treated as before.
3827 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3828 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3829 Unicode string.
3830
3831 I have not perpetrated such confusion in PCRE. Up till now, it took just
3832 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3833 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3834 translate to the appropriate multibyte character.
3835
3836 29. Applied some refactoring to reduce the number of warnings from Microsoft
3837 and Borland compilers. This has included removing the fudge introduced
3838 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3839 a warning about an unused variable.
3840
3841 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3842 characters since release 4.0, because Perl (from release 5.004) does not.
3843 [Or at least, is documented not to: some releases seem to be in conflict
3844 with the documentation.] However, when a pattern was studied with
3845 pcre_study() and all its branches started with \s, PCRE still included VT
3846 as a possible starting character. Of course, this did no harm; it just
3847 caused an unnecessary match attempt.
3848
3849 22. Removed a now-redundant internal flag bit that recorded the fact that case
3850 dependency changed within the pattern. This was once needed for "required
3851 byte" processing, but is no longer used. This recovers a now-scarce options
3852 bit. Also moved the least significant internal flag bit to the most-
3853 significant bit of the word, which was not previously used (hangover from
3854 the days when it was an int rather than a uint) to free up another bit for
3855 the future.
3856
3857 23. Added support for CRLF line endings as well as CR and LF. As well as the
3858 default being selectable at build time, it can now be changed at runtime
3859 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3860 specify that it is scanning data with non-default line endings.
3861
3862 24. Changed the definition of CXXLINK to make it agree with the definition of
3863 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3864
3865 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3866 recursions. This makes a big different to stack usage for some patterns.
3867
3868 26. If a subpattern containing a named recursion or subroutine reference such
3869 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3870 the space required for the compiled pattern went wrong and gave too small a
3871 value. Depending on the environment, this could lead to "Failed: internal
3872 error: code overflow at offset 49" or "glibc detected double free or
3873 corruption" errors.
3874
3875 27. Applied patches from Google (a) to support the new newline modes and (b) to
3876 advance over multibyte UTF-8 characters in GlobalReplace.
3877
3878 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3879 difference for some implementation of PCRE in some Windows version.
3880
3881 29. Added some extra testing facilities to pcretest:
3882
3883 \q<number> in a data line sets the "match limit" value
3884 \Q<number> in a data line sets the "match recursion limt" value
3885 -S <number> sets the stack size, where <number> is in megabytes
3886
3887 The -S option isn't available for Windows.
3888
3889
3890 Version 6.6 06-Feb-06
3891 ---------------------
3892
3893 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3894 in pcreposix.h. I have copied the definition from pcre.h.
3895
3896 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3897 because pcre.h is no longer a built file.
3898
3899 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3900 not normally included in the compiled code.
3901
3902
3903 Version 6.5 01-Feb-06
3904 ---------------------
3905
3906 1. When using the partial match feature with pcre_dfa_exec(), it was not
3907 anchoring the second and subsequent partial matches at the new starting
3908 point. This could lead to incorrect results. For example, with the pattern
3909 /1234/, partially matching against "123" and then "a4" gave a match.
3910
3911 2. Changes to pcregrep:
3912
3913 (a) All non-match returns from pcre_exec() were being treated as failures
3914 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3915 error message is output. Some extra information is given for the
3916 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3917 probably the only errors that are likely to be caused by users (by
3918 specifying a regex that has nested indefinite repeats, for instance).
3919 If there are more than 20 of these errors, pcregrep is abandoned.
3920
3921 (b) A binary zero was treated as data while matching, but terminated the
3922 output line if it was written out. This has been fixed: binary zeroes
3923 are now no different to any other data bytes.
3924
3925 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3926 used to set a locale for matching. The --locale=xxxx long option has
3927 been added (no short equivalent) to specify a locale explicitly on the
3928 pcregrep command, overriding the environment variables.
3929
3930 (d) When -B was used with -n, some line numbers in the output were one less
3931 than they should have been.
3932
3933 (e) Added the -o (--only-matching) option.
3934
3935 (f) If -A or -C was used with -c (count only), some lines of context were
3936 accidentally printed for the final match.
3937
3938 (g) Added the -H (--with-filename) option.
3939
3940 (h) The combination of options -rh failed to suppress file names for files
3941 that were found from directory arguments.
3942
3943 (i) Added the -D (--devices) and -d (--directories) options.
3944
3945 (j) Added the -F (--fixed-strings) option.
3946
3947 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3948
3949 (l) Added the --colo(u)r option.
3950
3951 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3952 is not present by default.
3953
3954 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3955 items such as (?R) or (?1), when the recursion could match a number of
3956 alternatives. If it matched one of the alternatives, but subsequently,
3957 outside the recursion, there was a failure, the code tried to back up into
3958 the recursion. However, because of the way PCRE is implemented, this is not
3959 possible, and the result was an incorrect result from the match.
3960
3961 In order to prevent this happening, the specification of recursion has
3962 been changed so that all such subpatterns are automatically treated as
3963 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3964
3965 4. I had overlooked the fact that, in some locales, there are characters for
3966 which isalpha() is true but neither isupper() nor islower() are true. In
3967 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3968 and ordfeminine) are like this. This affected the treatment of \w and \W
3969 when they appeared in character classes, but not when they appeared outside
3970 a character class. The bit map for "word" characters is now created
3971 separately from the results of isalnum() instead of just taking it from the
3972 upper, lower, and digit maps. (Plus the underscore character, of course.)
3973
3974 5. The above bug also affected the handling of POSIX character classes such as
3975 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3976 permanent tables. Instead, the bit maps for such a class were previously
3977 created as the appropriate unions of the upper, lower, and digit bitmaps.
3978 Now they are created by subtraction from the [[:word:]] class, which has
3979 its own bitmap.
3980
3981 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3982 It is created by subtracting the vertical space characters (\x09, \x0a,
3983 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3984 subtraction was done in the overall bitmap for a character class, meaning
3985 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3986 be recognized. This bug has been fixed.
3987
3988 7. Patches from the folks at Google:
3989
3990 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3991 real life, but is still worth protecting against".
3992
3993 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3994 regular expressions".
3995
3996 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3997 have it.
3998
3999 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
4000 "configure" and the latter not, in order to fix a problem somebody had
4001 with compiling the Arg class on HP-UX.
4002
4003 (e) Improve the error-handling of the C++ wrapper a little bit.
4004
4005 (f) New tests for checking recursion limiting.
4006
4007 8. The pcre_memmove() function, which is used only if the environment does not
4008 have a standard memmove() function (and is therefore rarely compiled),
4009 contained two bugs: (a) use of int instead of size_t, and (b) it was not
4010 returning a result (though PCRE never actually uses the result).
4011
4012 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
4013 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
4014 returned instead of calling malloc() with an overflowing number that would
4015 most likely cause subsequent chaos.
4016
4017 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
4018
4019 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
4020 with this option is matched, the nmatch and pmatch options of regexec() are
4021 ignored.
4022
4023 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
4024 provided in case anyone wants to the the POSIX interface with UTF-8
4025 strings.
4026
4027 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
4028 C++ linking (needed for some HP-UX environments).
4029
4030 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
4031 (unused parameter) and in the pcre_printint() function (omitted "default"
4032 switch label when the default is to do nothing).
4033
4034 15. Added some code to make it possible, when PCRE is compiled as a C++
4035 library, to replace subject pointers for pcre_exec() with a smart pointer
4036 class, thus making it possible to process discontinuous strings.
4037
4038 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
4039 much the same function. They were added by different people who were trying
4040 to make PCRE easy to compile on non-Unix systems. It has been suggested
4041 that PCRE_EXPORT be abolished now that there is more automatic apparatus
4042 for compiling on Windows systems. I have therefore replaced it with
4043 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
4044 defaults to "extern" for C or "extern C" for C++, which works fine on
4045 Unix-like systems. It is now possible to override the value of PCRE_DATA_
4046 SCOPE with something explicit in config.h. In addition:
4047
4048 (a) pcreposix.h still had just "extern" instead of either of these macros;
4049 I have replaced it with PCRE_DATA_SCOPE.
4050
4051 (b) Functions such as _pcre_xclass(), which are internal to the library,
4052 but external in the C sense, all had PCRE_EXPORT in their definitions.
4053 This is apparently wrong for the Windows case, so I have removed it.
4054 (It makes no difference on Unix-like systems.)
4055
4056 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
4057 of recursive calls to match(). This is different to MATCH_LIMIT because
4058 that limits the total number of calls to match(), not all of which increase
4059 the depth of recursion. Limiting the recursion depth limits the amount of
4060 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4061 when PCRE is compiled, and changed at run time. A patch from Google adds
4062 this functionality to the C++ interface.
4063
4064 18. Changes to the handling of Unicode character properties:
4065
4066 (a) Updated the table to Unicode 4.1.0.
4067
4068 (b) Recognize characters that are not in the table as "Cn" (undefined).
4069
4070 (c) I revised the way the table is implemented to a much improved format
4071 which includes recognition of ranges. It now supports the ranges that
4072 are defined in UnicodeData.txt, and it also amalgamates other
4073 characters into ranges. This has reduced the number of entries in the
4074 table from around 16,000 to around 3,000, thus reducing its size
4075 considerably. I realized I did not need to use a tree structure after
4076 all - a binary chop search is just as efficient. Having reduced the
4077 number of entries, I extended their size from 6 bytes to 8 bytes to
4078 allow for more data.
4079
4080 (d) Added support for Unicode script names via properties such as \p{Han}.
4081
4082 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4083 matching that character.
4084
4085 20. When matching a repeated Unicode property with a minimum greater than zero,
4086 (for example \pL{2,}), PCRE could look past the end of the subject if it
4087 reached it while seeking the minimum number of characters. This could
4088 happen only if some of the characters were more than one byte long, because
4089 there is a check for at least the minimum number of bytes.
4090
4091 21. Refactored the implementation of \p and \P so as to be more general, to
4092 allow for more different types of property in future. This has changed the
4093 compiled form incompatibly. Anybody with saved compiled patterns that use
4094 \p or \P will have to recompile them.
4095
4096 22. Added "Any" and "L&" to the supported property types.
4097
4098 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4099 but give a compile time error if the value is greater than 0xff.
4100
4101 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4102 accidentally not being installed or uninstalled.
4103
4104 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4105 made were to insert the current release number. This seemed silly, because
4106 it made things harder for people building PCRE on systems that don't run
4107 "configure". I have turned pcre.h into a distributed file, no longer built
4108 by "configure", with the version identification directly included. There is
4109 no longer a pcre.h.in file.
4110
4111 However, this change necessitated a change to the pcre-config script as
4112 well. It is built from pcre-config.in, and one of the substitutions was the
4113 release number. I have updated configure.ac so that ./configure now finds
4114 the release number by grepping pcre.h.
4115
4116 26. Added the ability to run the tests under valgrind.
4117
4118
4119 Version 6.4 05-Sep-05
4120 ---------------------
4121
4122 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4123 "--" to be printed when multiple files were scanned, even when none of the
4124 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4125 consider it to be a bug, and have restored the previous behaviour.
4126
4127 2. A couple of code tidies to get rid of compiler warnings.
4128
4129 3. The pcretest program used to cheat by referring to symbols in the library
4130 whose names begin with _pcre_. These are internal symbols that are not
4131 really supposed to be visible externally, and in some environments it is
4132 possible to suppress them. The cheating is now confined to including
4133 certain files from the library's source, which is a bit cleaner.
4134
4135 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4136 file's purpose clearer.
4137
4138 5. Reorganized pcre_ucp_findchar().
4139
4140
4141 Version 6.3 15-Aug-05
4142 ---------------------
4143
4144 1. The file libpcre.pc.in did not have general read permission in the tarball.
4145
4146 2. There were some problems when building without C++ support:
4147
4148 (a) If C++ support was not built, "make install" and "make test" still
4149 tried to test it.
4150
4151 (b) There were problems when the value of CXX was explicitly set. Some
4152 changes have been made to try to fix these, and ...
4153
4154 (c) --disable-cpp can now be used to explicitly disable C++ support.
4155
4156 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4157 backslash in a target when C++ was disabled. This confuses some
4158 versions of "make", apparently. Using an intermediate variable solves
4159 this. (Same for CPP_LOBJ.)
4160
4161 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4162 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4163 necessary on certain architectures.
4164
4165 4. Added a setting of -export-symbols-regex to the link command to remove
4166 those symbols that are exported in the C sense, but actually are local
4167 within the library, and not documented. Their names all begin with
4168 "_pcre_". This is not a perfect job, because (a) we have to except some
4169 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4170 available (and never for static libraries). I have made a note to try to
4171 find a way round (a) in the future.
4172
4173
4174 Version 6.2 01-Aug-05
4175 ---------------------
4176
4177 1. There was no test for integer overflow of quantifier values. A construction
4178 such as {1111111111111111} would give undefined results. What is worse, if
4179 a minimum quantifier for a parenthesized subpattern overflowed and became
4180 negative, the calculation of the memory size went wrong. This could have
4181 led to memory overwriting.
4182
4183 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4184
4185 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4186 operating environments where this matters.
4187
4188 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4189 PCRE options from within the C++ wrapper.
4190
4191 5. Named capturing subpatterns were not being correctly counted when a pattern
4192 was compiled. This caused two problems: (a) If there were more than 100
4193 such subpatterns, the calculation of the memory needed for the whole
4194 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4195 back references of the form \12, where the number was greater than 9, were
4196 not recognized as back references, even though there were sufficient
4197 previous subpatterns.
4198
4199 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4200 versions of gcc, e.g. 2.95.4.
4201
4202
4203 Version 6.1 21-Jun-05
4204 ---------------------
4205
4206 1. There was one reference to the variable "posix" in pcretest.c that was not
4207 surrounded by "#if !defined NOPOSIX".
4208
4209 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4210 the cross-check on the old pcre_info() function, for the benefit of the
4211 cut-down version of PCRE that is currently imported into Exim.
4212
4213 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4214 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4215 patterns that start (?i) but I don't think that matters. The use of (?i) is
4216 just an example; this all applies to the other options as well.
4217
4218 4. Since libtool seems to echo the compile commands it is issuing, the output
4219 from "make" can be reduced a bit by putting "@" in front of each libtool
4220 compile command.
4221
4222 5. Patch from the folks at Google for configure.in to be a bit more thorough
4223 in checking for a suitable C++ installation before trying to compile the
4224 C++ stuff. This should fix a reported problem when a compiler was present,
4225 but no suitable headers.
4226
4227 6. The man pages all had just "PCRE" as their title. I have changed them to
4228 be the relevant file name. I have also arranged that these names are
4229 retained in the file doc/pcre.txt, which is a concatenation in text format
4230 of all the man pages except the little individual ones for each function.
4231
4232 7. The NON-UNIX-USE file had not been updated for the different set of source
4233 files that come with release 6. I also added a few comments about the C++
4234 wrapper.
4235
4236
4237 Version 6.0 07-Jun-05
4238 ---------------------
4239
4240 1. Some minor internal re-organization to help with my DFA experiments.
4241
4242 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4243 didn't matter for the library itself when fully configured, but did matter
4244 when compiling without UCP support, or within Exim, where the ucp files are
4245 not imported.
4246
4247 3. Refactoring of the library code to split up the various functions into
4248 different source modules. The addition of the new DFA matching code (see
4249 below) to a single monolithic source would have made it really too
4250 unwieldy, quite apart from causing all the code to be include in a
4251 statically linked application, when only some functions are used. This is
4252 relevant even without the DFA addition now that patterns can be compiled in
4253 one application and matched in another.
4254
4255 The downside of splitting up is that there have to be some external
4256 functions and data tables that are used internally in different modules of
4257 the library but which are not part of the API. These have all had their
4258 names changed to start with "_pcre_" so that they are unlikely to clash
4259 with other external names.
4260
4261 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4262 a different (DFA) algorithm. Although it is slower than the original
4263 function, it does have some advantages for certain types of matching
4264 problem.
4265
4266 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4267 including restarting after a partial match.
4268
4269 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4270 defined when compiling for Windows was sent to me. I have put it into the
4271 code, though I have no means of testing or verifying it.
4272
4273 7. Added the pcre_refcount() auxiliary function.
4274
4275 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4276 match before or at the first newline in the subject string. In pcretest,
4277 the /f option on a pattern can be used to set this.
4278
4279 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4280 would behave wrongly. This has been present in PCRE since release 4.0.
4281
4282 10. A number of changes to the pcregrep command:
4283
4284 (a) Refactored how -x works; insert ^(...)$ instead of setting
4285 PCRE_ANCHORED and checking the length, in preparation for adding
4286 something similar for -w.
4287
4288 (b) Added the -w (match as a word) option.
4289
4290 (c) Refactored the way lines are read and buffered so as to have more
4291 than one at a time available.
4292
4293 (d) Implemented a pcregrep test script.
4294
4295 (e) Added the -M (multiline match) option. This allows patterns to match
4296 over several lines of the subject. The buffering ensures that at least
4297 8K, or the rest of the document (whichever is the shorter) is available
4298 for matching (and similarly the previous 8K for lookbehind assertions).
4299
4300 (f) Changed the --help output so that it now says
4301
4302 -w, --word-regex(p)
4303
4304 instead of two lines, one with "regex" and the other with "regexp"
4305 because that confused at least one person since the short forms are the
4306 same. (This required a bit of code, as the output is generated
4307 automatically from a table. It wasn't just a text change.)
4308
4309 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4310 option but starts with a hyphen. Could be a pattern or a path name
4311 starting with a hyphen, for instance.
4312
4313 (h) "-" can be given as a file name to represent stdin.
4314
4315 (i) When file names are being printed, "(standard input)" is used for
4316 the standard input, for compatibility with GNU grep. Previously
4317 "<stdin>" was used.
4318
4319 (j) The option --label=xxx can be used to supply a name to be used for
4320 stdin when file names are being printed. There is no short form.
4321
4322 (k) Re-factored the options decoding logic because we are going to add
4323 two more options that take data. Such options can now be given in four
4324 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4325
4326 (l) Added the -A, -B, and -C options for requesting that lines of context
4327 around matches be printed.
4328
4329 (m) Added the -L option to print the names of files that do not contain
4330 any matching lines, that is, the complement of -l.
4331
4332 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4333 continue to scan other files.
4334
4335 (o) The -s option was incorrectly implemented. For compatibility with other
4336 greps, it now suppresses the error message for a non-existent or non-
4337 accessible file (but not the return code). There is a new option called
4338 -q that suppresses the output of matching lines, which was what -s was
4339 previously doing.
4340
4341 (p) Added --include and --exclude options to specify files for inclusion
4342 and exclusion when recursing.
4343
4344 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4345 Hopefully, it now does.
4346
4347 12. Missing cast in pcre_study().
4348
4349 13. Added an "uninstall" target to the makefile.
4350
4351 14. Replaced "extern" in the function prototypes in Makefile.in with
4352 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4353 world, but is set differently for Windows.
4354
4355 15. Added a second compiling function called pcre_compile2(). The only
4356 difference is that it has an extra argument, which is a pointer to an
4357 integer error code. When there is a compile-time failure, this is set
4358 non-zero, in addition to the error test pointer being set to point to an
4359 error message. The new argument may be NULL if no error number is required
4360 (but then you may as well call pcre_compile(), which is now just a
4361 wrapper). This facility is provided because some applications need a
4362 numeric error indication, but it has also enabled me to tidy up the way
4363 compile-time errors are handled in the POSIX wrapper.
4364
4365 16. Added VPATH=.libs to the makefile; this should help when building with one
4366 prefix path and installing with another. (Or so I'm told by someone who
4367 knows more about this stuff than I do.)
4368
4369 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4370 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4371 match everything, including newlines. This is not POSIX-compatible, but
4372 somebody wanted the feature. From pcretest it can be activated by using
4373 both the P and the s flags.
4374
4375 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4376
4377 19. libpcre.pc was being incorrectly installed as executable.
4378
4379 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4380 it now also looks for '\r' so that it will work unmodified on Windows.
4381
4382 21. Added Google's contributed C++ wrapper to the distribution.
4383
4384 22. Added some untidy missing memory free() calls in pcretest, to keep
4385 Electric Fence happy when testing.
4386
4387
4388
4389 Version 5.0 13-Sep-04
4390 ---------------------
4391
4392 1. Internal change: literal characters are no longer packed up into items
4393 containing multiple characters in a single byte-string. Each character
4394 is now matched using a separate opcode. However, there may be more than one
4395 byte in the character in UTF-8 mode.
4396
4397 2. The pcre_callout_block structure has two new fields: pattern_position and
4398 next_item_length. These contain the offset in the pattern to the next match
4399 item, and its length, respectively.
4400
4401 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4402 insertion of callouts before each pattern item. Added the /C option to
4403 pcretest to make use of this.
4404
4405 4. On the advice of a Windows user, the lines
4406
4407 #if defined(_WIN32) || defined(WIN32)
4408 _setmode( _fileno( stdout ), 0x8000 );
4409 #endif /* defined(_WIN32) || defined(WIN32) */
4410
4411 have been added to the source of pcretest. This apparently does useful
4412 magic in relation to line terminators.
4413
4414 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4415 for the benefit of those environments where the "b" makes a difference.
4416
4417 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4418 to know about it. I have put a hack into configure.in that adds in code
4419 to set GCC=yes if CC=icc. This seems to end up at a point in the
4420 generated configure script that is early enough to affect the setting of
4421 compiler options, which is what is needed, but I have no means of testing
4422 whether it really works. (The user who reported this had patched the
4423 generated configure script, which of course I cannot do.)
4424
4425 LATER: After change 22 below (new libtool files), the configure script
4426 seems to know about icc (and also ecc). Therefore, I have commented out
4427 this hack in configure.in.
4428
4429 7. Added support for pkg-config (2 patches were sent in).
4430
4431 8. Negated POSIX character classes that used a combination of internal tables
4432 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4433 [[:^ascii]]. Typically, they would match almost any characters. The other
4434 POSIX classes were not broken in this way.
4435
4436 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4437 to find the match, as PCRE was deluded into thinking that the match had to
4438 start at the start point or following a newline. The same bug applied to
4439 patterns with negative forward assertions or any backward assertions
4440 preceding ".*" at the start, unless the pattern required a fixed first
4441 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4442
4443 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4444 starting at the last subject character, bytes beyond the end of the subject
4445 string were read.
4446
4447 11. Renamed the variable "class" as "classbits" to make life easier for C++
4448 users. (Previously there was a macro definition, but it apparently wasn't
4449 enough.)
4450
4451 12. Added the new field "tables" to the extra data so that tables can be passed
4452 in at exec time, or the internal tables can be re-selected. This allows
4453 a compiled regex to be saved and re-used at a later time by a different
4454 program that might have everything at different addresses.
4455
4456 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4457 -R library as well as a -L library.
4458
4459 14. The debugging options of pcretest (-d on the command line or D on a
4460 pattern) showed incorrect output for anything following an extended class
4461 that contained multibyte characters and which was followed by a quantifier.
4462
4463 15. Added optional support for general category Unicode character properties
4464 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4465 support. It adds about 90K to the size of the library. The meanings of the
4466 inbuilt class escapes such as \d and \s have NOT been changed.
4467
4468 16. Updated pcredemo.c to include calls to free() to release the memory for the
4469 compiled pattern.
4470
4471 17. The generated file chartables.c was being created in the source directory
4472 instead of in the building directory. This caused the build to fail if the
4473 source directory was different from the building directory, and was
4474 read-only.
4475
4476 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4477 file. No doubt somebody will tell me if they don't make sense... Also added
4478 Dan Mooney's comments about building on OpenVMS.
4479
4480 19. Added support for partial matching via the PCRE_PARTIAL option for
4481 pcre_exec() and the \P data escape in pcretest.
4482
4483 20. Extended pcretest with 3 new pattern features:
4484
4485 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4486 write the compiled pattern to the file whose name is "rest-of-line".
4487 This is a straight binary dump of the data, with the saved pointer to
4488 the character tables forced to be NULL. The study data, if any, is
4489 written too. After writing, pcretest reads a new pattern.
4490
4491 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4492 compiled pattern from the given file. There must not be any
4493 occurrences of "<" in the file name (pretty unlikely); if there are,
4494 pcretest will instead treat the initial "<" as a pattern delimiter.
4495 After reading in the pattern, pcretest goes on to read data lines as
4496 usual.
4497
4498 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4499 and 16-bit fields in a compiled pattern, to simulate a pattern that
4500 was compiled on a host of opposite endianness.
4501
4502 21. The pcre-exec() function can now cope with patterns that were compiled on
4503 hosts of opposite endianness, with this restriction:
4504
4505 As for any compiled expression that is saved and used later, the tables
4506 pointer field cannot be preserved; the extra_data field in the arguments
4507 to pcre_exec() should be used to pass in a tables address if a value
4508 other than the default internal tables were used at compile time.
4509
4510 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4511 now diagnosed as an error. Previously, most of the time, a negative number
4512 would have been treated as zero, but if in addition "ovector" was passed as
4513 NULL, a crash could occur.
4514
4515 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4516 new versions from the libtool 1.5 distribution (the last one is a copy of
4517 a file called libtool.m4). This seems to have fixed the need to patch
4518 "configure" to support Darwin 1.3 (which I used to do). However, I still
4519 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4520 workstation).
4521
4522 24. Changed the PCRE licence to be the more standard "BSD" licence.
4523
4524
4525 Version 4.5 01-Dec-03
4526 ---------------------
4527
4528 1. There has been some re-arrangement of the code for the match() function so
4529 that it can be compiled in a version that does not call itself recursively.
4530 Instead, it keeps those local variables that need separate instances for
4531 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4532 needs to "recurse". Keeping track of where control must go is done by means
4533 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4534 hide most of the details from the main code, and operates only if
4535 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4536 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4537 operating.
4538
4539 To make it easier for callers to provide specially tailored get/free
4540 functions for this usage, two new functions, pcre_stack_malloc, and
4541 pcre_stack_free, are used. They are always called in strict stacking order,
4542 and the size of block requested is always the same.
4543
4544 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4545 PCRE has been compiled to use the stack or the heap for recursion. The
4546 -C option of pcretest uses this to show which version is compiled.
4547
4548 A new data escape \S, is added to pcretest; it causes the amounts of store
4549 obtained and freed by both kinds of malloc/free at match time to be added
4550 to the output.
4551
4552 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4553 what's available on my current Linux desktop machine.
4554
4555 3. When matching a UTF-8 string, the test for a valid string at the start has
4556 been extended. If start_offset is not zero, PCRE now checks that it points
4557 to a byte that is the start of a UTF-8 character. If not, it returns
4558 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4559 this is necessary because there may be backward assertions in the pattern.
4560 When matching the same subject several times, it may save resources to use
4561 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4562
4563 4. The code for checking the validity of UTF-8 strings has been tightened so
4564 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4565 containing "overlong sequences".
4566
4567 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4568 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4569 should have been "&", but it just so happened that all the cases this let
4570 through by mistake were picked up later in the function.
4571
4572 6. I had used a variable called "isblank" - this is a C99 function, causing
4573 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4574
4575 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4576 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4577 so the version is shown; (c) stop "make test" echoing "./RunTest".
4578
4579 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4580
4581 9. The prototype for memmove() for systems that don't have it was using
4582 size_t, but the inclusion of the header that defines size_t was later. I've
4583 moved the #includes for the C headers earlier to avoid this.
4584
4585 10. Added some adjustments to the code to make it easier to compiler on certain
4586 special systems:
4587
4588 (a) Some "const" qualifiers were missing.
4589 (b) Added the macro EXPORT before all exported functions; by default this
4590 is defined to be empty.
4591 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4592 that it reads its output file name as an argument instead of writing
4593 to the standard output and assuming this can be redirected.
4594
4595 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4596 class containing characters with values greater than 255, PCRE compilation
4597 went into a loop.
4598
4599 12. A recursive reference to a subpattern that was within another subpattern
4600 that had a minimum quantifier of zero caused PCRE to crash. For example,
4601 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4602 recursion. If the recursively-called subpattern itself had a zero repeat,
4603 that was OK.
4604
4605 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4606 buffer into which it was copied (for escape processing) was still set at
4607 1024, so long lines caused crashes.
4608
4609 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4610 "internal error: code overflow...". This applied to any character class
4611 that was followed by a possessive quantifier.
4612
4613 15. Modified the Makefile to add libpcre.la as a prerequisite for
4614 libpcreposix.la because I was told this is needed for a parallel build to
4615 work.
4616
4617 16. If a pattern that contained .* following optional items at the start was
4618 studied, the wrong optimizing data was generated, leading to matching
4619 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4620 matching string must start with a or b or c. The correct conclusion for
4621 this pattern is that a match can start with any character.
4622
4623
4624 Version 4.4 13-Aug-03
4625 ---------------------
4626
4627 1. In UTF-8 mode, a character class containing characters with values between
4628 127 and 255 was not handled correctly if the compiled pattern was studied.
4629 In fixing this, I have also improved the studying algorithm for such
4630 classes (slightly).
4631
4632 2. Three internal functions had redundant arguments passed to them. Removal
4633 might give a very teeny performance improvement.
4634
4635 3. Documentation bug: the value of the capture_top field in a callout is *one
4636 more than* the number of the hightest numbered captured substring.
4637
4638 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4639 in incorrectly linking with a previously installed version. They now link
4640 explicitly with libpcre.la.
4641
4642 5. configure.in no longer needs to recognize Cygwin specially.
4643
4644 6. A problem in pcre.in for Windows platforms is fixed.
4645
4646 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4647 pcretest, it used to include the size of the study block as part of its
4648 output. Unfortunately, the structure contains a field that has a different
4649 size on different hardware architectures. This meant that the tests that
4650 showed this size failed. As the block is currently always of a fixed size,
4651 this information isn't actually particularly useful in pcretest output, so
4652 I have just removed it.
4653
4654 8. Three pre-processor statements accidentally did not start in column 1.
4655 Sadly, there are *still* compilers around that complain, even though
4656 standard C has not required this for well over a decade. Sigh.
4657
4658 9. In pcretest, the code for checking callouts passed small integers in the
4659 callout_data field, which is a void * field. However, some picky compilers
4660 complained about the casts involved for this on 64-bit systems. Now
4661 pcretest passes the address of the small integer instead, which should get
4662 rid of the warnings.
4663
4664 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4665 both compile and run time, and gives an error if an invalid UTF-8 sequence
4666 is found. There is a option for disabling this check in cases where the
4667 string is known to be correct and/or the maximum performance is wanted.
4668
4669 11. In response to a bug report, I changed one line in Makefile.in from
4670
4671 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4672 to
4673 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4674
4675 to look similar to other lines, but I have no way of telling whether this
4676 is the right thing to do, as I do not use Windows. No doubt I'll get told
4677 if it's wrong...
4678
4679
4680 Version 4.3 21-May-03
4681 ---------------------
4682
4683 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4684 Makefile.
4685
4686 2. Some refactoring to improve the quality of the code:
4687
4688 (i) The utf8_table... variables are now declared "const".
4689
4690 (ii) The code for \cx, which used the "case flipping" table to upper case
4691 lower case letters, now just substracts 32. This is ASCII-specific,
4692 but the whole concept of \cx is ASCII-specific, so it seems
4693 reasonable.
4694
4695 (iii) PCRE was using its character types table to recognize decimal and
4696 hexadecimal digits in the pattern. This is silly, because it handles
4697 only 0-9, a-f, and A-F, but the character types table is locale-
4698 specific, which means strange things might happen. A private
4699 table is now used for this - though it costs 256 bytes, a table is
4700 much faster than multiple explicit tests. Of course, the standard
4701 character types table is still used for matching digits in subject
4702 strings against \d.
4703
4704 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4705 ending in _t are). So I've renamed it as ESC_tee.
4706
4707 3. The first argument for regexec() in the POSIX wrapper should have been
4708 defined as "const".
4709
4710 4. Changed pcretest to use malloc() for its buffers so that they can be
4711 Electric Fenced for debugging.
4712
4713 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4714 to read one or more bytes before the start of the subject string. Often this
4715 had no effect on PCRE's behaviour, but in some circumstances it could
4716 provoke a segmentation fault.
4717
4718 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4719 to try to read one or more bytes before the start of the subject string.
4720
4721 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4722 UTF-8 support could misbehave in various ways if the subject string
4723 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4724 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4725 back over UTF-8 characters.)
4726
4727
4728 Version 4.2 14-Apr-03
4729 ---------------------
4730
4731 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4732
4733 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4734 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4735 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4736 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4737 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4738 and BUILD_EXEEXT
4739 Note: automatic setting of the BUILD variables is not yet working
4740 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4741 compile-time but not at link-time
4742 [LINK]: use for linking executables only
4743 make different versions for Windows and non-Windows
4744 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4745 libraries
4746 [LINK_FOR_BUILD]: new variable
4747 [OBJEXT]: use throughout
4748 [EXEEXT]: use throughout
4749 <winshared>: new target
4750 <wininstall>: new target
4751 <dftables.o>: use native compiler
4752 <dftables>: use native linker
4753 <install>: handle Windows platform correctly
4754 <clean>: ditto
4755 <check>: ditto
4756 copy DLL to top builddir before testing
4757
4758 As part of these changes, -no-undefined was removed again. This was reported
4759 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4760 in any case.
4761
4762 3. Some tidies to get rid of compiler warnings:
4763
4764 . In the match_data structure, match_limit was an unsigned long int, whereas
4765 match_call_count was an int. I've made them both unsigned long ints.
4766
4767 . In pcretest the fact that a const uschar * doesn't automatically cast to
4768 a void * provoked a warning.
4769
4770 . Turning on some more compiler warnings threw up some "shadow" variables
4771 and a few more missing casts.
4772
4773 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4774 option, a class that contained a single character with a value between 128
4775 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4776
4777 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4778 option, a class that contained several characters, but with at least one
4779 whose value was between 128 and 255 caused PCRE to crash.
4780
4781
4782 Version 4.1 12-Mar-03
4783 ---------------------
4784
4785 1. Compiling with gcc -pedantic found a couple of places where casts were
4786 needed, and a string in dftables.c that was longer than standard compilers are
4787 required to support.
4788
4789 2. Compiling with Sun's compiler found a few more places where the code could
4790 be tidied up in order to avoid warnings.
4791
4792 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4793 first of these names is deprecated in the latest Autoconf in favour of the name
4794 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4795 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4796 analogy I have changed it to CFLAGS_FOR_BUILD.
4797
4798 4. Added -no-undefined to the linking command in the Makefile, because this is
4799 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4800 linking step for the pcreposix library.
4801
4802 5. PCRE was failing to diagnose the case of two named groups with the same
4803 name.
4804
4805 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4806 literal character that is needed in the subject for a match, and scans along to
4807 ensure that it is present before embarking on the full matching process. This
4808 saves time in cases of nested unlimited repeats that are never going to match.
4809 Problem: the scan can take a lot of time if the subject is very long (e.g.
4810 megabytes), thus penalizing straightforward matches. It is now done only if the
4811 amount of subject to be scanned is less than 1000 bytes.
4812
4813 7. A lesser problem with the same optimization is that it was recording the
4814 first character of an anchored pattern as "needed", thus provoking a search
4815 right along the subject, even when the first match of the pattern was going to
4816 fail. The "needed" character is now not set for anchored patterns, unless it
4817 follows something in the pattern that is of non-fixed length. Thus, it still
4818 fulfils its original purpose of finding quick non-matches in cases of nested
4819 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4820
4821
4822 Version 4.0 17-Feb-03
4823 ---------------------
4824
4825 1. If a comment in an extended regex that started immediately after a meta-item
4826 extended to the end of string, PCRE compiled incorrect data. This could lead to
4827 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4828
4829 2. Moved to autoconf 2.53 and libtool 1.4.2.
4830
4831 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4832 the special perltest8 script is no longer needed - all the tests can be run
4833 from a single perltest script.
4834
4835 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4836 by \s. It has now been removed in PCRE. This means it isn't recognized as
4837 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4838 class [:space:] *does* include VT, thereby creating a mess.
4839
4840 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4841 space and tab.
4842
4843 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4844 its new features into the main test script, reducing the number of scripts.
4845
4846 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4847 were backward compatible, and made the (?i) apply to the whole pattern, as if
4848 /i were given. Now it behaves more logically, and applies the option setting
4849 only to what follows. PCRE has been changed to follow suit. However, if it
4850 finds options settings right at the start of the pattern, it extracts them into
4851 the global options, as before. Thus, they show up in the info data.
4852
4853 8. Added support for the \Q...\E escape sequence. Characters in between are
4854 treated as literals. This is slightly different from Perl in that $ and @ are
4855 also handled as literals inside the quotes. In Perl, they will cause variable
4856 interpolation. Note the following examples:
4857
4858 Pattern PCRE matches Perl matches
4859
4860 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4861 \Qabc\$xyz\E abc\$xyz abc\$xyz
4862 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4863
4864 For compatibility with Perl, \Q...\E sequences are recognized inside character
4865 classes as well as outside them.
4866
4867 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4868 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4869 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4870 signed/unsigned warnings.
4871
4872 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4873 option for pcretest, so I've replaced it by a simple function that does just
4874 that job.
4875
4876 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4877 "pcregrep -".
4878
4879 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4880 Java package. This provides some syntactic sugar for simple cases of what my
4881 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4882 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4883 item, you can use this simplified notation. Note that only makes sense with
4884 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4885 greediness, whatever the setting of the PCRE_UNGREEDY option.
4886
4887 13. A change of greediness default within a pattern was not taking effect at
4888 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4889 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4890 was abstracted outside.
4891
4892 14. PCRE now supports the \G assertion. It is true when the current matching
4893 position is at the start point of the match. This differs from \A when the
4894 starting offset is non-zero. Used with the /g option of pcretest (or similar
4895 code), it works in the same way as it does for Perl's /g option. If all
4896 alternatives of a regex begin with \G, the expression is anchored to the start
4897 match position, and the "anchored" flag is set in the compiled expression.
4898
4899 15. Some bugs concerning the handling of certain option changes within patterns
4900 have been fixed. These applied to options other than (?ims). For example,
4901 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4902 been the other way round. Some of this was related to change 7 above.
4903
4904 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4905 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4906 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4907 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4908
4909 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4910 mode. Unlike ".", it always matches newline, whatever the setting of
4911 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4912 assertions. Perl allows it, but it doesn't (in general) work because it can't
4913 calculate the length of the lookbehind. At least, that's the case for Perl
4914 5.8.0 - I've been told they are going to document that it doesn't work in
4915 future.
4916
4917 18. Added an error diagnosis for escapes that PCRE does not support: these are
4918 \L, \l, \N, \P, \p, \U, \u, and \X.
4919
4920 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4921 reading past the end of the pattern in cases such as /[abcd/.
4922
4923 20. PCRE was getting more memory than necessary for patterns with classes that
4924 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4925
4926 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4927 compiling PCRE for use with Virtual Pascal.
4928
4929 22. Small fix to the Makefile to make it work properly if the build is done
4930 outside the source tree.
4931
4932 23. Added a new extension: a condition to go with recursion. If a conditional
4933 subpattern starts with (?(R) the "true" branch is used if recursion has
4934 happened, whereas the "false" branch is used only at the top level.
4935
4936 24. When there was a very long string of literal characters (over 255 bytes
4937 without UTF support, over 250 bytes with UTF support), the computation of how
4938 much memory was required could be incorrect, leading to segfaults or other
4939 strange effects.
4940
4941 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4942 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4943 there was a subsequent back reference to those brackets. This meant that, for
4944 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4945 possible to check for precisely this case. All we can do is abandon the
4946 optimization if .* occurs inside capturing brackets when there are any back
4947 references whatsoever. (See below for a better fix that came later.)
4948
4949 26. The handling of the optimization for finding the first character of a
4950 non-anchored pattern, and for finding a character that is required later in the
4951 match were failing in some cases. This didn't break the matching; it just
4952 failed to optimize when it could. The way this is done has been re-implemented.
4953
4954 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4955
4956 28. Added a new feature that provides some of the functionality that Perl
4957 provides with (?{...}). The facility is termed a "callout". The way it is done
4958 in PCRE is for the caller to provide an optional function, by setting
4959 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4960 global variable. By default it is unset, which disables all calling out. To get
4961 the function called, the regex must include (?C) at appropriate points. This
4962 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4963 This provides a means of identifying different callout points. When PCRE
4964 reaches such a point in the regex, if pcre_callout has been set, the external
4965 function is called. It is provided with data in a structure called
4966 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4967 matching continues; if it returns a non-zero value, the match at the current
4968 point fails. However, backtracking will occur if possible. [This was changed
4969 later and other features added - see item 49 below.]
4970
4971 29. pcretest is upgraded to test the callout functionality. It provides a
4972 callout function that displays information. By default, it shows the start of
4973 the match and the current position in the text. There are some new data escapes
4974 to vary what happens:
4975
4976 \C+ in addition, show current contents of captured substrings
4977 \C- do not supply a callout function
4978 \C!n return 1 when callout number n is reached
4979 \C!n!m return 1 when callout number n is reached for the mth time
4980
4981 30. If pcregrep was called with the -l option and just a single file name, it
4982 output "<stdin>" if a match was found, instead of the file name.
4983
4984 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4985 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4986 pcre_exec(). This saves a malloc/free per call. The default value of
4987 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4988 when configuring.
4989
4990 32. The default maximum size of a compiled pattern is 64K. There have been a
4991 few cases of people hitting this limit. The code now uses macros to handle the
4992 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4993 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4994 configuring. Tests 2 and 5 work only with 2-byte links because they output
4995 debugging information about compiled patterns.
4996
4997 33. Internal code re-arrangements:
4998
4999 (a) Moved the debugging function for printing out a compiled regex into
5000 its own source file (printint.c) and used #include to pull it into
5001 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
5002 separate copies.
5003
5004 (b) Defined the list of op-code names for debugging as a macro in
5005 internal.h so that it is next to the definition of the opcodes.
5006
5007 (c) Defined a table of op-code lengths for simpler skipping along compiled
5008 code. This is again a macro in internal.h so that it is next to the
5009 definition of the opcodes.
5010
5011 34. Added support for recursive calls to individual subpatterns, along the
5012 lines of Robin Houston's patch (but implemented somewhat differently).
5013
5014 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
5015 allow it to read and process whole directories in Win32. This code was
5016 contributed by Lionel Fourquaux; it has not been tested by me.
5017
5018 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
5019 used to name a group. Names consist of alphanumerics and underscores, and must
5020 be unique. Back references use the syntax (?P=name) and recursive calls use
5021 (?P>name) which is a PCRE extension to the Python extension. Groups still have
5022 numbers. The function pcre_fullinfo() can be used after compilation to extract
5023 a name/number map. There are three relevant calls:
5024
5025 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
5026 PCRE_INFO_NAMECOUNT yields the number of entries
5027 PCRE_INFO_NAMETABLE yields a pointer to the map.
5028
5029 The map is a vector of fixed-size entries. The size of each entry depends on
5030 the length of the longest name used. The first two bytes of each entry are the
5031 group number, most significant byte first. There follows the corresponding
5032 name, zero terminated. The names are in alphabetical order.
5033
5034 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
5035 case instead of 255. Making it the same both with and without UTF-8 support
5036 means that the same test output works with both.
5037
5038 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
5039 calling malloc() with a zero argument.
5040
5041 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
5042 optimization. I've improved things by keeping a bitmap of backreferences with
5043 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
5044 fact referenced, the optimization can be applied. It is unlikely that a
5045 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
5046 the match to follow \n) will appear inside brackets with a number greater than
5047 31, but if it does, any back reference > 31 suppresses the optimization.
5048
5049 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
5050 of disabling numbered capturing parentheses. Any opening parenthesis that is
5051 not followed by ? behaves as if it were followed by ?: but named parentheses
5052 can still be used for capturing (and they will acquire numbers in the usual
5053 way).
5054
5055 41. Redesigned the return codes from the match() function into yes/no/error so
5056 that errors can be passed back from deep inside the nested calls. A malloc
5057 failure while inside a recursive subpattern call now causes the
5058 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
5059
5060 42. It is now possible to set a limit on the number of times the match()
5061 function is called in a call to pcre_exec(). This facility makes it possible to
5062 limit the amount of recursion and backtracking, though not in a directly
5063 obvious way, because the match() function is used in a number of different
5064 circumstances. The count starts from zero for each position in the subject
5065 string (for non-anchored patterns). The default limit is, for compatibility, a
5066 large number, namely 10 000 000. You can change this in two ways:
5067
5068 (a) When configuring PCRE before making, you can use --with-match-limit=n
5069 to set a default value for the compiled library.
5070
5071 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5072 a different value is set. See 45 below.
5073
5074 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5075
5076 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5077 of things that can be changed at compile time. The first argument specifies
5078 what is wanted and the second points to where the information is to be placed.
5079 The current list of available information is:
5080
5081 PCRE_CONFIG_UTF8
5082
5083 The output is an integer that is set to one if UTF-8 support is available;
5084 otherwise it is set to zero.
5085
5086 PCRE_CONFIG_NEWLINE
5087
5088 The output is an integer that it set to the value of the code that is used for
5089 newline. It is either LF (10) or CR (13).
5090
5091 PCRE_CONFIG_LINK_SIZE
5092
5093 The output is an integer that contains the number of bytes used for internal
5094 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5095
5096 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
5097
5098 The output is an integer that contains the threshold above which the POSIX
5099 interface uses malloc() for output vectors. See item 31 above.
5100
5101 PCRE_CONFIG_MATCH_LIMIT
5102
5103 The output is an unsigned integer that contains the default limit of the number
5104 of match() calls in a pcre_exec() execution. See 42 above.
5105
5106 44. pcretest has been upgraded by the addition of the -C option. This causes it
5107 to extract all the available output from the new pcre_config() function, and to
5108 output it. The program then exits immediately.
5109
5110 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5111 order to support additional features. One way would have been to define
5112 pcre_exec2() (for example) with extra arguments, but this would not have been
5113 extensible, and would also have required all calls to the original function to
5114 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5115 is used for passing in "extra" data from pcre_study().
5116
5117 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5118 contains the following fields:
5119
5120 flags a bitmap indicating which of the following fields are set
5121 study_data opaque data from pcre_study()
5122 match_limit a way of specifying a limit on match() calls for a specific
5123 call to pcre_exec()
5124 callout_data data for callouts (see 49 below)
5125
5126 The flag bits are also defined in pcre.h, and are
5127
5128 PCRE_EXTRA_STUDY_DATA
5129 PCRE_EXTRA_MATCH_LIMIT
5130 PCRE_EXTRA_CALLOUT_DATA
5131
5132 The pcre_study() function now returns one of these new pcre_extra blocks, with
5133 the actual study data pointed to by the study_data field, and the
5134 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5135 before. That is, this change is entirely upwards-compatible and requires no
5136 change to existing code.
5137
5138 If you want to pass in additional data to pcre_exec(), you can either place it
5139 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5140 block.
5141
5142 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5143 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5144 times with different match limits, until it finds the minimum value needed for
5145 pcre_exec() to complete. The value is then output. This can be instructive; for
5146 most simple matches the number is quite small, but for pathological cases it
5147 gets very large very quickly.
5148
5149 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5150 returns the size of the data block pointed to by the study_data field in a
5151 pcre_extra block, that is, the value that was passed as the argument to
5152 pcre_malloc() when PCRE was getting memory in which to place the information
5153 created by pcre_study(). The fourth argument should point to a size_t variable.
5154 pcretest has been extended so that this information is shown after a successful
5155 pcre_study() call when information about the compiled regex is being displayed.
5156
5157 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5158 because what follows is always an absolute path. (Later: it turns out that this
5159 is more than cosmetic for MinGW, because it doesn't like empty path
5160 components.)
5161
5162 49. Some changes have been made to the callout feature (see 28 above):
5163
5164 (i) A callout function now has three choices for what it returns:
5165
5166 0 => success, carry on matching
5167 > 0 => failure at this point, but backtrack if possible
5168 < 0 => serious error, return this value from pcre_exec()
5169
5170 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5171 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5172 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5173 use by callout functions. It will never be used by PCRE itself.
5174
5175 (ii) The pcre_extra structure (see 45 above) has a void * field called
5176 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5177 pcre_callout_block structure has a field of the same name. The contents of
5178 the field passed in the pcre_extra structure are passed to the callout
5179 function in the corresponding field in the callout block. This makes it
5180 easier to use the same callout-containing regex from multiple threads. For
5181 testing, the pcretest program has a new data escape
5182
5183 \C*n pass the number n (may be negative) as callout_data
5184
5185 If the callout function in pcretest receives a non-zero value as
5186 callout_data, it returns that value.
5187
5188 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5189 there were some redundant $(CFLAGS) in commands that are now specified as
5190 $(LINK), which already includes $(CFLAGS).
5191
5192 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5193 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5194 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5195 one-byte characters throughout. Note that case-insensitive matching applies
5196 only to characters whose values are less than 256. PCRE doesn't support the
5197 notion of cases for higher-valued characters.
5198
5199 (i) A character class whose characters are all within 0-255 is handled as
5200 a bit map, and the map is inverted for negative classes. Previously, a
5201 character > 255 always failed to match such a class; however it should
5202 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5203
5204 (ii) A negated character class with a single character < 255 is coded as
5205 "not this character" (OP_NOT). This wasn't working properly when the test
5206 character was multibyte, either singly or repeated.
5207
5208 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5209 mode, for example: \x{100}{2,3}.
5210
5211 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5212 singly or repeated) now correctly test multibyte characters. However,
5213 PCRE doesn't recognize any characters with values greater than 255 as
5214 digits, spaces, or word characters. Such characters always match \D, \S,
5215 and \W, and never match \d, \s, or \w.
5216
5217 (v) Classes may now contain characters and character ranges with values
5218 greater than 255. For example: [ab\x{100}-\x{400}].
5219
5220 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5221 PCRE in UTF-8 mode.
5222
5223 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5224 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5225 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5226 value.)
5227
5228 53. The single man page has become too large. I have therefore split it up into
5229 a number of separate man pages. These also give rise to individual HTML pages;
5230 these are now put in a separate directory, and there is an index.html page that
5231 lists them all. Some hyperlinking between the pages has been installed.
5232
5233 54. Added convenience functions for handling named capturing parentheses.
5234
5235 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5236 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5237 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5238 are faulted.
5239
5240 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5241 calling configure. These values are used when compiling the dftables.c program
5242 which is run to generate the source of the default character tables. They
5243 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5244 you will need to set these values.
5245
5246 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5247
5248
5249 Version 3.9 02-Jan-02
5250 ---------------------
5251
5252 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5253
5254 2. If --disable-static was given, the building process failed when trying to
5255 build pcretest and pcregrep. (For some reason it was using libtool to compile
5256 them, which is not right, as they aren't part of the library.)
5257
5258
5259 Version 3.8 18-Dec-01
5260 ---------------------
5261
5262 1. The experimental UTF-8 code was completely screwed up. It was packing the
5263 bytes in the wrong order. How dumb can you get?
5264
5265
5266 Version 3.7 29-Oct-01
5267 ---------------------
5268
5269 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5270 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5271 this didn't happen under Solaris 8, where I normally test things.
5272
5273 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5274 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5275 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5276 this fix an hour or so after the initial 3.7 release.)
5277
5278
5279 Version 3.6 23-Oct-01
5280 ---------------------
5281
5282 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5283 offsets passed as NULL with zero offset count.
5284
5285 2. The config.guess and config.sub files had not been updated when I moved to
5286 the latest autoconf.
5287
5288
5289 Version 3.5 15-Aug-01
5290 ---------------------
5291
5292 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5293 had been forgotten.
5294
5295 2. By using declared but undefined structures, we can avoid using "void"
5296 definitions in pcre.h while keeping the internal definitions of the structures
5297 private.
5298
5299 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5300 user point of view, this means that both static and shared libraries are built
5301 by default, but this can be individually controlled. More of the work of
5302 handling this static/shared cases is now inside libtool instead of PCRE's make
5303 file.
5304
5305 4. The pcretest utility is now installed along with pcregrep because it is
5306 useful for users (to test regexs) and by doing this, it automatically gets
5307 relinked by libtool. The documentation has been turned into a man page, so
5308 there are now .1, .txt, and .html versions in /doc.
5309
5310 5. Upgrades to pcregrep:
5311 (i) Added long-form option names like gnu grep.
5312 (ii) Added --help to list all options with an explanatory phrase.
5313 (iii) Added -r, --recursive to recurse into sub-directories.
5314 (iv) Added -f, --file to read patterns from a file.
5315
5316 6. pcre_exec() was referring to its "code" argument before testing that
5317 argument for NULL (and giving an error if it was NULL).
5318
5319 7. Upgraded Makefile.in to allow for compiling in a different directory from
5320 the source directory.
5321
5322 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5323 options bits, the pointer it was passed was to an int instead of to an unsigned
5324 long int. This mattered only on 64-bit systems.
5325
5326 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5327 generated) instead of pcre.in, which it its source. Also made the same change
5328 in several of the .c files.
5329
5330 10. A new release of gcc defines printf() as a macro, which broke pcretest
5331 because it had an ifdef in the middle of a string argument for printf(). Fixed
5332 by using separate calls to printf().
5333
5334 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5335 script, to force use of CR or LF instead of \n in the source. On non-Unix
5336 systems, the value can be set in config.h.
5337
5338 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5339 absolute limit. Changed the text of the error message to make this clear, and
5340 likewise updated the man page.
5341
5342 13. The limit of 99 on the number of capturing subpatterns has been removed.
5343 The new limit is 65535, which I hope will not be a "real" limit.
5344
5345
5346 Version 3.4 22-Aug-00
5347 ---------------------
5348
5349 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5350
5351 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5352
5353
5354 Version 3.3 01-Aug-00
5355 ---------------------
5356
5357 1. If an octal character was given, but the value was greater than \377, it
5358 was not getting masked to the least significant bits, as documented. This could
5359 lead to crashes in some systems.
5360
5361 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5362 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5363
5364 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5365 These just pass their arguments on to (pcre_free)(), but they are provided
5366 because some uses of PCRE bind it to non-C systems that can call its functions,
5367 but cannot call free() or pcre_free() directly.
5368
5369 4. Add "make test" as a synonym for "make check". Corrected some comments in
5370 the Makefile.
5371
5372 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5373 Makefile.
5374
5375 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5376 command called pgrep for grepping around the active processes.
5377
5378 7. Added the beginnings of support for UTF-8 character strings.
5379
5380 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5381 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5382 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5383 out for the ar command.)
5384
5385
5386 Version 3.2 12-May-00
5387 ---------------------
5388
5389 This is purely a bug fixing release.
5390
5391 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5392 of ZA. This was just one example of several cases that could provoke this bug,
5393 which was introduced by change 9 of version 2.00. The code for breaking
5394 infinite loops after an iteration that matches an empty string was't working
5395 correctly.
5396
5397 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5398 when matched against abbab (for example). After matching an empty string, it
5399 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5400 caused it to match further down the string than it should.
5401
5402 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5403 was there because it doesn't seem to be needed, and it causes trouble on some
5404 systems, as it is not a Standard C header. It has been removed.
5405
5406 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5407 were reported on the Macintosh. The changes were from
5408
5409 while ((c = *(++ptr)) != 0 && c != '\n');
5410 to
5411 while ((c = *(++ptr)) != 0 && c != '\n') ;
5412
5413 Totally extraordinary, but if that's what it takes...
5414
5415 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5416 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5417 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5418 assumes the way PCRE uses memmove() (always moving upwards).
5419
5420 6. PCRE is being used in one environment where strchr() is not available. There
5421 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5422 faster code anyway.
5423
5424
5425 Version 3.1 09-Feb-00
5426 ---------------------
5427
5428 The only change in this release is the fixing of some bugs in Makefile.in for
5429 the "install" target:
5430
5431 (1) It was failing to install pcreposix.h.
5432
5433 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5434
5435
5436 Version 3.0 01-Feb-00
5437 ---------------------
5438
5439 1. Add support for the /+ modifier to perltest (to output $` like it does in
5440 pcretest).
5441
5442 2. Add support for the /g modifier to perltest.
5443
5444 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5445 matches null strings.
5446
5447 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5448 pattern. Perl treats empty patterns specially - it reuses the most recent
5449 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5450 effect.
5451
5452 5. The POSIX interface was broken in that it was just handing over the POSIX
5453 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5454 required a bigger vector, with some working space on the end. This means that
5455 the POSIX wrapper now has to get and free some memory, and copy the results.
5456
5457 6. Added some simple autoconf support, placing the test data and the
5458 documentation in separate directories, re-organizing some of the
5459 information files, and making it build pcre-config (a GNU standard). Also added
5460 libtool support for building PCRE as a shared library, which is now the
5461 default.
5462
5463 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5464 09 are not valid octal constants. Single digits will be used for minor values
5465 less than 10.
5466
5467 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5468 existing programs that set these in the POSIX interface can use PCRE without
5469 modification.
5470
5471 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5472 return all that pcre_info() returns, plus additional data. The pcre_info()
5473 function is retained for compatibility, but is considered to be obsolete.
5474
5475 10. Added experimental recursion feature (?R) to handle one common case that
5476 Perl 5.6 will be able to do with (?p{...}).
5477
5478 11. Added support for POSIX character classes like [:alpha:], which Perl is
5479 adopting.
5480
5481
5482 Version 2.08 31-Aug-99
5483 ----------------------
5484
5485 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5486 trying to match at the startoffset position, but instead was moving forward to
5487 the next newline as if a previous match had failed.
5488
5489 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5490 and could get into a loop if a null string was matched other than at the start
5491 of the subject.
5492
5493 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5494 be distinguished at compile time, and for completeness also added PCRE_DATE.
5495
5496 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5497 in GnuWin32 environments.
5498
5499
5500 Version 2.07 29-Jul-99
5501 ----------------------
5502
5503 1. The documentation is now supplied in plain text form and HTML as well as in
5504 the form of man page sources.
5505
5506 2. C++ compilers don't like assigning (void *) values to other pointer types.
5507 In particular this affects malloc(). Although there is no problem in Standard
5508 C, I've put in casts to keep C++ compilers happy.
5509
5510 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5511 should be (const char *).
5512
5513 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5514 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5515 However, I haven't made this a standard facility. The documentation doesn't
5516 mention it, and the Makefile doesn't support it.
5517
5518 5. The Makefile now contains an "install" target, with editable destinations at
5519 the top of the file. The pcretest program is not installed.
5520
5521 6. pgrep -V now gives the PCRE version number and date.
5522
5523 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5524 causing the entire string to be ignored, instead of just the last character.
5525
5526 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5527 non-matching string, it can take a very, very long time, even for strings of
5528 quite modest length, because of the nested recursion. PCRE now does better in
5529 some of these cases. It does this by remembering the last required literal
5530 character in the pattern, and pre-searching the subject to ensure it is present
5531 before running the real match. In other words, it applies a heuristic to detect
5532 some types of certain failure quickly, and in the above example, if presented
5533 with a string that has no trailing " it gives "no match" very quickly.
5534
5535 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5536 other alternatives are tried instead.
5537
5538
5539 Version 2.06 09-Jun-99
5540 ----------------------
5541
5542 1. Change pcretest's output for amount of store used to show just the code
5543 space, because the remainder (the data block) varies in size between 32-bit and
5544 64-bit systems.
5545
5546 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5547 start matching at. This allows lookbehinds to work when searching for multiple
5548 occurrences in a string.
5549
5550 3. Added additional options to pcretest for testing multiple occurrences:
5551
5552 /+ outputs the rest of the string that follows a match
5553 /g loops for multiple occurrences, using the new startoffset argument
5554 /G loops for multiple occurrences by passing an incremented pointer
5555
5556 4. PCRE wasn't doing the "first character" optimization for patterns starting
5557 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5558 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5559 the letter 'x'. On long subject strings, this gives a significant speed-up.
5560
5561
5562 Version 2.05 21-Apr-99
5563 ----------------------
5564
5565 1. Changed the type of magic_number from int to long int so that it works
5566 properly on 16-bit systems.
5567
5568 2. Fixed a bug which caused patterns starting with .* not to work correctly
5569 when the subject string contained newline characters. PCRE was assuming
5570 anchoring for such patterns in all cases, which is not correct because .* will
5571 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5572 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5573 must be retried after every newline in the subject.
5574
5575
5576 Version 2.04 18-Feb-99
5577 ----------------------
5578
5579 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5580 computation of the store needed to hold the pattern was incorrect (too large).
5581 If such patterns were nested a few deep, this could multiply and become a real
5582 problem.
5583
5584 2. Added /M option to pcretest to show the memory requirement of a specific
5585 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5586
5587 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5588 compiled in such a way that the backtracking after subsequent failure was
5589 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5590 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5591
5592
5593 Version 2.03 02-Feb-99
5594 ----------------------
5595
5596 1. Fixed typo and small mistake in man page.
5597
5598 2. Added 4th condition (GPL supersedes if conflict) and created separate
5599 LICENCE file containing the conditions.
5600
5601 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5602 Perl, that is the internal \ allows the delimiter to be included in the
5603 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5604 the final delimiter, add \ to the end of the pattern (to test the error).
5605
5606 4. Added the convenience functions for extracting substrings after a successful
5607 match. Updated pcretest to make it able to test these functions.
5608
5609
5610 Version 2.02 14-Jan-99
5611 ----------------------
5612
5613 1. Initialized the working variables associated with each extraction so that
5614 their saving and restoring doesn't refer to uninitialized store.
5615
5616 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5617 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5618 fix the problem.
5619
5620 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5621 calls, and wasn't printing the correct value for compiling calls. Increased the
5622 default value of LOOPREPEAT, and the number of significant figures in the
5623 times.
5624
5625 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5626
5627 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5628 a building problem on Windows NT with a FAT file system.
5629
5630
5631 Version 2.01 21-Oct-98
5632 ----------------------
5633
5634 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5635 to character tables built by pcre_maketables() in the current locale. If NULL
5636 is passed, the default tables are used.
5637
5638
5639 Version 2.00 24-Sep-98
5640 ----------------------
5641
5642 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5643 it any more.
5644
5645 2. Allow quantification of (?>) groups, and make it work correctly.
5646
5647 3. The first character computation wasn't working for (?>) groups.
5648
5649 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5650 end of the subject) and add 5.005's \z, which really does match only at the
5651 very end of the subject.
5652
5653 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5654
5655 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5656 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5657 localized options. All options to pcre_study() were also removed.
5658
5659 7. Add other new features from 5.005:
5660
5661 $(?<= positive lookbehind
5662 $(?<! negative lookbehind
5663 (?imsx-imsx) added the unsetting capability
5664 such a setting is global if at outer level; local otherwise
5665 (?imsx-imsx:) non-capturing groups with option setting
5666 (?(cond)re|re) conditional pattern matching
5667
5668 A backreference to itself in a repeated group matches the previous
5669 captured string.
5670
5671 8. General tidying up of studying (both automatic and via "study")
5672 consequential on the addition of new assertions.
5673
5674 9. As in 5.005, unlimited repeated groups that could match an empty substring
5675 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5676 runtime if any iteration does actually match an empty substring.
5677
5678 10. Include the RunTest script in the distribution.
5679
5680 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5681 discrepancies, some of which were old and were also with respect to 5.004. They
5682 have now been fixed.
5683
5684
5685 Version 1.09 28-Apr-98
5686 ----------------------
5687
5688 1. A negated single character class followed by a quantifier with a minimum
5689 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5690 program crashes, or just wrong answers. This did not apply to negated classes
5691 containing more than one character, or to minima other than one.
5692
5693
5694 Version 1.08 27-Mar-98
5695 ----------------------
5696
5697 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5698
5699 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5700 latter must appear before anything that relies on it in the pattern.
5701
5702
5703 Version 1.07 16-Feb-98
5704 ----------------------
5705
5706 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5707 repeat of a potentially empty string).
5708
5709
5710 Version 1.06 23-Jan-98
5711 ----------------------
5712
5713 1. Added Markus Oberhumer's little patches for C++.
5714
5715 2. Literal strings longer than 255 characters were broken.
5716
5717
5718 Version 1.05 23-Dec-97
5719 ----------------------
5720
5721 1. Negated character classes containing more than one character were failing if
5722 PCRE_CASELESS was set at run time.
5723
5724
5725 Version 1.04 19-Dec-97
5726 ----------------------
5727
5728 1. Corrected the man page, where some "const" qualifiers had been omitted.
5729
5730 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5731 input syntax.
5732
5733 3. Fixed memory leak which occurred when a regex with back references was
5734 matched with an offsets vector that wasn't big enough. The temporary memory
5735 that is used in this case wasn't being freed if the match failed.
5736
5737 4. Tidied pcretest to ensure it frees memory that it gets.
5738
5739 5. Temporary memory was being obtained in the case where the passed offsets
5740 vector was exactly big enough.
5741
5742 6. Corrected definition of offsetof() from change 5 below.
5743
5744 7. I had screwed up change 6 below and broken the rules for the use of
5745 setjmp(). Now fixed.
5746
5747
5748 Version 1.03 18-Dec-97
5749 ----------------------
5750
5751 1. A erroneous regex with a missing opening parenthesis was correctly
5752 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5753 on some systems.
5754
5755 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5756 it was reported that one broken compiler failed on the former because "code" is
5757 also an independent variable.
5758
5759 3. The erroneous regex a[]b caused an array overrun reference.
5760
5761 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5762 fail on data ending with that character. (It was going on too far, and checking
5763 the next character, typically a binary zero.) This was specific to the
5764 optimized code for single-character negative classes.
5765
5766 5. Added a contributed patch from the TIN world which does the following:
5767
5768 + Add an undef for memmove, in case the the system defines a macro for it.
5769
5770 + Add a definition of offsetof(), in case there isn't one. (I don't know
5771 the reason behind this - offsetof() is part of the ANSI standard - but
5772 it does no harm).
5773
5774 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5775 most of the places where whitespace preceded '#'. I have given up and
5776 allowed the remaining 2 cases to be at the margin.
5777
5778 + Rename some variables in pcre to eliminate shadowing. This seems very
5779 pedantic, but does no harm, of course.
5780
5781 6. Moved the call to setjmp() into its own function, to get rid of warnings
5782 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5783
5784 7. Constructs such as \d{8,} were compiling into the equivalent of
5785 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5786 outcome, but in this particular case used more store than had been allocated,
5787 which caused the bug to be discovered because it threw up an internal error.
5788
5789 8. The debugging code in both pcre and pcretest for outputting the compiled
5790 form of a regex was going wrong in the case of back references followed by
5791 curly-bracketed repeats.
5792
5793
5794 Version 1.02 12-Dec-97
5795 ----------------------
5796
5797 1. Typos in pcre.3 and comments in the source fixed.
5798
5799 2. Applied a contributed patch to get rid of places where it used to remove
5800 'const' from variables, and fixed some signed/unsigned and uninitialized
5801 variable warnings.
5802
5803 3. Added the "runtest" target to Makefile.
5804
5805 4. Set default compiler flag to -O2 rather than just -O.
5806
5807
5808 Version 1.01 19-Nov-97
5809 ----------------------
5810
5811 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5812 like /([ab]*)*/, that is, for classes with more than one character in them.
5813
5814 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5815 as /((?>a*))*/ (a PCRE_EXTRA facility).
5816
5817
5818 Version 1.00 18-Nov-97
5819 ----------------------
5820
5821 1. Added compile-time macros to support systems such as SunOS4 which don't have
5822 memmove() or strerror() but have other things that can be used instead.
5823
5824 2. Arranged that "make clean" removes the executables.
5825
5826
5827 Version 0.99 27-Oct-97
5828 ----------------------
5829
5830 1. Fixed bug in code for optimizing classes with only one character. It was
5831 initializing a 32-byte map regardless, which could cause it to run off the end
5832 of the memory it had got.
5833
5834 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5835
5836
5837 Version 0.98 22-Oct-97
5838 ----------------------
5839
5840 1. Fixed bug in code for handling temporary memory usage when there are more
5841 back references than supplied space in the ovector. This could cause segfaults.
5842
5843
5844 Version 0.97 21-Oct-97
5845 ----------------------
5846
5847 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5848
5849 2. Optimized negated single characters not to use a bit map.
5850
5851 3. Brought error texts together as macro definitions; clarified some of them;
5852 fixed one that was wrong - it said "range out of order" when it meant "invalid
5853 escape sequence".
5854
5855 4. Changed some char * arguments to const char *.
5856
5857 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5858
5859 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5860 pcretest.
5861
5862
5863 Version 0.96 16-Oct-97
5864 ----------------------
5865
5866 1. Added a simple "pgrep" utility to the distribution.
5867
5868 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5869 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5870 where "ddd" means "one or more decimal digits".
5871
5872 3. Fixed serious bug. If a pattern had a back reference, but the call to
5873 pcre_exec() didn't supply a large enough ovector to record the related
5874 identifying subpattern, the match always failed. PCRE now remembers the number
5875 of the largest back reference, and gets some temporary memory in which to save
5876 the offsets during matching if necessary, in order to ensure that
5877 backreferences always work.
5878
5879 4. Increased the compatibility with Perl in a number of ways:
5880
5881 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5882 to request this handling. The option can be set at compile or exec time.
5883
5884 (b) $ matches before a terminating newline by default; an option
5885 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5886 mode). The option can be set at compile or exec time.
5887
5888 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5889 the same as Perl's. If the decimal number it represents is less than 10
5890 or there aren't that many previous left capturing parentheses, an octal
5891 escape is read. Inside a character class, it's always an octal escape,
5892 even if it is a single digit.
5893
5894 (d) An escaped but undefined alphabetic character is taken as a literal,
5895 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5896 escapes.
5897
5898 (e) {0} is now permitted. (The previous item is removed from the compiled
5899 pattern).
5900
5901 5. Changed all the names of code files so that the basic parts are no longer
5902 than 10 characters, and abolished the teeny "globals.c" file.
5903
5904 6. Changed the handling of character classes; they are now done with a 32-byte
5905 bit map always.
5906
5907 7. Added the -d and /D options to pcretest to make it possible to look at the
5908 internals of compilation without having to recompile pcre.
5909
5910
5911 Version 0.95 23-Sep-97
5912 ----------------------
5913
5914 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5915 \x20 at the start of a run of normal characters. These were being treated as
5916 real characters, instead of the source characters being re-checked.
5917
5918
5919 Version 0.94 18-Sep-97
5920 ----------------------
5921
5922 1. The functions are now thread-safe, with the caveat that the global variables
5923 containing pointers to malloc() and free() or alternative functions are the
5924 same for all threads.
5925
5926 2. Get pcre_study() to generate a bitmap of initial characters for non-
5927 anchored patterns when this is possible, and use it if passed to pcre_exec().
5928
5929
5930 Version 0.93 15-Sep-97
5931 ----------------------
5932
5933 1. /(b)|(:+)/ was computing an incorrect first character.
5934
5935 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5936 but not actually doing anything yet.
5937
5938 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5939 as Perl does (e.g. [-az] or [az-]).
5940
5941 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5942 all possible positions.
5943
5944 5. Split up into different modules to avoid including unneeded functions in a
5945 compiled binary. However, compile and exec are still in one module. The "study"
5946 function is split off.
5947
5948 6. The character tables are now in a separate module whose source is generated
5949 by an auxiliary program - but can then be edited by hand if required. There are
5950 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5951 toupper() in the code.
5952
5953 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5954 make them global. Abolish the function for setting them, as the caller can now
5955 set them directly.
5956
5957
5958 Version 0.92 11-Sep-97
5959 ----------------------
5960
5961 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5962 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5963
5964 2. Caseless matching was not working in character classes if the characters in
5965 the pattern were in upper case.
5966
5967 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5968
5969 4. Make PCRE_ANCHORED public and accept as a compile option.
5970
5971 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5972 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5973 pass them.
5974
5975 6. Give an error if bad option bits passed at compile or run time.
5976
5977 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5978 pcretest to cause it to pass that flag.
5979
5980 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5981 options, and the first character, if set.
5982
5983 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5984
5985
5986 Version 0.91 10-Sep-97
5987 ----------------------
5988
5989 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5990 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5991
5992 2. PCRE was looping on encountering an indefinitely repeated back reference to
5993 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5994 Perl does - treats the match as successful.
5995
5996 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5