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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1604 - (show annotations)
Tue Nov 17 17:27:17 2015 UTC (4 years ago) by ph10
File size: 284975 byte(s)
Error occurred while calculating annotation data.
Fix POSIX class single-character bug.
1 ChangeLog for PCRE
2 ------------------
3
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
6
7 Version 8.38 27-October-2015
8 ----------------------------
9
10 1. If a group that contained a recursive back reference also contained a
11 forward reference subroutine call followed by a non-forward-reference
12 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13 compile correct code, leading to undefined behaviour or an internally
14 detected error. This bug was discovered by the LLVM fuzzer.
15
16 2. Quantification of certain items (e.g. atomic back references) could cause
17 incorrect code to be compiled when recursive forward references were
18 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19 This bug was discovered by the LLVM fuzzer.
20
21 3. A repeated conditional group whose condition was a reference by name caused
22 a buffer overflow if there was more than one group with the given name.
23 This bug was discovered by the LLVM fuzzer.
24
25 4. A recursive back reference by name within a group that had the same name as
26 another group caused a buffer overflow. For example:
27 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28
29 5. A forward reference by name to a group whose number is the same as the
30 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31 a buffer overflow at compile time. This bug was discovered by the LLVM
32 fuzzer.
33
34 6. A lookbehind assertion within a set of mutually recursive subpatterns could
35 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36
37 7. Another buffer overflow bug involved duplicate named groups with a
38 reference between their definition, with a group that reset capture
39 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40 fixed by always allowing for more memory, even if not needed. (A proper fix
41 is implemented in PCRE2, but it involves more refactoring.)
42
43 8. There was no check for integer overflow in subroutine calls such as (?123).
44
45 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
46 being treated as a literal 'l' instead of causing an error.
47
48 10. There was a buffer overflow if pcre_exec() was called with an ovector of
49 size 1. This bug was found by american fuzzy lop.
50
51 11. If a non-capturing group containing a conditional group that could match
52 an empty string was repeated, it was not identified as matching an empty
53 string itself. For example: /^(?:(?(1)x|)+)+$()/.
54
55 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
56 \a and \e in test subject lines.
57
58 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
59 instead of the EBCDIC value.
60
61 14. The handling of \c in an EBCDIC environment has been revised so that it is
62 now compatible with the specification in Perl's perlebcdic page.
63
64 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
65 ASCII/Unicode. This has now been added to the list of characters that are
66 recognized as white space in EBCDIC.
67
68 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
69 error (correctly) when used outside a class, but did not give an error
70 within a class.
71
72 17. \h within a class was incorrectly compiled in EBCDIC environments.
73
74 18. A pattern with an unmatched closing parenthesis that contained a backward
75 assertion which itself contained a forward reference caused buffer
76 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
77
78 19. JIT should return with error when the compiled pattern requires more stack
79 space than the maximum.
80
81 20. A possessively repeated conditional group that could match an empty string,
82 for example, /(?(R))*+/, was incorrectly compiled.
83
84 21. Fix infinite recursion in the JIT compiler when certain patterns such as
85 /(?:|a|){100}x/ are analysed.
86
87 22. Some patterns with character classes involving [: and \\ were incorrectly
88 compiled and could cause reading from uninitialized memory or an incorrect
89 error diagnosis.
90
91 23. Pathological patterns containing many nested occurrences of [: caused
92 pcre_compile() to run for a very long time.
93
94 24. A conditional group with only one branch has an implicit empty alternative
95 branch and must therefore be treated as potentially matching an empty
96 string.
97
98 25. If (?R was followed by - or + incorrect behaviour happened instead of a
99 diagnostic.
100
101 26. Arrange to give up on finding the minimum matching length for overly
102 complex patterns.
103
104 27. Similar to (4) above: in a pattern with duplicated named groups and an
105 occurrence of (?| it is possible for an apparently non-recursive back
106 reference to become recursive if a later named group with the relevant
107 number is encountered. This could lead to a buffer overflow. Wen Guanxing
108 from Venustech ADLAB discovered this bug.
109
110 28. If pcregrep was given the -q option with -c or -l, or when handling a
111 binary file, it incorrectly wrote output to stdout.
112
113 29. The JIT compiler did not restore the control verb head in case of *THEN
114 control verbs. This issue was found by Karl Skomski with a custom LLVM
115 fuzzer.
116
117 30. Error messages for syntax errors following \g and \k were giving inaccurate
118 offsets in the pattern.
119
120 31. Added a check for integer overflow in conditions (?(<digits>) and
121 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
122 fuzzer.
123
124 32. Handling recursive references such as (?2) when the reference is to a group
125 later in the pattern uses code that is very hacked about and error-prone.
126 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
127 give an internal error if it is obvious that compiling has gone wrong.
128
129 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
130 This issue was found by Karl Skomski with a custom LLVM fuzzer.
131
132 34. The JIT compiler should restore the control chain for empty possessive
133 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
134
135 35. Match limit check added to JIT recursion. This issue was found by Karl
136 Skomski with a custom LLVM fuzzer.
137
138 36. Yet another case similar to 27 above has been circumvented by an
139 unconditional allocation of extra memory. This issue is fixed "properly" in
140 PCRE2 by refactoring the way references are handled. Wen Guanxing
141 from Venustech ADLAB discovered this bug.
142
143 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
144 with a custom LLVM fuzzer.
145
146 38. Fixed a corner case of range optimization in JIT.
147
148 39. An incorrect error "overran compiling workspace" was given if there were
149 exactly enough group forward references such that the last one extended
150 into the workspace safety margin. The next one would have expanded the
151 workspace. The test for overflow was not including the safety margin.
152
153 40. A match limit issue is fixed in JIT which was found by Karl Skomski
154 with a custom LLVM fuzzer.
155
156 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
157 work under Windows. (Why has it taken so long for anyone to notice?)
158
159 42. In a character class such as [\W\p{Any}] where both a negative-type escape
160 ("not a word character") and a property escape were present, the property
161 escape was being ignored.
162
163 43. Fix crash caused by very long (*MARK) or (*THEN) names.
164
165 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
166 by a single ASCII character in a class item, was incorrectly compiled in
167 UCP mode. The POSIX class got lost, but only if the single character
168 followed it.
169
170
171 Version 8.37 28-April-2015
172 --------------------------
173
174 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
175 for those parentheses to be closed with whatever has been captured so far.
176 However, it was failing to mark any other groups between the hightest
177 capture so far and the currrent group as "unset". Thus, the ovector for
178 those groups contained whatever was previously there. An example is the
179 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
180
181 2. If an assertion condition was quantified with a minimum of zero (an odd
182 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
183
184 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
185 unrecognized modifier, a crash could occur.
186
187 4. An attempt to do global matching in pcretest with a zero-length ovector
188 caused a crash.
189
190 5. Fixed a memory leak during matching that could occur for a subpattern
191 subroutine call (recursive or otherwise) if the number of captured groups
192 that had to be saved was greater than ten.
193
194 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
195 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
196 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
197
198 7. A UTF pattern containing a "not" match of a non-ASCII character and a
199 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
200
201 8. When a pattern is compiled, it remembers the highest back reference so that
202 when matching, if the ovector is too small, extra memory can be obtained to
203 use instead. A conditional subpattern whose condition is a check on a
204 capture having happened, such as, for example in the pattern
205 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
206 setting the highest backreference number. This mattered only if pcre_exec()
207 was called with an ovector that was too small to hold the capture, and there
208 was no other kind of back reference (a situation which is probably quite
209 rare). The effect of the bug was that the condition was always treated as
210 FALSE when the capture could not be consulted, leading to a incorrect
211 behaviour by pcre_exec(). This bug has been fixed.
212
213 9. A reference to a duplicated named group (either a back reference or a test
214 for being set in a conditional) that occurred in a part of the pattern where
215 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
216 to be incorrectly calculated, leading to overwriting.
217
218 10. A mutually recursive set of back references such as (\2)(\1) caused a
219 segfault at study time (while trying to find the minimum matching length).
220 The infinite loop is now broken (with the minimum length unset, that is,
221 zero).
222
223 11. If an assertion that was used as a condition was quantified with a minimum
224 of zero, matching went wrong. In particular, if the whole group had
225 unlimited repetition and could match an empty string, a segfault was
226 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
227 assertions to be quantified, but not if they are being used as conditions,
228 so the above pattern is faulted by Perl. PCRE has now been changed so that
229 it also rejects such patterns.
230
231 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
232 failed to allow the zero-repeat case if pcre2_exec() was called with an
233 ovector too small to capture the group.
234
235 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
236 Red Hat Product Security:
237
238 (a) A crash if /K and /F were both set with the option to save the compiled
239 pattern.
240
241 (b) Another crash if the option to print captured substrings in a callout
242 was combined with setting a null ovector, for example \O\C+ as a subject
243 string.
244
245 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
246 forward reference repeated a large (but limited) number of times within a
247 repeated outer group that has a zero minimum quantifier, caused incorrect
248 code to be compiled, leading to the error "internal error:
249 previously-checked referenced subpattern not found" when an incorrect
250 memory address was read. This bug was reported as "heap overflow",
251 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
252 CVE-2015-2325.
253
254 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
255 call within a group that also contained a recursive back reference caused
256 incorrect code to be compiled. This bug was reported as "heap overflow",
257 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
258 number CVE-2015-2326.
259
260 24. Computing the size of the JIT read-only data in advance has been a source
261 of various issues, and new ones are still appear unfortunately. To fix
262 existing and future issues, size computation is eliminated from the code,
263 and replaced by on-demand memory allocation.
264
265 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
266 adjacent to the end of the range, and the range contained characters with
267 more than one other case, caused incorrect behaviour when compiled in UTF
268 mode. In that example, the range a-j was left out of the class.
269
270 26. Fix JIT compilation of conditional blocks, which assertion
271 is converted to (*FAIL). E.g: /(?(?!))/.
272
273 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
274 discovered by the LLVM fuzzer.
275
276 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
277 when this assertion was used as a condition, for example (?(?!)a|b). In
278 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
279 error about an unsupported item.
280
281 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
282 possessification code could take exponential time to complete. A recursion
283 depth limit of 1000 has been imposed to limit the resources used by this
284 optimization.
285
286 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
287 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
288 because \S ensures they are all in the class. The code for doing this was
289 interacting badly with the code for computing the amount of space needed to
290 compile the pattern, leading to a buffer overflow. This bug was discovered
291 by the LLVM fuzzer.
292
293 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
294 other kinds of group caused stack overflow at compile time. This bug was
295 discovered by the LLVM fuzzer.
296
297 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
298 between a subroutine call and its quantifier was incorrectly compiled,
299 leading to buffer overflow or other errors. This bug was discovered by the
300 LLVM fuzzer.
301
302 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
303 assertion after (?(. The code was failing to check the character after
304 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
305 was discovered by the LLVM fuzzer.
306
307 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
308 a fixed maximum following a group that contains a subroutine reference was
309 incorrectly compiled and could trigger buffer overflow. This bug was
310 discovered by the LLVM fuzzer.
311
312 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
313 caused a stack overflow instead of the diagnosis of a non-fixed length
314 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
315
316 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
317 (e.g. /(?<=\Ka)/) could make pcregrep loop.
318
319 37. There was a similar problem to 36 in pcretest for global matches.
320
321 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
322 and a subsequent item in the pattern caused a non-match, backtracking over
323 the repeated \X did not stop, but carried on past the start of the subject,
324 causing reference to random memory and/or a segfault. There were also some
325 other cases where backtracking after \C could crash. This set of bugs was
326 discovered by the LLVM fuzzer.
327
328 39. The function for finding the minimum length of a matching string could take
329 a very long time if mutual recursion was present many times in a pattern,
330 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
331 method has been implemented. This infelicity was discovered by the LLVM
332 fuzzer.
333
334 40. Static linking against the PCRE library using the pkg-config module was
335 failing on missing pthread symbols.
336
337
338 Version 8.36 26-September-2014
339 ------------------------------
340
341 1. Got rid of some compiler warnings in the C++ modules that were shown up by
342 -Wmissing-field-initializers and -Wunused-parameter.
343
344 2. The tests for quantifiers being too big (greater than 65535) were being
345 applied after reading the number, and stupidly assuming that integer
346 overflow would give a negative number. The tests are now applied as the
347 numbers are read.
348
349 3. Tidy code in pcre_exec.c where two branches that used to be different are
350 now the same.
351
352 4. The JIT compiler did not generate match limit checks for certain
353 bracketed expressions with quantifiers. This may lead to exponential
354 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
355 issue should be resolved now.
356
357 5. Fixed an issue, which occures when nested alternatives are optimized
358 with table jumps.
359
360 6. Inserted two casts and changed some ints to size_t in the light of some
361 reported 64-bit compiler warnings (Bugzilla 1477).
362
363 7. Fixed a bug concerned with zero-minimum possessive groups that could match
364 an empty string, which sometimes were behaving incorrectly in the
365 interpreter (though correctly in the JIT matcher). This pcretest input is
366 an example:
367
368 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
369 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
370
371 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
372 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
373 for an empty string was breaking the inner loop and carrying on at a lower
374 level, when possessive repeated groups should always return to a higher
375 level as they have no backtrack points in them. The empty string test now
376 occurs at the outer level.
377
378 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
379 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
380
381 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
382 doesn't).
383
384 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
385 to hold a pattern to be incorrectly computed (too small) when there were
386 named back references to duplicated names. This could cause "internal
387 error: code overflow" or "double free or corruption" or other memory
388 handling errors.
389
390 11. When named subpatterns had the same prefixes, back references could be
391 confused. For example, in this pattern:
392
393 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
394
395 the reference to 'Name' was incorrectly treated as a reference to a
396 duplicate name.
397
398 12. A pattern such as /^s?c/mi8 where the optional character has more than
399 one "other case" was incorrectly compiled such that it would only try to
400 match starting at "c".
401
402 13. When a pattern starting with \s was studied, VT was not included in the
403 list of possible starting characters; this should have been part of the
404 8.34/18 patch.
405
406 14. If a character class started [\Qx]... where x is any character, the class
407 was incorrectly terminated at the ].
408
409 15. If a pattern that started with a caseless match for a character with more
410 than one "other case" was studied, PCRE did not set up the starting code
411 unit bit map for the list of possible characters. Now it does. This is an
412 optimization improvement, not a bug fix.
413
414 16. The Unicode data tables have been updated to Unicode 7.0.0.
415
416 17. Fixed a number of memory leaks in pcregrep.
417
418 18. Avoid a compiler warning (from some compilers) for a function call with
419 a cast that removes "const" from an lvalue by using an intermediate
420 variable (to which the compiler does not object).
421
422 19. Incorrect code was compiled if a group that contained an internal recursive
423 back reference was optional (had quantifier with a minimum of zero). This
424 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
425 caused segmentation faults because of stack overflows at compile time.
426
427 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
428 group that is quantified with an indefinite repeat, caused a compile-time
429 loop which used up all the system stack and provoked a segmentation fault.
430 This was not the same bug as 19 above.
431
432 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
433 Patch by Mike Frysinger.
434
435
436 Version 8.35 04-April-2014
437 --------------------------
438
439 1. A new flag is set, when property checks are present in an XCLASS.
440 When this flag is not set, PCRE can perform certain optimizations
441 such as studying these XCLASS-es.
442
443 2. The auto-possessification of character sets were improved: a normal
444 and an extended character set can be compared now. Furthermore
445 the JIT compiler optimizes more character set checks.
446
447 3. Got rid of some compiler warnings for potentially uninitialized variables
448 that show up only when compiled with -O2.
449
450 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
451 of a match later then the end of the match. The pcretest program was not
452 handling the case sensibly - it was outputting from the start to the next
453 binary zero. It now reports this situation in a message, and outputs the
454 text from the end to the start.
455
456 5. Fast forward search is improved in JIT. Instead of the first three
457 characters, any three characters with fixed position can be searched.
458 Search order: first, last, middle.
459
460 6. Improve character range checks in JIT. Characters are read by an inprecise
461 function now, which returns with an unknown value if the character code is
462 above a certain threshold (e.g: 256). The only limitation is that the value
463 must be bigger than the threshold as well. This function is useful when
464 the characters above the threshold are handled in the same way.
465
466 7. The macros whose names start with RAWUCHAR are placeholders for a future
467 mode in which only the bottom 21 bits of 32-bit data items are used. To
468 make this more memorable for those maintaining the code, the names have
469 been changed to start with UCHAR21, and an extensive comment has been added
470 to their definition.
471
472 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
473 to the export list in Makefile.am (they were accidentally omitted from the
474 8.34 tarball).
475
476 9. The informational output from pcretest used the phrase "starting byte set"
477 which is inappropriate for the 16-bit and 32-bit libraries. As the output
478 for "first char" and "need char" really means "non-UTF-char", I've changed
479 "byte" to "char", and slightly reworded the output. The documentation about
480 these values has also been (I hope) clarified.
481
482 10. Another JIT related optimization: use table jumps for selecting the correct
483 backtracking path, when more than four alternatives are present inside a
484 bracket.
485
486 11. Empty match is not possible, when the minimum length is greater than zero,
487 and there is no \K in the pattern. JIT should avoid empty match checks in
488 such cases.
489
490 12. In a caseless character class with UCP support, when a character with more
491 than one alternative case was not the first character of a range, not all
492 the alternative cases were added to the class. For example, s and \x{17f}
493 are both alternative cases for S: the class [RST] was handled correctly,
494 but [R-T] was not.
495
496 13. The configure.ac file always checked for pthread support when JIT was
497 enabled. This is not used in Windows, so I have put this test inside a
498 check for the presence of windows.h (which was already tested for).
499
500 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
501 The algorithm provides a way to skip certain starting offsets, and usually
502 faster than linear prefix searches.
503
504 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
505 as for 'fr_FR' and 'french'. For some reason, however, it then used the
506 Windows-specific input and output files, which have 'french' screwed in.
507 So this could never have worked. One of the problems with locales is that
508 they aren't always the same. I have now updated RunTest so that it checks
509 the output of the locale test (test 3) against three different output
510 files, and it allows the test to pass if any one of them matches. With luck
511 this should make the test pass on some versions of Solaris where it was
512 failing. Because of the uncertainty, the script did not used to stop if
513 test 3 failed; it now does. If further versions of a French locale ever
514 come to light, they can now easily be added.
515
516 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
517 there was a message during ./configure, but it did not stop. This now
518 provokes an error. The invalid example in README has been corrected.
519 If a value less than the minimum is given, the minimum value has always
520 been used, but now a warning is given.
521
522 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
523 was a bug in the test system, which is now fixed. Also, the list of various
524 configurations that are tested for each release did not have one with both
525 16/32 bits and --enable-bar-anycrlf. It now does.
526
527 18. pcretest was missing "-C bsr" for displaying the \R default setting.
528
529 19. Little endian PowerPC systems are supported now by the JIT compiler.
530
531 20. The fast forward newline mechanism could enter to an infinite loop on
532 certain invalid UTF-8 input. Although we don't support these cases
533 this issue can be fixed by a performance optimization.
534
535 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
536 not take account if existing stack usage. There is now a new global
537 variable called pcre_stack_guard that can be set to point to an external
538 function to check stack availability. It is called at the start of
539 processing every parenthesized group.
540
541 22. A typo in the code meant that in ungreedy mode the max/min qualifier
542 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
543 not match "ab".
544
545 23. When UTF was disabled, the JIT program reported some incorrect compile
546 errors. These messages are silenced now.
547
548 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
549 compiler.
550
551 25. Change all the temporary files used in RunGrepTest to be different to those
552 used by RunTest so that the tests can be run simultaneously, for example by
553 "make -j check".
554
555
556 Version 8.34 15-December-2013
557 -----------------------------
558
559 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
560 executable memory. Patch inspired by Carsten Klein.
561
562 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
563 this macro is never tested and has no effect, because the work to support
564 coverage involves only compiling and linking options and special targets in
565 the Makefile. The comment in config.h implied that defining the macro would
566 enable coverage support, which is totally false. There was also support for
567 setting this macro in the CMake files (my fault, I just copied it from
568 configure). SUPPORT_GCOV has now been removed.
569
570 3. Make a small performance improvement in strlen16() and strlen32() in
571 pcretest.
572
573 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
574 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
575 these cases). There was also one in pcretest.c.
576
577 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
578
579 6. In UTF mode, the code for checking whether a group could match an empty
580 string (which is used for indefinitely repeated groups to allow for
581 breaking an infinite loop) was broken when the group contained a repeated
582 negated single-character class with a character that occupied more than one
583 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
584 UTF-8 mode). The effect was undefined: the group might or might not be
585 deemed as matching an empty string, or the program might have crashed.
586
587 7. The code for checking whether a group could match an empty string was not
588 recognizing that \h, \H, \v, \V, and \R must match a character.
589
590 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
591 an empty string. If it can, pcretest shows this in its information output.
592
593 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
594 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
595 matched by pcre_exec() without using JIT:
596
597 (a) If the rest of the pattern did not match after a maximal run of
598 grapheme clusters, the code for backing up to try with fewer of them
599 did not always back up over a full grapheme when characters that do not
600 have the modifier quality were involved, e.g. Hangul syllables.
601
602 (b) If the match point in a subject started with modifier character, and
603 there was no match, the code could incorrectly back up beyond the match
604 point, and potentially beyond the first character in the subject,
605 leading to a segfault or an incorrect match result.
606
607 10. A conditional group with an assertion condition could lead to PCRE
608 recording an incorrect first data item for a match if no other first data
609 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
610 first data item, and therefore matched "ca" after "c" instead of at the
611 start.
612
613 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
614 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
615
616 12. The source of pcregrep now includes z/OS-specific code so that it can be
617 compiled for z/OS as part of the special z/OS distribution.
618
619 13. Added the -T and -TM options to pcretest.
620
621 14. The code in pcre_compile.c for creating the table of named capturing groups
622 has been refactored. Instead of creating the table dynamically during the
623 actual compiling pass, the information is remembered during the pre-compile
624 pass (on the stack unless there are more than 20 named groups, in which
625 case malloc() is used) and the whole table is created before the actual
626 compile happens. This has simplified the code (it is now nearly 150 lines
627 shorter) and prepared the way for better handling of references to groups
628 with duplicate names.
629
630 15. A back reference to a named subpattern when there is more than one of the
631 same name now checks them in the order in which they appear in the pattern.
632 The first one that is set is used for the reference. Previously only the
633 first one was inspected. This change makes PCRE more compatible with Perl.
634
635 16. Unicode character properties were updated from Unicode 6.3.0.
636
637 17. The compile-time code for auto-possessification has been refactored, based
638 on a patch by Zoltan Herczeg. It now happens after instead of during
639 compilation. The code is cleaner, and more cases are handled. The option
640 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
641 options in pcretest are provided to set it. It can also be set by
642 (*NO_AUTO_POSSESS) at the start of a pattern.
643
644 18. The character VT has been added to the default ("C" locale) set of
645 characters that match \s and are generally treated as white space,
646 following this same change in Perl 5.18. There is now no difference between
647 "Perl space" and "POSIX space". Whether VT is treated as white space in
648 other locales depends on the locale.
649
650 19. The code for checking named groups as conditions, either for being set or
651 for being recursed, has been refactored (this is related to 14 and 15
652 above). Processing unduplicated named groups should now be as fast at
653 numerical groups, and processing duplicated groups should be faster than
654 before.
655
656 20. Two patches to the CMake build system, by Alexander Barkov:
657
658 (1) Replace the "source" command by "." in CMakeLists.txt because
659 "source" is a bash-ism.
660
661 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
662 without these the CMake build does not work on Solaris.
663
664 21. Perl has changed its handling of \8 and \9. If there is no previously
665 encountered capturing group of those numbers, they are treated as the
666 literal characters 8 and 9 instead of a binary zero followed by the
667 literals. PCRE now does the same.
668
669 22. Following Perl, added \o{} to specify codepoints in octal, making it
670 possible to specify values greater than 0777 and also making them
671 unambiguous.
672
673 23. Perl now gives an error for missing closing braces after \x{... instead of
674 treating the string as literal. PCRE now does the same.
675
676 24. RunTest used to grumble if an inappropriate test was selected explicitly,
677 but just skip it when running all tests. This make it awkward to run ranges
678 of tests when one of them was inappropriate. Now it just skips any
679 inappropriate tests, as it always did when running all tests.
680
681 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
682 character types such as \d or \w, too many callouts were inserted, and the
683 data that they returned was rubbish.
684
685 26. In UCP mode, \s was not matching two of the characters that Perl matches,
686 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
687 were matched by \h. The code has now been refactored so that the lists of
688 the horizontal and vertical whitespace characters used for \h and \v (which
689 are defined only in one place) are now also used for \s.
690
691 27. Add JIT support for the 64 bit TileGX architecture.
692 Patch by Jiong Wang (Tilera Corporation).
693
694 28. Possessive quantifiers for classes (both explicit and automatically
695 generated) now use special opcodes instead of wrapping in ONCE brackets.
696
697 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
698 (because it's meaningless), this was not happening when PCRE_CASELESS was
699 set. Not wrong, but inefficient.
700
701 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
702 properties for \w, \d, etc) is present in a test regex. Otherwise if the
703 test contains no characters greater than 255, Perl doesn't realise it
704 should be using Unicode semantics.
705
706 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
707 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
708 does in Unicode mode.
709
710 32. Added the "forbid" facility to pcretest so that putting tests into the
711 wrong test files can sometimes be quickly detected.
712
713 33. There is now a limit (default 250) on the depth of nesting of parentheses.
714 This limit is imposed to control the amount of system stack used at compile
715 time. It can be changed at build time by --with-parens-nest-limit=xxx or
716 the equivalent in CMake.
717
718 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
719 errors. Perl warns for these when in warning mode, but PCRE has no facility
720 for giving warnings.
721
722 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
723 However, this was not working for (?!) because it is optimized to (*FAIL),
724 for which PCRE does not allow quantifiers. The optimization is now disabled
725 when a quantifier follows (?!). I can't see any use for this, but it makes
726 things uniform.
727
728 36. Perl no longer allows group names to start with digits, so I have made this
729 change also in PCRE. It simplifies the code a bit.
730
731 37. In extended mode, Perl ignores spaces before a + that indicates a
732 possessive quantifier. PCRE allowed a space before the quantifier, but not
733 before the possessive +. It now does.
734
735 38. The use of \K (reset reported match start) within a repeated possessive
736 group such as (a\Kb)*+ was not working.
737
738 40. Document that the same character tables must be used at compile time and
739 run time, and that the facility to pass tables to pcre_exec() and
740 pcre_dfa_exec() is for use only with saved/restored patterns.
741
742 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
743 features for Builds with MSVC:
744
745 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
746 builds). The .rc files can be used to set FileDescription and many other
747 attributes.
748
749 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
750 This allows higher-level build scripts which want .pdb files to avoid
751 hard-coding the exact files needed."
752
753 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
754 mean "start of word" and "end of word", respectively, as a transition aid.
755
756 43. A minimizing repeat of a class containing codepoints greater than 255 in
757 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
758 compiled to use the heap for recursion.
759
760 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
761 is configured.
762
763
764 Version 8.33 28-May-2013
765 ------------------------
766
767 1. Added 'U' to some constants that are compared to unsigned integers, to
768 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
769 variables that are added to signed variables, to ensure the result is
770 signed and can be negated.
771
772 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
773 CMake config files.
774
775 3. Revise the creation of config.h.generic so that all boolean macros are
776 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
777 overriding via -D on the command line possible.
778
779 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
780 to unsigned int is reported to make a quite noticeable speed difference in
781 a specific Windows environment. Testing on Linux did also appear to show
782 some benefit (and it is clearly not harmful). Also fixed the definition of
783 Xop which should be unsigned.
784
785 5. Related to (4), changing the definition of the intermediate variable cc
786 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
787 improvements.
788
789 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
790 unnecessary spaces.
791
792 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
793 and later.
794
795 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
796
797 9. Optimizing fast_forward_start_bits in JIT.
798
799 10. Adding support for callouts in JIT, and fixing some issues revealed
800 during this work. Namely:
801
802 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
803
804 (b) Minimum length was not checked before the matching is started.
805
806 11. The value of capture_last that is passed to callouts was incorrect in some
807 cases when there was a capture on one path that was subsequently abandoned
808 after a backtrack. Also, the capture_last value is now reset after a
809 recursion, since all captures are also reset in this case.
810
811 12. The interpreter no longer returns the "too many substrings" error in the
812 case when an overflowing capture is in a branch that is subsequently
813 abandoned after a backtrack.
814
815 13. In the pathological case when an offset vector of size 2 is used, pcretest
816 now prints out the matched string after a yield of 0 or 1.
817
818 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
819 Only supported by the JIT compiler at the moment.
820
821 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
822
823 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
824 offset of the starting point of the matching process, provided the offsets
825 vector is large enough.
826
827 17. The \A escape now records a lookbehind value of 1, though its execution
828 does not actually inspect the previous character. This is to ensure that,
829 in partial multi-segment matching, at least one character from the old
830 segment is retained when a new segment is processed. Otherwise, if there
831 are no lookbehinds in the pattern, \A might match incorrectly at the start
832 of a new segment.
833
834 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
835
836 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
837 gives some modest performance improvement in 8-bit mode.
838
839 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
840 be expressed in certain programming languages using Universal Character
841 Names.
842
843 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
844 which points out that "non characters" are not "characters that may not
845 appear in Unicode strings" but rather "characters that are reserved for
846 internal use and have only local meaning".
847
848 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
849 there was a conditional group that depended on an assertion, if the
850 assertion was false, the callout that immediately followed the alternation
851 in the condition was skipped when pcre_exec() was used for matching.
852
853 23. Allow an explicit callout to be inserted before an assertion that is the
854 condition for a conditional group, for compatibility with automatic
855 callouts, which always insert a callout at this point.
856
857 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
858 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
859
860 25. (*PRUNE) is now supported by the JIT compiler.
861
862 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
863
864 27. Fix the case where there are two or more SKIPs with arguments that may be
865 ignored.
866
867 28. (*SKIP) is now supported by the JIT compiler.
868
869 29. (*THEN) is now supported by the JIT compiler.
870
871 30. Update RunTest with additional test selector options.
872
873 31. The way PCRE handles backtracking verbs has been changed in two ways.
874
875 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
876 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
877 backtracking. In some cases this makes it more Perl-compatible, but Perl's
878 rather obscure rules do not always do the same thing.
879
880 (2) Previously, backtracking verbs were confined within assertions. This is
881 no longer the case for positive assertions, except for (*ACCEPT). Again,
882 this sometimes improves Perl compatibility, and sometimes does not.
883
884 32. A number of tests that were in test 2 because Perl did things differently
885 have been moved to test 1, because either Perl or PCRE has changed, and
886 these tests are now compatible.
887
888 32. Backtracking control verbs are now handled in the same way in JIT and
889 interpreter.
890
891 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
892 contained a forward subroutine reference caused a compile error.
893
894 34. Auto-detect and optimize limited repetitions in JIT.
895
896 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
897 blocking (*UTF) etc.
898
899 36. In the interpreter, maximizing pattern repetitions for characters and
900 character types now use tail recursion, which reduces stack usage.
901
902 37. The value of the max lookbehind was not correctly preserved if a compiled
903 and saved regex was reloaded on a host of different endianness.
904
905 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
906 of the compiled pattern block, expand the flags field from 16 to 32 bits
907 because it was almost full.
908
909 39. Try madvise first before posix_madvise.
910
911 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
912 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
913 original change locked out all matches of empty strings. This has been
914 changed so that one match of an empty string per line is recognized.
915 Subsequent searches on the same line (for colouring or for --only-matching,
916 for example) do not recognize empty strings.
917
918 41. Applied a user patch to fix a number of spelling mistakes in comments.
919
920 42. Data lines longer than 65536 caused pcretest to crash.
921
922 43. Clarified the data type for length and startoffset arguments for pcre_exec
923 and pcre_dfa_exec in the function-specific man pages, where they were
924 explicitly stated to be in bytes, never having been updated. I also added
925 some clarification to the pcreapi man page.
926
927 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
928 a segmentation fault.
929
930
931 Version 8.32 30-November-2012
932 -----------------------------
933
934 1. Improved JIT compiler optimizations for first character search and single
935 character iterators.
936
937 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
938 Patch by Daniel Richard G.
939
940 3. Single character iterator optimizations in the JIT compiler.
941
942 4. Improved JIT compiler optimizations for character ranges.
943
944 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
945 Reported by Giuseppe D'Angelo.
946
947 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
948 of a line, was being set incorrectly in cases where .* appeared inside
949 atomic brackets at the start of a pattern, or where there was a subsequent
950 *PRUNE or *SKIP.
951
952 7. Improved instruction cache flush for POWER/PowerPC.
953 Patch by Daniel Richard G.
954
955 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
956 grep:
957
958 (a) There is now no limit to the number of patterns to be matched.
959
960 (b) An error is given if a pattern is too long.
961
962 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
963 are now supported.
964
965 (d) --exclude-from and --include-from (multiple use) have been added.
966
967 (e) Exclusions and inclusions now apply to all files and directories, not
968 just to those obtained from scanning a directory recursively.
969
970 (f) Multiple uses of -f and --file-list are now supported.
971
972 (g) In a Windows environment, the default for -d has been changed from
973 "read" (the GNU grep default) to "skip", because otherwise the presence
974 of a directory in the file list provokes an error.
975
976 (h) The documentation has been revised and clarified in places.
977
978 9. Improve the matching speed of capturing brackets.
979
980 10. Changed the meaning of \X so that it now matches a Unicode extended
981 grapheme cluster.
982
983 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
984 out POSIX threads when JIT support is configured.
985
986 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
987
988 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
989 structure could go wrong in environments where size_t is not the same size
990 as int.
991
992 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
993 set.
994
995 15. The EBCDIC support had decayed; later updates to the code had included
996 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
997 general tidy up of EBCDIC-related issues, and the documentation was also
998 not quite right. There is now a test that can be run on ASCII systems to
999 check some of the EBCDIC-related things (but is it not a full test).
1000
1001 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1002 in a small tidy to the code.
1003
1004 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1005
1006 18. If the --only-matching (-o) option in pcregrep is specified multiple
1007 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1008 substrings matched by the 1st and 2nd capturing parentheses. A separating
1009 string can be specified by --om-separator (default empty).
1010
1011 19. Improving the first n character searches.
1012
1013 20. Turn case lists for horizontal and vertical white space into macros so that
1014 they are defined only once.
1015
1016 21. This set of changes together give more compatible Unicode case-folding
1017 behaviour for characters that have more than one other case when UCP
1018 support is available.
1019
1020 (a) The Unicode property table now has offsets into a new table of sets of
1021 three or more characters that are case-equivalent. The MultiStage2.py
1022 script that generates these tables (the pcre_ucd.c file) now scans
1023 CaseFolding.txt instead of UnicodeData.txt for character case
1024 information.
1025
1026 (b) The code for adding characters or ranges of characters to a character
1027 class has been abstracted into a generalized function that also handles
1028 case-independence. In UTF-mode with UCP support, this uses the new data
1029 to handle characters with more than one other case.
1030
1031 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1032 whose other case is greater than 256 are now correctly matched
1033 caselessly. Previously, the high codepoint matched the low one, but not
1034 vice versa.
1035
1036 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1037 of the new class addition function, using character lists defined as
1038 macros alongside the case definitions of 20 above.
1039
1040 (e) Caseless back references now work with characters that have more than
1041 one other case.
1042
1043 (f) General caseless matching of characters with more than one other case
1044 is supported.
1045
1046 22. Unicode character properties were updated from Unicode 6.2.0
1047
1048 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1049
1050 24. Add support for 32-bit character strings, and UTF-32
1051
1052 25. Major JIT compiler update (code refactoring and bugfixing).
1053 Experimental Sparc 32 support is added.
1054
1055 26. Applied a modified version of Daniel Richard G's patch to create
1056 pcre.h.generic and config.h.generic by "make" instead of in the
1057 PrepareRelease script.
1058
1059 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1060 pcre_compile.c when checking for a zero character.
1061
1062 28. Introducing a native interface for JIT. Through this interface, the compiled
1063 machine code can be directly executed. The purpose of this interface is to
1064 provide fast pattern matching, so several sanity checks are not performed.
1065 However, feature tests are still performed. The new interface provides
1066 1.4x speedup compared to the old one.
1067
1068 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1069 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1070 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1071
1072 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1073 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1074 the "old" RFC 2279). Instead, it ended up passing a negative length to
1075 pcre_exec().
1076
1077 31. Add support for GCC's visibility feature to hide internal functions.
1078
1079 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1080 "unknown -C option" after outputting 0 or 1.
1081
1082 33. There is now support for generating a code coverage report for the test
1083 suite in environments where gcc is the compiler and lcov is installed. This
1084 is mainly for the benefit of the developers.
1085
1086 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1087 unaddressable using valgrind annotations, allowing valgrind to detect
1088 invalid memory accesses. This is mainly for the benefit of the developers.
1089
1090 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1091
1092 26. Give configure error if --enable-cpp but no C++ compiler found.
1093
1094
1095 Version 8.31 06-July-2012
1096 -------------------------
1097
1098 1. Fixing a wrong JIT test case and some compiler warnings.
1099
1100 2. Removed a bashism from the RunTest script.
1101
1102 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1103 to unsigned type, result still unsigned" that was given by an MS compiler
1104 on encountering the code "-sizeof(xxx)".
1105
1106 4. Partial matching support is added to the JIT compiler.
1107
1108 5. Fixed several bugs concerned with partial matching of items that consist
1109 of more than one character:
1110
1111 (a) /^(..)\1/ did not partially match "aba" because checking references was
1112 done on an "all or nothing" basis. This also applied to repeated
1113 references.
1114
1115 (b) \R did not give a hard partial match if \r was found at the end of the
1116 subject.
1117
1118 (c) \X did not give a hard partial match after matching one or more
1119 characters at the end of the subject.
1120
1121 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1122 a partial match for the string "\r".
1123
1124 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1125 a partial match for a CR character at the end of the subject string.
1126
1127 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1128 running pcretest, the text "(JIT)" added to the output whenever JIT is
1129 actually used to run the match.
1130
1131 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1132 or /S+[+] with a digit between 1 and 7.
1133
1134 8. OP_NOT now supports any UTF character not just single-byte ones.
1135
1136 9. (*MARK) control verb is now supported by the JIT compiler.
1137
1138 10. The command "./RunTest list" lists the available tests without actually
1139 running any of them. (Because I keep forgetting what they all are.)
1140
1141 11. Add PCRE_INFO_MAXLOOKBEHIND.
1142
1143 12. Applied a (slightly modified) user-supplied patch that improves performance
1144 when the heap is used for recursion (compiled with --disable-stack-for-
1145 recursion). Instead of malloc and free for each heap frame each time a
1146 logical recursion happens, frames are retained on a chain and re-used where
1147 possible. This sometimes gives as much as 30% improvement.
1148
1149 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1150 call.
1151
1152 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1153
1154 15. It is now possible to link pcretest with libedit as an alternative to
1155 libreadline.
1156
1157 16. (*COMMIT) control verb is now supported by the JIT compiler.
1158
1159 17. The Unicode data tables have been updated to Unicode 6.1.0.
1160
1161 18. Added --file-list option to pcregrep.
1162
1163 19. Added binary file support to pcregrep, including the -a, --binary-files,
1164 -I, and --text options.
1165
1166 20. The madvise function is renamed for posix_madvise for QNX compatibility
1167 reasons. Fixed by Giuseppe D'Angelo.
1168
1169 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1170
1171 22. Changed the option for creating symbolic links for 16-bit man pages from
1172 -s to -sf so that re-installing does not cause issues.
1173
1174 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1175
1176 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1177 matches in certain environments (the workspace was not being correctly
1178 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1179 some of the workspace data at the beginning of a restart.
1180
1181 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1182 was not doing so when it should - probably a typo introduced by SVN 528
1183 (change 8.10/14).
1184
1185 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1186 \w+ when the character tables indicated that \x{c4} was a word character.
1187 There were several related cases, all because the tests for doing a table
1188 lookup were testing for characters less than 127 instead of 255.
1189
1190 27. If a pattern contains capturing parentheses that are not used in a match,
1191 their slots in the ovector are set to -1. For those that are higher than
1192 any matched groups, this happens at the end of processing. In the case when
1193 there were back references that the ovector was too small to contain
1194 (causing temporary malloc'd memory to be used during matching), and the
1195 highest capturing number was not used, memory off the end of the ovector
1196 was incorrectly being set to -1. (It was using the size of the temporary
1197 memory instead of the true size.)
1198
1199 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1200 ovector size, it uses memory at the end of the block that it has got.
1201
1202 29. Check for an overlong MARK name and give an error at compile time. The
1203 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1204
1205 30. JIT compiler update.
1206
1207 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1208 Rill for the patch.
1209
1210 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1211
1212 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1213
1214 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1215 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1216 was enabled.
1217
1218 35. Improve JIT code generation for greedy plus quantifier.
1219
1220 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1221 1 to "aa" instead of to an empty string. The bug affected repeated groups
1222 that could potentially match an empty string.
1223
1224 37. Optimizing single character iterators in JIT.
1225
1226 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1227 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1228 codepoints that are too big for the mode are faulted, and in a UTF mode,
1229 disallowed codepoints are also faulted.
1230
1231 39. If PCRE was compiled with UTF support, in three places in the DFA
1232 matcher there was code that should only have been obeyed in UTF mode, but
1233 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1234 processing when bytes with values greater than 127 were present. In 16-bit
1235 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1236 both cases the values are those that cannot be the first data item in a UTF
1237 character. The three items that might have provoked this were recursions,
1238 possessively repeated groups, and atomic groups.
1239
1240 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1241 and pcregrep, because some OS require shared objects to be explicitly
1242 passed to ld, causing the link step to fail if they are not.
1243
1244 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1245 mode, patterns that started with \h* or \R* might be incorrectly matched.
1246
1247
1248 Version 8.30 04-February-2012
1249 -----------------------------
1250
1251 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1252 name is defined in ctype.h.
1253
1254 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1255 only in quite long subpatterns.
1256
1257 3. Removed the function pcre_info(), which has been obsolete and deprecated
1258 since it was replaced by pcre_fullinfo() in February 2000.
1259
1260 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1261 match a (*MARK), and the match failed at the start of the subject, a
1262 reference to memory before the start of the subject could occur. This bug
1263 was introduced by fix 17 of release 8.21.
1264
1265 5. A reference to an unset group with zero minimum repetition was giving
1266 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1267 /(another)?(\1?)test/ matched against "hello world test". This bug was
1268 introduced in release 8.13.
1269
1270 6. Add support for 16-bit character strings (a large amount of work involving
1271 many changes and refactorings).
1272
1273 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1274 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1275 from a file.
1276
1277 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1278 rounding is not applied in this particular case).
1279
1280 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1281 if they appear, or are escaped, in patterns.
1282
1283 10. Get rid of a number of -Wunused-but-set-variable warnings.
1284
1285 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1286 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1287 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1288 also returns the mark "x". This bug applied to capturing parentheses,
1289 non-capturing parentheses, and atomic parentheses. It also applied to some
1290 assertions.
1291
1292 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1293 information out of configure.ac instead of relying on pcre.h.generic, which
1294 is not stored in the repository.
1295
1296 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1297 -lreadline.
1298
1299 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1300
1301 15. Applied Graycode's patch to put the top-level frame on the stack rather
1302 than the heap when not using the stack for recursion. This gives a
1303 performance improvement in many cases when recursion is not deep.
1304
1305 16. Experimental code added to "pcretest -C" to output the stack frame size.
1306
1307
1308 Version 8.21 12-Dec-2011
1309 ------------------------
1310
1311 1. Updating the JIT compiler.
1312
1313 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1314 are added as well.
1315
1316 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1317 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1318 calling _pcre_jit_exec. Some extra comments are added.
1319
1320 4. (*MARK) settings inside atomic groups that do not contain any capturing
1321 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1322 was introduced by change 18 for 8.20.
1323
1324 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1325 ECMA-262 standard.
1326
1327 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1328 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1329 This bug was probably introduced by change 9 of 8.13.
1330
1331 7. While fixing 6 above, I noticed that a number of other items were being
1332 incorrectly rejected as "not fixed length". This arose partly because newer
1333 opcodes had not been added to the fixed-length checking code. I have (a)
1334 corrected the bug and added tests for these items, and (b) arranged for an
1335 error to occur if an unknown opcode is encountered while checking for fixed
1336 length instead of just assuming "not fixed length". The items that were
1337 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1338 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1339 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1340
1341 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1342 being incorrectly compiled and would have given unpredicatble results.
1343
1344 9. A possessively repeated subpattern with minimum repeat count greater than
1345 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1346 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1347 the first (A) could occur when it should not.
1348
1349 10. Add a cast and remove a redundant test from the code.
1350
1351 11. JIT should use pcre_malloc/pcre_free for allocation.
1352
1353 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1354 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1355 is anything other than /usr, -L is still shown).
1356
1357 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1358
1359 14. Perl does not support \N without a following name in a [] class; PCRE now
1360 also gives an error.
1361
1362 15. If a forward reference was repeated with an upper limit of around 2000,
1363 it caused the error "internal error: overran compiling workspace". The
1364 maximum number of forward references (including repeats) was limited by the
1365 internal workspace, and dependent on the LINK_SIZE. The code has been
1366 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1367 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1368 of around 200,000 forward references. While doing this, I also speeded up
1369 the filling in of repeated forward references.
1370
1371 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1372 incorrectly expecting the subject to contain another "a" after the start.
1373
1374 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1375 in the match, the SKIP should be ignored. This was not happening; instead
1376 the SKIP was being treated as NOMATCH. For patterns such as
1377 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1378 tested.
1379
1380 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1381 now much more compatible with Perl, in particular in cases where the result
1382 is a non-match for a non-anchored pattern. For example, if
1383 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1384 "m", where previously it did not return a name. A side effect of this
1385 change is that for partial matches, the last encountered mark name is
1386 returned, as for non matches. A number of tests that were previously not
1387 Perl-compatible have been moved into the Perl-compatible test files. The
1388 refactoring has had the pleasing side effect of removing one argument from
1389 the match() function, thus reducing its stack requirements.
1390
1391 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1392 subsequent uses of /S (without +) incorrectly behaved like /S+.
1393
1394 21. Retrieve executable code size support for the JIT compiler and fixing
1395 some warnings.
1396
1397 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1398 not work when the shorter character appeared right at the end of the
1399 subject string.
1400
1401 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1402 systems.
1403
1404 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1405 output it when the /M option is used in pcretest.
1406
1407 25. The CheckMan script was not being included in the distribution. Also, added
1408 an explicit "perl" to run Perl scripts from the PrepareRelease script
1409 because this is reportedly needed in Windows.
1410
1411 26. If study data was being save in a file and studying had not found a set of
1412 "starts with" bytes for the pattern, the data written to the file (though
1413 never used) was taken from uninitialized memory and so caused valgrind to
1414 complain.
1415
1416 27. Updated RunTest.bat as provided by Sheri Pierce.
1417
1418 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1419
1420 29. Computation of memory usage for the table of capturing group names was
1421 giving an unnecessarily large value.
1422
1423
1424 Version 8.20 21-Oct-2011
1425 ------------------------
1426
1427 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1428 a POSIX class. After further experiments with Perl, which convinced me that
1429 Perl has bugs and confusions, a closing square bracket is no longer allowed
1430 in a POSIX name. This bug also affected patterns with classes that started
1431 with full stops.
1432
1433 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1434 captured substring, but while checking the failing first alternative,
1435 substring 1 is temporarily captured. If the output vector supplied to
1436 pcre_exec() was not big enough for this capture, the yield of the function
1437 was still zero ("insufficient space for captured substrings"). This cannot
1438 be totally fixed without adding another stack variable, which seems a lot
1439 of expense for a edge case. However, I have improved the situation in cases
1440 such as /(a)(b)x|abc/ matched against "abc", where the return code
1441 indicates that fewer than the maximum number of slots in the ovector have
1442 been set.
1443
1444 3. Related to (2) above: when there are more back references in a pattern than
1445 slots in the output vector, pcre_exec() uses temporary memory during
1446 matching, and copies in the captures as far as possible afterwards. It was
1447 using the entire output vector, but this conflicts with the specification
1448 that only 2/3 is used for passing back captured substrings. Now it uses
1449 only the first 2/3, for compatibility. This is, of course, another edge
1450 case.
1451
1452 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1453 main code base, and can be used by building with --enable-jit. When this is
1454 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1455 runtime --no-jit option is given.
1456
1457 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1458 ovector, the return from the function was zero, implying that there were
1459 other matches that did not fit. The correct "exactly full" value is now
1460 returned.
1461
1462 6. If a subpattern that was called recursively or as a subroutine contained
1463 (*PRUNE) or any other control that caused it to give a non-standard return,
1464 invalid errors such as "Error -26 (nested recursion at the same subject
1465 position)" or even infinite loops could occur.
1466
1467 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1468 computing the minimum length on reaching *ACCEPT, and so ended up with the
1469 wrong value of 1 rather than 0. Further investigation indicates that
1470 computing a minimum subject length in the presence of *ACCEPT is difficult
1471 (think back references, subroutine calls), and so I have changed the code
1472 so that no minimum is registered for a pattern that contains *ACCEPT.
1473
1474 8. If (*THEN) was present in the first (true) branch of a conditional group,
1475 it was not handled as intended. [But see 16 below.]
1476
1477 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1478 Sheri Pierce.
1479
1480 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1481 the first byte in a match must be "a".
1482
1483 11. Change 17 for 8.13 increased the recursion depth for patterns like
1484 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1485 pattern contains any instances of (*THEN). If it does not, the old
1486 optimizations are restored. It would be nice to do this on a per-group
1487 basis, but at the moment that is not feasible.
1488
1489 12. In some environments, the output of pcretest -C is CRLF terminated. This
1490 broke RunTest's code that checks for the link size. A single white space
1491 character after the value is now allowed for.
1492
1493 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1494 For "fr", it uses the Windows-specific input and output files.
1495
1496 14. If (*THEN) appeared in a group that was called recursively or as a
1497 subroutine, it did not work as intended. [But see next item.]
1498
1499 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1500 pattern fragments (but not containing any | characters). If A and B are
1501 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1502 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1503 D. In other words, Perl considers parentheses that do not contain any |
1504 characters to be part of a surrounding alternative, whereas PCRE was
1505 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1506 differently. PCRE now behaves in the same way as Perl, except in the case
1507 of subroutine/recursion calls such as (?1) which have in any case always
1508 been different (but PCRE had them first :-).
1509
1510 16. Related to 15 above: Perl does not treat the | in a conditional group as
1511 creating alternatives. Such a group is treated in the same way as an
1512 ordinary group without any | characters when processing (*THEN). PCRE has
1513 been changed to match Perl's behaviour.
1514
1515 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1516 RunGrepTest script failed.
1517
1518 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1519 inevitable for groups that contain captures, but it can lead to a lot of
1520 stack use in large patterns. The old behaviour has been restored for atomic
1521 groups that do not contain any capturing parentheses.
1522
1523 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1524 suppress the check for a minimum subject length at run time. (If it was
1525 given to pcre_exec() or pcre_dfa_exec() it did work.)
1526
1527 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1528 fail to work when decoding hex characters in data strings in EBCDIC
1529 environments.
1530
1531 21. It appears that in at least one Mac OS environment, the isxdigit() function
1532 is implemented as a macro that evaluates to its argument more than once,
1533 contravening the C 90 Standard (I haven't checked a later standard). There
1534 was an instance in pcretest which caused it to go wrong when processing
1535 \x{...} escapes in subject strings. The has been rewritten to avoid using
1536 things like p++ in the argument of isxdigit().
1537
1538
1539 Version 8.13 16-Aug-2011
1540 ------------------------
1541
1542 1. The Unicode data tables have been updated to Unicode 6.0.0.
1543
1544 2. Two minor typos in pcre_internal.h have been fixed.
1545
1546 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1547 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1548 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1549
1550 4. There were a number of related bugs in the code for matching backrefences
1551 caselessly in UTF-8 mode when codes for the characters concerned were
1552 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1553 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1554 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1555 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1556 2-byte code at the end of the subject (it thought there wasn't enough data
1557 left).
1558
1559 5. Comprehensive information about what went wrong is now returned by
1560 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1561 as the output vector has at least 2 elements. The offset of the start of
1562 the failing character and a reason code are placed in the vector.
1563
1564 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1565 now returned is for the first byte of the failing character, instead of the
1566 last byte inspected. This is an incompatible change, but I hope it is small
1567 enough not to be a problem. It makes the returned offset consistent with
1568 pcre_exec() and pcre_dfa_exec().
1569
1570 7. pcretest now gives a text phrase as well as the error number when
1571 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1572 failure, the offset and reason code are output.
1573
1574 8. When \R was used with a maximizing quantifier it failed to skip backwards
1575 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1576 back over a single character (\n). This seems wrong (because it treated the
1577 two characters as a single entity when going forwards), conflicts with the
1578 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1579 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1580 has been changed.
1581
1582 9. Some internal refactoring has changed the processing so that the handling
1583 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1584 time (the PCRE_DOTALL option was changed this way some time ago: version
1585 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1586 which was always a bit of a fudge. It also means that there is one less
1587 argument for the match() function, which reduces its stack requirements
1588 slightly. This change also fixes an incompatibility with Perl: the pattern
1589 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1590
1591 10. More internal refactoring has drastically reduced the number of recursive
1592 calls to match() for possessively repeated groups such as (abc)++ when
1593 using pcre_exec().
1594
1595 11. While implementing 10, a number of bugs in the handling of groups were
1596 discovered and fixed:
1597
1598 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1599 (a|)*(?1) gave a compile-time internal error.
1600 ((a|)+)+ did not notice that the outer group could match an empty string.
1601 (^a|^)+ was not marked as anchored.
1602 (.*a|.*)+ was not marked as matching at start or after a newline.
1603
1604 12. Yet more internal refactoring has removed another argument from the match()
1605 function. Special calls to this function are now indicated by setting a
1606 value in a variable in the "match data" data block.
1607
1608 13. Be more explicit in pcre_study() instead of relying on "default" for
1609 opcodes that mean there is no starting character; this means that when new
1610 ones are added and accidentally left out of pcre_study(), testing should
1611 pick them up.
1612
1613 14. The -s option of pcretest has been documented for ages as being an old
1614 synonym of -m (show memory usage). I have changed it to mean "force study
1615 for every regex", that is, assume /S for every regex. This is similar to -i
1616 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1617 using it. It makes it easier to run collections of tests with and without
1618 study enabled, and thereby test pcre_study() more easily. All the standard
1619 tests are now run with and without -s (but some patterns can be marked as
1620 "never study" - see 20 below).
1621
1622 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1623 restoration of the capturing data to the outer values was not happening
1624 correctly.
1625
1626 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1627 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1628 pattern had matched an empty string, and so incorrectly returned a no
1629 match.
1630
1631 17. There was optimizing code for the last branch of non-capturing parentheses,
1632 and also for the obeyed branch of a conditional subexpression, which used
1633 tail recursion to cut down on stack usage. Unfortunately, now that there is
1634 the possibility of (*THEN) occurring in these branches, tail recursion is
1635 no longer possible because the return has to be checked for (*THEN). These
1636 two optimizations have therefore been removed. [But see 8.20/11 above.]
1637
1638 18. If a pattern containing \R was studied, it was assumed that \R always
1639 matched two bytes, thus causing the minimum subject length to be
1640 incorrectly computed because \R can also match just one byte.
1641
1642 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1643 was incorrectly computed.
1644
1645 20. If /S is present twice on a test pattern in pcretest input, it now
1646 *disables* studying, thereby overriding the use of -s on the command line
1647 (see 14 above). This is necessary for one or two tests to keep the output
1648 identical in both cases.
1649
1650 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1651 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1652
1653 22. When an atomic group that contained a capturing parenthesis was
1654 successfully matched, but the branch in which it appeared failed, the
1655 capturing was not being forgotten if a higher numbered group was later
1656 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1657 group 1 to "a", when in fact it should be unset. This applied to multi-
1658 branched capturing and non-capturing groups, repeated or not, and also to
1659 positive assertions (capturing in negative assertions does not happen
1660 in PCRE) and also to nested atomic groups.
1661
1662 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1663 subject after a captured substring, to make it easier to tell which of a
1664 number of identical substrings has been captured.
1665
1666 24. The way atomic groups are processed by pcre_exec() has been changed so that
1667 if they are repeated, backtracking one repetition now resets captured
1668 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1669 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1670 "aaa". Previously, it would have been "a". As part of this code
1671 refactoring, the way recursive calls are handled has also been changed.
1672
1673 25. If an assertion condition captured any substrings, they were not passed
1674 back unless some other capturing happened later. For example, if
1675 (?(?=(a))a) was matched against "a", no capturing was returned.
1676
1677 26. When studying a pattern that contained subroutine calls or assertions,
1678 the code for finding the minimum length of a possible match was handling
1679 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1680 group 1 called group 2 while simultaneously a separate group 2 called group
1681 1). A stack overflow occurred in this case. I have fixed this by limiting
1682 the recursion depth to 10.
1683
1684 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1685 Fortmann. This supports explicit test numbers on the command line, and has
1686 argument validation and error reporting.
1687
1688 28. An instance of \X with an unlimited repeat could fail if at any point the
1689 first character it looked at was a mark character.
1690
1691 29. Some minor code refactoring concerning Unicode properties and scripts
1692 should reduce the stack requirement of match() slightly.
1693
1694 30. Added the '=' option to pcretest to check the setting of unused capturing
1695 slots at the end of the pattern, which are documented as being -1, but are
1696 not included in the return count.
1697
1698 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1699 compiled something random. Now it gives a compile-time error (as does
1700 Perl).
1701
1702 32. A *MARK encountered during the processing of a positive assertion is now
1703 recorded and passed back (compatible with Perl).
1704
1705 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1706 had alternative anchored branches, the search for a second match in a line
1707 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1708 matched the line "0102" twice. The same bug affected patterns that started
1709 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1710 twice.
1711
1712 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1713 does, and because of capturing effects, quantifying parenthesized
1714 assertions may at times be useful. Quantifiers are now allowed for
1715 parenthesized assertions.
1716
1717 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1718
1719 36. \g was being checked for fancy things in a character class, when it should
1720 just be a literal "g".
1721
1722 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1723 appearance of a nested POSIX class supersedes an apparent external class.
1724 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1725 unescaped square brackets may also appear as part of class names. For
1726 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1727 more like Perl. (But see 8.20/1 above.)
1728
1729 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1730 was because it thought it was \N{name}, which is not supported).
1731
1732 39. Add minix to OS list not supporting the -S option in pcretest.
1733
1734 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1735 cannot analyze patterns in sufficient detail to catch mutual recursions
1736 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1737 subgroup is called recursively as a subpattern for a second time at the
1738 same position in the subject string. In previous releases this might have
1739 been caught by the recursion limit, or it might have run out of stack.
1740
1741 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1742 happen only once. PCRE was, however incorrectly giving a compile time error
1743 "recursive call could loop indefinitely" because it cannot analyze the
1744 pattern in sufficient detail. The compile time test no longer happens when
1745 PCRE is compiling a conditional subpattern, but actual runaway loops are
1746 now caught at runtime (see 40 above).
1747
1748 42. It seems that Perl allows any characters other than a closing parenthesis
1749 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1750 has been changed to be the same.
1751
1752 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1753 as not to get warnings when autogen.sh is called. Also changed
1754 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1755
1756 44. To help people who use pcregrep to scan files containing exceedingly long
1757 lines, the following changes have been made:
1758
1759 (a) The default value of the buffer size parameter has been increased from
1760 8K to 20K. (The actual buffer used is three times this size.)
1761
1762 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1763 PCRE is built.
1764
1765 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1766 to be set at run time.
1767
1768 (d) Numerical values in pcregrep options can be followed by K or M, for
1769 example --buffer-size=50K.
1770
1771 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1772 given and the return code is set to 2.
1773
1774 45. Add a pointer to the latest mark to the callout data block.
1775
1776 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1777 partial match of an empty string instead of no match. This was specific to
1778 the use of ".".
1779
1780 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1781 complete match instead of a partial match. This bug was dependent on both
1782 the PCRE_UTF8 and PCRE_DOTALL options being set.
1783
1784 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1785 starting byte set, because \b was not being ignored.
1786
1787
1788 Version 8.12 15-Jan-2011
1789 ------------------------
1790
1791 1. Fixed some typos in the markup of the man pages, and wrote a script that
1792 checks for such things as part of the documentation building process.
1793
1794 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1795 --match-limit and --recursion-limit options (added for 8.11). In
1796 particular, this made one of the standard tests fail. (The integer value
1797 went into the wrong half of a long int.)
1798
1799 3. If the --colour option was given to pcregrep with -v (invert match), it
1800 did strange things, either producing crazy output, or crashing. It should,
1801 of course, ignore a request for colour when reporting lines that do not
1802 match.
1803
1804 4. Another pcregrep bug caused similar problems if --colour was specified with
1805 -M (multiline) and the pattern match finished with a line ending.
1806
1807 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1808 matched in multiline mode, the following line was shown as part of the
1809 match. This seems wrong, so I have changed it.
1810
1811 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1812 the check for further matches in the same line (so they could be coloured)
1813 to overrun the end of the current line. If another match was found, it was
1814 incorrectly shown (and then shown again when found in the next line).
1815
1816 7. If pcregrep was compiled under Windows, there was a reference to the
1817 function pcregrep_exit() before it was defined. I am assuming this was
1818 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1819 reported by a user. I've moved the definition above the reference.
1820
1821
1822 Version 8.11 10-Dec-2010
1823 ------------------------
1824
1825 1. (*THEN) was not working properly if there were untried alternatives prior
1826 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1827 backtracked to try for "b" instead of moving to the next alternative branch
1828 at the same level (in this case, to look for "c"). The Perl documentation
1829 is clear that when (*THEN) is backtracked onto, it goes to the "next
1830 alternative in the innermost enclosing group".
1831
1832 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1833 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1834 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1835 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1836 (*THEN).
1837
1838 3. If \s appeared in a character class, it removed the VT character from
1839 the class, even if it had been included by some previous item, for example
1840 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1841 of \s, but is part of the POSIX "space" class.)
1842
1843 4. A partial match never returns an empty string (because you can always
1844 match an empty string at the end of the subject); however the checking for
1845 an empty string was starting at the "start of match" point. This has been
1846 changed to the "earliest inspected character" point, because the returned
1847 data for a partial match starts at this character. This means that, for
1848 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1849 (previously it gave "no match").
1850
1851 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1852 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1853 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1854 has an implication that the given string is incomplete (because a partial
1855 match is preferred over a full match). For this reason, these items now
1856 give a partial match in this situation. [Aside: previously, the one case
1857 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1858 match rather than a full match, which was wrong by the old rules, but is
1859 now correct.]
1860
1861 6. There was a bug in the handling of #-introduced comments, recognized when
1862 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1863 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1864 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1865 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1866 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1867 places in pcre_compile().
1868
1869 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1870 comments when looking ahead for named forward references to subpatterns,
1871 the only newline sequence it recognized was NL. It now handles newlines
1872 according to the set newline convention.
1873
1874 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1875 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1876 cater for a lack of strerror(). These oversights have been fixed.
1877
1878 9. Added --match-limit and --recursion-limit to pcregrep.
1879
1880 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1881
1882 11. When the -o option was used, pcregrep was setting a return code of 1, even
1883 when matches were found, and --line-buffered was not being honoured.
1884
1885 12. Added an optional parentheses number to the -o and --only-matching options
1886 of pcregrep.
1887
1888 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1889 can match an empty string. The code to do it in pcretest and pcredemo
1890 needed fixing:
1891
1892 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1893 only one byte after an empty string match just before CRLF (this case
1894 just got forgotten; "any" and "anycrlf" were OK).
1895
1896 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1897 mode when an empty string match preceded an ASCII character followed by
1898 a non-ASCII character. (The code for advancing by one character rather
1899 than one byte was nonsense.)
1900
1901 (c) The pcredemo.c sample program did not have any code at all to handle
1902 the cases when CRLF is a valid newline sequence.
1903
1904 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1905 as a starting offset was within the subject string. There is now a new
1906 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1907 negative or greater than the length of the string. In order to test this,
1908 pcretest is extended to allow the setting of negative starting offsets.
1909
1910 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1911 starting offset points to the beginning of a UTF-8 character was
1912 unnecessarily clumsy. I tidied it up.
1913
1914 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1915 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1916
1917 17. Nobody had reported that the --include_dir option, which was added in
1918 release 7.7 should have been called --include-dir (hyphen, not underscore)
1919 for compatibility with GNU grep. I have changed it to --include-dir, but
1920 left --include_dir as an undocumented synonym, and the same for
1921 --exclude-dir, though that is not available in GNU grep, at least as of
1922 release 2.5.4.
1923
1924 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1925 characters from a string of bytes) have been redefined so as not to use
1926 loops, in order to improve performance in some environments. At the same
1927 time, I abstracted some of the common code into auxiliary macros to save
1928 repetition (this should not affect the compiled code).
1929
1930 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1931 compile-time error is now given if \c is not followed by an ASCII
1932 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1933 different, and any byte value is allowed.)
1934
1935 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1936 START_OPTIMIZE option, which is now allowed at compile time - but just
1937 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1938 to pcregrep and other applications that have no direct access to PCRE
1939 options. The new /Y option in pcretest sets this option when calling
1940 pcre_compile().
1941
1942 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1943 back references. Groups containing recursive back references were forced to
1944 be atomic by that change, but in the case of named groups, the amount of
1945 memory required was incorrectly computed, leading to "Failed: internal
1946 error: code overflow". This has been fixed.
1947
1948 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1949 pcretest.c, to avoid build problems in some Borland environments.
1950
1951
1952 Version 8.10 25-Jun-2010
1953 ------------------------
1954
1955 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1956 THEN.
1957
1958 2. (*ACCEPT) was not working when inside an atomic group.
1959
1960 3. Inside a character class, \B is treated as a literal by default, but
1961 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1962 causes the error). The code is unchanged, but I tidied the documentation.
1963
1964 4. Inside a character class, PCRE always treated \R and \X as literals,
1965 whereas Perl faults them if its -w option is set. I have changed PCRE so
1966 that it faults them when PCRE_EXTRA is set.
1967
1968 5. Added support for \N, which always matches any character other than
1969 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1970
1971 6. When compiling pcregrep with newer versions of gcc which may have
1972 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1973 declared with attribute warn_unused_result" were given. Just casting the
1974 result to (void) does not stop the warnings; a more elaborate fudge is
1975 needed. I've used a macro to implement this.
1976
1977 7. Minor change to pcretest.c to avoid a compiler warning.
1978
1979 8. Added four artifical Unicode properties to help with an option to make
1980 \s etc use properties (see next item). The new properties are: Xan
1981 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1982
1983 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1984 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1985 this option. Modified pcretest to add /W to test this facility. Added
1986 REG_UCP to make it available via the POSIX interface.
1987
1988 10. Added --line-buffered to pcregrep.
1989
1990 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1991 studied, and the match started with a letter with a code point greater than
1992 127 whose first byte was different to the first byte of the other case of
1993 the letter, the other case of this starting letter was not recognized
1994 (#976).
1995
1996 12. If a pattern that was studied started with a repeated Unicode property
1997 test, for example, \p{Nd}+, there was the theoretical possibility of
1998 setting up an incorrect bitmap of starting bytes, but fortunately it could
1999 not have actually happened in practice until change 8 above was made (it
2000 added property types that matched character-matching opcodes).
2001
2002 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2003 possible starting bytes for non-anchored patterns.
2004
2005 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2006 \R, and also a number of cases that involve Unicode properties, both
2007 explicit and implicit when PCRE_UCP is set.
2008
2009 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2010 input, it could crash or give wrong results if characters with values
2011 greater than 0xc0 were present in the subject string. (Detail: it assumed
2012 UTF-8 input when processing these items.)
2013
2014 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2015 size_t is 64-bit (#991).
2016
2017 17. Added a check for running out of memory when PCRE is compiled with
2018 --disable-stack-for-recursion (#990).
2019
2020 18. If the last data line in a file for pcretest does not have a newline on
2021 the end, a newline was missing in the output.
2022
2023 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2024 less than 128) in its various bitmaps. However, there is a facility for
2025 generating tables according to the current locale when PCRE is compiled. It
2026 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2027 characters, are recognized by isspace() and therefore were getting set in
2028 these tables, and indeed these tables seem to approximate to ISO 8859. This
2029 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2030 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2031 which of course cannot start UTF-8 characters. I have changed the code so
2032 that only real ASCII characters (less than 128) and the correct starting
2033 bytes for UTF-8 encodings are set for characters greater than 127 when in
2034 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2035 altogether.)
2036
2037 20. Added the /T option to pcretest so as to be able to run tests with non-
2038 standard character tables, thus making it possible to include the tests
2039 used for 19 above in the standard set of tests.
2040
2041 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2042 reference to a subpattern the other side of a comment that contains an
2043 opening parenthesis caused either an internal compiling error, or a
2044 reference to the wrong subpattern.
2045
2046
2047 Version 8.02 19-Mar-2010
2048 ------------------------
2049
2050 1. The Unicode data tables have been updated to Unicode 5.2.0.
2051
2052 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2053 configured.
2054
2055 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2056 original author of that file, following a query about its status.
2057
2058 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2059 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2060
2061 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2062 quantifier applied to a forward-referencing subroutine call, could compile
2063 incorrect code or give the error "internal error: previously-checked
2064 referenced subpattern not found".
2065
2066 6. Both MS Visual Studio and Symbian OS have problems with initializing
2067 variables to point to external functions. For these systems, therefore,
2068 pcre_malloc etc. are now initialized to local functions that call the
2069 relevant global functions.
2070
2071 7. There were two entries missing in the vectors called coptable and poptable
2072 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2073 I've fixed the data, and added a kludgy way of testing at compile time that
2074 the lengths are correct (equal to the number of opcodes).
2075
2076 8. Following on from 7, I added a similar kludge to check the length of the
2077 eint vector in pcreposix.c.
2078
2079 9. Error texts for pcre_compile() are held as one long string to avoid too
2080 much relocation at load time. To find a text, the string is searched,
2081 counting zeros. There was no check for running off the end of the string,
2082 which could happen if a new error number was added without updating the
2083 string.
2084
2085 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2086
2087 11. \K was not working if it appeared in an atomic group or in a group that
2088 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2089 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2090 the assertion is positive, but not if it is negative.
2091
2092 12. Change 11 fortuitously reduced the size of the stack frame used in the
2093 "match()" function of pcre_exec.c by one pointer. Forthcoming
2094 implementation of support for (*MARK) will need an extra pointer on the
2095 stack; I have reserved it now, so that the stack frame size does not
2096 decrease.
2097
2098 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2099 item in branch that calls a recursion is a subroutine call - as in the
2100 second branch in the above example - was incorrectly given the compile-
2101 time error "recursive call could loop indefinitely" because pcre_compile()
2102 was not correctly checking the subroutine for matching a non-empty string.
2103
2104 14. The checks for overrunning compiling workspace could trigger after an
2105 overrun had occurred. This is a "should never occur" error, but it can be
2106 triggered by pathological patterns such as hundreds of nested parentheses.
2107 The checks now trigger 100 bytes before the end of the workspace.
2108
2109 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2110
2111
2112 Version 8.01 19-Jan-2010
2113 ------------------------
2114
2115 1. If a pattern contained a conditional subpattern with only one branch (in
2116 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2117 computed the wrong minimum data length (which is of course zero for such
2118 subpatterns). This could cause incorrect "no match" results.
2119
2120 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2121 the pattern is reset in the first branch, pcre_compile() failed with
2122 "internal error: code overflow at offset...". This happened only when
2123 the reset was to the original external option setting. (An optimization
2124 abstracts leading options settings into an external setting, which was the
2125 cause of this.)
2126
2127 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2128 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2129 assertion pattern did not match (meaning that the assertion was true), it
2130 was incorrectly treated as false if the SKIP had been reached during the
2131 matching. This also applied to assertions used as conditions.
2132
2133 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2134 assertion subpattern, including such a pattern used as a condition,
2135 unpredictable results occurred, instead of the error return
2136 PCRE_ERROR_DFA_UITEM.
2137
2138 5. The C++ GlobalReplace function was not working like Perl for the special
2139 situation when an empty string is matched. It now does the fancy magic
2140 stuff that is necessary.
2141
2142 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2143 removed. (These were left over from very, very early versions of PCRE.)
2144
2145 7. Some cosmetic changes to the code to make life easier when compiling it
2146 as part of something else:
2147
2148 (a) Change DEBUG to PCRE_DEBUG.
2149
2150 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2151 called "current" as "current_branch", to prevent a collision with the
2152 Linux macro when compiled as a kernel module.
2153
2154 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2155 prevent a collision with the Linux macro when compiled as a kernel
2156 module.
2157
2158 8. In pcre_compile() there are some checks for integer overflows that used to
2159 cast potentially large values to (double). This has been changed to that
2160 when building, a check for int64_t is made, and if it is found, it is used
2161 instead, thus avoiding the use of floating point arithmetic. (There is no
2162 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2163 double.
2164
2165 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2166 2005 (difference between two addresses compared to an unsigned value).
2167
2168 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2169 custom one, because of the following reported problem in Windows:
2170
2171 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2172 under Win32.
2173 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2174 therefore missing the function definition.
2175 - The compiler thus generates a "C" signature for the test function.
2176 - The linker fails to find the "C" function.
2177 - PCRE fails to configure if asked to do so against libbz2.
2178
2179 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2180 messages were output:
2181
2182 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2183 rerunning libtoolize, to keep the correct libtool macros in-tree.
2184 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2185
2186 I have done both of these things.
2187
2188 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2189 most of the time, it *can* run out if it is given a pattern that contains a
2190 runaway infinite recursion. I updated the discussion in the pcrestack man
2191 page.
2192
2193 13. Now that we have gone to the x.xx style of version numbers, the minor
2194 version may start with zero. Using 08 or 09 is a bad idea because users
2195 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2196 interpreted as invalid octal numbers. I've updated the previous comment in
2197 configure.ac, and also added a check that gives an error if 08 or 09 are
2198 used.
2199
2200 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2201 causing partial matching to fail when the end of the subject matched \W
2202 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2203
2204 15. There were some discrepancies between the declarations in pcre_internal.h
2205 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2206 their definitions. The declarations used "const uschar *" and the
2207 definitions used USPTR. Even though USPTR is normally defined as "const
2208 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2209 reported that: "This difference in casting confuses some C++ compilers, for
2210 example, SunCC recognizes above declarations as different functions and
2211 generates broken code for hbpcre." I have changed the declarations to use
2212 USPTR.
2213
2214 16. GNU libtool is named differently on some systems. The autogen.sh script now
2215 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2216 (FreeBSD).
2217
2218 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2219 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2220 comment: "Figure out how to create a longlong from a string: strtoll and
2221 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2222 instance, but it only takes 2 args instead of 3!"
2223
2224 18. A subtle bug concerned with back references has been fixed by a change of
2225 specification, with a corresponding code fix. A pattern such as
2226 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2227 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2228 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2229 same bug. Such groups have to be quantified to be useful, or contained
2230 inside another quantified group. (If there's no repetition, the reference
2231 can never match.) The problem arises because, having left the group and
2232 moved on to the rest of the pattern, a later failure that backtracks into
2233 the group uses the captured value from the final iteration of the group
2234 rather than the correct earlier one. I have fixed this in PCRE by forcing
2235 any group that contains a reference to itself to be an atomic group; that
2236 is, there cannot be any backtracking into it once it has completed. This is
2237 similar to recursive and subroutine calls.
2238
2239
2240 Version 8.00 19-Oct-09
2241 ----------------------
2242
2243 1. The table for translating pcre_compile() error codes into POSIX error codes
2244 was out-of-date, and there was no check on the pcre_compile() error code
2245 being within the table. This could lead to an OK return being given in
2246 error.
2247
2248 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2249 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2250 in a Windows environment.
2251
2252 3. The pcregrep --count option prints the count for each file even when it is
2253 zero, as does GNU grep. However, pcregrep was also printing all files when
2254 --files-with-matches was added. Now, when both options are given, it prints
2255 counts only for those files that have at least one match. (GNU grep just
2256 prints the file name in this circumstance, but including the count seems
2257 more useful - otherwise, why use --count?) Also ensured that the
2258 combination -clh just lists non-zero counts, with no names.
2259
2260 4. The long form of the pcregrep -F option was incorrectly implemented as
2261 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2262 but it seems right to fix it, and I didn't think it was worth preserving
2263 the old behaviour.
2264
2265 5. The command line items --regex=pattern and --regexp=pattern were not
2266 recognized by pcregrep, which required --regex pattern or --regexp pattern
2267 (with a space rather than an '='). The man page documented the '=' forms,
2268 which are compatible with GNU grep; these now work.
2269
2270 6. No libpcreposix.pc file was created for pkg-config; there was just
2271 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2272
2273 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2274 when UCP support is not needed, by modifying the Python script that
2275 generates it from Unicode data files. This should not matter if the module
2276 is correctly used as a library, but I received one complaint about 50K of
2277 unwanted data. My guess is that the person linked everything into his
2278 program rather than using a library. Anyway, it does no harm.
2279
2280 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2281 was a minimum greater than 1 for a wide character in a possessive
2282 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2283 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2284 character. Chaos in the form of incorrect output or a compiling loop could
2285 result.
2286
2287 9. The restrictions on what a pattern can contain when partial matching is
2288 requested for pcre_exec() have been removed. All patterns can now be
2289 partially matched by this function. In addition, if there are at least two
2290 slots in the offset vector, the offset of the earliest inspected character
2291 for the match and the offset of the end of the subject are set in them when
2292 PCRE_ERROR_PARTIAL is returned.
2293
2294 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2295 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2296 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2297 and may be more useful for multi-segment matching.
2298
2299 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2300 used to be given if ever the end of the subject was reached; now it is
2301 given only if matching could not proceed because another character was
2302 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2303 string "Z", which now yields "no match" instead of "partial match". In the
2304 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2305 final character ended with (*FAIL).
2306
2307 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2308 if the pattern had a "must contain" character that was already found in the
2309 earlier partial match, unless partial matching was again requested. For
2310 example, with the pattern /dog.(body)?/, the "must contain" character is
2311 "g". If the first part-match was for the string "dog", restarting with
2312 "sbody" failed. This bug has been fixed.
2313
2314 13. The string returned by pcre_dfa_exec() after a partial match has been
2315 changed so that it starts at the first inspected character rather than the
2316 first character of the match. This makes a difference only if the pattern
2317 starts with a lookbehind assertion or \b or \B (\K is not supported by
2318 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2319 matching functions compatible, and I think it's the right thing to do.
2320
2321 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2322 so that the demonstration program is easily available in environments where
2323 PCRE has not been installed from source.
2324
2325 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2326 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2327 library.
2328
2329 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2330 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2331 is not the first non-POSIX option to be added. Clearly some people find
2332 these options useful.
2333
2334 17. If a caller to the POSIX matching function regexec() passes a non-zero
2335 value for nmatch with a NULL value for pmatch, the value of
2336 nmatch is forced to zero.
2337
2338 18. RunGrepTest did not have a test for the availability of the -u option of
2339 the diff command, as RunTest does. It now checks in the same way as
2340 RunTest, and also checks for the -b option.
2341
2342 19. If an odd number of negated classes containing just a single character
2343 interposed, within parentheses, between a forward reference to a named
2344 subpattern and the definition of the subpattern, compilation crashed with
2345 an internal error, complaining that it could not find the referenced
2346 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2347 [The bug was that it was starting one character too far in when skipping
2348 over the character class, thus treating the ] as data rather than
2349 terminating the class. This meant it could skip too much.]
2350
2351 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2352 /g option in pcretest when the pattern contains \K, which makes it possible
2353 to have an empty string match not at the start, even when the pattern is
2354 anchored. Updated pcretest and pcredemo to use this option.
2355
2356 21. If the maximum number of capturing subpatterns in a recursion was greater
2357 than the maximum at the outer level, the higher number was returned, but
2358 with unset values at the outer level. The correct (outer level) value is
2359 now given.
2360
2361 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2362 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2363 make it do so. The string so far is captured, making this feature
2364 compatible with Perl.
2365
2366 23. The tests have been re-organized, adding tests 11 and 12, to make it
2367 possible to check the Perl 5.10 features against Perl 5.10.
2368
2369 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2370 pattern matches a fixed length string. PCRE did not allow this; now it
2371 does. Neither allows recursion.
2372
2373 25. I finally figured out how to implement a request to provide the minimum
2374 length of subject string that was needed in order to match a given pattern.
2375 (It was back references and recursion that I had previously got hung up
2376 on.) This code has now been added to pcre_study(); it finds a lower bound
2377 to the length of subject needed. It is not necessarily the greatest lower
2378 bound, but using it to avoid searching strings that are too short does give
2379 some useful speed-ups. The value is available to calling programs via
2380 pcre_fullinfo().
2381
2382 26. While implementing 25, I discovered to my embarrassment that pcretest had
2383 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2384 study optimizations had never been tested with that matching function.
2385 Oops. What is worse, even when it was passed study data, there was a bug in
2386 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2387 were also very few tests of studied patterns with pcre_dfa_exec().
2388
2389 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2390 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2391 on the other side of the coin, they are no longer allowed to have different
2392 names, because these cannot be distinguished in PCRE, and this has caused
2393 confusion. (This is a difference from Perl.)
2394
2395 28. When duplicate subpattern names are present (necessarily with different
2396 numbers, as required by 27 above), and a test is made by name in a
2397 conditional pattern, either for a subpattern having been matched, or for
2398 recursion in such a pattern, all the associated numbered subpatterns are
2399 tested, and the overall condition is true if the condition is true for any
2400 one of them. This is the way Perl works, and is also more like the way
2401 testing by number works.
2402
2403
2404 Version 7.9 11-Apr-09
2405 ---------------------
2406
2407 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2408 (pcretest), all targets were linked against these libraries. This included
2409 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2410 libraries. This caused unwanted dependencies to be created. This problem
2411 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2412 pcretest is linked with readline.
2413
2414 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2415 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2416 moved outside it again, because FALSE and TRUE are already defined in AIX,
2417 but BOOL is not.
2418
2419 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2420 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2421
2422 4. The pcregrep documentation said spaces were inserted as well as colons (or
2423 hyphens) following file names and line numbers when outputting matching
2424 lines. This is not true; no spaces are inserted. I have also clarified the
2425 wording for the --colour (or --color) option.
2426
2427 5. In pcregrep, when --colour was used with -o, the list of matching strings
2428 was not coloured; this is different to GNU grep, so I have changed it to be
2429 the same.
2430
2431 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2432 each matching line was coloured. Now it goes on to look for further matches
2433 of any of the test patterns, which is the same behaviour as GNU grep.
2434
2435 7. A pattern that could match an empty string could cause pcregrep to loop; it
2436 doesn't make sense to accept an empty string match in pcregrep, so I have
2437 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2438 seems to be how GNU grep behaves. [But see later change 40 for release
2439 8.33.]
2440
2441 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2442 start or after a newline", because the conditional assertion was not being
2443 correctly handled. The rule now is that both the assertion and what follows
2444 in the first alternative must satisfy the test.
2445
2446 9. If auto-callout was enabled in a pattern with a conditional group whose
2447 condition was an assertion, PCRE could crash during matching, both with
2448 pcre_exec() and pcre_dfa_exec().
2449
2450 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2451 used for matching.
2452
2453 11. Unicode property support in character classes was not working for
2454 characters (bytes) greater than 127 when not in UTF-8 mode.
2455
2456 12. Added the -M command line option to pcretest.
2457
2458 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2459
2460 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2461
2462 16. Added comments and documentation about mis-use of no_arg in the C++
2463 wrapper.
2464
2465 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2466 from Martin Jerabek that uses macro names for all relevant character and
2467 string constants.
2468
2469 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2470 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2471 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2472 these, but not everybody uses configure.
2473
2474 19. A conditional group that had only one branch was not being correctly
2475 recognized as an item that could match an empty string. This meant that an
2476 enclosing group might also not be so recognized, causing infinite looping
2477 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2478 with the subject "ab", where knowledge that the repeated group can match
2479 nothing is needed in order to break the loop.
2480
2481 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2482 exec(), but without supplying a callout function, matching went wrong.
2483
2484 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2485 leak if the size of the offset vector was greater than 30. When the vector
2486 is smaller, the saved offsets during recursion go onto a local stack
2487 vector, but for larger vectors malloc() is used. It was failing to free
2488 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2489 error, in fact).
2490
2491 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2492 heapframe that is used only when UTF-8 support is enabled. This caused no
2493 problem, but was untidy.
2494
2495 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2496 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2497 included within another project.
2498
2499 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2500 slightly modified by me:
2501
2502 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2503 not building pcregrep.
2504
2505 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2506 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2507
2508 25. Forward references, both numeric and by name, in patterns that made use of
2509 duplicate group numbers, could behave incorrectly or give incorrect errors,
2510 because when scanning forward to find the reference group, PCRE was not
2511 taking into account the duplicate group numbers. A pattern such as
2512 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2513
2514 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2515 the feature of a custom pointer more persuasive (as requested by a user).
2516
2517 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2518 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2519 pre-defined.
2520
2521 28. Added support for (*UTF8) at the start of a pattern.
2522
2523 29. Arrange for flags added by the "release type" setting in CMake to be shown
2524 in the configuration summary.
2525
2526
2527 Version 7.8 05-Sep-08
2528 ---------------------
2529
2530 1. Replaced UCP searching code with optimized version as implemented for Ad
2531 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2532 stage table and inline lookup instead of a function, giving speed ups of 2
2533 to 5 times on some simple patterns that I tested. Permission was given to
2534 distribute the MultiStage2.py script that generates the tables (it's not in
2535 the tarball, but is in the Subversion repository).
2536
2537 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2538 scripts.
2539
2540 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2541 a group with a zero qualifier. The result of the study could be incorrect,
2542 or the function might crash, depending on the pattern.
2543
2544 4. Caseless matching was not working for non-ASCII characters in back
2545 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2546 It now works when Unicode Property Support is available.
2547
2548 5. In pcretest, an escape such as \x{de} in the data was always generating
2549 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2550 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2551 truncation.
2552
2553 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2554
2555 7. Added two (int) casts to pcregrep when printing the difference of two
2556 pointers, in case they are 64-bit values.
2557
2558 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2559 test 2 if it fails.
2560
2561 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2562 and a #define of that name to empty if it is not externally set. This is to
2563 allow users of MSVC to set it if necessary.
2564
2565 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2566 the convenience functions in the pcre_get.c source file.
2567
2568 11. An option change at the start of a pattern that had top-level alternatives
2569 could cause overwriting and/or a crash. This command provoked a crash in
2570 some environments:
2571
2572 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2573
2574 This potential security problem was recorded as CVE-2008-2371.
2575
2576 12. For a pattern where the match had to start at the beginning or immediately
2577 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2578 pcre_dfa_exec() could read past the end of the passed subject if there was
2579 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2580 pcretest so that it places the subject at the end of its malloc-ed buffer.
2581
2582 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2583 exec() might read past the end of the data buffer in UTF-8 mode.
2584
2585 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2586 the data contained the byte 0x85 as part of a UTF-8 character within its
2587 first line. This applied both to normal and DFA matching.
2588
2589 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2590 /^[^d]*?$/8 failed to match "abc".
2591
2592 16. Added a missing copyright notice to pcrecpp_internal.h.
2593
2594 17. Make it more clear in the documentation that values returned from
2595 pcre_exec() in ovector are byte offsets, not character counts.
2596
2597 18. Tidied a few places to stop certain compilers from issuing warnings.
2598
2599 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2600 supplied by Stefan Weber. I made a further small update for 7.8 because
2601 there is a change of source arrangements: the pcre_searchfuncs.c module is
2602 replaced by pcre_ucd.c.
2603
2604
2605 Version 7.7 07-May-08
2606 ---------------------
2607
2608 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2609 a string to a long long, pretend we don't even have a long long." This is
2610 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2611
2612 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2613 pre-7.6 versions, which defined a global no_arg variable instead of putting
2614 it in the RE class. (See also #8 below.)
2615
2616 3. Remove a line of dead code, identified by coverity and reported by Nuno
2617 Lopes.
2618
2619 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2620
2621 (1) The include/exclude patterns were being applied to the whole pathnames
2622 of files, instead of just to the final components.
2623
2624 (2) If there was more than one level of directory, the subdirectories were
2625 skipped unless they satisfied the include/exclude conditions. This is
2626 inconsistent with GNU grep (and could even be seen as contrary to the
2627 pcregrep specification - which I improved to make it absolutely clear).
2628 The action now is always to scan all levels of directory, and just
2629 apply the include/exclude patterns to regular files.
2630
2631 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2632 --exclude_dir in the tests to avoid scanning .svn directories.
2633
2634 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2635 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2636 doesn't support NULs in patterns.
2637
2638 7. Added some missing "const"s to declarations of static tables in
2639 pcre_compile.c and pcre_dfa_exec.c.
2640
2641 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2642 caused by fix #2 above. (Subsequently also a second patch to fix the
2643 first patch. And a third patch - this was a messy problem.)
2644
2645 9. Applied Craig's patch to remove the use of push_back().
2646
2647 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2648 matching function regexec().
2649
2650 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2651 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2652 references. PCRE supports relative numbers with this syntax (I don't think
2653 Oniguruma does).
2654
2655 12. Previously, a group with a zero repeat such as (...){0} was completely
2656 omitted from the compiled regex. However, this means that if the group
2657 was called as a subroutine from elsewhere in the pattern, things went wrong
2658 (an internal error was given). Such groups are now left in the compiled
2659 pattern, with a new opcode that causes them to be skipped at execution
2660 time.
2661
2662 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2663 to the way PCRE behaves:
2664
2665 (a) A lone ] character is dis-allowed (Perl treats it as data).
2666
2667 (b) A back reference to an unmatched subpattern matches an empty string
2668 (Perl fails the current match path).
2669
2670 (c) A data ] in a character class must be notated as \] because if the
2671 first data character in a class is ], it defines an empty class. (In
2672 Perl it is not possible to have an empty class.) The empty class []
2673 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2674 The negative empty class [^] matches any one character, independently
2675 of the DOTALL setting.
2676
2677 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2678 non-existent subpattern following a character class starting with ']' and
2679 containing () gave an internal compiling error instead of "reference to
2680 non-existent subpattern". Fortunately, when the pattern did exist, the
2681 compiled code was correct. (When scanning forwards to check for the
2682 existence of the subpattern, it was treating the data ']' as terminating
2683 the class, so got the count wrong. When actually compiling, the reference
2684 was subsequently set up correctly.)
2685
2686 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2687 it was being rejected as not supported by pcre_dfa_exec(), even though
2688 other assertions are supported. I have made pcre_dfa_exec() support
2689 (*FAIL).
2690
2691 16. The implementation of 13c above involved the invention of a new opcode,
2692 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2693 cannot be changed at match time, I realized I could make a small
2694 improvement to matching performance by compiling OP_ALLANY instead of
2695 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2696 on the OP_ANY path.
2697
2698 17. Compiling pcretest on Windows with readline support failed without the
2699 following two fixes: (1) Make the unistd.h include conditional on
2700 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2701
2702 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2703 ncurses library to be included for pcretest when ReadLine support is
2704 requested, but also to allow for it to be overridden. This patch came from
2705 Daniel Bergström.
2706
2707 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2708 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2709 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2710 spotting this.
2711
2712
2713 Version 7.6 28-Jan-08
2714 ---------------------
2715
2716 1. A character class containing a very large number of characters with
2717 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2718 overflow.
2719
2720 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2721 HAVE_LONG_LONG is not defined.
2722
2723 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2724 bring them up to date and include new features. This patch includes:
2725
2726 - Fixed PH's badly added libz and libbz2 support.
2727 - Fixed a problem with static linking.
2728 - Added pcredemo. [But later removed - see 7 below.]
2729 - Fixed dftables problem and added an option.
2730 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2731 HAVE_LONG_LONG.
2732 - Added readline support for pcretest.
2733 - Added an listing of the option settings after cmake has run.
2734
2735 4. A user submitted a patch to Makefile that makes it easy to create
2736 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2737 Makefile.am that cause it to include this special target, without
2738 affecting anything else. Note that the same mingw target plus all
2739 the other distribution libraries and programs are now supported
2740 when configuring with CMake (see 6 below) instead of with
2741 Configure/Make.
2742
2743 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2744 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2745 exported in the Windows port". It has not yet been confirmed that the patch
2746 solves the problem, but it does no harm.
2747
2748 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2749 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2750 with CMake, and also correct the comment about stack recursion.
2751
2752 7. Remove the automatic building of pcredemo from the ./configure system and
2753 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2754 of a program that users should build themselves after PCRE is installed, so
2755 building it automatically is not really right. What is more, it gave
2756 trouble in some build environments.
2757
2758 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2759
2760
2761 Version 7.5 10-Jan-08
2762 ---------------------
2763
2764 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2765 values in parens when parsing an RE using the C++ wrapper."
2766
2767 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2768 Characters greater than 255 were excluded from the class instead of being
2769 included.
2770
2771 3. The same bug as (2) above applied to negated POSIX classes such as
2772 [:^space:].
2773
2774 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2775 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2776 I have changed it.
2777
2778 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2779 first named subpattern) and a construct such as (?&a) would reference the
2780 first named subpattern whose name started with "a" (in other words, the
2781 length check was missing). Both these problems are fixed. "Subpattern name
2782 expected" is now given for (?&) (a zero-length name), and this patch also
2783 makes it give the same error for \k'' (previously it complained that that
2784 was a reference to a non-existent subpattern).
2785
2786 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2787 this is right because (?- can be followed by option settings as well as by
2788 digits. I have, however, made the messages clearer.
2789
2790 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2791 than the number used in the conditional) now cause a compile-time error.
2792 This is actually not compatible with Perl, which accepts such patterns, but
2793 treats the conditional as always being FALSE (as PCRE used to), but it
2794 seems to me that giving a diagnostic is better.
2795
2796 8. Change "alphameric" to the more common word "alphanumeric" in comments
2797 and messages.
2798
2799 9. Fix two occurrences of "backslash" in comments that should have been
2800 "backspace".
2801
2802 10. Remove two redundant lines of code that can never be obeyed (their function
2803 was moved elsewhere).
2804
2805 11. The program that makes PCRE's Unicode character property table had a bug
2806 which caused it to generate incorrect table entries for sequences of
2807 characters that have the same character type, but are in different scripts.
2808 It amalgamated them into a single range, with the script of the first of
2809 them. In other words, some characters were in the wrong script. There were
2810 thirteen such cases, affecting characters in the following ranges:
2811
2812 U+002b0 - U+002c1
2813 U+0060c - U+0060d
2814 U+0061e - U+00612
2815 U+0064b - U+0065e
2816 U+0074d - U+0076d
2817 U+01800 - U+01805
2818 U+01d00 - U+01d77
2819 U+01d9b - U+01dbf
2820 U+0200b - U+0200f
2821 U+030fc - U+030fe
2822 U+03260 - U+0327f
2823 U+0fb46 - U+0fbb1
2824 U+10450 - U+1049d
2825
2826 12. The -o option (show only the matching part of a line) for pcregrep was not
2827 compatible with GNU grep in that, if there was more than one match in a
2828 line, it showed only the first of them. It now behaves in the same way as
2829 GNU grep.
2830
2831 13. If the -o and -v options were combined for pcregrep, it printed a blank
2832 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2833 does the same. The return code can be used to tell if there were any
2834 non-matching lines.
2835
2836 14. Added --file-offsets and --line-offsets to pcregrep.
2837
2838 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2839 infinitely looping recursion. The bug was that positive lookaheads were not
2840 being skipped when checking for a possible empty match (negative lookaheads
2841 and both kinds of lookbehind were skipped).
2842
2843 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2844 inclusion of <windows.h> to before rather than after the definition of
2845 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2846
2847 17. Specifying a possessive quantifier with a specific limit for a Unicode
2848 character property caused pcre_compile() to compile bad code, which led at
2849 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2850 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2851 caused the error; without that there was no problem.
2852
2853 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2854
2855 19. Added --enable-pcretest-libreadline.
2856
2857 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2858 RE::GlobalReplace(). As a result, the number of replacements returned was
2859 double what it should be. I removed one of the increments, but Craig sent a
2860 later patch that removed the other one (the right fix) and added unit tests
2861 that check the return values (which was not done before).
2862
2863 21. Several CMake things:
2864
2865 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2866 the names libpcre and libpcreposix, not just pcre and pcreposix.
2867
2868 (2) The above change means that pcretest and pcregrep are now correctly
2869 linked with the newly-built libraries, not previously installed ones.
2870
2871 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2872
2873 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2874 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2875 UTF-8 newline character). The key issue is that the pattern starts .*;
2876 this means that the match must be either at the beginning, or after a
2877 newline. The bug was in the code for advancing after a failed match and
2878 checking that the new position followed a newline. It was not taking
2879 account of UTF-8 characters correctly.
2880
2881 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2882 character classes. PCRE was not treating the sequence [:...:] as a
2883 character class unless the ... were all letters. Perl, however, seems to
2884 allow any characters between [: and :], though of course it rejects as
2885 unknown any "names" that contain non-letters, because all the known class
2886 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2887 for example, whereas PCRE did not - it did not recognize a POSIX character
2888 class. This seemed a bit dangerous, so the code has been changed to be
2889 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2890 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2891 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2892 Perl does, and where it didn't before.
2893
2894 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2895 Windows environments %n is disabled by default.
2896
2897
2898 Version 7.4 21-Sep-07
2899 ---------------------
2900
2901 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2902 means that a class such as [\s] counted as "explicit reference to CR or
2903 LF". That isn't really right - the whole point of the change was to try to
2904 help when there was an actual mention of one of the two characters. So now
2905 the change happens only if \r or \n (or a literal CR or LF) character is
2906 encountered.
2907
2908 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2909 of both had grown to the point where there were only 3 bits left.
2910 Fortunately, there was spare space in the data structure, and so I have
2911 moved the internal flags into a new 16-bit field to free up more option
2912 bits.
2913
2914 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2915 but did not set the internal JCHANGED flag - either of these is enough to
2916 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2917 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2918 start sets both bits.
2919
2920 4. Added options (at build time, compile time, exec time) to change \R from
2921 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2922
2923 5. doc/pcresyntax.html was missing from the distribution.
2924
2925 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2926 compatibility, even though it is no longer used.
2927
2928 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2929 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2930 windows.h file is present (where different names are used). [This was
2931 reversed later after testing - see 16 below.]
2932
2933 8. Changed all #include <config.h> to #include "config.h". There were also
2934 some further <pcre.h> cases that I changed to "pcre.h".
2935
2936 9. When pcregrep was used with the --colour option, it missed the line ending
2937 sequence off the lines that it output.
2938
2939 10. It was pointed out to me that arrays of string pointers cause lots of
2940 relocations when a shared library is dynamically loaded. A technique of
2941 using a single long string with a table of offsets can drastically reduce
2942 these. I have refactored PCRE in four places to do this. The result is
2943 dramatic:
2944
2945 Originally: 290
2946 After changing UCP table: 187
2947 After changing error message table: 43
2948 After changing table of "verbs" 36
2949 After changing table of Posix names 22
2950
2951 Thanks to the folks working on Gregex for glib for this insight.
2952
2953 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2954 unicode-properties was also set.
2955
2956 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2957
2958 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2959 checked only for CRLF.
2960
2961 14. Added casts to pcretest.c to avoid compiler warnings.
2962
2963 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2964
2965 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2966 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2967 entirely. This removes changes made in 7 above.
2968
2969 17. The CMake files have been updated, and there is now more information about
2970 building with CMake in the NON-UNIX-USE document.
2971
2972
2973 Version 7.3 28-Aug-07
2974 ---------------------
2975
2976 1. In the rejigging of the build system that eventually resulted in 7.1, the
2977 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2978 brackets there is not right, since it causes compilers to look for an
2979 installed pcre.h, not the version that is in the source that is being
2980 compiled (which of course may be different). I have changed it back to:
2981
2982 #include "pcre.h"
2983
2984 I have a vague recollection that the change was concerned with compiling in
2985 different directories, but in the new build system, that is taken care of
2986 by the VPATH setting the Makefile.
2987
2988 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2989 when the subject happened to end in the byte 0x85 (e.g. if the last
2990 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2991 characters but of course it shouldn't be taken as a newline when it is part
2992 of another character. The bug was that, for an unlimited repeat of . in
2993 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2994 characters when looking for a newline.
2995
2996 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2997
2998 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2999 in debug output.
3000
3001 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3002 long printing in the pcrecpp unittest when running under MinGW.
3003
3004 6. ESC_K was left out of the EBCDIC table.
3005
3006 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3007 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3008 limit also applies to "virtual nesting" when a pattern is recursive, and in
3009 this case 1000 isn't so big. I have been able to remove this limit at the
3010 expense of backing off one optimization in certain circumstances. Normally,
3011 when pcre_exec() would call its internal match() function recursively and
3012 immediately return the result unconditionally, it uses a "tail recursion"
3013 feature to save stack. However, when a subpattern that can match an empty
3014 string has an unlimited repetition quantifier, it no longer makes this
3015 optimization. That gives it a stack frame in which to save the data for
3016 checking that an empty string has been matched. Previously this was taken
3017 from the 1000-entry workspace that had been reserved. So now there is no
3018 explicit limit, but more stack is used.
3019
3020 8. Applied Daniel's patches to solve problems with the import/export magic
3021 syntax that is required for Windows, and which was going wrong for the
3022 pcreposix and pcrecpp parts of the library. These were overlooked when this
3023 problem was solved for the main library.
3024
3025 9. There were some crude static tests to avoid integer overflow when computing
3026 the size of patterns that contain repeated groups with explicit upper
3027 limits. As the maximum quantifier is 65535, the maximum group length was
3028 set at 30,000 so that the product of these two numbers did not overflow a
3029 32-bit integer. However, it turns out that people want to use groups that
3030 are longer than 30,000 bytes (though not repeat them that many times).
3031 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3032 made it possible to implement the integer overflow checks in a much more
3033 dynamic way, which I have now done. The artificial limitation on group
3034 length has been removed - we now have only the limit on the total length of
3035 the compiled pattern, which depends on the LINK_SIZE setting.
3036
3037 10. Fixed a bug in the documentation for get/copy named substring when
3038 duplicate names are permitted. If none of the named substrings are set, the
3039 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3040 empty string.
3041
3042 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3043 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3044 because the ] is interpreted as the first data character and the
3045 terminating ] is not found. PCRE has been made compatible with Perl in this
3046 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3047 cause memory overwriting.
3048
3049 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3050 string has been matched (to stop an infinite loop). It was not recognizing
3051 a conditional subpattern that could match an empty string if that
3052 subpattern was within another subpattern. For example, it looped when
3053 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3054 condition was not nested. This bug has been fixed.
3055
3056 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3057 past the start of the subject in the presence of bytes with the top bit
3058 set, for example "\x8aBCD".
3059
3060 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3061 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3062
3063 14. Optimized (?!) to (*FAIL).
3064
3065 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3066 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3067 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3068 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3069 does: it's just the validity check that is more restrictive.
3070
3071 16. Inserted checks for integer overflows during escape sequence (backslash)
3072 processing, and also fixed erroneous offset values for syntax errors during
3073 backslash processing.
3074
3075 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3076 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3077
3078 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3079 caused an overrun.
3080
3081 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3082 something other than just ASCII characters) inside a group that had an
3083 unlimited repeat caused a loop at compile time (while checking to see
3084 whether the group could match an empty string).
3085
3086 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3087 [\P{Any}] did so. (Error in the code for printing property names.)
3088
3089 21. An orphan \E inside a character class could cause a crash.
3090
3091 22. A repeated capturing bracket such as (A)? could cause a wild memory
3092 reference during compilation.
3093
3094 23. There are several functions in pcre_compile() that scan along a compiled
3095 expression for various reasons (e.g. to see if it's fixed length for look
3096 behind). There were bugs in these functions when a repeated \p or \P was
3097 present in the pattern. These operators have additional parameters compared
3098 with \d, etc, and these were not being taken into account when moving along
3099 the compiled data. Specifically:
3100
3101 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3102 length.
3103
3104 (b) An item such as \pL+ within a repeated group could cause crashes or
3105 loops.
3106
3107 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3108 "reference to non-existent subpattern" error.
3109
3110 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3111
3112 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3113 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3114
3115 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3116
3117 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3118 character were causing crashes (broken optimization).
3119
3120 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3121 \p or \P) caused a compile-time loop.
3122
3123 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3124 break. For example, the unstudied pattern [\r\n]A does not match the string
3125 "\r\nA" because change 7.0/46 below moves the current point on by two
3126 characters after failing to match at the start. However, the pattern \nA
3127 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3128 the same is true. There doesn't seem any very clean way out of this, but
3129 what I have chosen to do makes the common cases work: PCRE now takes note
3130 of whether there can be an explicit match for \r or \n anywhere in the
3131 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3132 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3133 pattern has explicit CR or LF references.
3134
3135 29. Added (*CR) etc for changing newline setting at start of pattern.
3136
3137
3138 Version 7.2 19-Jun-07
3139 ---------------------
3140
3141 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3142 which is apparently normally available under Windows.
3143
3144 2. Re-jig the pcregrep tests with different newline settings in an attempt
3145 to make them independent of the local environment's newline setting.
3146
3147 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3148
3149 4. Some of the "internals" tests were previously cut out when the link size
3150 was not 2, because the output contained actual offsets. The recent new
3151 "Z" feature of pcretest means that these can be cut out, making the tests
3152 usable with all link sizes.
3153
3154 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3155 stack recursion. This gives a massive performance boost under BSD, but just
3156 a small improvement under Linux. However, it saves one field in the frame
3157 in all cases.
3158
3159 6. Added more features from the forthcoming Perl 5.10:
3160
3161 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3162 recursion call. It refers to the nth most recently opened parentheses.
3163
3164 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3165 to be opened parentheses.
3166
3167 (c) Conditions that refer to capturing parentheses can be specified
3168 relatively, for example, (?(-2)... or (?(+3)...
3169
3170 (d) \K resets the start of the current match so that everything before
3171 is not part of it.
3172
3173 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3174
3175 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3176 reference syntax.
3177
3178 (g) (?| introduces a group in which the numbering of parentheses in each
3179 alternative starts with the same number.
3180
3181 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3182
3183 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3184 PCRE_INFO_JCHANGED.
3185
3186 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3187 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3188 for detecting groups that can match an empty string.
3189
3190 9. A pattern with a very large number of alternatives (more than several
3191 hundred) was running out of internal workspace during the pre-compile
3192 phase, where pcre_compile() figures out how much memory will be needed. A
3193 bit of new cunning has reduced the workspace needed for groups with
3194 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3195 workspace instead of running out of the 4096 that are available.
3196
3197 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3198
3199 11. Applied patch from Google to remove an optimization that didn't quite work.
3200 The report of the bug said:
3201
3202 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3203 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3204 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3205
3206 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3207 it matched the wrong number of bytes.
3208
3209
3210 Version 7.1 24-Apr-07
3211 ---------------------
3212
3213 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3214 that is more "standard", making use of automake and other Autotools. There
3215 is some re-arrangement of the files and adjustment of comments consequent
3216 on this.
3217
3218 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3219 for recursive directory scanning broke on some systems because the files
3220 are not scanned in any specific order and on different systems the order
3221 was different. A call to "sort" has been inserted into RunGrepTest for the
3222 approprate test as a short-term fix. In the longer term there may be an
3223 alternative.
3224
3225 3. I had an email from Eric Raymond about problems translating some of PCRE's
3226 man pages to HTML (despite the fact that I distribute HTML pages, some
3227 people do their own conversions for various reasons). The problems
3228 concerned the use of low-level troff macros .br and .in. I have therefore
3229 removed all such uses from the man pages (some were redundant, some could
3230 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3231 HTML has been updated to handle .nf/.fi and to complain if it encounters
3232 .br or .in.
3233
3234 4. Updated comments in configure.ac that get placed in config.h.in and also
3235 arranged for config.h to be included in the distribution, with the name
3236 config.h.generic, for the benefit of those who have to compile without
3237 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3238
3239 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3240 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3241 makevp.bat for the current PCRE, using the additional files
3242 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3243
3244 6. A Windows user reported a minor discrepancy with test 2, which turned out
3245 to be caused by a trailing space on an input line that had got lost in his
3246 copy. The trailing space was an accident, so I've just removed it.
3247
3248 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3249 that is needed.
3250
3251 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3252 as "const" (a) because they are and (b) because it helps the PHP
3253 maintainers who have recently made a script to detect big data structures
3254 in the php code that should be moved to the .rodata section. I remembered
3255 to update Builducptable as well, so it won't revert if ucptable.h is ever
3256 re-created.
3257
3258 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3259 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3260 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3261 support is not required. This saves 1.5-2K of code, which is important in
3262 some applications.
3263
3264 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3265 so as not to refer to the tables, even though these functions will never be
3266 called when UTF-8 support is disabled. Otherwise there are problems with a
3267 shared library.
3268
3269 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3270
3271 (a) It was defining its arguments as char * instead of void *.
3272
3273 (b) It was assuming that all moves were upwards in memory; this was true
3274 a long time ago when I wrote it, but is no longer the case.
3275
3276 The emulated memove() is provided for those environments that have neither
3277 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3278 is clearly not the case, as these two bugs were recently reported.
3279
3280 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3281 and Detrail to create the HTML documentation, the .txt form of the man
3282 pages, and it removes trailing spaces from listed files. It also creates
3283 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3284 case, it wraps all the #defines with #ifndefs. This script should be run
3285 before "make dist".
3286
3287 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3288 with Unicode property support.
3289
3290 (a) For a maximizing quantifier, if the two different cases of the
3291 character were of different lengths in their UTF-8 codings (there are
3292 some cases like this - I found 11), and the matching function had to
3293 back up over a mixture of the two cases, it incorrectly assumed they
3294 were both the same length.
3295
3296 (b) When PCRE was configured to use the heap rather than the stack for
3297 recursion during matching, it was not correctly preserving the data for
3298 the other case of a UTF-8 character when checking ahead for a match
3299 while processing a minimizing repeat. If the check also involved
3300 matching a wide character, but failed, corruption could cause an
3301 erroneous result when trying to check for a repeat of the original
3302 character.
3303
3304 13. Some tidying changes to the testing mechanism:
3305
3306 (a) The RunTest script now detects the internal link size and whether there
3307 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3308 values substituted by "configure". (The RunGrepTest script already did
3309 this for UTF-8.) The configure.ac script no longer substitutes the
3310 relevant variables.
3311
3312 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3313 with length and offset values. This means that the output is different
3314 for different internal link sizes. Test 2 is skipped for link sizes
3315 other than 2 because of this, bypassing the problem. Unfortunately,
3316 there was also a test in test 3 (the locale tests) that used /B and
3317 failed for link sizes other than 2. Rather than cut the whole test out,
3318 I have added a new /Z option to pcretest that replaces the length and
3319 offset values with spaces. This is now used to make test 3 independent
3320 of link size. (Test 2 will be tidied up later.)
3321
3322 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3323 segmentation fault instead of returning the appropriate error message.
3324
3325 15. In multiline mode when the newline sequence was set to "any", the pattern
3326 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3327 This doesn't seem right; it now treats the CRLF combination as the line
3328 ending, and so does not match in that case. It's only a pattern such as ^$
3329 that would hit this one: something like ^ABC$ would have failed after \r
3330 and then tried again after \r\n.
3331
3332 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3333 in an attempt to make files that differ only in their line terminators
3334 compare equal. This works on Linux.
3335
3336 17. Under certain error circumstances pcregrep might try to free random memory
3337 as it exited. This is now fixed, thanks to valgrind.
3338
3339 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3340 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3341 was because its rules for how to advance for /g after matching an empty
3342 string at the end of a line did not allow for this case. They now check for
3343 it specially.
3344
3345 20. pcretest is supposed to handle patterns and data of any length, by
3346 extending its buffers when necessary. It was getting this wrong when the
3347 buffer for a data line had to be extended.
3348
3349 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3350 CRLF as a newline sequence.
3351
3352 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3353 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3354 I have nevertheless tidied it up.
3355
3356 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3357
3358 24. Added a man page for pcre-config.
3359
3360
3361 Version 7.0 19-Dec-06
3362 ---------------------
3363
3364 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3365 moving to gcc 4.1.1.
3366
3367 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3368 sys/time.h, which is documented as needed for this function. It doesn't
3369 seem to matter on Linux, but it showed up on some releases of OS X.
3370
3371 3. It seems that there are systems where bytes whose values are greater than
3372 127 match isprint() in the "C" locale. The "C" locale should be the
3373 default when a C program starts up. In most systems, only ASCII printing
3374 characters match isprint(). This difference caused the output from pcretest
3375 to vary, making some of the tests fail. I have changed pcretest so that:
3376
3377 (a) When it is outputting text in the compiled version of a pattern, bytes
3378 other than 32-126 are always shown as hex escapes.
3379
3380 (b) When it is outputting text that is a matched part of a subject string,
3381 it does the same, unless a different locale has been set for the match
3382 (using the /L modifier). In this case, it uses isprint() to decide.
3383
3384 4. Fixed a major bug that caused incorrect computation of the amount of memory
3385 required for a compiled pattern when options that changed within the
3386 pattern affected the logic of the preliminary scan that determines the
3387 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3388 that the computed length was too small. The symptoms of this bug were
3389 either the PCRE error "internal error: code overflow" from pcre_compile(),
3390 or a glibc crash with a message such as "pcretest: free(): invalid next
3391 size (fast)". Examples of patterns that provoked this bug (shown in
3392 pcretest format) are:
3393
3394 /(?-x: )/x
3395 /(?x)(?-x: \s*#\s*)/
3396 /((?i)[\x{c0}])/8
3397 /(?i:[\x{c0}])/8
3398
3399 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3400 is now done differently.
3401
3402 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3403 wrapper classes; (b) implement a new function in the C++ scanner that is
3404 more efficient than the old way of doing things because it avoids levels of
3405 recursion in the regex matching; (c) add a paragraph to the documentation
3406 for the FullMatch() function.
3407
3408 6. The escape sequence \n was being treated as whatever was defined as
3409 "newline". Not only was this contrary to the documentation, which states
3410 that \n is character 10 (hex 0A), but it also went horribly wrong when
3411 "newline" was defined as CRLF. This has been fixed.
3412
3413 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3414 was being set to -1 for the "end of line" case (supposedly a value that no
3415 character can have). Though this value is never used (the check for end of
3416 line is "zero bytes in current character"), it caused compiler complaints.
3417 I've changed it to 0xffffffff.
3418
3419 8. In pcre_version.c, the version string was being built by a sequence of
3420 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3421 string (as it is for production releases) called a macro with an empty
3422 argument. The C standard says the result of this is undefined. The gcc
3423 compiler treats it as an empty string (which was what was wanted) but it is
3424 reported that Visual C gives an error. The source has been hacked around to
3425 avoid this problem.
3426
3427 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3428 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3429 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3430 of them did).
3431
3432 10. Originally, pcretest opened its input and output without "b"; then I was
3433 told that "b" was needed in some environments, so it was added for release
3434 5.0 to both the input and output. (It makes no difference on Unix-like
3435 systems.) Later I was told that it is wrong for the input on Windows. I've
3436 now abstracted the modes into two macros, to make it easier to fiddle with
3437 them, and removed "b" from the input mode under Windows.
3438
3439 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3440
3441 12. Added -help and --help to pcretest as an official way of being reminded
3442 of the options.
3443
3444 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3445 and pcrecpp.cc because they annoy compilers at high warning levels.
3446
3447 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3448
3449 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3450 scripts are not C programs :-) and which was not noticed because it works
3451 on Linux.
3452
3453 16. pcretest is supposed to handle any length of pattern and data line (as one
3454 line or as a continued sequence of lines) by extending its input buffer if
3455 necessary. This feature was broken for very long pattern lines, leading to
3456 a string of junk being passed to pcre_compile() if the pattern was longer
3457 than about 50K.
3458
3459 17. I have done a major re-factoring of the way pcre_compile() computes the
3460 amount of memory needed for a compiled pattern. Previously, there was code
3461 that made a preliminary scan of the pattern in order to do this. That was
3462 OK when PCRE was new, but as the facilities have expanded, it has become
3463 harder and harder to keep it in step with the real compile phase, and there
3464 have been a number of bugs (see for example, 4 above). I have now found a
3465 cunning way of running the real compile function in a "fake" mode that
3466 enables it to compute how much memory it would need, while actually only
3467 ever using a few hundred bytes of working memory and without too many
3468 tests of the mode. This should make future maintenance and development
3469 easier. A side effect of this work is that the limit of 200 on the nesting
3470 depth of parentheses has been removed (though this was never a serious
3471 limitation, I suspect). However, there is a downside: pcre_compile() now
3472 runs more slowly than before (30% or more, depending on the pattern). I
3473 hope this isn't a big issue. There is no effect on runtime performance.
3474
3475 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3476 newline (only possible for the last line of a file) and it was a
3477 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3478
3479 19. Added additional timing features to pcretest. (1) The -tm option now times
3480 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3481 separate command line item, by a number that specifies the number of
3482 repeats to use when timing. The default is 50000; this gives better
3483 precision, but takes uncomfortably long for very large patterns.
3484
3485 20. Extended pcre_study() to be more clever in cases where a branch of a
3486 subpattern has no definite first character. For example, (a*|b*)[cd] would
3487 previously give no result from pcre_study(). Now it recognizes that the
3488 first character must be a, b, c, or d.
3489
3490 21. There was an incorrect error "recursive call could loop indefinitely" if
3491 a subpattern (or the entire pattern) that was being tested for matching an
3492 empty string contained only one non-empty item after a nested subpattern.
3493 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3494 incorrectly, because the \d was being skipped in the check.
3495
3496 22. The pcretest program now has a new pattern option /B and a command line
3497 option -b, which is equivalent to adding /B to every pattern. This causes
3498 it to show the compiled bytecode, without the additional information that
3499 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3500 is the same as /B/I).
3501
3502 23. A new optimization is now able automatically to treat some sequences such
3503 as a*b as a*+b. More specifically, if something simple (such as a character
3504 or a simple class like \d) has an unlimited quantifier, and is followed by
3505 something that cannot possibly match the quantified thing, the quantifier
3506 is automatically "possessified".
3507
3508 24. A recursive reference to a subpattern whose number was greater than 39
3509 went wrong under certain circumstances in UTF-8 mode. This bug could also
3510 have affected the operation of pcre_study().
3511
3512 25. Realized that a little bit of performance could be had by replacing
3513 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3514
3515 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3516
3517 27. Possessive quantifiers such as a++ were previously implemented by turning
3518 them into atomic groups such as ($>a+). Now they have their own opcodes,
3519 which improves performance. This includes the automatically created ones
3520 from 23 above.
3521
3522 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3523 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3524 the first matched character to be a colon. This applied both to named and
3525 numbered groups.
3526
3527 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3528
3529 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3530 building PCRE on Windows easier, so I have included it in the distribution.
3531
3532 31. There is now a check in pcretest against a ridiculously large number being
3533 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3534 loop, the loop is abandoned.
3535
3536 32. Forward references to subpatterns in conditions such as (?(2)...) where
3537 subpattern 2 is defined later cause pcre_compile() to search forwards in
3538 the pattern for the relevant set of parentheses. This search went wrong
3539 when there were unescaped parentheses in a character class, parentheses
3540 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3541
3542 33. "Subroutine" calls and backreferences were previously restricted to
3543 referencing subpatterns earlier in the regex. This restriction has now
3544 been removed.
3545
3546 34. Added a number of extra features that are going to be in Perl 5.10. On the
3547 whole, these are just syntactic alternatives for features that PCRE had
3548 previously implemented using the Python syntax or my own invention. The
3549 other formats are all retained for compatibility.
3550
3551 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3552 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3553 also .NET compatible.
3554
3555 (b) A recursion or subroutine call to a named group can now be defined as
3556 (?&name) as well as (?P>name).
3557
3558 (c) A backreference to a named group can now be defined as \k<name> or
3559 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3560 5.10, are also .NET compatible.
3561
3562 (d) A conditional reference to a named group can now use the syntax
3563 (?(<name>) or (?('name') as well as (?(name).
3564
3565 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3566 groups (named and numbered) that are never evaluated inline, but can be
3567 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3568 is always false. There may be only one alternative in such a group.
3569
3570 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3571 as the simple (?(R). The condition is true only if the most recent
3572 recursion is that of the given number or name. It does not search out
3573 through the entire recursion stack.
3574
3575 (g) The escape \gN or \g{N} has been added, where N is a positive or
3576 negative number, specifying an absolute or relative reference.
3577
3578 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3579 some "unreachable code" warnings.
3580
3581 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3582 things, this adds five new scripts.
3583
3584 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3585 There were also incompatibilities regarding the handling of \Q..\E inside
3586 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3587 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3588
3589 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3590 matches an empty string, and forcibly breaks the loop. There were bugs in
3591 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3592 against aaaa the result was just "a" rather than "aaaa", for example. Two
3593 separate and independent bugs (that affected different cases) have been
3594 fixed.
3595
3596 39. Refactored the code to abolish the use of different opcodes for small
3597 capturing bracket numbers. This is a tidy that I avoided doing when I
3598 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3599 The new approach is not only tidier, it makes it possible to reduce the
3600 memory needed to fix the previous bug (38).
3601
3602 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3603 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3604 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3605 mode.
3606
3607 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3608 report.
3609
3610 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3611 copy construction and assignment in the C++ wrapper.
3612
3613 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3614 couple of bugs that could have given wrong results in the "--newline=crlf"
3615 case.
3616
3617 44. Added a number of casts and did some reorganization of signed/unsigned int
3618 variables following suggestions from Dair Grant. Also renamed the variable
3619 "this" as "item" because it is a C++ keyword.
3620
3621 45. Arranged for dftables to add
3622
3623 #include "pcre_internal.h"
3624
3625 to pcre_chartables.c because without it, gcc 4.x may remove the array
3626 definition from the final binary if PCRE is built into a static library and
3627 dead code stripping is activated.
3628
3629 46. For an unanchored pattern, if a match attempt fails at the start of a
3630 newline sequence, and the newline setting is CRLF or ANY, and the next two
3631 characters are CRLF, advance by two characters instead of one.
3632
3633
3634 Version 6.7 04-Jul-06
3635 ---------------------
3636
3637 1. In order to handle tests when input lines are enormously long, pcretest has
3638 been re-factored so that it automatically extends its buffers when
3639 necessary. The code is crude, but this _is_ just a test program. The
3640 default size has been increased from 32K to 50K.
3641
3642 2. The code in pcre_study() was using the value of the re argument before
3643 testing it for NULL. (Of course, in any sensible call of the function, it
3644 won't be NULL.)
3645
3646 3. The memmove() emulation function in pcre_internal.h, which is used on
3647 systems that lack both memmove() and bcopy() - that is, hardly ever -
3648 was missing a "static" storage class specifier.
3649
3650 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3651 containing an extended class (one that cannot be represented by a bitmap
3652 because it contains high-valued characters or Unicode property items, e.g.
3653 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3654 pattern, but PCRE should not loop if you do not (it no longer does).
3655 [Detail: two cases were found: (a) a repeated subpattern containing an
3656 extended class; (b) a recursive reference to a subpattern that followed a
3657 previous extended class. It wasn't skipping over the extended class
3658 correctly when UTF-8 mode was not set.]
3659
3660 5. A negated single-character class was not being recognized as fixed-length
3661 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3662 compile error "lookbehind assertion is not fixed length".
3663
3664 6. The RunPerlTest auxiliary script was showing an unexpected difference
3665 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3666 write a Perl script that can interpret lines of an input file either as
3667 byte characters or as UTF-8, which is what "perltest" was being required to
3668 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3669 can't do is switch easily at run time between having the "use utf8;" pragma
3670 or not. In the end, I fudged it by using the RunPerlTest script to insert
3671 "use utf8;" explicitly for the UTF-8 tests.
3672
3673 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3674 the end of the subject string, contrary to the documentation and to what
3675 Perl does. This was true of both matching functions. Now it matches only at
3676 the start of the subject and immediately after *internal* newlines.
3677
3678 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3679 a pointer to an int instead of a pointer to an unsigned long int. This
3680 caused problems on 64-bit systems.
3681
3682 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3683 instance of the 'standard' template library not being so standard".
3684
3685 10. There was no check on the number of named subpatterns nor the maximum
3686 length of a subpattern name. The product of these values is used to compute
3687 the size of the memory block for a compiled pattern. By supplying a very
3688 long subpattern name and a large number of named subpatterns, the size
3689 computation could be caused to overflow. This is now prevented by limiting
3690 the length of names to 32 characters, and the number of named subpatterns
3691 to 10,000.
3692
3693 11. Subpatterns that are repeated with specific counts have to be replicated in
3694 the compiled pattern. The size of memory for this was computed from the
3695 length of the subpattern and the repeat count. The latter is limited to
3696 65535, but there was no limit on the former, meaning that integer overflow
3697 could in principle occur. The compiled length of a repeated subpattern is
3698 now limited to 30,000 bytes in order to prevent this.
3699
3700 12. Added the optional facility to have named substrings with the same name.
3701
3702 13. Added the ability to use a named substring as a condition, using the
3703 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3704 are numbers (not recommended). Forward references are permitted.
3705
3706 14. Added forward references in named backreferences (if you see what I mean).
3707
3708 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3709 pattern could run off the end of the subject. For example, the pattern
3710 "(?s)(.{1,5})"8 did this with the subject "ab".
3711
3712 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3713 PCRE_CASELESS was set when matching characters that were quantified with ?
3714 or *.
3715
3716 17. A character class other than a single negated character that had a minimum
3717 but no maximum quantifier - for example [ab]{6,} - was not handled
3718 correctly by pce_dfa_exec(). It would match only one character.
3719
3720 18. A valid (though odd) pattern that looked like a POSIX character
3721 class but used an invalid character after [ (for example [[,abc,]]) caused
3722 pcre_compile() to give the error "Failed: internal error: code overflow" or
3723 in some cases to crash with a glibc free() error. This could even happen if
3724 the pattern terminated after [[ but there just happened to be a sequence of
3725 letters, a binary zero, and a closing ] in the memory that followed.
3726
3727 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3728 over the years. Originally (before any Unicode support), just the bottom 8
3729 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3730 output from "man perlunicode" includes this:
3731
3732 The regular expression compiler produces polymorphic opcodes. That
3733 is, the pattern adapts to the data and automatically switches to
3734 the Unicode character scheme when presented with Unicode data--or
3735 instead uses a traditional byte scheme when presented with byte
3736 data.
3737
3738 Sadly, a wide octal escape does not cause a switch, and in a string with
3739 no other multibyte characters, these octal escapes are treated as before.
3740 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3741 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3742 Unicode string.
3743
3744 I have not perpetrated such confusion in PCRE. Up till now, it took just
3745 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3746 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3747 translate to the appropriate multibyte character.
3748
3749 29. Applied some refactoring to reduce the number of warnings from Microsoft
3750 and Borland compilers. This has included removing the fudge introduced
3751 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3752 a warning about an unused variable.
3753
3754 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3755 characters since release 4.0, because Perl (from release 5.004) does not.
3756 [Or at least, is documented not to: some releases seem to be in conflict
3757 with the documentation.] However, when a pattern was studied with
3758 pcre_study() and all its branches started with \s, PCRE still included VT
3759 as a possible starting character. Of course, this did no harm; it just
3760 caused an unnecessary match attempt.
3761
3762 22. Removed a now-redundant internal flag bit that recorded the fact that case
3763 dependency changed within the pattern. This was once needed for "required
3764 byte" processing, but is no longer used. This recovers a now-scarce options
3765 bit. Also moved the least significant internal flag bit to the most-
3766 significant bit of the word, which was not previously used (hangover from
3767 the days when it was an int rather than a uint) to free up another bit for
3768 the future.
3769
3770 23. Added support for CRLF line endings as well as CR and LF. As well as the
3771 default being selectable at build time, it can now be changed at runtime
3772 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3773 specify that it is scanning data with non-default line endings.
3774
3775 24. Changed the definition of CXXLINK to make it agree with the definition of
3776 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3777
3778 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3779 recursions. This makes a big different to stack usage for some patterns.
3780
3781 26. If a subpattern containing a named recursion or subroutine reference such
3782 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3783 the space required for the compiled pattern went wrong and gave too small a
3784 value. Depending on the environment, this could lead to "Failed: internal
3785 error: code overflow at offset 49" or "glibc detected double free or
3786 corruption" errors.
3787
3788 27. Applied patches from Google (a) to support the new newline modes and (b) to
3789 advance over multibyte UTF-8 characters in GlobalReplace.
3790
3791 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3792 difference for some implementation of PCRE in some Windows version.
3793
3794 29. Added some extra testing facilities to pcretest:
3795
3796 \q<number> in a data line sets the "match limit" value
3797 \Q<number> in a data line sets the "match recursion limt" value
3798 -S <number> sets the stack size, where <number> is in megabytes
3799
3800 The -S option isn't available for Windows.
3801
3802
3803 Version 6.6 06-Feb-06
3804 ---------------------
3805
3806 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3807 in pcreposix.h. I have copied the definition from pcre.h.
3808
3809 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3810 because pcre.h is no longer a built file.
3811
3812 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3813 not normally included in the compiled code.
3814
3815
3816 Version 6.5 01-Feb-06
3817 ---------------------
3818
3819 1. When using the partial match feature with pcre_dfa_exec(), it was not
3820 anchoring the second and subsequent partial matches at the new starting
3821 point. This could lead to incorrect results. For example, with the pattern
3822 /1234/, partially matching against "123" and then "a4" gave a match.
3823
3824 2. Changes to pcregrep:
3825
3826 (a) All non-match returns from pcre_exec() were being treated as failures
3827 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3828 error message is output. Some extra information is given for the
3829 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3830 probably the only errors that are likely to be caused by users (by
3831 specifying a regex that has nested indefinite repeats, for instance).
3832 If there are more than 20 of these errors, pcregrep is abandoned.
3833
3834 (b) A binary zero was treated as data while matching, but terminated the
3835 output line if it was written out. This has been fixed: binary zeroes
3836 are now no different to any other data bytes.
3837
3838 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3839 used to set a locale for matching. The --locale=xxxx long option has
3840 been added (no short equivalent) to specify a locale explicitly on the
3841 pcregrep command, overriding the environment variables.
3842
3843 (d) When -B was used with -n, some line numbers in the output were one less
3844 than they should have been.
3845
3846 (e) Added the -o (--only-matching) option.
3847
3848 (f) If -A or -C was used with -c (count only), some lines of context were
3849 accidentally printed for the final match.
3850
3851 (g) Added the -H (--with-filename) option.
3852
3853 (h) The combination of options -rh failed to suppress file names for files
3854 that were found from directory arguments.
3855
3856 (i) Added the -D (--devices) and -d (--directories) options.
3857
3858 (j) Added the -F (--fixed-strings) option.
3859
3860 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3861
3862 (l) Added the --colo(u)r option.
3863
3864 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3865 is not present by default.
3866
3867 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3868 items such as (?R) or (?1), when the recursion could match a number of
3869 alternatives. If it matched one of the alternatives, but subsequently,
3870 outside the recursion, there was a failure, the code tried to back up into
3871 the recursion. However, because of the way PCRE is implemented, this is not
3872 possible, and the result was an incorrect result from the match.
3873
3874 In order to prevent this happening, the specification of recursion has
3875 been changed so that all such subpatterns are automatically treated as
3876 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3877
3878 4. I had overlooked the fact that, in some locales, there are characters for
3879 which isalpha() is true but neither isupper() nor islower() are true. In
3880 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3881 and ordfeminine) are like this. This affected the treatment of \w and \W
3882 when they appeared in character classes, but not when they appeared outside
3883 a character class. The bit map for "word" characters is now created
3884 separately from the results of isalnum() instead of just taking it from the
3885 upper, lower, and digit maps. (Plus the underscore character, of course.)
3886
3887 5. The above bug also affected the handling of POSIX character classes such as
3888 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3889 permanent tables. Instead, the bit maps for such a class were previously
3890 created as the appropriate unions of the upper, lower, and digit bitmaps.
3891 Now they are created by subtraction from the [[:word:]] class, which has
3892 its own bitmap.
3893
3894 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3895 It is created by subtracting the vertical space characters (\x09, \x0a,
3896 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3897 subtraction was done in the overall bitmap for a character class, meaning
3898 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3899 be recognized. This bug has been fixed.
3900
3901 7. Patches from the folks at Google:
3902
3903 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3904 real life, but is still worth protecting against".
3905
3906 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3907 regular expressions".
3908
3909 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3910 have it.
3911
3912 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3913 "configure" and the latter not, in order to fix a problem somebody had
3914 with compiling the Arg class on HP-UX.
3915
3916 (e) Improve the error-handling of the C++ wrapper a little bit.
3917
3918 (f) New tests for checking recursion limiting.
3919
3920 8. The pcre_memmove() function, which is used only if the environment does not
3921 have a standard memmove() function (and is therefore rarely compiled),
3922 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3923 returning a result (though PCRE never actually uses the result).
3924
3925 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3926 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3927 returned instead of calling malloc() with an overflowing number that would
3928 most likely cause subsequent chaos.
3929
3930 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3931
3932 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3933 with this option is matched, the nmatch and pmatch options of regexec() are
3934 ignored.
3935
3936 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3937 provided in case anyone wants to the the POSIX interface with UTF-8
3938 strings.
3939
3940 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3941 C++ linking (needed for some HP-UX environments).
3942
3943 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3944 (unused parameter) and in the pcre_printint() function (omitted "default"
3945 switch label when the default is to do nothing).
3946
3947 15. Added some code to make it possible, when PCRE is compiled as a C++
3948 library, to replace subject pointers for pcre_exec() with a smart pointer
3949 class, thus making it possible to process discontinuous strings.
3950
3951 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3952 much the same function. They were added by different people who were trying
3953 to make PCRE easy to compile on non-Unix systems. It has been suggested
3954 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3955 for compiling on Windows systems. I have therefore replaced it with
3956 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3957 defaults to "extern" for C or "extern C" for C++, which works fine on
3958 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3959 SCOPE with something explicit in config.h. In addition:
3960
3961 (a) pcreposix.h still had just "extern" instead of either of these macros;
3962 I have replaced it with PCRE_DATA_SCOPE.
3963
3964 (b) Functions such as _pcre_xclass(), which are internal to the library,
3965 but external in the C sense, all had PCRE_EXPORT in their definitions.
3966 This is apparently wrong for the Windows case, so I have removed it.
3967 (It makes no difference on Unix-like systems.)
3968
3969 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3970 of recursive calls to match(). This is different to MATCH_LIMIT because
3971 that limits the total number of calls to match(), not all of which increase
3972 the depth of recursion. Limiting the recursion depth limits the amount of
3973 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3974 when PCRE is compiled, and changed at run time. A patch from Google adds
3975 this functionality to the C++ interface.
3976
3977 18. Changes to the handling of Unicode character properties:
3978
3979 (a) Updated the table to Unicode 4.1.0.
3980
3981 (b) Recognize characters that are not in the table as "Cn" (undefined).
3982
3983 (c) I revised the way the table is implemented to a much improved format
3984 which includes recognition of ranges. It now supports the ranges that
3985 are defined in UnicodeData.txt, and it also amalgamates other
3986 characters into ranges. This has reduced the number of entries in the
3987 table from around 16,000 to around 3,000, thus reducing its size
3988 considerably. I realized I did not need to use a tree structure after
3989 all - a binary chop search is just as efficient. Having reduced the
3990 number of entries, I extended their size from 6 bytes to 8 bytes to
3991 allow for more data.
3992
3993 (d) Added support for Unicode script names via properties such as \p{Han}.
3994
3995 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3996 matching that character.
3997
3998 20. When matching a repeated Unicode property with a minimum greater than zero,
3999 (for example \pL{2,}), PCRE could look past the end of the subject if it
4000 reached it while seeking the minimum number of characters. This could
4001 happen only if some of the characters were more than one byte long, because
4002 there is a check for at least the minimum number of bytes.
4003
4004 21. Refactored the implementation of \p and \P so as to be more general, to
4005 allow for more different types of property in future. This has changed the
4006 compiled form incompatibly. Anybody with saved compiled patterns that use
4007 \p or \P will have to recompile them.
4008
4009 22. Added "Any" and "L&" to the supported property types.
4010
4011 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4012 but give a compile time error if the value is greater than 0xff.
4013
4014 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4015 accidentally not being installed or uninstalled.
4016
4017 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4018 made were to insert the current release number. This seemed silly, because
4019 it made things harder for people building PCRE on systems that don't run
4020 "configure". I have turned pcre.h into a distributed file, no longer built
4021 by "configure", with the version identification directly included. There is
4022 no longer a pcre.h.in file.
4023
4024 However, this change necessitated a change to the pcre-config script as
4025 well. It is built from pcre-config.in, and one of the substitutions was the
4026 release number. I have updated configure.ac so that ./configure now finds
4027 the release number by grepping pcre.h.
4028
4029 26. Added the ability to run the tests under valgrind.
4030
4031
4032 Version 6.4 05-Sep-05
4033 ---------------------
4034
4035 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4036 "--" to be printed when multiple files were scanned, even when none of the
4037 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4038 consider it to be a bug, and have restored the previous behaviour.
4039
4040 2. A couple of code tidies to get rid of compiler warnings.
4041
4042 3. The pcretest program used to cheat by referring to symbols in the library
4043 whose names begin with _pcre_. These are internal symbols that are not
4044 really supposed to be visible externally, and in some environments it is
4045 possible to suppress them. The cheating is now confined to including
4046 certain files from the library's source, which is a bit cleaner.
4047
4048 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4049 file's purpose clearer.
4050
4051 5. Reorganized pcre_ucp_findchar().
4052
4053
4054 Version 6.3 15-Aug-05
4055 ---------------------
4056
4057 1. The file libpcre.pc.in did not have general read permission in the tarball.
4058
4059 2. There were some problems when building without C++ support:
4060
4061 (a) If C++ support was not built, "make install" and "make test" still
4062 tried to test it.
4063
4064 (b) There were problems when the value of CXX was explicitly set. Some
4065 changes have been made to try to fix these, and ...
4066
4067 (c) --disable-cpp can now be used to explicitly disable C++ support.
4068
4069 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4070 backslash in a target when C++ was disabled. This confuses some
4071 versions of "make", apparently. Using an intermediate variable solves
4072 this. (Same for CPP_LOBJ.)
4073
4074 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4075 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4076 necessary on certain architectures.
4077
4078 4. Added a setting of -export-symbols-regex to the link command to remove
4079 those symbols that are exported in the C sense, but actually are local
4080 within the library, and not documented. Their names all begin with
4081 "_pcre_". This is not a perfect job, because (a) we have to except some
4082 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4083 available (and never for static libraries). I have made a note to try to
4084 find a way round (a) in the future.
4085
4086
4087 Version 6.2 01-Aug-05
4088 ---------------------
4089
4090 1. There was no test for integer overflow of quantifier values. A construction
4091 such as {1111111111111111} would give undefined results. What is worse, if
4092 a minimum quantifier for a parenthesized subpattern overflowed and became
4093 negative, the calculation of the memory size went wrong. This could have
4094 led to memory overwriting.
4095
4096 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4097
4098 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4099 operating environments where this matters.
4100
4101 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4102 PCRE options from within the C++ wrapper.
4103
4104 5. Named capturing subpatterns were not being correctly counted when a pattern
4105 was compiled. This caused two problems: (a) If there were more than 100
4106 such subpatterns, the calculation of the memory needed for the whole
4107 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4108 back references of the form \12, where the number was greater than 9, were
4109 not recognized as back references, even though there were sufficient
4110 previous subpatterns.
4111
4112 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4113 versions of gcc, e.g. 2.95.4.
4114
4115
4116 Version 6.1 21-Jun-05
4117 ---------------------
4118
4119 1. There was one reference to the variable "posix" in pcretest.c that was not
4120 surrounded by "#if !defined NOPOSIX".
4121
4122 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4123 the cross-check on the old pcre_info() function, for the benefit of the
4124 cut-down version of PCRE that is currently imported into Exim.
4125
4126 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4127 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4128 patterns that start (?i) but I don't think that matters. The use of (?i) is
4129 just an example; this all applies to the other options as well.
4130
4131 4. Since libtool seems to echo the compile commands it is issuing, the output
4132 from "make" can be reduced a bit by putting "@" in front of each libtool
4133 compile command.
4134
4135 5. Patch from the folks at Google for configure.in to be a bit more thorough
4136 in checking for a suitable C++ installation before trying to compile the
4137 C++ stuff. This should fix a reported problem when a compiler was present,
4138 but no suitable headers.
4139
4140 6. The man pages all had just "PCRE" as their title. I have changed them to
4141 be the relevant file name. I have also arranged that these names are
4142 retained in the file doc/pcre.txt, which is a concatenation in text format
4143 of all the man pages except the little individual ones for each function.
4144
4145 7. The NON-UNIX-USE file had not been updated for the different set of source
4146 files that come with release 6. I also added a few comments about the C++
4147 wrapper.
4148
4149
4150 Version 6.0 07-Jun-05
4151 ---------------------
4152
4153 1. Some minor internal re-organization to help with my DFA experiments.
4154
4155 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4156 didn't matter for the library itself when fully configured, but did matter
4157 when compiling without UCP support, or within Exim, where the ucp files are
4158 not imported.
4159
4160 3. Refactoring of the library code to split up the various functions into
4161 different source modules. The addition of the new DFA matching code (see
4162 below) to a single monolithic source would have made it really too
4163 unwieldy, quite apart from causing all the code to be include in a
4164 statically linked application, when only some functions are used. This is
4165 relevant even without the DFA addition now that patterns can be compiled in
4166 one application and matched in another.
4167
4168 The downside of splitting up is that there have to be some external
4169 functions and data tables that are used internally in different modules of
4170 the library but which are not part of the API. These have all had their
4171 names changed to start with "_pcre_" so that they are unlikely to clash
4172 with other external names.
4173
4174 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4175 a different (DFA) algorithm. Although it is slower than the original
4176 function, it does have some advantages for certain types of matching
4177 problem.
4178
4179 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4180 including restarting after a partial match.
4181
4182 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4183 defined when compiling for Windows was sent to me. I have put it into the
4184 code, though I have no means of testing or verifying it.
4185
4186 7. Added the pcre_refcount() auxiliary function.
4187
4188 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4189 match before or at the first newline in the subject string. In pcretest,
4190 the /f option on a pattern can be used to set this.
4191
4192 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4193 would behave wrongly. This has been present in PCRE since release 4.0.
4194
4195 10. A number of changes to the pcregrep command:
4196
4197 (a) Refactored how -x works; insert ^(...)$ instead of setting
4198 PCRE_ANCHORED and checking the length, in preparation for adding
4199 something similar for -w.
4200
4201 (b) Added the -w (match as a word) option.
4202
4203 (c) Refactored the way lines are read and buffered so as to have more
4204 than one at a time available.
4205
4206 (d) Implemented a pcregrep test script.
4207
4208 (e) Added the -M (multiline match) option. This allows patterns to match
4209 over several lines of the subject. The buffering ensures that at least
4210 8K, or the rest of the document (whichever is the shorter) is available
4211 for matching (and similarly the previous 8K for lookbehind assertions).
4212
4213 (f) Changed the --help output so that it now says
4214
4215 -w, --word-regex(p)
4216
4217 instead of two lines, one with "regex" and the other with "regexp"
4218 because that confused at least one person since the short forms are the
4219 same. (This required a bit of code, as the output is generated
4220 automatically from a table. It wasn't just a text change.)
4221
4222 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4223 option but starts with a hyphen. Could be a pattern or a path name
4224 starting with a hyphen, for instance.
4225
4226 (h) "-" can be given as a file name to represent stdin.
4227
4228 (i) When file names are being printed, "(standard input)" is used for
4229 the standard input, for compatibility with GNU grep. Previously
4230 "<stdin>" was used.
4231
4232 (j) The option --label=xxx can be used to supply a name to be used for
4233 stdin when file names are being printed. There is no short form.
4234
4235 (k) Re-factored the options decoding logic because we are going to add
4236 two more options that take data. Such options can now be given in four
4237 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4238
4239 (l) Added the -A, -B, and -C options for requesting that lines of context
4240 around matches be printed.
4241
4242 (m) Added the -L option to print the names of files that do not contain
4243 any matching lines, that is, the complement of -l.
4244
4245 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4246 continue to scan other files.
4247
4248 (o) The -s option was incorrectly implemented. For compatibility with other
4249 greps, it now suppresses the error message for a non-existent or non-
4250 accessible file (but not the return code). There is a new option called
4251 -q that suppresses the output of matching lines, which was what -s was
4252 previously doing.
4253
4254 (p) Added --include and --exclude options to specify files for inclusion
4255 and exclusion when recursing.
4256
4257 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4258 Hopefully, it now does.
4259
4260 12. Missing cast in pcre_study().
4261
4262 13. Added an "uninstall" target to the makefile.
4263
4264 14. Replaced "extern" in the function prototypes in Makefile.in with
4265 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4266 world, but is set differently for Windows.
4267
4268 15. Added a second compiling function called pcre_compile2(). The only
4269 difference is that it has an extra argument, which is a pointer to an
4270 integer error code. When there is a compile-time failure, this is set
4271 non-zero, in addition to the error test pointer being set to point to an
4272 error message. The new argument may be NULL if no error number is required
4273 (but then you may as well call pcre_compile(), which is now just a
4274 wrapper). This facility is provided because some applications need a
4275 numeric error indication, but it has also enabled me to tidy up the way
4276 compile-time errors are handled in the POSIX wrapper.
4277
4278 16. Added VPATH=.libs to the makefile; this should help when building with one
4279 prefix path and installing with another. (Or so I'm told by someone who
4280 knows more about this stuff than I do.)
4281
4282 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4283 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4284 match everything, including newlines. This is not POSIX-compatible, but
4285 somebody wanted the feature. From pcretest it can be activated by using
4286 both the P and the s flags.
4287
4288 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4289
4290 19. libpcre.pc was being incorrectly installed as executable.
4291
4292 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4293 it now also looks for '\r' so that it will work unmodified on Windows.
4294
4295 21. Added Google's contributed C++ wrapper to the distribution.
4296
4297 22. Added some untidy missing memory free() calls in pcretest, to keep
4298 Electric Fence happy when testing.
4299
4300
4301
4302 Version 5.0 13-Sep-04
4303 ---------------------
4304
4305 1. Internal change: literal characters are no longer packed up into items
4306 containing multiple characters in a single byte-string. Each character
4307 is now matched using a separate opcode. However, there may be more than one
4308 byte in the character in UTF-8 mode.
4309
4310 2. The pcre_callout_block structure has two new fields: pattern_position and
4311 next_item_length. These contain the offset in the pattern to the next match
4312 item, and its length, respectively.
4313
4314 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4315 insertion of callouts before each pattern item. Added the /C option to
4316 pcretest to make use of this.
4317
4318 4. On the advice of a Windows user, the lines
4319
4320 #if defined(_WIN32) || defined(WIN32)
4321 _setmode( _fileno( stdout ), 0x8000 );
4322 #endif /* defined(_WIN32) || defined(WIN32) */
4323
4324 have been added to the source of pcretest. This apparently does useful
4325 magic in relation to line terminators.
4326
4327 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4328 for the benefit of those environments where the "b" makes a difference.
4329
4330 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4331 to know about it. I have put a hack into configure.in that adds in code
4332 to set GCC=yes if CC=icc. This seems to end up at a point in the
4333 generated configure script that is early enough to affect the setting of
4334 compiler options, which is what is needed, but I have no means of testing
4335 whether it really works. (The user who reported this had patched the
4336 generated configure script, which of course I cannot do.)
4337
4338 LATER: After change 22 below (new libtool files), the configure script
4339 seems to know about icc (and also ecc). Therefore, I have commented out
4340 this hack in configure.in.
4341
4342 7. Added support for pkg-config (2 patches were sent in).
4343
4344 8. Negated POSIX character classes that used a combination of internal tables
4345 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4346 [[:^ascii]]. Typically, they would match almost any characters. The other
4347 POSIX classes were not broken in this way.
4348
4349 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4350 to find the match, as PCRE was deluded into thinking that the match had to
4351 start at the start point or following a newline. The same bug applied to
4352 patterns with negative forward assertions or any backward assertions
4353 preceding ".*" at the start, unless the pattern required a fixed first
4354 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4355
4356 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4357 starting at the last subject character, bytes beyond the end of the subject
4358 string were read.
4359
4360 11. Renamed the variable "class" as "classbits" to make life easier for C++
4361 users. (Previously there was a macro definition, but it apparently wasn't
4362 enough.)
4363
4364 12. Added the new field "tables" to the extra data so that tables can be passed
4365 in at exec time, or the internal tables can be re-selected. This allows
4366 a compiled regex to be saved and re-used at a later time by a different
4367 program that might have everything at different addresses.
4368
4369 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4370 -R library as well as a -L library.
4371
4372 14. The debugging options of pcretest (-d on the command line or D on a
4373 pattern) showed incorrect output for anything following an extended class
4374 that contained multibyte characters and which was followed by a quantifier.
4375
4376 15. Added optional support for general category Unicode character properties
4377 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4378 support. It adds about 90K to the size of the library. The meanings of the
4379 inbuilt class escapes such as \d and \s have NOT been changed.
4380
4381 16. Updated pcredemo.c to include calls to free() to release the memory for the
4382 compiled pattern.
4383
4384 17. The generated file chartables.c was being created in the source directory
4385 instead of in the building directory. This caused the build to fail if the
4386 source directory was different from the building directory, and was
4387 read-only.
4388
4389 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4390 file. No doubt somebody will tell me if they don't make sense... Also added
4391 Dan Mooney's comments about building on OpenVMS.
4392
4393 19. Added support for partial matching via the PCRE_PARTIAL option for
4394 pcre_exec() and the \P data escape in pcretest.
4395
4396 20. Extended pcretest with 3 new pattern features:
4397
4398 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4399 write the compiled pattern to the file whose name is "rest-of-line".
4400 This is a straight binary dump of the data, with the saved pointer to
4401 the character tables forced to be NULL. The study data, if any, is
4402 written too. After writing, pcretest reads a new pattern.
4403
4404 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4405 compiled pattern from the given file. There must not be any
4406 occurrences of "<" in the file name (pretty unlikely); if there are,
4407 pcretest will instead treat the initial "<" as a pattern delimiter.
4408 After reading in the pattern, pcretest goes on to read data lines as
4409 usual.
4410
4411 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4412 and 16-bit fields in a compiled pattern, to simulate a pattern that
4413 was compiled on a host of opposite endianness.
4414
4415 21. The pcre-exec() function can now cope with patterns that were compiled on
4416 hosts of opposite endianness, with this restriction:
4417
4418 As for any compiled expression that is saved and used later, the tables
4419 pointer field cannot be preserved; the extra_data field in the arguments
4420 to pcre_exec() should be used to pass in a tables address if a value
4421 other than the default internal tables were used at compile time.
4422
4423 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4424 now diagnosed as an error. Previously, most of the time, a negative number
4425 would have been treated as zero, but if in addition "ovector" was passed as
4426 NULL, a crash could occur.
4427
4428 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4429 new versions from the libtool 1.5 distribution (the last one is a copy of
4430 a file called libtool.m4). This seems to have fixed the need to patch
4431 "configure" to support Darwin 1.3 (which I used to do). However, I still
4432 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4433 workstation).
4434
4435 24. Changed the PCRE licence to be the more standard "BSD" licence.
4436
4437
4438 Version 4.5 01-Dec-03
4439 ---------------------
4440
4441 1. There has been some re-arrangement of the code for the match() function so
4442 that it can be compiled in a version that does not call itself recursively.
4443 Instead, it keeps those local variables that need separate instances for
4444 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4445 needs to "recurse". Keeping track of where control must go is done by means
4446 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4447 hide most of the details from the main code, and operates only if
4448 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4449 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4450 operating.
4451
4452 To make it easier for callers to provide specially tailored get/free
4453 functions for this usage, two new functions, pcre_stack_malloc, and
4454 pcre_stack_free, are used. They are always called in strict stacking order,
4455 and the size of block requested is always the same.
4456
4457 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4458 PCRE has been compiled to use the stack or the heap for recursion. The
4459 -C option of pcretest uses this to show which version is compiled.
4460
4461 A new data escape \S, is added to pcretest; it causes the amounts of store
4462 obtained and freed by both kinds of malloc/free at match time to be added
4463 to the output.
4464
4465 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4466 what's available on my current Linux desktop machine.
4467
4468 3. When matching a UTF-8 string, the test for a valid string at the start has
4469 been extended. If start_offset is not zero, PCRE now checks that it points
4470 to a byte that is the start of a UTF-8 character. If not, it returns
4471 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4472 this is necessary because there may be backward assertions in the pattern.
4473 When matching the same subject several times, it may save resources to use
4474 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4475
4476 4. The code for checking the validity of UTF-8 strings has been tightened so
4477 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4478 containing "overlong sequences".
4479
4480 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4481 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4482 should have been "&", but it just so happened that all the cases this let
4483 through by mistake were picked up later in the function.
4484
4485 6. I had used a variable called "isblank" - this is a C99 function, causing
4486 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4487
4488 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4489 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4490 so the version is shown; (c) stop "make test" echoing "./RunTest".
4491
4492 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4493
4494 9. The prototype for memmove() for systems that don't have it was using
4495 size_t, but the inclusion of the header that defines size_t was later. I've
4496 moved the #includes for the C headers earlier to avoid this.
4497
4498 10. Added some adjustments to the code to make it easier to compiler on certain
4499 special systems:
4500
4501 (a) Some "const" qualifiers were missing.
4502 (b) Added the macro EXPORT before all exported functions; by default this
4503 is defined to be empty.
4504 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4505 that it reads its output file name as an argument instead of writing
4506 to the standard output and assuming this can be redirected.
4507
4508 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4509 class containing characters with values greater than 255, PCRE compilation
4510 went into a loop.
4511
4512 12. A recursive reference to a subpattern that was within another subpattern
4513 that had a minimum quantifier of zero caused PCRE to crash. For example,
4514 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4515 recursion. If the recursively-called subpattern itself had a zero repeat,
4516 that was OK.
4517
4518 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4519 buffer into which it was copied (for escape processing) was still set at
4520 1024, so long lines caused crashes.
4521
4522 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4523 "internal error: code overflow...". This applied to any character class
4524 that was followed by a possessive quantifier.
4525
4526 15. Modified the Makefile to add libpcre.la as a prerequisite for
4527 libpcreposix.la because I was told this is needed for a parallel build to
4528 work.
4529
4530 16. If a pattern that contained .* following optional items at the start was
4531 studied, the wrong optimizing data was generated, leading to matching
4532 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4533 matching string must start with a or b or c. The correct conclusion for
4534 this pattern is that a match can start with any character.
4535
4536
4537 Version 4.4 13-Aug-03
4538 ---------------------
4539
4540 1. In UTF-8 mode, a character class containing characters with values between
4541 127 and 255 was not handled correctly if the compiled pattern was studied.
4542 In fixing this, I have also improved the studying algorithm for such
4543 classes (slightly).
4544
4545 2. Three internal functions had redundant arguments passed to them. Removal
4546 might give a very teeny performance improvement.
4547
4548 3. Documentation bug: the value of the capture_top field in a callout is *one
4549 more than* the number of the hightest numbered captured substring.
4550
4551 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4552 in incorrectly linking with a previously installed version. They now link
4553 explicitly with libpcre.la.
4554
4555 5. configure.in no longer needs to recognize Cygwin specially.
4556
4557 6. A problem in pcre.in for Windows platforms is fixed.
4558
4559 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4560 pcretest, it used to include the size of the study block as part of its
4561 output. Unfortunately, the structure contains a field that has a different
4562 size on different hardware architectures. This meant that the tests that
4563 showed this size failed. As the block is currently always of a fixed size,
4564 this information isn't actually particularly useful in pcretest output, so
4565 I have just removed it.
4566
4567 8. Three pre-processor statements accidentally did not start in column 1.
4568 Sadly, there are *still* compilers around that complain, even though
4569 standard C has not required this for well over a decade. Sigh.
4570
4571 9. In pcretest, the code for checking callouts passed small integers in the
4572 callout_data field, which is a void * field. However, some picky compilers
4573 complained about the casts involved for this on 64-bit systems. Now
4574 pcretest passes the address of the small integer instead, which should get
4575 rid of the warnings.
4576
4577 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4578 both compile and run time, and gives an error if an invalid UTF-8 sequence
4579 is found. There is a option for disabling this check in cases where the
4580 string is known to be correct and/or the maximum performance is wanted.
4581
4582 11. In response to a bug report, I changed one line in Makefile.in from
4583
4584 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4585 to
4586 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4587
4588 to look similar to other lines, but I have no way of telling whether this
4589 is the right thing to do, as I do not use Windows. No doubt I'll get told
4590 if it's wrong...
4591
4592
4593 Version 4.3 21-May-03
4594 ---------------------
4595
4596 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4597 Makefile.
4598
4599 2. Some refactoring to improve the quality of the code:
4600
4601 (i) The utf8_table... variables are now declared "const".
4602
4603 (ii) The code for \cx, which used the "case flipping" table to upper case
4604 lower case letters, now just substracts 32. This is ASCII-specific,
4605 but the whole concept of \cx is ASCII-specific, so it seems
4606 reasonable.
4607
4608 (iii) PCRE was using its character types table to recognize decimal and
4609 hexadecimal digits in the pattern. This is silly, because it handles
4610 only 0-9, a-f, and A-F, but the character types table is locale-
4611 specific, which means strange things might happen. A private
4612 table is now used for this - though it costs 256 bytes, a table is
4613 much faster than multiple explicit tests. Of course, the standard
4614 character types table is still used for matching digits in subject
4615 strings against \d.
4616
4617 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4618 ending in _t are). So I've renamed it as ESC_tee.
4619
4620 3. The first argument for regexec() in the POSIX wrapper should have been
4621 defined as "const".
4622
4623 4. Changed pcretest to use malloc() for its buffers so that they can be
4624 Electric Fenced for debugging.
4625
4626 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4627 to read one or more bytes before the start of the subject string. Often this
4628 had no effect on PCRE's behaviour, but in some circumstances it could
4629 provoke a segmentation fault.
4630
4631 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4632 to try to read one or more bytes before the start of the subject string.
4633
4634 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4635 UTF-8 support could misbehave in various ways if the subject string
4636 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4637 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4638 back over UTF-8 characters.)
4639
4640
4641 Version 4.2 14-Apr-03
4642 ---------------------
4643
4644 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4645
4646 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4647 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4648 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4649 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4650 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4651 and BUILD_EXEEXT
4652 Note: automatic setting of the BUILD variables is not yet working
4653 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4654 compile-time but not at link-time
4655 [LINK]: use for linking executables only
4656 make different versions for Windows and non-Windows
4657 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4658 libraries
4659 [LINK_FOR_BUILD]: new variable
4660 [OBJEXT]: use throughout
4661 [EXEEXT]: use throughout
4662 <winshared>: new target
4663 <wininstall>: new target
4664 <dftables.o>: use native compiler
4665 <dftables>: use native linker
4666 <install>: handle Windows platform correctly
4667 <clean>: ditto
4668 <check>: ditto
4669 copy DLL to top builddir before testing
4670
4671 As part of these changes, -no-undefined was removed again. This was reported
4672 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4673 in any case.
4674
4675 3. Some tidies to get rid of compiler warnings:
4676
4677 . In the match_data structure, match_limit was an unsigned long int, whereas
4678 match_call_count was an int. I've made them both unsigned long ints.
4679
4680 . In pcretest the fact that a const uschar * doesn't automatically cast to
4681 a void * provoked a warning.
4682
4683 . Turning on some more compiler warnings threw up some "shadow" variables
4684 and a few more missing casts.
4685
4686 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4687 option, a class that contained a single character with a value between 128
4688 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4689
4690 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4691 option, a class that contained several characters, but with at least one
4692 whose value was between 128 and 255 caused PCRE to crash.
4693
4694
4695 Version 4.1 12-Mar-03
4696 ---------------------
4697
4698 1. Compiling with gcc -pedantic found a couple of places where casts were
4699 needed, and a string in dftables.c that was longer than standard compilers are
4700 required to support.
4701
4702 2. Compiling with Sun's compiler found a few more places where the code could
4703 be tidied up in order to avoid warnings.
4704
4705 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4706 first of these names is deprecated in the latest Autoconf in favour of the name
4707 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4708 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4709 analogy I have changed it to CFLAGS_FOR_BUILD.
4710
4711 4. Added -no-undefined to the linking command in the Makefile, because this is
4712 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4713 linking step for the pcreposix library.
4714
4715 5. PCRE was failing to diagnose the case of two named groups with the same
4716 name.
4717
4718 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4719 literal character that is needed in the subject for a match, and scans along to
4720 ensure that it is present before embarking on the full matching process. This
4721 saves time in cases of nested unlimited repeats that are never going to match.
4722 Problem: the scan can take a lot of time if the subject is very long (e.g.
4723 megabytes), thus penalizing straightforward matches. It is now done only if the
4724 amount of subject to be scanned is less than 1000 bytes.
4725
4726 7. A lesser problem with the same optimization is that it was recording the
4727 first character of an anchored pattern as "needed", thus provoking a search
4728 right along the subject, even when the first match of the pattern was going to
4729 fail. The "needed" character is now not set for anchored patterns, unless it
4730 follows something in the pattern that is of non-fixed length. Thus, it still
4731 fulfils its original purpose of finding quick non-matches in cases of nested
4732 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4733
4734
4735 Version 4.0 17-Feb-03
4736 ---------------------
4737
4738 1. If a comment in an extended regex that started immediately after a meta-item
4739 extended to the end of string, PCRE compiled incorrect data. This could lead to
4740 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4741
4742 2. Moved to autoconf 2.53 and libtool 1.4.2.
4743
4744 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4745 the special perltest8 script is no longer needed - all the tests can be run
4746 from a single perltest script.
4747
4748 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4749 by \s. It has now been removed in PCRE. This means it isn't recognized as
4750 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4751 class [:space:] *does* include VT, thereby creating a mess.
4752
4753 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4754 space and tab.
4755
4756 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4757 its new features into the main test script, reducing the number of scripts.
4758
4759 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4760 were backward compatible, and made the (?i) apply to the whole pattern, as if
4761 /i were given. Now it behaves more logically, and applies the option setting
4762 only to what follows. PCRE has been changed to follow suit. However, if it
4763 finds options settings right at the start of the pattern, it extracts them into
4764 the global options, as before. Thus, they show up in the info data.
4765
4766 8. Added support for the \Q...\E escape sequence. Characters in between are
4767 treated as literals. This is slightly different from Perl in that $ and @ are
4768 also handled as literals inside the quotes. In Perl, they will cause variable
4769 interpolation. Note the following examples:
4770
4771 Pattern PCRE matches Perl matches
4772
4773 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4774 \Qabc\$xyz\E abc\$xyz abc\$xyz
4775 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4776
4777 For compatibility with Perl, \Q...\E sequences are recognized inside character
4778 classes as well as outside them.
4779
4780 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4781 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4782 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4783 signed/unsigned warnings.
4784
4785 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4786 option for pcretest, so I've replaced it by a simple function that does just
4787 that job.
4788
4789 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4790 "pcregrep -".
4791
4792 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4793 Java package. This provides some syntactic sugar for simple cases of what my
4794 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4795 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4796 item, you can use this simplified notation. Note that only makes sense with
4797 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4798 greediness, whatever the setting of the PCRE_UNGREEDY option.
4799
4800 13. A change of greediness default within a pattern was not taking effect at
4801 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4802 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4803 was abstracted outside.
4804
4805 14. PCRE now supports the \G assertion. It is true when the current matching
4806 position is at the start point of the match. This differs from \A when the
4807 starting offset is non-zero. Used with the /g option of pcretest (or similar
4808 code), it works in the same way as it does for Perl's /g option. If all
4809 alternatives of a regex begin with \G, the expression is anchored to the start
4810 match position, and the "anchored" flag is set in the compiled expression.
4811
4812 15. Some bugs concerning the handling of certain option changes within patterns
4813 have been fixed. These applied to options other than (?ims). For example,
4814 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4815 been the other way round. Some of this was related to change 7 above.
4816
4817 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4818 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4819 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4820 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4821
4822 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4823 mode. Unlike ".", it always matches newline, whatever the setting of
4824 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4825 assertions. Perl allows it, but it doesn't (in general) work because it can't
4826 calculate the length of the lookbehind. At least, that's the case for Perl
4827 5.8.0 - I've been told they are going to document that it doesn't work in
4828 future.
4829
4830 18. Added an error diagnosis for escapes that PCRE does not support: these are
4831 \L, \l, \N, \P, \p, \U, \u, and \X.
4832
4833 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4834 reading past the end of the pattern in cases such as /[abcd/.
4835
4836 20. PCRE was getting more memory than necessary for patterns with classes that
4837 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4838
4839 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4840 compiling PCRE for use with Virtual Pascal.
4841
4842 22. Small fix to the Makefile to make it work properly if the build is done
4843 outside the source tree.
4844
4845 23. Added a new extension: a condition to go with recursion. If a conditional
4846 subpattern starts with (?(R) the "true" branch is used if recursion has
4847 happened, whereas the "false" branch is used only at the top level.
4848
4849 24. When there was a very long string of literal characters (over 255 bytes
4850 without UTF support, over 250 bytes with UTF support), the computation of how
4851 much memory was required could be incorrect, leading to segfaults or other
4852 strange effects.
4853
4854 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4855 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4856 there was a subsequent back reference to those brackets. This meant that, for
4857 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4858 possible to check for precisely this case. All we can do is abandon the
4859 optimization if .* occurs inside capturing brackets when there are any back
4860 references whatsoever. (See below for a better fix that came later.)
4861
4862 26. The handling of the optimization for finding the first character of a
4863 non-anchored pattern, and for finding a character that is required later in the
4864 match were failing in some cases. This didn't break the matching; it just
4865 failed to optimize when it could. The way this is done has been re-implemented.
4866
4867 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4868
4869 28. Added a new feature that provides some of the functionality that Perl
4870 provides with (?{...}). The facility is termed a "callout". The way it is done
4871 in PCRE is for the caller to provide an optional function, by setting
4872 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4873 global variable. By default it is unset, which disables all calling out. To get
4874 the function called, the regex must include (?C) at appropriate points. This
4875 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4876 This provides a means of identifying different callout points. When PCRE
4877 reaches such a point in the regex, if pcre_callout has been set, the external
4878 function is called. It is provided with data in a structure called
4879 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4880 matching continues; if it returns a non-zero value, the match at the current
4881 point fails. However, backtracking will occur if possible. [This was changed
4882 later and other features added - see item 49 below.]
4883
4884 29. pcretest is upgraded to test the callout functionality. It provides a
4885 callout function that displays information. By default, it shows the start of
4886 the match and the current position in the text. There are some new data escapes
4887 to vary what happens:
4888
4889 \C+ in addition, show current contents of captured substrings
4890 \C- do not supply a callout function
4891 \C!n return 1 when callout number n is reached
4892 \C!n!m return 1 when callout number n is reached for the mth time
4893
4894 30. If pcregrep was called with the -l option and just a single file name, it
4895 output "<stdin>" if a match was found, instead of the file name.
4896
4897 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4898 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4899 pcre_exec(). This saves a malloc/free per call. The default value of
4900 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4901 when configuring.
4902
4903 32. The default maximum size of a compiled pattern is 64K. There have been a
4904 few cases of people hitting this limit. The code now uses macros to handle the
4905 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4906 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4907 configuring. Tests 2 and 5 work only with 2-byte links because they output
4908 debugging information about compiled patterns.
4909
4910 33. Internal code re-arrangements:
4911
4912 (a) Moved the debugging function for printing out a compiled regex into
4913 its own source file (printint.c) and used #include to pull it into
4914 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4915 separate copies.
4916
4917 (b) Defined the list of op-code names for debugging as a macro in
4918 internal.h so that it is next to the definition of the opcodes.
4919
4920 (c) Defined a table of op-code lengths for simpler skipping along compiled
4921 code. This is again a macro in internal.h so that it is next to the
4922 definition of the opcodes.
4923
4924 34. Added support for recursive calls to individual subpatterns, along the
4925 lines of Robin Houston's patch (but implemented somewhat differently).
4926
4927 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4928 allow it to read and process whole directories in Win32. This code was
4929 contributed by Lionel Fourquaux; it has not been tested by me.
4930
4931 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4932 used to name a group. Names consist of alphanumerics and underscores, and must
4933 be unique. Back references use the syntax (?P=name) and recursive calls use
4934 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4935 numbers. The function pcre_fullinfo() can be used after compilation to extract
4936 a name/number map. There are three relevant calls:
4937
4938 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4939 PCRE_INFO_NAMECOUNT yields the number of entries
4940 PCRE_INFO_NAMETABLE yields a pointer to the map.
4941
4942 The map is a vector of fixed-size entries. The size of each entry depends on
4943 the length of the longest name used. The first two bytes of each entry are the
4944 group number, most significant byte first. There follows the corresponding
4945 name, zero terminated. The names are in alphabetical order.
4946
4947 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4948 case instead of 255. Making it the same both with and without UTF-8 support
4949 means that the same test output works with both.
4950
4951 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4952 calling malloc() with a zero argument.
4953
4954 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4955 optimization. I've improved things by keeping a bitmap of backreferences with
4956 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4957 fact referenced, the optimization can be applied. It is unlikely that a
4958 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4959 the match to follow \n) will appear inside brackets with a number greater than
4960 31, but if it does, any back reference > 31 suppresses the optimization.
4961
4962 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4963 of disabling numbered capturing parentheses. Any opening parenthesis that is
4964 not followed by ? behaves as if it were followed by ?: but named parentheses
4965 can still be used for capturing (and they will acquire numbers in the usual
4966 way).
4967
4968 41. Redesigned the return codes from the match() function into yes/no/error so
4969 that errors can be passed back from deep inside the nested calls. A malloc
4970 failure while inside a recursive subpattern call now causes the
4971 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4972
4973 42. It is now possible to set a limit on the number of times the match()
4974 function is called in a call to pcre_exec(). This facility makes it possible to
4975 limit the amount of recursion and backtracking, though not in a directly
4976 obvious way, because the match() function is used in a number of different
4977 circumstances. The count starts from zero for each position in the subject
4978 string (for non-anchored patterns). The default limit is, for compatibility, a
4979 large number, namely 10 000 000. You can change this in two ways:
4980
4981 (a) When configuring PCRE before making, you can use --with-match-limit=n
4982 to set a default value for the compiled library.
4983
4984 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4985 a different value is set. See 45 below.
4986
4987 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4988
4989 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4990 of things that can be changed at compile time. The first argument specifies
4991 what is wanted and the second points to where the information is to be placed.
4992 The current list of available information is:
4993
4994 PCRE_CONFIG_UTF8
4995
4996 The output is an integer that is set to one if UTF-8 support is available;
4997 otherwise it is set to zero.
4998
4999 PCRE_CONFIG_NEWLINE
5000
5001 The output is an integer that it set to the value of the code that is used for
5002 newline. It is either LF (10) or CR (13).
5003
5004 PCRE_CONFIG_LINK_SIZE
5005
5006 The output is an integer that contains the number of bytes used for internal
5007 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5008
5009 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
5010
5011 The output is an integer that contains the threshold above which the POSIX
5012 interface uses malloc() for output vectors. See item 31 above.
5013
5014 PCRE_CONFIG_MATCH_LIMIT
5015
5016 The output is an unsigned integer that contains the default limit of the number
5017 of match() calls in a pcre_exec() execution. See 42 above.
5018
5019 44. pcretest has been upgraded by the addition of the -C option. This causes it
5020 to extract all the available output from the new pcre_config() function, and to
5021 output it. The program then exits immediately.
5022
5023 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5024 order to support additional features. One way would have been to define
5025 pcre_exec2() (for example) with extra arguments, but this would not have been
5026 extensible, and would also have required all calls to the original function to
5027 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5028 is used for passing in "extra" data from pcre_study().
5029
5030 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5031 contains the following fields:
5032
5033 flags a bitmap indicating which of the following fields are set
5034 study_data opaque data from pcre_study()
5035 match_limit a way of specifying a limit on match() calls for a specific
5036 call to pcre_exec()
5037 callout_data data for callouts (see 49 below)
5038
5039 The flag bits are also defined in pcre.h, and are
5040
5041 PCRE_EXTRA_STUDY_DATA
5042 PCRE_EXTRA_MATCH_LIMIT
5043 PCRE_EXTRA_CALLOUT_DATA
5044
5045 The pcre_study() function now returns one of these new pcre_extra blocks, with
5046 the actual study data pointed to by the study_data field, and the
5047 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5048 before. That is, this change is entirely upwards-compatible and requires no
5049 change to existing code.
5050
5051 If you want to pass in additional data to pcre_exec(), you can either place it
5052 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5053 block.
5054
5055 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5056 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5057 times with different match limits, until it finds the minimum value needed for
5058 pcre_exec() to complete. The value is then output. This can be instructive; for
5059 most simple matches the number is quite small, but for pathological cases it
5060 gets very large very quickly.
5061
5062 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5063 returns the size of the data block pointed to by the study_data field in a
5064 pcre_extra block, that is, the value that was passed as the argument to
5065 pcre_malloc() when PCRE was getting memory in which to place the information
5066 created by pcre_study(). The fourth argument should point to a size_t variable.
5067 pcretest has been extended so that this information is shown after a successful
5068 pcre_study() call when information about the compiled regex is being displayed.
5069
5070 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5071 because what follows is always an absolute path. (Later: it turns out that this
5072 is more than cosmetic for MinGW, because it doesn't like empty path
5073 components.)
5074
5075 49. Some changes have been made to the callout feature (see 28 above):
5076
5077 (i) A callout function now has three choices for what it returns:
5078
5079 0 => success, carry on matching
5080 > 0 => failure at this point, but backtrack if possible
5081 < 0 => serious error, return this value from pcre_exec()
5082
5083 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5084 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5085 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5086 use by callout functions. It will never be used by PCRE itself.
5087
5088 (ii) The pcre_extra structure (see 45 above) has a void * field called
5089 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5090 pcre_callout_block structure has a field of the same name. The contents of
5091 the field passed in the pcre_extra structure are passed to the callout
5092 function in the corresponding field in the callout block. This makes it
5093 easier to use the same callout-containing regex from multiple threads. For
5094 testing, the pcretest program has a new data escape
5095
5096 \C*n pass the number n (may be negative) as callout_data
5097
5098 If the callout function in pcretest receives a non-zero value as
5099 callout_data, it returns that value.
5100
5101 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5102 there were some redundant $(CFLAGS) in commands that are now specified as
5103 $(LINK), which already includes $(CFLAGS).
5104
5105 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5106 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5107 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5108 one-byte characters throughout. Note that case-insensitive matching applies
5109 only to characters whose values are less than 256. PCRE doesn't support the
5110 notion of cases for higher-valued characters.
5111
5112 (i) A character class whose characters are all within 0-255 is handled as
5113 a bit map, and the map is inverted for negative classes. Previously, a
5114 character > 255 always failed to match such a class; however it should
5115 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5116
5117 (ii) A negated character class with a single character < 255 is coded as
5118 "not this character" (OP_NOT). This wasn't working properly when the test
5119 character was multibyte, either singly or repeated.
5120
5121 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5122 mode, for example: \x{100}{2,3}.
5123
5124 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5125 singly or repeated) now correctly test multibyte characters. However,
5126 PCRE doesn't recognize any characters with values greater than 255 as
5127 digits, spaces, or word characters. Such characters always match \D, \S,
5128 and \W, and never match \d, \s, or \w.
5129
5130 (v) Classes may now contain characters and character ranges with values
5131 greater than 255. For example: [ab\x{100}-\x{400}].
5132
5133 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5134 PCRE in UTF-8 mode.
5135
5136 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5137 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5138 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5139 value.)
5140
5141 53. The single man page has become too large. I have therefore split it up into
5142 a number of separate man pages. These also give rise to individual HTML pages;
5143 these are now put in a separate directory, and there is an index.html page that
5144 lists them all. Some hyperlinking between the pages has been installed.
5145
5146 54. Added convenience functions for handling named capturing parentheses.
5147
5148 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5149 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5150 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5151 are faulted.
5152
5153 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5154 calling configure. These values are used when compiling the dftables.c program
5155 which is run to generate the source of the default character tables. They
5156 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5157 you will need to set these values.
5158
5159 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5160
5161
5162 Version 3.9 02-Jan-02
5163 ---------------------
5164
5165 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5166
5167 2. If --disable-static was given, the building process failed when trying to
5168 build pcretest and pcregrep. (For some reason it was using libtool to compile
5169 them, which is not right, as they aren't part of the library.)
5170
5171
5172 Version 3.8 18-Dec-01
5173 ---------------------
5174
5175 1. The experimental UTF-8 code was completely screwed up. It was packing the
5176 bytes in the wrong order. How dumb can you get?
5177
5178
5179 Version 3.7 29-Oct-01
5180 ---------------------
5181
5182 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5183 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5184 this didn't happen under Solaris 8, where I normally test things.
5185
5186 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5187 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5188 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5189 this fix an hour or so after the initial 3.7 release.)
5190
5191
5192 Version 3.6 23-Oct-01
5193 ---------------------
5194
5195 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5196 offsets passed as NULL with zero offset count.
5197
5198 2. The config.guess and config.sub files had not been updated when I moved to
5199 the latest autoconf.
5200
5201
5202 Version 3.5 15-Aug-01
5203 ---------------------
5204
5205 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5206 had been forgotten.
5207
5208 2. By using declared but undefined structures, we can avoid using "void"
5209 definitions in pcre.h while keeping the internal definitions of the structures
5210 private.
5211
5212 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5213 user point of view, this means that both static and shared libraries are built
5214 by default, but this can be individually controlled. More of the work of
5215 handling this static/shared cases is now inside libtool instead of PCRE's make
5216 file.
5217
5218 4. The pcretest utility is now installed along with pcregrep because it is
5219 useful for users (to test regexs) and by doing this, it automatically gets
5220 relinked by libtool. The documentation has been turned into a man page, so
5221 there are now .1, .txt, and .html versions in /doc.
5222
5223 5. Upgrades to pcregrep:
5224 (i) Added long-form option names like gnu grep.
5225 (ii) Added --help to list all options with an explanatory phrase.
5226 (iii) Added -r, --recursive to recurse into sub-directories.
5227 (iv) Added -f, --file to read patterns from a file.
5228
5229 6. pcre_exec() was referring to its "code" argument before testing that
5230 argument for NULL (and giving an error if it was NULL).
5231
5232 7. Upgraded Makefile.in to allow for compiling in a different directory from
5233 the source directory.
5234
5235 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5236 options bits, the pointer it was passed was to an int instead of to an unsigned
5237 long int. This mattered only on 64-bit systems.
5238
5239 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5240 generated) instead of pcre.in, which it its source. Also made the same change
5241 in several of the .c files.
5242
5243 10. A new release of gcc defines printf() as a macro, which broke pcretest
5244 because it had an ifdef in the middle of a string argument for printf(). Fixed
5245 by using separate calls to printf().
5246
5247 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5248 script, to force use of CR or LF instead of \n in the source. On non-Unix
5249 systems, the value can be set in config.h.
5250
5251 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5252 absolute limit. Changed the text of the error message to make this clear, and
5253 likewise updated the man page.
5254
5255 13. The limit of 99 on the number of capturing subpatterns has been removed.
5256 The new limit is 65535, which I hope will not be a "real" limit.
5257
5258
5259 Version 3.4 22-Aug-00
5260 ---------------------
5261
5262 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5263
5264 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5265
5266
5267 Version 3.3 01-Aug-00
5268 ---------------------
5269
5270 1. If an octal character was given, but the value was greater than \377, it
5271 was not getting masked to the least significant bits, as documented. This could
5272 lead to crashes in some systems.
5273
5274 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5275 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5276
5277 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5278 These just pass their arguments on to (pcre_free)(), but they are provided
5279 because some uses of PCRE bind it to non-C systems that can call its functions,
5280 but cannot call free() or pcre_free() directly.
5281
5282 4. Add "make test" as a synonym for "make check". Corrected some comments in
5283 the Makefile.
5284
5285 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5286 Makefile.
5287
5288 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5289 command called pgrep for grepping around the active processes.
5290
5291 7. Added the beginnings of support for UTF-8 character strings.
5292
5293 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5294 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5295 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5296 out for the ar command.)
5297
5298
5299 Version 3.2 12-May-00
5300 ---------------------
5301
5302 This is purely a bug fixing release.
5303
5304 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5305 of ZA. This was just one example of several cases that could provoke this bug,
5306 which was introduced by change 9 of version 2.00. The code for breaking
5307 infinite loops after an iteration that matches an empty string was't working
5308 correctly.
5309
5310 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5311 when matched against abbab (for example). After matching an empty string, it
5312 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5313 caused it to match further down the string than it should.
5314
5315 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5316 was there because it doesn't seem to be needed, and it causes trouble on some
5317 systems, as it is not a Standard C header. It has been removed.
5318
5319 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5320 were reported on the Macintosh. The changes were from
5321
5322 while ((c = *(++ptr)) != 0 && c != '\n');
5323 to
5324 while ((c = *(++ptr)) != 0 && c != '\n') ;
5325
5326 Totally extraordinary, but if that's what it takes...
5327
5328 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5329 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5330 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5331 assumes the way PCRE uses memmove() (always moving upwards).
5332
5333 6. PCRE is being used in one environment where strchr() is not available. There
5334 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5335 faster code anyway.
5336
5337
5338 Version 3.1 09-Feb-00
5339 ---------------------
5340
5341 The only change in this release is the fixing of some bugs in Makefile.in for
5342 the "install" target:
5343
5344 (1) It was failing to install pcreposix.h.
5345
5346 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5347
5348
5349 Version 3.0 01-Feb-00
5350 ---------------------
5351
5352 1. Add support for the /+ modifier to perltest (to output $` like it does in
5353 pcretest).
5354
5355 2. Add support for the /g modifier to perltest.
5356
5357 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5358 matches null strings.
5359
5360 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5361 pattern. Perl treats empty patterns specially - it reuses the most recent
5362 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5363 effect.
5364
5365 5. The POSIX interface was broken in that it was just handing over the POSIX
5366 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5367 required a bigger vector, with some working space on the end. This means that
5368 the POSIX wrapper now has to get and free some memory, and copy the results.
5369
5370 6. Added some simple autoconf support, placing the test data and the
5371 documentation in separate directories, re-organizing some of the
5372 information files, and making it build pcre-config (a GNU standard). Also added
5373 libtool support for building PCRE as a shared library, which is now the
5374 default.
5375
5376 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5377 09 are not valid octal constants. Single digits will be used for minor values
5378 less than 10.
5379
5380 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5381 existing programs that set these in the POSIX interface can use PCRE without
5382 modification.
5383
5384 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5385 return all that pcre_info() returns, plus additional data. The pcre_info()
5386 function is retained for compatibility, but is considered to be obsolete.
5387
5388 10. Added experimental recursion feature (?R) to handle one common case that
5389 Perl 5.6 will be able to do with (?p{...}).
5390
5391 11. Added support for POSIX character classes like [:alpha:], which Perl is
5392 adopting.
5393
5394
5395 Version 2.08 31-Aug-99
5396 ----------------------
5397
5398 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5399 trying to match at the startoffset position, but instead was moving forward to
5400 the next newline as if a previous match had failed.
5401
5402 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5403 and could get into a loop if a null string was matched other than at the start
5404 of the subject.
5405
5406 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5407 be distinguished at compile time, and for completeness also added PCRE_DATE.
5408
5409 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5410 in GnuWin32 environments.
5411
5412
5413 Version 2.07 29-Jul-99
5414 ----------------------
5415
5416 1. The documentation is now supplied in plain text form and HTML as well as in
5417 the form of man page sources.
5418
5419 2. C++ compilers don't like assigning (void *) values to other pointer types.
5420 In particular this affects malloc(). Although there is no problem in Standard
5421 C, I've put in casts to keep C++ compilers happy.
5422
5423 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5424 should be (const char *).
5425
5426 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5427 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5428 However, I haven't made this a standard facility. The documentation doesn't
5429 mention it, and the Makefile doesn't support it.
5430
5431 5. The Makefile now contains an "install" target, with editable destinations at
5432 the top of the file. The pcretest program is not installed.
5433
5434 6. pgrep -V now gives the PCRE version number and date.
5435
5436 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5437 causing the entire string to be ignored, instead of just the last character.
5438
5439 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5440 non-matching string, it can take a very, very long time, even for strings of
5441 quite modest length, because of the nested recursion. PCRE now does better in
5442 some of these cases. It does this by remembering the last required literal
5443 character in the pattern, and pre-searching the subject to ensure it is present
5444 before running the real match. In other words, it applies a heuristic to detect
5445 some types of certain failure quickly, and in the above example, if presented
5446 with a string that has no trailing " it gives "no match" very quickly.
5447
5448 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5449 other alternatives are tried instead.
5450
5451
5452 Version 2.06 09-Jun-99
5453 ----------------------
5454
5455 1. Change pcretest's output for amount of store used to show just the code
5456 space, because the remainder (the data block) varies in size between 32-bit and
5457 64-bit systems.
5458
5459 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5460 start matching at. This allows lookbehinds to work when searching for multiple
5461 occurrences in a string.
5462
5463 3. Added additional options to pcretest for testing multiple occurrences:
5464
5465 /+ outputs the rest of the string that follows a match
5466 /g loops for multiple occurrences, using the new startoffset argument
5467 /G loops for multiple occurrences by passing an incremented pointer
5468
5469 4. PCRE wasn't doing the "first character" optimization for patterns starting
5470 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5471 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5472 the letter 'x'. On long subject strings, this gives a significant speed-up.
5473
5474
5475 Version 2.05 21-Apr-99
5476 ----------------------
5477
5478 1. Changed the type of magic_number from int to long int so that it works
5479 properly on 16-bit systems.
5480
5481 2. Fixed a bug which caused patterns starting with .* not to work correctly
5482 when the subject string contained newline characters. PCRE was assuming
5483 anchoring for such patterns in all cases, which is not correct because .* will
5484 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5485 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5486 must be retried after every newline in the subject.
5487
5488
5489 Version 2.04 18-Feb-99
5490 ----------------------
5491
5492 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5493 computation of the store needed to hold the pattern was incorrect (too large).
5494 If such patterns were nested a few deep, this could multiply and become a real
5495 problem.
5496
5497 2. Added /M option to pcretest to show the memory requirement of a specific
5498 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5499
5500 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5501 compiled in such a way that the backtracking after subsequent failure was
5502 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5503 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5504
5505
5506 Version 2.03 02-Feb-99
5507 ----------------------
5508
5509 1. Fixed typo and small mistake in man page.
5510
5511 2. Added 4th condition (GPL supersedes if conflict) and created separate
5512 LICENCE file containing the conditions.
5513
5514 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5515 Perl, that is the internal \ allows the delimiter to be included in the
5516 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5517 the final delimiter, add \ to the end of the pattern (to test the error).
5518
5519 4. Added the convenience functions for extracting substrings after a successful
5520 match. Updated pcretest to make it able to test these functions.
5521
5522
5523 Version 2.02 14-Jan-99
5524 ----------------------
5525
5526 1. Initialized the working variables associated with each extraction so that
5527 their saving and restoring doesn't refer to uninitialized store.
5528
5529 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5530 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5531 fix the problem.
5532
5533 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5534 calls, and wasn't printing the correct value for compiling calls. Increased the
5535 default value of LOOPREPEAT, and the number of significant figures in the
5536 times.
5537
5538 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5539
5540 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5541 a building problem on Windows NT with a FAT file system.
5542
5543
5544 Version 2.01 21-Oct-98
5545 ----------------------
5546
5547 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5548 to character tables built by pcre_maketables() in the current locale. If NULL
5549 is passed, the default tables are used.
5550
5551
5552 Version 2.00 24-Sep-98
5553 ----------------------
5554
5555 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5556 it any more.
5557
5558 2. Allow quantification of (?>) groups, and make it work correctly.
5559
5560 3. The first character computation wasn't working for (?>) groups.
5561
5562 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5563 end of the subject) and add 5.005's \z, which really does match only at the
5564 very end of the subject.
5565
5566 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5567
5568 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5569 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5570 localized options. All options to pcre_study() were also removed.
5571
5572 7. Add other new features from 5.005:
5573
5574 $(?<= positive lookbehind
5575 $(?<! negative lookbehind
5576 (?imsx-imsx) added the unsetting capability
5577 such a setting is global if at outer level; local otherwise
5578 (?imsx-imsx:) non-capturing groups with option setting
5579 (?(cond)re|re) conditional pattern matching
5580
5581 A backreference to itself in a repeated group matches the previous
5582 captured string.
5583
5584 8. General tidying up of studying (both automatic and via "study")
5585 consequential on the addition of new assertions.
5586
5587 9. As in 5.005, unlimited repeated groups that could match an empty substring
5588 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5589 runtime if any iteration does actually match an empty substring.
5590
5591 10. Include the RunTest script in the distribution.
5592
5593 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5594 discrepancies, some of which were old and were also with respect to 5.004. They
5595 have now been fixed.
5596
5597
5598 Version 1.09 28-Apr-98
5599 ----------------------
5600
5601 1. A negated single character class followed by a quantifier with a minimum
5602 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5603 program crashes, or just wrong answers. This did not apply to negated classes
5604 containing more than one character, or to minima other than one.
5605
5606
5607 Version 1.08 27-Mar-98
5608 ----------------------
5609
5610 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5611
5612 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5613 latter must appear before anything that relies on it in the pattern.
5614
5615
5616 Version 1.07 16-Feb-98
5617 ----------------------
5618
5619 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5620 repeat of a potentially empty string).
5621
5622
5623 Version 1.06 23-Jan-98
5624 ----------------------
5625
5626 1. Added Markus Oberhumer's little patches for C++.
5627
5628 2. Literal strings longer than 255 characters were broken.
5629
5630
5631 Version 1.05 23-Dec-97
5632 ----------------------
5633
5634 1. Negated character classes containing more than one character were failing if
5635 PCRE_CASELESS was set at run time.
5636
5637
5638 Version 1.04 19-Dec-97
5639 ----------------------
5640
5641 1. Corrected the man page, where some "const" qualifiers had been omitted.
5642
5643 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5644 input syntax.
5645
5646 3. Fixed memory leak which occurred when a regex with back references was
5647 matched with an offsets vector that wasn't big enough. The temporary memory
5648 that is used in this case wasn't being freed if the match failed.
5649
5650 4. Tidied pcretest to ensure it frees memory that it gets.
5651
5652 5. Temporary memory was being obtained in the case where the passed offsets
5653 vector was exactly big enough.
5654
5655 6. Corrected definition of offsetof() from change 5 below.
5656
5657 7. I had screwed up change 6 below and broken the rules for the use of
5658 setjmp(). Now fixed.
5659
5660
5661 Version 1.03 18-Dec-97
5662 ----------------------
5663
5664 1. A erroneous regex with a missing opening parenthesis was correctly
5665 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5666 on some systems.
5667
5668 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5669 it was reported that one broken compiler failed on the former because "code" is
5670 also an independent variable.
5671
5672 3. The erroneous regex a[]b caused an array overrun reference.
5673
5674 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5675 fail on data ending with that character. (It was going on too far, and checking
5676 the next character, typically a binary zero.) This was specific to the
5677 optimized code for single-character negative classes.
5678
5679 5. Added a contributed patch from the TIN world which does the following:
5680
5681 + Add an undef for memmove, in case the the system defines a macro for it.
5682
5683 + Add a definition of offsetof(), in case there isn't one. (I don't know
5684 the reason behind this - offsetof() is part of the ANSI standard - but
5685 it does no harm).
5686
5687 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5688 most of the places where whitespace preceded '#'. I have given up and
5689 allowed the remaining 2 cases to be at the margin.
5690
5691 + Rename some variables in pcre to eliminate shadowing. This seems very
5692 pedantic, but does no harm, of course.
5693
5694 6. Moved the call to setjmp() into its own function, to get rid of warnings
5695 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5696
5697 7. Constructs such as \d{8,} were compiling into the equivalent of
5698 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5699 outcome, but in this particular case used more store than had been allocated,
5700 which caused the bug to be discovered because it threw up an internal error.
5701
5702 8. The debugging code in both pcre and pcretest for outputting the compiled
5703 form of a regex was going wrong in the case of back references followed by
5704 curly-bracketed repeats.
5705
5706
5707 Version 1.02 12-Dec-97
5708 ----------------------
5709
5710 1. Typos in pcre.3 and comments in the source fixed.
5711
5712 2. Applied a contributed patch to get rid of places where it used to remove
5713 'const' from variables, and fixed some signed/unsigned and uninitialized
5714 variable warnings.
5715
5716 3. Added the "runtest" target to Makefile.
5717
5718 4. Set default compiler flag to -O2 rather than just -O.
5719
5720
5721 Version 1.01 19-Nov-97
5722 ----------------------
5723
5724 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5725 like /([ab]*)*/, that is, for classes with more than one character in them.
5726
5727 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5728 as /((?>a*))*/ (a PCRE_EXTRA facility).
5729
5730
5731 Version 1.00 18-Nov-97
5732 ----------------------
5733
5734 1. Added compile-time macros to support systems such as SunOS4 which don't have
5735 memmove() or strerror() but have other things that can be used instead.
5736
5737 2. Arranged that "make clean" removes the executables.
5738
5739
5740 Version 0.99 27-Oct-97
5741 ----------------------
5742
5743 1. Fixed bug in code for optimizing classes with only one character. It was
5744 initializing a 32-byte map regardless, which could cause it to run off the end
5745 of the memory it had got.
5746
5747 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5748
5749
5750 Version 0.98 22-Oct-97
5751 ----------------------
5752
5753 1. Fixed bug in code for handling temporary memory usage when there are more
5754 back references than supplied space in the ovector. This could cause segfaults.
5755
5756
5757 Version 0.97 21-Oct-97
5758 ----------------------
5759
5760 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5761
5762 2. Optimized negated single characters not to use a bit map.
5763
5764 3. Brought error texts together as macro definitions; clarified some of them;
5765 fixed one that was wrong - it said "range out of order" when it meant "invalid
5766 escape sequence".
5767
5768 4. Changed some char * arguments to const char *.
5769
5770 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5771
5772 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5773 pcretest.
5774
5775
5776 Version 0.96 16-Oct-97
5777 ----------------------
5778
5779 1. Added a simple "pgrep" utility to the distribution.
5780
5781 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5782 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5783 where "ddd" means "one or more decimal digits".
5784
5785 3. Fixed serious bug. If a pattern had a back reference, but the call to
5786 pcre_exec() didn't supply a large enough ovector to record the related
5787 identifying subpattern, the match always failed. PCRE now remembers the number
5788 of the largest back reference, and gets some temporary memory in which to save
5789 the offsets during matching if necessary, in order to ensure that
5790 backreferences always work.
5791
5792 4. Increased the compatibility with Perl in a number of ways:
5793
5794 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5795 to request this handling. The option can be set at compile or exec time.
5796
5797 (b) $ matches before a terminating newline by default; an option
5798 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5799 mode). The option can be set at compile or exec time.
5800
5801 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5802 the same as Perl's. If the decimal number it represents is less than 10
5803 or there aren't that many previous left capturing parentheses, an octal
5804 escape is read. Inside a character class, it's always an octal escape,
5805 even if it is a single digit.
5806
5807 (d) An escaped but undefined alphabetic character is taken as a literal,
5808 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5809 escapes.
5810
5811 (e) {0} is now permitted. (The previous item is removed from the compiled
5812 pattern).
5813
5814 5. Changed all the names of code files so that the basic parts are no longer
5815 than 10 characters, and abolished the teeny "globals.c" file.
5816
5817 6. Changed the handling of character classes; they are now done with a 32-byte
5818 bit map always.
5819
5820 7. Added the -d and /D options to pcretest to make it possible to look at the
5821 internals of compilation without having to recompile pcre.
5822
5823
5824 Version 0.95 23-Sep-97
5825 ----------------------
5826
5827 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5828 \x20 at the start of a run of normal characters. These were being treated as
5829 real characters, instead of the source characters being re-checked.
5830
5831
5832 Version 0.94 18-Sep-97
5833 ----------------------
5834
5835 1. The functions are now thread-safe, with the caveat that the global variables
5836 containing pointers to malloc() and free() or alternative functions are the
5837 same for all threads.
5838
5839 2. Get pcre_study() to generate a bitmap of initial characters for non-
5840 anchored patterns when this is possible, and use it if passed to pcre_exec().
5841
5842
5843 Version 0.93 15-Sep-97
5844 ----------------------
5845
5846 1. /(b)|(:+)/ was computing an incorrect first character.
5847
5848 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5849 but not actually doing anything yet.
5850
5851 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5852 as Perl does (e.g. [-az] or [az-]).
5853
5854 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5855 all possible positions.
5856
5857 5. Split up into different modules to avoid including unneeded functions in a
5858 compiled binary. However, compile and exec are still in one module. The "study"
5859 function is split off.
5860
5861 6. The character tables are now in a separate module whose source is generated
5862 by an auxiliary program - but can then be edited by hand if required. There are
5863 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5864 toupper() in the code.
5865
5866 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5867 make them global. Abolish the function for setting them, as the caller can now
5868 set them directly.
5869
5870
5871 Version 0.92 11-Sep-97
5872 ----------------------
5873
5874 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5875 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5876
5877 2. Caseless matching was not working in character classes if the characters in
5878 the pattern were in upper case.
5879
5880 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5881
5882 4. Make PCRE_ANCHORED public and accept as a compile option.
5883
5884 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5885 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5886 pass them.
5887
5888 6. Give an error if bad option bits passed at compile or run time.
5889
5890 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5891 pcretest to cause it to pass that flag.
5892
5893 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5894 options, and the first character, if set.
5895
5896 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5897
5898
5899 Version 0.91 10-Sep-97
5900 ----------------------
5901
5902 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5903 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5904
5905 2. PCRE was looping on encountering an indefinitely repeated back reference to
5906 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5907 Perl does - treats the match as successful.
5908
5909 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5