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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1599 - (show annotations)
Wed Sep 2 08:50:31 2015 UTC (4 years, 1 month ago) by zherczeg
File size: 284438 byte(s)
Error occurred while calculating annotation data.
A match limit issue is fixed in JIT.
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 xx-xxx-xxxx
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
157 Version 8.37 28-April-2015
158 --------------------------
159
160 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
161 for those parentheses to be closed with whatever has been captured so far.
162 However, it was failing to mark any other groups between the hightest
163 capture so far and the currrent group as "unset". Thus, the ovector for
164 those groups contained whatever was previously there. An example is the
165 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
166
167 2. If an assertion condition was quantified with a minimum of zero (an odd
168 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
169
170 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
171 unrecognized modifier, a crash could occur.
172
173 4. An attempt to do global matching in pcretest with a zero-length ovector
174 caused a crash.
175
176 5. Fixed a memory leak during matching that could occur for a subpattern
177 subroutine call (recursive or otherwise) if the number of captured groups
178 that had to be saved was greater than ten.
179
180 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
181 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
182 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
183
184 7. A UTF pattern containing a "not" match of a non-ASCII character and a
185 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
186
187 8. When a pattern is compiled, it remembers the highest back reference so that
188 when matching, if the ovector is too small, extra memory can be obtained to
189 use instead. A conditional subpattern whose condition is a check on a
190 capture having happened, such as, for example in the pattern
191 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
192 setting the highest backreference number. This mattered only if pcre_exec()
193 was called with an ovector that was too small to hold the capture, and there
194 was no other kind of back reference (a situation which is probably quite
195 rare). The effect of the bug was that the condition was always treated as
196 FALSE when the capture could not be consulted, leading to a incorrect
197 behaviour by pcre_exec(). This bug has been fixed.
198
199 9. A reference to a duplicated named group (either a back reference or a test
200 for being set in a conditional) that occurred in a part of the pattern where
201 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
202 to be incorrectly calculated, leading to overwriting.
203
204 10. A mutually recursive set of back references such as (\2)(\1) caused a
205 segfault at study time (while trying to find the minimum matching length).
206 The infinite loop is now broken (with the minimum length unset, that is,
207 zero).
208
209 11. If an assertion that was used as a condition was quantified with a minimum
210 of zero, matching went wrong. In particular, if the whole group had
211 unlimited repetition and could match an empty string, a segfault was
212 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
213 assertions to be quantified, but not if they are being used as conditions,
214 so the above pattern is faulted by Perl. PCRE has now been changed so that
215 it also rejects such patterns.
216
217 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
218 failed to allow the zero-repeat case if pcre2_exec() was called with an
219 ovector too small to capture the group.
220
221 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
222 Red Hat Product Security:
223
224 (a) A crash if /K and /F were both set with the option to save the compiled
225 pattern.
226
227 (b) Another crash if the option to print captured substrings in a callout
228 was combined with setting a null ovector, for example \O\C+ as a subject
229 string.
230
231 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
232 forward reference repeated a large (but limited) number of times within a
233 repeated outer group that has a zero minimum quantifier, caused incorrect
234 code to be compiled, leading to the error "internal error:
235 previously-checked referenced subpattern not found" when an incorrect
236 memory address was read. This bug was reported as "heap overflow",
237 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
238 CVE-2015-2325.
239
240 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
241 call within a group that also contained a recursive back reference caused
242 incorrect code to be compiled. This bug was reported as "heap overflow",
243 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
244 number CVE-2015-2326.
245
246 24. Computing the size of the JIT read-only data in advance has been a source
247 of various issues, and new ones are still appear unfortunately. To fix
248 existing and future issues, size computation is eliminated from the code,
249 and replaced by on-demand memory allocation.
250
251 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
252 adjacent to the end of the range, and the range contained characters with
253 more than one other case, caused incorrect behaviour when compiled in UTF
254 mode. In that example, the range a-j was left out of the class.
255
256 26. Fix JIT compilation of conditional blocks, which assertion
257 is converted to (*FAIL). E.g: /(?(?!))/.
258
259 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
260 discovered by the LLVM fuzzer.
261
262 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
263 when this assertion was used as a condition, for example (?(?!)a|b). In
264 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
265 error about an unsupported item.
266
267 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
268 possessification code could take exponential time to complete. A recursion
269 depth limit of 1000 has been imposed to limit the resources used by this
270 optimization.
271
272 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
273 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
274 because \S ensures they are all in the class. The code for doing this was
275 interacting badly with the code for computing the amount of space needed to
276 compile the pattern, leading to a buffer overflow. This bug was discovered
277 by the LLVM fuzzer.
278
279 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
280 other kinds of group caused stack overflow at compile time. This bug was
281 discovered by the LLVM fuzzer.
282
283 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
284 between a subroutine call and its quantifier was incorrectly compiled,
285 leading to buffer overflow or other errors. This bug was discovered by the
286 LLVM fuzzer.
287
288 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
289 assertion after (?(. The code was failing to check the character after
290 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
291 was discovered by the LLVM fuzzer.
292
293 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
294 a fixed maximum following a group that contains a subroutine reference was
295 incorrectly compiled and could trigger buffer overflow. This bug was
296 discovered by the LLVM fuzzer.
297
298 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
299 caused a stack overflow instead of the diagnosis of a non-fixed length
300 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
301
302 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
303 (e.g. /(?<=\Ka)/) could make pcregrep loop.
304
305 37. There was a similar problem to 36 in pcretest for global matches.
306
307 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
308 and a subsequent item in the pattern caused a non-match, backtracking over
309 the repeated \X did not stop, but carried on past the start of the subject,
310 causing reference to random memory and/or a segfault. There were also some
311 other cases where backtracking after \C could crash. This set of bugs was
312 discovered by the LLVM fuzzer.
313
314 39. The function for finding the minimum length of a matching string could take
315 a very long time if mutual recursion was present many times in a pattern,
316 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
317 method has been implemented. This infelicity was discovered by the LLVM
318 fuzzer.
319
320 40. Static linking against the PCRE library using the pkg-config module was
321 failing on missing pthread symbols.
322
323
324 Version 8.36 26-September-2014
325 ------------------------------
326
327 1. Got rid of some compiler warnings in the C++ modules that were shown up by
328 -Wmissing-field-initializers and -Wunused-parameter.
329
330 2. The tests for quantifiers being too big (greater than 65535) were being
331 applied after reading the number, and stupidly assuming that integer
332 overflow would give a negative number. The tests are now applied as the
333 numbers are read.
334
335 3. Tidy code in pcre_exec.c where two branches that used to be different are
336 now the same.
337
338 4. The JIT compiler did not generate match limit checks for certain
339 bracketed expressions with quantifiers. This may lead to exponential
340 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
341 issue should be resolved now.
342
343 5. Fixed an issue, which occures when nested alternatives are optimized
344 with table jumps.
345
346 6. Inserted two casts and changed some ints to size_t in the light of some
347 reported 64-bit compiler warnings (Bugzilla 1477).
348
349 7. Fixed a bug concerned with zero-minimum possessive groups that could match
350 an empty string, which sometimes were behaving incorrectly in the
351 interpreter (though correctly in the JIT matcher). This pcretest input is
352 an example:
353
354 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
355 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
356
357 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
358 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
359 for an empty string was breaking the inner loop and carrying on at a lower
360 level, when possessive repeated groups should always return to a higher
361 level as they have no backtrack points in them. The empty string test now
362 occurs at the outer level.
363
364 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
365 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
366
367 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
368 doesn't).
369
370 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
371 to hold a pattern to be incorrectly computed (too small) when there were
372 named back references to duplicated names. This could cause "internal
373 error: code overflow" or "double free or corruption" or other memory
374 handling errors.
375
376 11. When named subpatterns had the same prefixes, back references could be
377 confused. For example, in this pattern:
378
379 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
380
381 the reference to 'Name' was incorrectly treated as a reference to a
382 duplicate name.
383
384 12. A pattern such as /^s?c/mi8 where the optional character has more than
385 one "other case" was incorrectly compiled such that it would only try to
386 match starting at "c".
387
388 13. When a pattern starting with \s was studied, VT was not included in the
389 list of possible starting characters; this should have been part of the
390 8.34/18 patch.
391
392 14. If a character class started [\Qx]... where x is any character, the class
393 was incorrectly terminated at the ].
394
395 15. If a pattern that started with a caseless match for a character with more
396 than one "other case" was studied, PCRE did not set up the starting code
397 unit bit map for the list of possible characters. Now it does. This is an
398 optimization improvement, not a bug fix.
399
400 16. The Unicode data tables have been updated to Unicode 7.0.0.
401
402 17. Fixed a number of memory leaks in pcregrep.
403
404 18. Avoid a compiler warning (from some compilers) for a function call with
405 a cast that removes "const" from an lvalue by using an intermediate
406 variable (to which the compiler does not object).
407
408 19. Incorrect code was compiled if a group that contained an internal recursive
409 back reference was optional (had quantifier with a minimum of zero). This
410 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
411 caused segmentation faults because of stack overflows at compile time.
412
413 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
414 group that is quantified with an indefinite repeat, caused a compile-time
415 loop which used up all the system stack and provoked a segmentation fault.
416 This was not the same bug as 19 above.
417
418 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
419 Patch by Mike Frysinger.
420
421
422 Version 8.35 04-April-2014
423 --------------------------
424
425 1. A new flag is set, when property checks are present in an XCLASS.
426 When this flag is not set, PCRE can perform certain optimizations
427 such as studying these XCLASS-es.
428
429 2. The auto-possessification of character sets were improved: a normal
430 and an extended character set can be compared now. Furthermore
431 the JIT compiler optimizes more character set checks.
432
433 3. Got rid of some compiler warnings for potentially uninitialized variables
434 that show up only when compiled with -O2.
435
436 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
437 of a match later then the end of the match. The pcretest program was not
438 handling the case sensibly - it was outputting from the start to the next
439 binary zero. It now reports this situation in a message, and outputs the
440 text from the end to the start.
441
442 5. Fast forward search is improved in JIT. Instead of the first three
443 characters, any three characters with fixed position can be searched.
444 Search order: first, last, middle.
445
446 6. Improve character range checks in JIT. Characters are read by an inprecise
447 function now, which returns with an unknown value if the character code is
448 above a certain threshold (e.g: 256). The only limitation is that the value
449 must be bigger than the threshold as well. This function is useful when
450 the characters above the threshold are handled in the same way.
451
452 7. The macros whose names start with RAWUCHAR are placeholders for a future
453 mode in which only the bottom 21 bits of 32-bit data items are used. To
454 make this more memorable for those maintaining the code, the names have
455 been changed to start with UCHAR21, and an extensive comment has been added
456 to their definition.
457
458 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
459 to the export list in Makefile.am (they were accidentally omitted from the
460 8.34 tarball).
461
462 9. The informational output from pcretest used the phrase "starting byte set"
463 which is inappropriate for the 16-bit and 32-bit libraries. As the output
464 for "first char" and "need char" really means "non-UTF-char", I've changed
465 "byte" to "char", and slightly reworded the output. The documentation about
466 these values has also been (I hope) clarified.
467
468 10. Another JIT related optimization: use table jumps for selecting the correct
469 backtracking path, when more than four alternatives are present inside a
470 bracket.
471
472 11. Empty match is not possible, when the minimum length is greater than zero,
473 and there is no \K in the pattern. JIT should avoid empty match checks in
474 such cases.
475
476 12. In a caseless character class with UCP support, when a character with more
477 than one alternative case was not the first character of a range, not all
478 the alternative cases were added to the class. For example, s and \x{17f}
479 are both alternative cases for S: the class [RST] was handled correctly,
480 but [R-T] was not.
481
482 13. The configure.ac file always checked for pthread support when JIT was
483 enabled. This is not used in Windows, so I have put this test inside a
484 check for the presence of windows.h (which was already tested for).
485
486 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
487 The algorithm provides a way to skip certain starting offsets, and usually
488 faster than linear prefix searches.
489
490 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
491 as for 'fr_FR' and 'french'. For some reason, however, it then used the
492 Windows-specific input and output files, which have 'french' screwed in.
493 So this could never have worked. One of the problems with locales is that
494 they aren't always the same. I have now updated RunTest so that it checks
495 the output of the locale test (test 3) against three different output
496 files, and it allows the test to pass if any one of them matches. With luck
497 this should make the test pass on some versions of Solaris where it was
498 failing. Because of the uncertainty, the script did not used to stop if
499 test 3 failed; it now does. If further versions of a French locale ever
500 come to light, they can now easily be added.
501
502 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
503 there was a message during ./configure, but it did not stop. This now
504 provokes an error. The invalid example in README has been corrected.
505 If a value less than the minimum is given, the minimum value has always
506 been used, but now a warning is given.
507
508 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
509 was a bug in the test system, which is now fixed. Also, the list of various
510 configurations that are tested for each release did not have one with both
511 16/32 bits and --enable-bar-anycrlf. It now does.
512
513 18. pcretest was missing "-C bsr" for displaying the \R default setting.
514
515 19. Little endian PowerPC systems are supported now by the JIT compiler.
516
517 20. The fast forward newline mechanism could enter to an infinite loop on
518 certain invalid UTF-8 input. Although we don't support these cases
519 this issue can be fixed by a performance optimization.
520
521 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
522 not take account if existing stack usage. There is now a new global
523 variable called pcre_stack_guard that can be set to point to an external
524 function to check stack availability. It is called at the start of
525 processing every parenthesized group.
526
527 22. A typo in the code meant that in ungreedy mode the max/min qualifier
528 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
529 not match "ab".
530
531 23. When UTF was disabled, the JIT program reported some incorrect compile
532 errors. These messages are silenced now.
533
534 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
535 compiler.
536
537 25. Change all the temporary files used in RunGrepTest to be different to those
538 used by RunTest so that the tests can be run simultaneously, for example by
539 "make -j check".
540
541
542 Version 8.34 15-December-2013
543 -----------------------------
544
545 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
546 executable memory. Patch inspired by Carsten Klein.
547
548 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
549 this macro is never tested and has no effect, because the work to support
550 coverage involves only compiling and linking options and special targets in
551 the Makefile. The comment in config.h implied that defining the macro would
552 enable coverage support, which is totally false. There was also support for
553 setting this macro in the CMake files (my fault, I just copied it from
554 configure). SUPPORT_GCOV has now been removed.
555
556 3. Make a small performance improvement in strlen16() and strlen32() in
557 pcretest.
558
559 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
560 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
561 these cases). There was also one in pcretest.c.
562
563 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
564
565 6. In UTF mode, the code for checking whether a group could match an empty
566 string (which is used for indefinitely repeated groups to allow for
567 breaking an infinite loop) was broken when the group contained a repeated
568 negated single-character class with a character that occupied more than one
569 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
570 UTF-8 mode). The effect was undefined: the group might or might not be
571 deemed as matching an empty string, or the program might have crashed.
572
573 7. The code for checking whether a group could match an empty string was not
574 recognizing that \h, \H, \v, \V, and \R must match a character.
575
576 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
577 an empty string. If it can, pcretest shows this in its information output.
578
579 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
580 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
581 matched by pcre_exec() without using JIT:
582
583 (a) If the rest of the pattern did not match after a maximal run of
584 grapheme clusters, the code for backing up to try with fewer of them
585 did not always back up over a full grapheme when characters that do not
586 have the modifier quality were involved, e.g. Hangul syllables.
587
588 (b) If the match point in a subject started with modifier character, and
589 there was no match, the code could incorrectly back up beyond the match
590 point, and potentially beyond the first character in the subject,
591 leading to a segfault or an incorrect match result.
592
593 10. A conditional group with an assertion condition could lead to PCRE
594 recording an incorrect first data item for a match if no other first data
595 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
596 first data item, and therefore matched "ca" after "c" instead of at the
597 start.
598
599 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
600 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
601
602 12. The source of pcregrep now includes z/OS-specific code so that it can be
603 compiled for z/OS as part of the special z/OS distribution.
604
605 13. Added the -T and -TM options to pcretest.
606
607 14. The code in pcre_compile.c for creating the table of named capturing groups
608 has been refactored. Instead of creating the table dynamically during the
609 actual compiling pass, the information is remembered during the pre-compile
610 pass (on the stack unless there are more than 20 named groups, in which
611 case malloc() is used) and the whole table is created before the actual
612 compile happens. This has simplified the code (it is now nearly 150 lines
613 shorter) and prepared the way for better handling of references to groups
614 with duplicate names.
615
616 15. A back reference to a named subpattern when there is more than one of the
617 same name now checks them in the order in which they appear in the pattern.
618 The first one that is set is used for the reference. Previously only the
619 first one was inspected. This change makes PCRE more compatible with Perl.
620
621 16. Unicode character properties were updated from Unicode 6.3.0.
622
623 17. The compile-time code for auto-possessification has been refactored, based
624 on a patch by Zoltan Herczeg. It now happens after instead of during
625 compilation. The code is cleaner, and more cases are handled. The option
626 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
627 options in pcretest are provided to set it. It can also be set by
628 (*NO_AUTO_POSSESS) at the start of a pattern.
629
630 18. The character VT has been added to the default ("C" locale) set of
631 characters that match \s and are generally treated as white space,
632 following this same change in Perl 5.18. There is now no difference between
633 "Perl space" and "POSIX space". Whether VT is treated as white space in
634 other locales depends on the locale.
635
636 19. The code for checking named groups as conditions, either for being set or
637 for being recursed, has been refactored (this is related to 14 and 15
638 above). Processing unduplicated named groups should now be as fast at
639 numerical groups, and processing duplicated groups should be faster than
640 before.
641
642 20. Two patches to the CMake build system, by Alexander Barkov:
643
644 (1) Replace the "source" command by "." in CMakeLists.txt because
645 "source" is a bash-ism.
646
647 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
648 without these the CMake build does not work on Solaris.
649
650 21. Perl has changed its handling of \8 and \9. If there is no previously
651 encountered capturing group of those numbers, they are treated as the
652 literal characters 8 and 9 instead of a binary zero followed by the
653 literals. PCRE now does the same.
654
655 22. Following Perl, added \o{} to specify codepoints in octal, making it
656 possible to specify values greater than 0777 and also making them
657 unambiguous.
658
659 23. Perl now gives an error for missing closing braces after \x{... instead of
660 treating the string as literal. PCRE now does the same.
661
662 24. RunTest used to grumble if an inappropriate test was selected explicitly,
663 but just skip it when running all tests. This make it awkward to run ranges
664 of tests when one of them was inappropriate. Now it just skips any
665 inappropriate tests, as it always did when running all tests.
666
667 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
668 character types such as \d or \w, too many callouts were inserted, and the
669 data that they returned was rubbish.
670
671 26. In UCP mode, \s was not matching two of the characters that Perl matches,
672 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
673 were matched by \h. The code has now been refactored so that the lists of
674 the horizontal and vertical whitespace characters used for \h and \v (which
675 are defined only in one place) are now also used for \s.
676
677 27. Add JIT support for the 64 bit TileGX architecture.
678 Patch by Jiong Wang (Tilera Corporation).
679
680 28. Possessive quantifiers for classes (both explicit and automatically
681 generated) now use special opcodes instead of wrapping in ONCE brackets.
682
683 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
684 (because it's meaningless), this was not happening when PCRE_CASELESS was
685 set. Not wrong, but inefficient.
686
687 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
688 properties for \w, \d, etc) is present in a test regex. Otherwise if the
689 test contains no characters greater than 255, Perl doesn't realise it
690 should be using Unicode semantics.
691
692 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
693 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
694 does in Unicode mode.
695
696 32. Added the "forbid" facility to pcretest so that putting tests into the
697 wrong test files can sometimes be quickly detected.
698
699 33. There is now a limit (default 250) on the depth of nesting of parentheses.
700 This limit is imposed to control the amount of system stack used at compile
701 time. It can be changed at build time by --with-parens-nest-limit=xxx or
702 the equivalent in CMake.
703
704 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
705 errors. Perl warns for these when in warning mode, but PCRE has no facility
706 for giving warnings.
707
708 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
709 However, this was not working for (?!) because it is optimized to (*FAIL),
710 for which PCRE does not allow quantifiers. The optimization is now disabled
711 when a quantifier follows (?!). I can't see any use for this, but it makes
712 things uniform.
713
714 36. Perl no longer allows group names to start with digits, so I have made this
715 change also in PCRE. It simplifies the code a bit.
716
717 37. In extended mode, Perl ignores spaces before a + that indicates a
718 possessive quantifier. PCRE allowed a space before the quantifier, but not
719 before the possessive +. It now does.
720
721 38. The use of \K (reset reported match start) within a repeated possessive
722 group such as (a\Kb)*+ was not working.
723
724 40. Document that the same character tables must be used at compile time and
725 run time, and that the facility to pass tables to pcre_exec() and
726 pcre_dfa_exec() is for use only with saved/restored patterns.
727
728 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
729 features for Builds with MSVC:
730
731 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
732 builds). The .rc files can be used to set FileDescription and many other
733 attributes.
734
735 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
736 This allows higher-level build scripts which want .pdb files to avoid
737 hard-coding the exact files needed."
738
739 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
740 mean "start of word" and "end of word", respectively, as a transition aid.
741
742 43. A minimizing repeat of a class containing codepoints greater than 255 in
743 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
744 compiled to use the heap for recursion.
745
746 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
747 is configured.
748
749
750 Version 8.33 28-May-2013
751 ------------------------
752
753 1. Added 'U' to some constants that are compared to unsigned integers, to
754 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
755 variables that are added to signed variables, to ensure the result is
756 signed and can be negated.
757
758 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
759 CMake config files.
760
761 3. Revise the creation of config.h.generic so that all boolean macros are
762 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
763 overriding via -D on the command line possible.
764
765 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
766 to unsigned int is reported to make a quite noticeable speed difference in
767 a specific Windows environment. Testing on Linux did also appear to show
768 some benefit (and it is clearly not harmful). Also fixed the definition of
769 Xop which should be unsigned.
770
771 5. Related to (4), changing the definition of the intermediate variable cc
772 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
773 improvements.
774
775 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
776 unnecessary spaces.
777
778 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
779 and later.
780
781 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
782
783 9. Optimizing fast_forward_start_bits in JIT.
784
785 10. Adding support for callouts in JIT, and fixing some issues revealed
786 during this work. Namely:
787
788 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
789
790 (b) Minimum length was not checked before the matching is started.
791
792 11. The value of capture_last that is passed to callouts was incorrect in some
793 cases when there was a capture on one path that was subsequently abandoned
794 after a backtrack. Also, the capture_last value is now reset after a
795 recursion, since all captures are also reset in this case.
796
797 12. The interpreter no longer returns the "too many substrings" error in the
798 case when an overflowing capture is in a branch that is subsequently
799 abandoned after a backtrack.
800
801 13. In the pathological case when an offset vector of size 2 is used, pcretest
802 now prints out the matched string after a yield of 0 or 1.
803
804 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
805 Only supported by the JIT compiler at the moment.
806
807 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
808
809 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
810 offset of the starting point of the matching process, provided the offsets
811 vector is large enough.
812
813 17. The \A escape now records a lookbehind value of 1, though its execution
814 does not actually inspect the previous character. This is to ensure that,
815 in partial multi-segment matching, at least one character from the old
816 segment is retained when a new segment is processed. Otherwise, if there
817 are no lookbehinds in the pattern, \A might match incorrectly at the start
818 of a new segment.
819
820 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
821
822 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
823 gives some modest performance improvement in 8-bit mode.
824
825 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
826 be expressed in certain programming languages using Universal Character
827 Names.
828
829 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
830 which points out that "non characters" are not "characters that may not
831 appear in Unicode strings" but rather "characters that are reserved for
832 internal use and have only local meaning".
833
834 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
835 there was a conditional group that depended on an assertion, if the
836 assertion was false, the callout that immediately followed the alternation
837 in the condition was skipped when pcre_exec() was used for matching.
838
839 23. Allow an explicit callout to be inserted before an assertion that is the
840 condition for a conditional group, for compatibility with automatic
841 callouts, which always insert a callout at this point.
842
843 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
844 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
845
846 25. (*PRUNE) is now supported by the JIT compiler.
847
848 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
849
850 27. Fix the case where there are two or more SKIPs with arguments that may be
851 ignored.
852
853 28. (*SKIP) is now supported by the JIT compiler.
854
855 29. (*THEN) is now supported by the JIT compiler.
856
857 30. Update RunTest with additional test selector options.
858
859 31. The way PCRE handles backtracking verbs has been changed in two ways.
860
861 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
862 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
863 backtracking. In some cases this makes it more Perl-compatible, but Perl's
864 rather obscure rules do not always do the same thing.
865
866 (2) Previously, backtracking verbs were confined within assertions. This is
867 no longer the case for positive assertions, except for (*ACCEPT). Again,
868 this sometimes improves Perl compatibility, and sometimes does not.
869
870 32. A number of tests that were in test 2 because Perl did things differently
871 have been moved to test 1, because either Perl or PCRE has changed, and
872 these tests are now compatible.
873
874 32. Backtracking control verbs are now handled in the same way in JIT and
875 interpreter.
876
877 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
878 contained a forward subroutine reference caused a compile error.
879
880 34. Auto-detect and optimize limited repetitions in JIT.
881
882 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
883 blocking (*UTF) etc.
884
885 36. In the interpreter, maximizing pattern repetitions for characters and
886 character types now use tail recursion, which reduces stack usage.
887
888 37. The value of the max lookbehind was not correctly preserved if a compiled
889 and saved regex was reloaded on a host of different endianness.
890
891 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
892 of the compiled pattern block, expand the flags field from 16 to 32 bits
893 because it was almost full.
894
895 39. Try madvise first before posix_madvise.
896
897 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
898 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
899 original change locked out all matches of empty strings. This has been
900 changed so that one match of an empty string per line is recognized.
901 Subsequent searches on the same line (for colouring or for --only-matching,
902 for example) do not recognize empty strings.
903
904 41. Applied a user patch to fix a number of spelling mistakes in comments.
905
906 42. Data lines longer than 65536 caused pcretest to crash.
907
908 43. Clarified the data type for length and startoffset arguments for pcre_exec
909 and pcre_dfa_exec in the function-specific man pages, where they were
910 explicitly stated to be in bytes, never having been updated. I also added
911 some clarification to the pcreapi man page.
912
913 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
914 a segmentation fault.
915
916
917 Version 8.32 30-November-2012
918 -----------------------------
919
920 1. Improved JIT compiler optimizations for first character search and single
921 character iterators.
922
923 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
924 Patch by Daniel Richard G.
925
926 3. Single character iterator optimizations in the JIT compiler.
927
928 4. Improved JIT compiler optimizations for character ranges.
929
930 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
931 Reported by Giuseppe D'Angelo.
932
933 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
934 of a line, was being set incorrectly in cases where .* appeared inside
935 atomic brackets at the start of a pattern, or where there was a subsequent
936 *PRUNE or *SKIP.
937
938 7. Improved instruction cache flush for POWER/PowerPC.
939 Patch by Daniel Richard G.
940
941 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
942 grep:
943
944 (a) There is now no limit to the number of patterns to be matched.
945
946 (b) An error is given if a pattern is too long.
947
948 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
949 are now supported.
950
951 (d) --exclude-from and --include-from (multiple use) have been added.
952
953 (e) Exclusions and inclusions now apply to all files and directories, not
954 just to those obtained from scanning a directory recursively.
955
956 (f) Multiple uses of -f and --file-list are now supported.
957
958 (g) In a Windows environment, the default for -d has been changed from
959 "read" (the GNU grep default) to "skip", because otherwise the presence
960 of a directory in the file list provokes an error.
961
962 (h) The documentation has been revised and clarified in places.
963
964 9. Improve the matching speed of capturing brackets.
965
966 10. Changed the meaning of \X so that it now matches a Unicode extended
967 grapheme cluster.
968
969 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
970 out POSIX threads when JIT support is configured.
971
972 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
973
974 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
975 structure could go wrong in environments where size_t is not the same size
976 as int.
977
978 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
979 set.
980
981 15. The EBCDIC support had decayed; later updates to the code had included
982 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
983 general tidy up of EBCDIC-related issues, and the documentation was also
984 not quite right. There is now a test that can be run on ASCII systems to
985 check some of the EBCDIC-related things (but is it not a full test).
986
987 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
988 in a small tidy to the code.
989
990 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
991
992 18. If the --only-matching (-o) option in pcregrep is specified multiple
993 times, each one causes appropriate output. For example, -o1 -o2 outputs the
994 substrings matched by the 1st and 2nd capturing parentheses. A separating
995 string can be specified by --om-separator (default empty).
996
997 19. Improving the first n character searches.
998
999 20. Turn case lists for horizontal and vertical white space into macros so that
1000 they are defined only once.
1001
1002 21. This set of changes together give more compatible Unicode case-folding
1003 behaviour for characters that have more than one other case when UCP
1004 support is available.
1005
1006 (a) The Unicode property table now has offsets into a new table of sets of
1007 three or more characters that are case-equivalent. The MultiStage2.py
1008 script that generates these tables (the pcre_ucd.c file) now scans
1009 CaseFolding.txt instead of UnicodeData.txt for character case
1010 information.
1011
1012 (b) The code for adding characters or ranges of characters to a character
1013 class has been abstracted into a generalized function that also handles
1014 case-independence. In UTF-mode with UCP support, this uses the new data
1015 to handle characters with more than one other case.
1016
1017 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1018 whose other case is greater than 256 are now correctly matched
1019 caselessly. Previously, the high codepoint matched the low one, but not
1020 vice versa.
1021
1022 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1023 of the new class addition function, using character lists defined as
1024 macros alongside the case definitions of 20 above.
1025
1026 (e) Caseless back references now work with characters that have more than
1027 one other case.
1028
1029 (f) General caseless matching of characters with more than one other case
1030 is supported.
1031
1032 22. Unicode character properties were updated from Unicode 6.2.0
1033
1034 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1035
1036 24. Add support for 32-bit character strings, and UTF-32
1037
1038 25. Major JIT compiler update (code refactoring and bugfixing).
1039 Experimental Sparc 32 support is added.
1040
1041 26. Applied a modified version of Daniel Richard G's patch to create
1042 pcre.h.generic and config.h.generic by "make" instead of in the
1043 PrepareRelease script.
1044
1045 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1046 pcre_compile.c when checking for a zero character.
1047
1048 28. Introducing a native interface for JIT. Through this interface, the compiled
1049 machine code can be directly executed. The purpose of this interface is to
1050 provide fast pattern matching, so several sanity checks are not performed.
1051 However, feature tests are still performed. The new interface provides
1052 1.4x speedup compared to the old one.
1053
1054 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1055 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1056 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1057
1058 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1059 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1060 the "old" RFC 2279). Instead, it ended up passing a negative length to
1061 pcre_exec().
1062
1063 31. Add support for GCC's visibility feature to hide internal functions.
1064
1065 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1066 "unknown -C option" after outputting 0 or 1.
1067
1068 33. There is now support for generating a code coverage report for the test
1069 suite in environments where gcc is the compiler and lcov is installed. This
1070 is mainly for the benefit of the developers.
1071
1072 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1073 unaddressable using valgrind annotations, allowing valgrind to detect
1074 invalid memory accesses. This is mainly for the benefit of the developers.
1075
1076 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1077
1078 26. Give configure error if --enable-cpp but no C++ compiler found.
1079
1080
1081 Version 8.31 06-July-2012
1082 -------------------------
1083
1084 1. Fixing a wrong JIT test case and some compiler warnings.
1085
1086 2. Removed a bashism from the RunTest script.
1087
1088 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1089 to unsigned type, result still unsigned" that was given by an MS compiler
1090 on encountering the code "-sizeof(xxx)".
1091
1092 4. Partial matching support is added to the JIT compiler.
1093
1094 5. Fixed several bugs concerned with partial matching of items that consist
1095 of more than one character:
1096
1097 (a) /^(..)\1/ did not partially match "aba" because checking references was
1098 done on an "all or nothing" basis. This also applied to repeated
1099 references.
1100
1101 (b) \R did not give a hard partial match if \r was found at the end of the
1102 subject.
1103
1104 (c) \X did not give a hard partial match after matching one or more
1105 characters at the end of the subject.
1106
1107 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1108 a partial match for the string "\r".
1109
1110 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1111 a partial match for a CR character at the end of the subject string.
1112
1113 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1114 running pcretest, the text "(JIT)" added to the output whenever JIT is
1115 actually used to run the match.
1116
1117 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1118 or /S+[+] with a digit between 1 and 7.
1119
1120 8. OP_NOT now supports any UTF character not just single-byte ones.
1121
1122 9. (*MARK) control verb is now supported by the JIT compiler.
1123
1124 10. The command "./RunTest list" lists the available tests without actually
1125 running any of them. (Because I keep forgetting what they all are.)
1126
1127 11. Add PCRE_INFO_MAXLOOKBEHIND.
1128
1129 12. Applied a (slightly modified) user-supplied patch that improves performance
1130 when the heap is used for recursion (compiled with --disable-stack-for-
1131 recursion). Instead of malloc and free for each heap frame each time a
1132 logical recursion happens, frames are retained on a chain and re-used where
1133 possible. This sometimes gives as much as 30% improvement.
1134
1135 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1136 call.
1137
1138 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1139
1140 15. It is now possible to link pcretest with libedit as an alternative to
1141 libreadline.
1142
1143 16. (*COMMIT) control verb is now supported by the JIT compiler.
1144
1145 17. The Unicode data tables have been updated to Unicode 6.1.0.
1146
1147 18. Added --file-list option to pcregrep.
1148
1149 19. Added binary file support to pcregrep, including the -a, --binary-files,
1150 -I, and --text options.
1151
1152 20. The madvise function is renamed for posix_madvise for QNX compatibility
1153 reasons. Fixed by Giuseppe D'Angelo.
1154
1155 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1156
1157 22. Changed the option for creating symbolic links for 16-bit man pages from
1158 -s to -sf so that re-installing does not cause issues.
1159
1160 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1161
1162 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1163 matches in certain environments (the workspace was not being correctly
1164 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1165 some of the workspace data at the beginning of a restart.
1166
1167 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1168 was not doing so when it should - probably a typo introduced by SVN 528
1169 (change 8.10/14).
1170
1171 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1172 \w+ when the character tables indicated that \x{c4} was a word character.
1173 There were several related cases, all because the tests for doing a table
1174 lookup were testing for characters less than 127 instead of 255.
1175
1176 27. If a pattern contains capturing parentheses that are not used in a match,
1177 their slots in the ovector are set to -1. For those that are higher than
1178 any matched groups, this happens at the end of processing. In the case when
1179 there were back references that the ovector was too small to contain
1180 (causing temporary malloc'd memory to be used during matching), and the
1181 highest capturing number was not used, memory off the end of the ovector
1182 was incorrectly being set to -1. (It was using the size of the temporary
1183 memory instead of the true size.)
1184
1185 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1186 ovector size, it uses memory at the end of the block that it has got.
1187
1188 29. Check for an overlong MARK name and give an error at compile time. The
1189 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1190
1191 30. JIT compiler update.
1192
1193 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1194 Rill for the patch.
1195
1196 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1197
1198 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1199
1200 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1201 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1202 was enabled.
1203
1204 35. Improve JIT code generation for greedy plus quantifier.
1205
1206 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1207 1 to "aa" instead of to an empty string. The bug affected repeated groups
1208 that could potentially match an empty string.
1209
1210 37. Optimizing single character iterators in JIT.
1211
1212 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1213 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1214 codepoints that are too big for the mode are faulted, and in a UTF mode,
1215 disallowed codepoints are also faulted.
1216
1217 39. If PCRE was compiled with UTF support, in three places in the DFA
1218 matcher there was code that should only have been obeyed in UTF mode, but
1219 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1220 processing when bytes with values greater than 127 were present. In 16-bit
1221 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1222 both cases the values are those that cannot be the first data item in a UTF
1223 character. The three items that might have provoked this were recursions,
1224 possessively repeated groups, and atomic groups.
1225
1226 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1227 and pcregrep, because some OS require shared objects to be explicitly
1228 passed to ld, causing the link step to fail if they are not.
1229
1230 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1231 mode, patterns that started with \h* or \R* might be incorrectly matched.
1232
1233
1234 Version 8.30 04-February-2012
1235 -----------------------------
1236
1237 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1238 name is defined in ctype.h.
1239
1240 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1241 only in quite long subpatterns.
1242
1243 3. Removed the function pcre_info(), which has been obsolete and deprecated
1244 since it was replaced by pcre_fullinfo() in February 2000.
1245
1246 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1247 match a (*MARK), and the match failed at the start of the subject, a
1248 reference to memory before the start of the subject could occur. This bug
1249 was introduced by fix 17 of release 8.21.
1250
1251 5. A reference to an unset group with zero minimum repetition was giving
1252 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1253 /(another)?(\1?)test/ matched against "hello world test". This bug was
1254 introduced in release 8.13.
1255
1256 6. Add support for 16-bit character strings (a large amount of work involving
1257 many changes and refactorings).
1258
1259 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1260 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1261 from a file.
1262
1263 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1264 rounding is not applied in this particular case).
1265
1266 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1267 if they appear, or are escaped, in patterns.
1268
1269 10. Get rid of a number of -Wunused-but-set-variable warnings.
1270
1271 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1272 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1273 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1274 also returns the mark "x". This bug applied to capturing parentheses,
1275 non-capturing parentheses, and atomic parentheses. It also applied to some
1276 assertions.
1277
1278 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1279 information out of configure.ac instead of relying on pcre.h.generic, which
1280 is not stored in the repository.
1281
1282 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1283 -lreadline.
1284
1285 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1286
1287 15. Applied Graycode's patch to put the top-level frame on the stack rather
1288 than the heap when not using the stack for recursion. This gives a
1289 performance improvement in many cases when recursion is not deep.
1290
1291 16. Experimental code added to "pcretest -C" to output the stack frame size.
1292
1293
1294 Version 8.21 12-Dec-2011
1295 ------------------------
1296
1297 1. Updating the JIT compiler.
1298
1299 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1300 are added as well.
1301
1302 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1303 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1304 calling _pcre_jit_exec. Some extra comments are added.
1305
1306 4. (*MARK) settings inside atomic groups that do not contain any capturing
1307 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1308 was introduced by change 18 for 8.20.
1309
1310 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1311 ECMA-262 standard.
1312
1313 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1314 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1315 This bug was probably introduced by change 9 of 8.13.
1316
1317 7. While fixing 6 above, I noticed that a number of other items were being
1318 incorrectly rejected as "not fixed length". This arose partly because newer
1319 opcodes had not been added to the fixed-length checking code. I have (a)
1320 corrected the bug and added tests for these items, and (b) arranged for an
1321 error to occur if an unknown opcode is encountered while checking for fixed
1322 length instead of just assuming "not fixed length". The items that were
1323 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1324 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1325 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1326
1327 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1328 being incorrectly compiled and would have given unpredicatble results.
1329
1330 9. A possessively repeated subpattern with minimum repeat count greater than
1331 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1332 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1333 the first (A) could occur when it should not.
1334
1335 10. Add a cast and remove a redundant test from the code.
1336
1337 11. JIT should use pcre_malloc/pcre_free for allocation.
1338
1339 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1340 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1341 is anything other than /usr, -L is still shown).
1342
1343 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1344
1345 14. Perl does not support \N without a following name in a [] class; PCRE now
1346 also gives an error.
1347
1348 15. If a forward reference was repeated with an upper limit of around 2000,
1349 it caused the error "internal error: overran compiling workspace". The
1350 maximum number of forward references (including repeats) was limited by the
1351 internal workspace, and dependent on the LINK_SIZE. The code has been
1352 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1353 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1354 of around 200,000 forward references. While doing this, I also speeded up
1355 the filling in of repeated forward references.
1356
1357 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1358 incorrectly expecting the subject to contain another "a" after the start.
1359
1360 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1361 in the match, the SKIP should be ignored. This was not happening; instead
1362 the SKIP was being treated as NOMATCH. For patterns such as
1363 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1364 tested.
1365
1366 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1367 now much more compatible with Perl, in particular in cases where the result
1368 is a non-match for a non-anchored pattern. For example, if
1369 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1370 "m", where previously it did not return a name. A side effect of this
1371 change is that for partial matches, the last encountered mark name is
1372 returned, as for non matches. A number of tests that were previously not
1373 Perl-compatible have been moved into the Perl-compatible test files. The
1374 refactoring has had the pleasing side effect of removing one argument from
1375 the match() function, thus reducing its stack requirements.
1376
1377 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1378 subsequent uses of /S (without +) incorrectly behaved like /S+.
1379
1380 21. Retrieve executable code size support for the JIT compiler and fixing
1381 some warnings.
1382
1383 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1384 not work when the shorter character appeared right at the end of the
1385 subject string.
1386
1387 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1388 systems.
1389
1390 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1391 output it when the /M option is used in pcretest.
1392
1393 25. The CheckMan script was not being included in the distribution. Also, added
1394 an explicit "perl" to run Perl scripts from the PrepareRelease script
1395 because this is reportedly needed in Windows.
1396
1397 26. If study data was being save in a file and studying had not found a set of
1398 "starts with" bytes for the pattern, the data written to the file (though
1399 never used) was taken from uninitialized memory and so caused valgrind to
1400 complain.
1401
1402 27. Updated RunTest.bat as provided by Sheri Pierce.
1403
1404 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1405
1406 29. Computation of memory usage for the table of capturing group names was
1407 giving an unnecessarily large value.
1408
1409
1410 Version 8.20 21-Oct-2011
1411 ------------------------
1412
1413 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1414 a POSIX class. After further experiments with Perl, which convinced me that
1415 Perl has bugs and confusions, a closing square bracket is no longer allowed
1416 in a POSIX name. This bug also affected patterns with classes that started
1417 with full stops.
1418
1419 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1420 captured substring, but while checking the failing first alternative,
1421 substring 1 is temporarily captured. If the output vector supplied to
1422 pcre_exec() was not big enough for this capture, the yield of the function
1423 was still zero ("insufficient space for captured substrings"). This cannot
1424 be totally fixed without adding another stack variable, which seems a lot
1425 of expense for a edge case. However, I have improved the situation in cases
1426 such as /(a)(b)x|abc/ matched against "abc", where the return code
1427 indicates that fewer than the maximum number of slots in the ovector have
1428 been set.
1429
1430 3. Related to (2) above: when there are more back references in a pattern than
1431 slots in the output vector, pcre_exec() uses temporary memory during
1432 matching, and copies in the captures as far as possible afterwards. It was
1433 using the entire output vector, but this conflicts with the specification
1434 that only 2/3 is used for passing back captured substrings. Now it uses
1435 only the first 2/3, for compatibility. This is, of course, another edge
1436 case.
1437
1438 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1439 main code base, and can be used by building with --enable-jit. When this is
1440 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1441 runtime --no-jit option is given.
1442
1443 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1444 ovector, the return from the function was zero, implying that there were
1445 other matches that did not fit. The correct "exactly full" value is now
1446 returned.
1447
1448 6. If a subpattern that was called recursively or as a subroutine contained
1449 (*PRUNE) or any other control that caused it to give a non-standard return,
1450 invalid errors such as "Error -26 (nested recursion at the same subject
1451 position)" or even infinite loops could occur.
1452
1453 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1454 computing the minimum length on reaching *ACCEPT, and so ended up with the
1455 wrong value of 1 rather than 0. Further investigation indicates that
1456 computing a minimum subject length in the presence of *ACCEPT is difficult
1457 (think back references, subroutine calls), and so I have changed the code
1458 so that no minimum is registered for a pattern that contains *ACCEPT.
1459
1460 8. If (*THEN) was present in the first (true) branch of a conditional group,
1461 it was not handled as intended. [But see 16 below.]
1462
1463 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1464 Sheri Pierce.
1465
1466 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1467 the first byte in a match must be "a".
1468
1469 11. Change 17 for 8.13 increased the recursion depth for patterns like
1470 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1471 pattern contains any instances of (*THEN). If it does not, the old
1472 optimizations are restored. It would be nice to do this on a per-group
1473 basis, but at the moment that is not feasible.
1474
1475 12. In some environments, the output of pcretest -C is CRLF terminated. This
1476 broke RunTest's code that checks for the link size. A single white space
1477 character after the value is now allowed for.
1478
1479 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1480 For "fr", it uses the Windows-specific input and output files.
1481
1482 14. If (*THEN) appeared in a group that was called recursively or as a
1483 subroutine, it did not work as intended. [But see next item.]
1484
1485 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1486 pattern fragments (but not containing any | characters). If A and B are
1487 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1488 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1489 D. In other words, Perl considers parentheses that do not contain any |
1490 characters to be part of a surrounding alternative, whereas PCRE was
1491 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1492 differently. PCRE now behaves in the same way as Perl, except in the case
1493 of subroutine/recursion calls such as (?1) which have in any case always
1494 been different (but PCRE had them first :-).
1495
1496 16. Related to 15 above: Perl does not treat the | in a conditional group as
1497 creating alternatives. Such a group is treated in the same way as an
1498 ordinary group without any | characters when processing (*THEN). PCRE has
1499 been changed to match Perl's behaviour.
1500
1501 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1502 RunGrepTest script failed.
1503
1504 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1505 inevitable for groups that contain captures, but it can lead to a lot of
1506 stack use in large patterns. The old behaviour has been restored for atomic
1507 groups that do not contain any capturing parentheses.
1508
1509 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1510 suppress the check for a minimum subject length at run time. (If it was
1511 given to pcre_exec() or pcre_dfa_exec() it did work.)
1512
1513 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1514 fail to work when decoding hex characters in data strings in EBCDIC
1515 environments.
1516
1517 21. It appears that in at least one Mac OS environment, the isxdigit() function
1518 is implemented as a macro that evaluates to its argument more than once,
1519 contravening the C 90 Standard (I haven't checked a later standard). There
1520 was an instance in pcretest which caused it to go wrong when processing
1521 \x{...} escapes in subject strings. The has been rewritten to avoid using
1522 things like p++ in the argument of isxdigit().
1523
1524
1525 Version 8.13 16-Aug-2011
1526 ------------------------
1527
1528 1. The Unicode data tables have been updated to Unicode 6.0.0.
1529
1530 2. Two minor typos in pcre_internal.h have been fixed.
1531
1532 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1533 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1534 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1535
1536 4. There were a number of related bugs in the code for matching backrefences
1537 caselessly in UTF-8 mode when codes for the characters concerned were
1538 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1539 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1540 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1541 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1542 2-byte code at the end of the subject (it thought there wasn't enough data
1543 left).
1544
1545 5. Comprehensive information about what went wrong is now returned by
1546 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1547 as the output vector has at least 2 elements. The offset of the start of
1548 the failing character and a reason code are placed in the vector.
1549
1550 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1551 now returned is for the first byte of the failing character, instead of the
1552 last byte inspected. This is an incompatible change, but I hope it is small
1553 enough not to be a problem. It makes the returned offset consistent with
1554 pcre_exec() and pcre_dfa_exec().
1555
1556 7. pcretest now gives a text phrase as well as the error number when
1557 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1558 failure, the offset and reason code are output.
1559
1560 8. When \R was used with a maximizing quantifier it failed to skip backwards
1561 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1562 back over a single character (\n). This seems wrong (because it treated the
1563 two characters as a single entity when going forwards), conflicts with the
1564 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1565 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1566 has been changed.
1567
1568 9. Some internal refactoring has changed the processing so that the handling
1569 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1570 time (the PCRE_DOTALL option was changed this way some time ago: version
1571 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1572 which was always a bit of a fudge. It also means that there is one less
1573 argument for the match() function, which reduces its stack requirements
1574 slightly. This change also fixes an incompatibility with Perl: the pattern
1575 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1576
1577 10. More internal refactoring has drastically reduced the number of recursive
1578 calls to match() for possessively repeated groups such as (abc)++ when
1579 using pcre_exec().
1580
1581 11. While implementing 10, a number of bugs in the handling of groups were
1582 discovered and fixed:
1583
1584 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1585 (a|)*(?1) gave a compile-time internal error.
1586 ((a|)+)+ did not notice that the outer group could match an empty string.
1587 (^a|^)+ was not marked as anchored.
1588 (.*a|.*)+ was not marked as matching at start or after a newline.
1589
1590 12. Yet more internal refactoring has removed another argument from the match()
1591 function. Special calls to this function are now indicated by setting a
1592 value in a variable in the "match data" data block.
1593
1594 13. Be more explicit in pcre_study() instead of relying on "default" for
1595 opcodes that mean there is no starting character; this means that when new
1596 ones are added and accidentally left out of pcre_study(), testing should
1597 pick them up.
1598
1599 14. The -s option of pcretest has been documented for ages as being an old
1600 synonym of -m (show memory usage). I have changed it to mean "force study
1601 for every regex", that is, assume /S for every regex. This is similar to -i
1602 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1603 using it. It makes it easier to run collections of tests with and without
1604 study enabled, and thereby test pcre_study() more easily. All the standard
1605 tests are now run with and without -s (but some patterns can be marked as
1606 "never study" - see 20 below).
1607
1608 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1609 restoration of the capturing data to the outer values was not happening
1610 correctly.
1611
1612 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1613 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1614 pattern had matched an empty string, and so incorrectly returned a no
1615 match.
1616
1617 17. There was optimizing code for the last branch of non-capturing parentheses,
1618 and also for the obeyed branch of a conditional subexpression, which used
1619 tail recursion to cut down on stack usage. Unfortunately, now that there is
1620 the possibility of (*THEN) occurring in these branches, tail recursion is
1621 no longer possible because the return has to be checked for (*THEN). These
1622 two optimizations have therefore been removed. [But see 8.20/11 above.]
1623
1624 18. If a pattern containing \R was studied, it was assumed that \R always
1625 matched two bytes, thus causing the minimum subject length to be
1626 incorrectly computed because \R can also match just one byte.
1627
1628 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1629 was incorrectly computed.
1630
1631 20. If /S is present twice on a test pattern in pcretest input, it now
1632 *disables* studying, thereby overriding the use of -s on the command line
1633 (see 14 above). This is necessary for one or two tests to keep the output
1634 identical in both cases.
1635
1636 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1637 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1638
1639 22. When an atomic group that contained a capturing parenthesis was
1640 successfully matched, but the branch in which it appeared failed, the
1641 capturing was not being forgotten if a higher numbered group was later
1642 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1643 group 1 to "a", when in fact it should be unset. This applied to multi-
1644 branched capturing and non-capturing groups, repeated or not, and also to
1645 positive assertions (capturing in negative assertions does not happen
1646 in PCRE) and also to nested atomic groups.
1647
1648 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1649 subject after a captured substring, to make it easier to tell which of a
1650 number of identical substrings has been captured.
1651
1652 24. The way atomic groups are processed by pcre_exec() has been changed so that
1653 if they are repeated, backtracking one repetition now resets captured
1654 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1655 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1656 "aaa". Previously, it would have been "a". As part of this code
1657 refactoring, the way recursive calls are handled has also been changed.
1658
1659 25. If an assertion condition captured any substrings, they were not passed
1660 back unless some other capturing happened later. For example, if
1661 (?(?=(a))a) was matched against "a", no capturing was returned.
1662
1663 26. When studying a pattern that contained subroutine calls or assertions,
1664 the code for finding the minimum length of a possible match was handling
1665 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1666 group 1 called group 2 while simultaneously a separate group 2 called group
1667 1). A stack overflow occurred in this case. I have fixed this by limiting
1668 the recursion depth to 10.
1669
1670 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1671 Fortmann. This supports explicit test numbers on the command line, and has
1672 argument validation and error reporting.
1673
1674 28. An instance of \X with an unlimited repeat could fail if at any point the
1675 first character it looked at was a mark character.
1676
1677 29. Some minor code refactoring concerning Unicode properties and scripts
1678 should reduce the stack requirement of match() slightly.
1679
1680 30. Added the '=' option to pcretest to check the setting of unused capturing
1681 slots at the end of the pattern, which are documented as being -1, but are
1682 not included in the return count.
1683
1684 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1685 compiled something random. Now it gives a compile-time error (as does
1686 Perl).
1687
1688 32. A *MARK encountered during the processing of a positive assertion is now
1689 recorded and passed back (compatible with Perl).
1690
1691 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1692 had alternative anchored branches, the search for a second match in a line
1693 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1694 matched the line "0102" twice. The same bug affected patterns that started
1695 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1696 twice.
1697
1698 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1699 does, and because of capturing effects, quantifying parenthesized
1700 assertions may at times be useful. Quantifiers are now allowed for
1701 parenthesized assertions.
1702
1703 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1704
1705 36. \g was being checked for fancy things in a character class, when it should
1706 just be a literal "g".
1707
1708 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1709 appearance of a nested POSIX class supersedes an apparent external class.
1710 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1711 unescaped square brackets may also appear as part of class names. For
1712 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1713 more like Perl. (But see 8.20/1 above.)
1714
1715 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1716 was because it thought it was \N{name}, which is not supported).
1717
1718 39. Add minix to OS list not supporting the -S option in pcretest.
1719
1720 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1721 cannot analyze patterns in sufficient detail to catch mutual recursions
1722 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1723 subgroup is called recursively as a subpattern for a second time at the
1724 same position in the subject string. In previous releases this might have
1725 been caught by the recursion limit, or it might have run out of stack.
1726
1727 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1728 happen only once. PCRE was, however incorrectly giving a compile time error
1729 "recursive call could loop indefinitely" because it cannot analyze the
1730 pattern in sufficient detail. The compile time test no longer happens when
1731 PCRE is compiling a conditional subpattern, but actual runaway loops are
1732 now caught at runtime (see 40 above).
1733
1734 42. It seems that Perl allows any characters other than a closing parenthesis
1735 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1736 has been changed to be the same.
1737
1738 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1739 as not to get warnings when autogen.sh is called. Also changed
1740 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1741
1742 44. To help people who use pcregrep to scan files containing exceedingly long
1743 lines, the following changes have been made:
1744
1745 (a) The default value of the buffer size parameter has been increased from
1746 8K to 20K. (The actual buffer used is three times this size.)
1747
1748 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1749 PCRE is built.
1750
1751 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1752 to be set at run time.
1753
1754 (d) Numerical values in pcregrep options can be followed by K or M, for
1755 example --buffer-size=50K.
1756
1757 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1758 given and the return code is set to 2.
1759
1760 45. Add a pointer to the latest mark to the callout data block.
1761
1762 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1763 partial match of an empty string instead of no match. This was specific to
1764 the use of ".".
1765
1766 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1767 complete match instead of a partial match. This bug was dependent on both
1768 the PCRE_UTF8 and PCRE_DOTALL options being set.
1769
1770 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1771 starting byte set, because \b was not being ignored.
1772
1773
1774 Version 8.12 15-Jan-2011
1775 ------------------------
1776
1777 1. Fixed some typos in the markup of the man pages, and wrote a script that
1778 checks for such things as part of the documentation building process.
1779
1780 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1781 --match-limit and --recursion-limit options (added for 8.11). In
1782 particular, this made one of the standard tests fail. (The integer value
1783 went into the wrong half of a long int.)
1784
1785 3. If the --colour option was given to pcregrep with -v (invert match), it
1786 did strange things, either producing crazy output, or crashing. It should,
1787 of course, ignore a request for colour when reporting lines that do not
1788 match.
1789
1790 4. Another pcregrep bug caused similar problems if --colour was specified with
1791 -M (multiline) and the pattern match finished with a line ending.
1792
1793 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1794 matched in multiline mode, the following line was shown as part of the
1795 match. This seems wrong, so I have changed it.
1796
1797 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1798 the check for further matches in the same line (so they could be coloured)
1799 to overrun the end of the current line. If another match was found, it was
1800 incorrectly shown (and then shown again when found in the next line).
1801
1802 7. If pcregrep was compiled under Windows, there was a reference to the
1803 function pcregrep_exit() before it was defined. I am assuming this was
1804 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1805 reported by a user. I've moved the definition above the reference.
1806
1807
1808 Version 8.11 10-Dec-2010
1809 ------------------------
1810
1811 1. (*THEN) was not working properly if there were untried alternatives prior
1812 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1813 backtracked to try for "b" instead of moving to the next alternative branch
1814 at the same level (in this case, to look for "c"). The Perl documentation
1815 is clear that when (*THEN) is backtracked onto, it goes to the "next
1816 alternative in the innermost enclosing group".
1817
1818 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1819 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1820 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1821 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1822 (*THEN).
1823
1824 3. If \s appeared in a character class, it removed the VT character from
1825 the class, even if it had been included by some previous item, for example
1826 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1827 of \s, but is part of the POSIX "space" class.)
1828
1829 4. A partial match never returns an empty string (because you can always
1830 match an empty string at the end of the subject); however the checking for
1831 an empty string was starting at the "start of match" point. This has been
1832 changed to the "earliest inspected character" point, because the returned
1833 data for a partial match starts at this character. This means that, for
1834 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1835 (previously it gave "no match").
1836
1837 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1838 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1839 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1840 has an implication that the given string is incomplete (because a partial
1841 match is preferred over a full match). For this reason, these items now
1842 give a partial match in this situation. [Aside: previously, the one case
1843 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1844 match rather than a full match, which was wrong by the old rules, but is
1845 now correct.]
1846
1847 6. There was a bug in the handling of #-introduced comments, recognized when
1848 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1849 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1850 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1851 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1852 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1853 places in pcre_compile().
1854
1855 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1856 comments when looking ahead for named forward references to subpatterns,
1857 the only newline sequence it recognized was NL. It now handles newlines
1858 according to the set newline convention.
1859
1860 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1861 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1862 cater for a lack of strerror(). These oversights have been fixed.
1863
1864 9. Added --match-limit and --recursion-limit to pcregrep.
1865
1866 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1867
1868 11. When the -o option was used, pcregrep was setting a return code of 1, even
1869 when matches were found, and --line-buffered was not being honoured.
1870
1871 12. Added an optional parentheses number to the -o and --only-matching options
1872 of pcregrep.
1873
1874 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1875 can match an empty string. The code to do it in pcretest and pcredemo
1876 needed fixing:
1877
1878 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1879 only one byte after an empty string match just before CRLF (this case
1880 just got forgotten; "any" and "anycrlf" were OK).
1881
1882 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1883 mode when an empty string match preceded an ASCII character followed by
1884 a non-ASCII character. (The code for advancing by one character rather
1885 than one byte was nonsense.)
1886
1887 (c) The pcredemo.c sample program did not have any code at all to handle
1888 the cases when CRLF is a valid newline sequence.
1889
1890 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1891 as a starting offset was within the subject string. There is now a new
1892 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1893 negative or greater than the length of the string. In order to test this,
1894 pcretest is extended to allow the setting of negative starting offsets.
1895
1896 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1897 starting offset points to the beginning of a UTF-8 character was
1898 unnecessarily clumsy. I tidied it up.
1899
1900 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1901 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1902
1903 17. Nobody had reported that the --include_dir option, which was added in
1904 release 7.7 should have been called --include-dir (hyphen, not underscore)
1905 for compatibility with GNU grep. I have changed it to --include-dir, but
1906 left --include_dir as an undocumented synonym, and the same for
1907 --exclude-dir, though that is not available in GNU grep, at least as of
1908 release 2.5.4.
1909
1910 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1911 characters from a string of bytes) have been redefined so as not to use
1912 loops, in order to improve performance in some environments. At the same
1913 time, I abstracted some of the common code into auxiliary macros to save
1914 repetition (this should not affect the compiled code).
1915
1916 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1917 compile-time error is now given if \c is not followed by an ASCII
1918 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1919 different, and any byte value is allowed.)
1920
1921 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1922 START_OPTIMIZE option, which is now allowed at compile time - but just
1923 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1924 to pcregrep and other applications that have no direct access to PCRE
1925 options. The new /Y option in pcretest sets this option when calling
1926 pcre_compile().
1927
1928 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1929 back references. Groups containing recursive back references were forced to
1930 be atomic by that change, but in the case of named groups, the amount of
1931 memory required was incorrectly computed, leading to "Failed: internal
1932 error: code overflow". This has been fixed.
1933
1934 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1935 pcretest.c, to avoid build problems in some Borland environments.
1936
1937
1938 Version 8.10 25-Jun-2010
1939 ------------------------
1940
1941 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1942 THEN.
1943
1944 2. (*ACCEPT) was not working when inside an atomic group.
1945
1946 3. Inside a character class, \B is treated as a literal by default, but
1947 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1948 causes the error). The code is unchanged, but I tidied the documentation.
1949
1950 4. Inside a character class, PCRE always treated \R and \X as literals,
1951 whereas Perl faults them if its -w option is set. I have changed PCRE so
1952 that it faults them when PCRE_EXTRA is set.
1953
1954 5. Added support for \N, which always matches any character other than
1955 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1956
1957 6. When compiling pcregrep with newer versions of gcc which may have
1958 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1959 declared with attribute warn_unused_result" were given. Just casting the
1960 result to (void) does not stop the warnings; a more elaborate fudge is
1961 needed. I've used a macro to implement this.
1962
1963 7. Minor change to pcretest.c to avoid a compiler warning.
1964
1965 8. Added four artifical Unicode properties to help with an option to make
1966 \s etc use properties (see next item). The new properties are: Xan
1967 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1968
1969 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1970 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1971 this option. Modified pcretest to add /W to test this facility. Added
1972 REG_UCP to make it available via the POSIX interface.
1973
1974 10. Added --line-buffered to pcregrep.
1975
1976 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1977 studied, and the match started with a letter with a code point greater than
1978 127 whose first byte was different to the first byte of the other case of
1979 the letter, the other case of this starting letter was not recognized
1980 (#976).
1981
1982 12. If a pattern that was studied started with a repeated Unicode property
1983 test, for example, \p{Nd}+, there was the theoretical possibility of
1984 setting up an incorrect bitmap of starting bytes, but fortunately it could
1985 not have actually happened in practice until change 8 above was made (it
1986 added property types that matched character-matching opcodes).
1987
1988 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1989 possible starting bytes for non-anchored patterns.
1990
1991 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1992 \R, and also a number of cases that involve Unicode properties, both
1993 explicit and implicit when PCRE_UCP is set.
1994
1995 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1996 input, it could crash or give wrong results if characters with values
1997 greater than 0xc0 were present in the subject string. (Detail: it assumed
1998 UTF-8 input when processing these items.)
1999
2000 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2001 size_t is 64-bit (#991).
2002
2003 17. Added a check for running out of memory when PCRE is compiled with
2004 --disable-stack-for-recursion (#990).
2005
2006 18. If the last data line in a file for pcretest does not have a newline on
2007 the end, a newline was missing in the output.
2008
2009 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2010 less than 128) in its various bitmaps. However, there is a facility for
2011 generating tables according to the current locale when PCRE is compiled. It
2012 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2013 characters, are recognized by isspace() and therefore were getting set in
2014 these tables, and indeed these tables seem to approximate to ISO 8859. This
2015 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2016 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2017 which of course cannot start UTF-8 characters. I have changed the code so
2018 that only real ASCII characters (less than 128) and the correct starting
2019 bytes for UTF-8 encodings are set for characters greater than 127 when in
2020 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2021 altogether.)
2022
2023 20. Added the /T option to pcretest so as to be able to run tests with non-
2024 standard character tables, thus making it possible to include the tests
2025 used for 19 above in the standard set of tests.
2026
2027 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2028 reference to a subpattern the other side of a comment that contains an
2029 opening parenthesis caused either an internal compiling error, or a
2030 reference to the wrong subpattern.
2031
2032
2033 Version 8.02 19-Mar-2010
2034 ------------------------
2035
2036 1. The Unicode data tables have been updated to Unicode 5.2.0.
2037
2038 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2039 configured.
2040
2041 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2042 original author of that file, following a query about its status.
2043
2044 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2045 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2046
2047 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2048 quantifier applied to a forward-referencing subroutine call, could compile
2049 incorrect code or give the error "internal error: previously-checked
2050 referenced subpattern not found".
2051
2052 6. Both MS Visual Studio and Symbian OS have problems with initializing
2053 variables to point to external functions. For these systems, therefore,
2054 pcre_malloc etc. are now initialized to local functions that call the
2055 relevant global functions.
2056
2057 7. There were two entries missing in the vectors called coptable and poptable
2058 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2059 I've fixed the data, and added a kludgy way of testing at compile time that
2060 the lengths are correct (equal to the number of opcodes).
2061
2062 8. Following on from 7, I added a similar kludge to check the length of the
2063 eint vector in pcreposix.c.
2064
2065 9. Error texts for pcre_compile() are held as one long string to avoid too
2066 much relocation at load time. To find a text, the string is searched,
2067 counting zeros. There was no check for running off the end of the string,
2068 which could happen if a new error number was added without updating the
2069 string.
2070
2071 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2072
2073 11. \K was not working if it appeared in an atomic group or in a group that
2074 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2075 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2076 the assertion is positive, but not if it is negative.
2077
2078 12. Change 11 fortuitously reduced the size of the stack frame used in the
2079 "match()" function of pcre_exec.c by one pointer. Forthcoming
2080 implementation of support for (*MARK) will need an extra pointer on the
2081 stack; I have reserved it now, so that the stack frame size does not
2082 decrease.
2083
2084 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2085 item in branch that calls a recursion is a subroutine call - as in the
2086 second branch in the above example - was incorrectly given the compile-
2087 time error "recursive call could loop indefinitely" because pcre_compile()
2088 was not correctly checking the subroutine for matching a non-empty string.
2089
2090 14. The checks for overrunning compiling workspace could trigger after an
2091 overrun had occurred. This is a "should never occur" error, but it can be
2092 triggered by pathological patterns such as hundreds of nested parentheses.
2093 The checks now trigger 100 bytes before the end of the workspace.
2094
2095 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2096
2097
2098 Version 8.01 19-Jan-2010
2099 ------------------------
2100
2101 1. If a pattern contained a conditional subpattern with only one branch (in
2102 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2103 computed the wrong minimum data length (which is of course zero for such
2104 subpatterns). This could cause incorrect "no match" results.
2105
2106 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2107 the pattern is reset in the first branch, pcre_compile() failed with
2108 "internal error: code overflow at offset...". This happened only when
2109 the reset was to the original external option setting. (An optimization
2110 abstracts leading options settings into an external setting, which was the
2111 cause of this.)
2112
2113 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2114 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2115 assertion pattern did not match (meaning that the assertion was true), it
2116 was incorrectly treated as false if the SKIP had been reached during the
2117 matching. This also applied to assertions used as conditions.
2118
2119 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2120 assertion subpattern, including such a pattern used as a condition,
2121 unpredictable results occurred, instead of the error return
2122 PCRE_ERROR_DFA_UITEM.
2123
2124 5. The C++ GlobalReplace function was not working like Perl for the special
2125 situation when an empty string is matched. It now does the fancy magic
2126 stuff that is necessary.
2127
2128 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2129 removed. (These were left over from very, very early versions of PCRE.)
2130
2131 7. Some cosmetic changes to the code to make life easier when compiling it
2132 as part of something else:
2133
2134 (a) Change DEBUG to PCRE_DEBUG.
2135
2136 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2137 called "current" as "current_branch", to prevent a collision with the
2138 Linux macro when compiled as a kernel module.
2139
2140 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2141 prevent a collision with the Linux macro when compiled as a kernel
2142 module.
2143
2144 8. In pcre_compile() there are some checks for integer overflows that used to
2145 cast potentially large values to (double). This has been changed to that
2146 when building, a check for int64_t is made, and if it is found, it is used
2147 instead, thus avoiding the use of floating point arithmetic. (There is no
2148 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2149 double.
2150
2151 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2152 2005 (difference between two addresses compared to an unsigned value).
2153
2154 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2155 custom one, because of the following reported problem in Windows:
2156
2157 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2158 under Win32.
2159 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2160 therefore missing the function definition.
2161 - The compiler thus generates a "C" signature for the test function.
2162 - The linker fails to find the "C" function.
2163 - PCRE fails to configure if asked to do so against libbz2.
2164
2165 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2166 messages were output:
2167
2168 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2169 rerunning libtoolize, to keep the correct libtool macros in-tree.
2170 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2171
2172 I have done both of these things.
2173
2174 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2175 most of the time, it *can* run out if it is given a pattern that contains a
2176 runaway infinite recursion. I updated the discussion in the pcrestack man
2177 page.
2178
2179 13. Now that we have gone to the x.xx style of version numbers, the minor
2180 version may start with zero. Using 08 or 09 is a bad idea because users
2181 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2182 interpreted as invalid octal numbers. I've updated the previous comment in
2183 configure.ac, and also added a check that gives an error if 08 or 09 are
2184 used.
2185
2186 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2187 causing partial matching to fail when the end of the subject matched \W
2188 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2189
2190 15. There were some discrepancies between the declarations in pcre_internal.h
2191 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2192 their definitions. The declarations used "const uschar *" and the
2193 definitions used USPTR. Even though USPTR is normally defined as "const
2194 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2195 reported that: "This difference in casting confuses some C++ compilers, for
2196 example, SunCC recognizes above declarations as different functions and
2197 generates broken code for hbpcre." I have changed the declarations to use
2198 USPTR.
2199
2200 16. GNU libtool is named differently on some systems. The autogen.sh script now
2201 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2202 (FreeBSD).
2203
2204 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2205 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2206 comment: "Figure out how to create a longlong from a string: strtoll and
2207 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2208 instance, but it only takes 2 args instead of 3!"
2209
2210 18. A subtle bug concerned with back references has been fixed by a change of
2211 specification, with a corresponding code fix. A pattern such as
2212 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2213 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2214 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2215 same bug. Such groups have to be quantified to be useful, or contained
2216 inside another quantified group. (If there's no repetition, the reference
2217 can never match.) The problem arises because, having left the group and
2218 moved on to the rest of the pattern, a later failure that backtracks into
2219 the group uses the captured value from the final iteration of the group
2220 rather than the correct earlier one. I have fixed this in PCRE by forcing
2221 any group that contains a reference to itself to be an atomic group; that
2222 is, there cannot be any backtracking into it once it has completed. This is
2223 similar to recursive and subroutine calls.
2224
2225
2226 Version 8.00 19-Oct-09
2227 ----------------------
2228
2229 1. The table for translating pcre_compile() error codes into POSIX error codes
2230 was out-of-date, and there was no check on the pcre_compile() error code
2231 being within the table. This could lead to an OK return being given in
2232 error.
2233
2234 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2235 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2236 in a Windows environment.
2237
2238 3. The pcregrep --count option prints the count for each file even when it is
2239 zero, as does GNU grep. However, pcregrep was also printing all files when
2240 --files-with-matches was added. Now, when both options are given, it prints
2241 counts only for those files that have at least one match. (GNU grep just
2242 prints the file name in this circumstance, but including the count seems
2243 more useful - otherwise, why use --count?) Also ensured that the
2244 combination -clh just lists non-zero counts, with no names.
2245
2246 4. The long form of the pcregrep -F option was incorrectly implemented as
2247 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2248 but it seems right to fix it, and I didn't think it was worth preserving
2249 the old behaviour.
2250
2251 5. The command line items --regex=pattern and --regexp=pattern were not
2252 recognized by pcregrep, which required --regex pattern or --regexp pattern
2253 (with a space rather than an '='). The man page documented the '=' forms,
2254 which are compatible with GNU grep; these now work.
2255
2256 6. No libpcreposix.pc file was created for pkg-config; there was just
2257 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2258
2259 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2260 when UCP support is not needed, by modifying the Python script that
2261 generates it from Unicode data files. This should not matter if the module
2262 is correctly used as a library, but I received one complaint about 50K of
2263 unwanted data. My guess is that the person linked everything into his
2264 program rather than using a library. Anyway, it does no harm.
2265
2266 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2267 was a minimum greater than 1 for a wide character in a possessive
2268 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2269 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2270 character. Chaos in the form of incorrect output or a compiling loop could
2271 result.
2272
2273 9. The restrictions on what a pattern can contain when partial matching is
2274 requested for pcre_exec() have been removed. All patterns can now be
2275 partially matched by this function. In addition, if there are at least two
2276 slots in the offset vector, the offset of the earliest inspected character
2277 for the match and the offset of the end of the subject are set in them when
2278 PCRE_ERROR_PARTIAL is returned.
2279
2280 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2281 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2282 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2283 and may be more useful for multi-segment matching.
2284
2285 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2286 used to be given if ever the end of the subject was reached; now it is
2287 given only if matching could not proceed because another character was
2288 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2289 string "Z", which now yields "no match" instead of "partial match". In the
2290 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2291 final character ended with (*FAIL).
2292
2293 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2294 if the pattern had a "must contain" character that was already found in the
2295 earlier partial match, unless partial matching was again requested. For
2296 example, with the pattern /dog.(body)?/, the "must contain" character is
2297 "g". If the first part-match was for the string "dog", restarting with
2298 "sbody" failed. This bug has been fixed.
2299
2300 13. The string returned by pcre_dfa_exec() after a partial match has been
2301 changed so that it starts at the first inspected character rather than the
2302 first character of the match. This makes a difference only if the pattern
2303 starts with a lookbehind assertion or \b or \B (\K is not supported by
2304 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2305 matching functions compatible, and I think it's the right thing to do.
2306
2307 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2308 so that the demonstration program is easily available in environments where
2309 PCRE has not been installed from source.
2310
2311 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2312 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2313 library.
2314
2315 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2316 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2317 is not the first non-POSIX option to be added. Clearly some people find
2318 these options useful.
2319
2320 17. If a caller to the POSIX matching function regexec() passes a non-zero
2321 value for nmatch with a NULL value for pmatch, the value of
2322 nmatch is forced to zero.
2323
2324 18. RunGrepTest did not have a test for the availability of the -u option of
2325 the diff command, as RunTest does. It now checks in the same way as
2326 RunTest, and also checks for the -b option.
2327
2328 19. If an odd number of negated classes containing just a single character
2329 interposed, within parentheses, between a forward reference to a named
2330 subpattern and the definition of the subpattern, compilation crashed with
2331 an internal error, complaining that it could not find the referenced
2332 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2333 [The bug was that it was starting one character too far in when skipping
2334 over the character class, thus treating the ] as data rather than
2335 terminating the class. This meant it could skip too much.]
2336
2337 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2338 /g option in pcretest when the pattern contains \K, which makes it possible
2339 to have an empty string match not at the start, even when the pattern is
2340 anchored. Updated pcretest and pcredemo to use this option.
2341
2342 21. If the maximum number of capturing subpatterns in a recursion was greater
2343 than the maximum at the outer level, the higher number was returned, but
2344 with unset values at the outer level. The correct (outer level) value is
2345 now given.
2346
2347 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2348 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2349 make it do so. The string so far is captured, making this feature
2350 compatible with Perl.
2351
2352 23. The tests have been re-organized, adding tests 11 and 12, to make it
2353 possible to check the Perl 5.10 features against Perl 5.10.
2354
2355 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2356 pattern matches a fixed length string. PCRE did not allow this; now it
2357 does. Neither allows recursion.
2358
2359 25. I finally figured out how to implement a request to provide the minimum
2360 length of subject string that was needed in order to match a given pattern.
2361 (It was back references and recursion that I had previously got hung up
2362 on.) This code has now been added to pcre_study(); it finds a lower bound
2363 to the length of subject needed. It is not necessarily the greatest lower
2364 bound, but using it to avoid searching strings that are too short does give
2365 some useful speed-ups. The value is available to calling programs via
2366 pcre_fullinfo().
2367
2368 26. While implementing 25, I discovered to my embarrassment that pcretest had
2369 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2370 study optimizations had never been tested with that matching function.
2371 Oops. What is worse, even when it was passed study data, there was a bug in
2372 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2373 were also very few tests of studied patterns with pcre_dfa_exec().
2374
2375 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2376 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2377 on the other side of the coin, they are no longer allowed to have different
2378 names, because these cannot be distinguished in PCRE, and this has caused
2379 confusion. (This is a difference from Perl.)
2380
2381 28. When duplicate subpattern names are present (necessarily with different
2382 numbers, as required by 27 above), and a test is made by name in a
2383 conditional pattern, either for a subpattern having been matched, or for
2384 recursion in such a pattern, all the associated numbered subpatterns are
2385 tested, and the overall condition is true if the condition is true for any
2386 one of them. This is the way Perl works, and is also more like the way
2387 testing by number works.
2388
2389
2390 Version 7.9 11-Apr-09
2391 ---------------------
2392
2393 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2394 (pcretest), all targets were linked against these libraries. This included
2395 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2396 libraries. This caused unwanted dependencies to be created. This problem
2397 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2398 pcretest is linked with readline.
2399
2400 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2401 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2402 moved outside it again, because FALSE and TRUE are already defined in AIX,
2403 but BOOL is not.
2404
2405 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2406 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2407
2408 4. The pcregrep documentation said spaces were inserted as well as colons (or
2409 hyphens) following file names and line numbers when outputting matching
2410 lines. This is not true; no spaces are inserted. I have also clarified the
2411 wording for the --colour (or --color) option.
2412
2413 5. In pcregrep, when --colour was used with -o, the list of matching strings
2414 was not coloured; this is different to GNU grep, so I have changed it to be
2415 the same.
2416
2417 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2418 each matching line was coloured. Now it goes on to look for further matches
2419 of any of the test patterns, which is the same behaviour as GNU grep.
2420
2421 7. A pattern that could match an empty string could cause pcregrep to loop; it
2422 doesn't make sense to accept an empty string match in pcregrep, so I have
2423 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2424 seems to be how GNU grep behaves. [But see later change 40 for release
2425 8.33.]
2426
2427 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2428 start or after a newline", because the conditional assertion was not being
2429 correctly handled. The rule now is that both the assertion and what follows
2430 in the first alternative must satisfy the test.
2431
2432 9. If auto-callout was enabled in a pattern with a conditional group whose
2433 condition was an assertion, PCRE could crash during matching, both with
2434 pcre_exec() and pcre_dfa_exec().
2435
2436 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2437 used for matching.
2438
2439 11. Unicode property support in character classes was not working for
2440 characters (bytes) greater than 127 when not in UTF-8 mode.
2441
2442 12. Added the -M command line option to pcretest.
2443
2444 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2445
2446 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2447
2448 16. Added comments and documentation about mis-use of no_arg in the C++
2449 wrapper.
2450
2451 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2452 from Martin Jerabek that uses macro names for all relevant character and
2453 string constants.
2454
2455 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2456 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2457 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2458 these, but not everybody uses configure.
2459
2460 19. A conditional group that had only one branch was not being correctly
2461 recognized as an item that could match an empty string. This meant that an
2462 enclosing group might also not be so recognized, causing infinite looping
2463 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2464 with the subject "ab", where knowledge that the repeated group can match
2465 nothing is needed in order to break the loop.
2466
2467 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2468 exec(), but without supplying a callout function, matching went wrong.
2469
2470 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2471 leak if the size of the offset vector was greater than 30. When the vector
2472 is smaller, the saved offsets during recursion go onto a local stack
2473 vector, but for larger vectors malloc() is used. It was failing to free
2474 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2475 error, in fact).
2476
2477 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2478 heapframe that is used only when UTF-8 support is enabled. This caused no
2479 problem, but was untidy.
2480
2481 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2482 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2483 included within another project.
2484
2485 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2486 slightly modified by me:
2487
2488 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2489 not building pcregrep.
2490
2491 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2492 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2493
2494 25. Forward references, both numeric and by name, in patterns that made use of
2495 duplicate group numbers, could behave incorrectly or give incorrect errors,
2496 because when scanning forward to find the reference group, PCRE was not
2497 taking into account the duplicate group numbers. A pattern such as
2498 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2499
2500 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2501 the feature of a custom pointer more persuasive (as requested by a user).
2502
2503 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2504 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2505 pre-defined.
2506
2507 28. Added support for (*UTF8) at the start of a pattern.
2508
2509 29. Arrange for flags added by the "release type" setting in CMake to be shown
2510 in the configuration summary.
2511
2512
2513 Version 7.8 05-Sep-08
2514 ---------------------
2515
2516 1. Replaced UCP searching code with optimized version as implemented for Ad
2517 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2518 stage table and inline lookup instead of a function, giving speed ups of 2
2519 to 5 times on some simple patterns that I tested. Permission was given to
2520 distribute the MultiStage2.py script that generates the tables (it's not in
2521 the tarball, but is in the Subversion repository).
2522
2523 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2524 scripts.
2525
2526 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2527 a group with a zero qualifier. The result of the study could be incorrect,
2528 or the function might crash, depending on the pattern.
2529
2530 4. Caseless matching was not working for non-ASCII characters in back
2531 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2532 It now works when Unicode Property Support is available.
2533
2534 5. In pcretest, an escape such as \x{de} in the data was always generating
2535 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2536 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2537 truncation.
2538
2539 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2540
2541 7. Added two (int) casts to pcregrep when printing the difference of two
2542 pointers, in case they are 64-bit values.
2543
2544 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2545 test 2 if it fails.
2546
2547 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2548 and a #define of that name to empty if it is not externally set. This is to
2549 allow users of MSVC to set it if necessary.
2550
2551 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2552 the convenience functions in the pcre_get.c source file.
2553
2554 11. An option change at the start of a pattern that had top-level alternatives
2555 could cause overwriting and/or a crash. This command provoked a crash in
2556 some environments:
2557
2558 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2559
2560 This potential security problem was recorded as CVE-2008-2371.
2561
2562 12. For a pattern where the match had to start at the beginning or immediately
2563 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2564 pcre_dfa_exec() could read past the end of the passed subject if there was
2565 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2566 pcretest so that it places the subject at the end of its malloc-ed buffer.
2567
2568 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2569 exec() might read past the end of the data buffer in UTF-8 mode.
2570
2571 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2572 the data contained the byte 0x85 as part of a UTF-8 character within its
2573 first line. This applied both to normal and DFA matching.
2574
2575 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2576 /^[^d]*?$/8 failed to match "abc".
2577
2578 16. Added a missing copyright notice to pcrecpp_internal.h.
2579
2580 17. Make it more clear in the documentation that values returned from
2581 pcre_exec() in ovector are byte offsets, not character counts.
2582
2583 18. Tidied a few places to stop certain compilers from issuing warnings.
2584
2585 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2586 supplied by Stefan Weber. I made a further small update for 7.8 because
2587 there is a change of source arrangements: the pcre_searchfuncs.c module is
2588 replaced by pcre_ucd.c.
2589
2590
2591 Version 7.7 07-May-08
2592 ---------------------
2593
2594 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2595 a string to a long long, pretend we don't even have a long long." This is
2596 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2597
2598 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2599 pre-7.6 versions, which defined a global no_arg variable instead of putting
2600 it in the RE class. (See also #8 below.)
2601
2602 3. Remove a line of dead code, identified by coverity and reported by Nuno
2603 Lopes.
2604
2605 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2606
2607 (1) The include/exclude patterns were being applied to the whole pathnames
2608 of files, instead of just to the final components.
2609
2610 (2) If there was more than one level of directory, the subdirectories were
2611 skipped unless they satisfied the include/exclude conditions. This is
2612 inconsistent with GNU grep (and could even be seen as contrary to the
2613 pcregrep specification - which I improved to make it absolutely clear).
2614 The action now is always to scan all levels of directory, and just
2615 apply the include/exclude patterns to regular files.
2616
2617 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2618 --exclude_dir in the tests to avoid scanning .svn directories.
2619
2620 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2621 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2622 doesn't support NULs in patterns.
2623
2624 7. Added some missing "const"s to declarations of static tables in
2625 pcre_compile.c and pcre_dfa_exec.c.
2626
2627 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2628 caused by fix #2 above. (Subsequently also a second patch to fix the
2629 first patch. And a third patch - this was a messy problem.)
2630
2631 9. Applied Craig's patch to remove the use of push_back().
2632
2633 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2634 matching function regexec().
2635
2636 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2637 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2638 references. PCRE supports relative numbers with this syntax (I don't think
2639 Oniguruma does).
2640
2641 12. Previously, a group with a zero repeat such as (...){0} was completely
2642 omitted from the compiled regex. However, this means that if the group
2643 was called as a subroutine from elsewhere in the pattern, things went wrong
2644 (an internal error was given). Such groups are now left in the compiled
2645 pattern, with a new opcode that causes them to be skipped at execution
2646 time.
2647
2648 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2649 to the way PCRE behaves:
2650
2651 (a) A lone ] character is dis-allowed (Perl treats it as data).
2652
2653 (b) A back reference to an unmatched subpattern matches an empty string
2654 (Perl fails the current match path).
2655
2656 (c) A data ] in a character class must be notated as \] because if the
2657 first data character in a class is ], it defines an empty class. (In
2658 Perl it is not possible to have an empty class.) The empty class []
2659 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2660 The negative empty class [^] matches any one character, independently
2661 of the DOTALL setting.
2662
2663 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2664 non-existent subpattern following a character class starting with ']' and
2665 containing () gave an internal compiling error instead of "reference to
2666 non-existent subpattern". Fortunately, when the pattern did exist, the
2667 compiled code was correct. (When scanning forwards to check for the
2668 existence of the subpattern, it was treating the data ']' as terminating
2669 the class, so got the count wrong. When actually compiling, the reference
2670 was subsequently set up correctly.)
2671
2672 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2673 it was being rejected as not supported by pcre_dfa_exec(), even though
2674 other assertions are supported. I have made pcre_dfa_exec() support
2675 (*FAIL).
2676
2677 16. The implementation of 13c above involved the invention of a new opcode,
2678 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2679 cannot be changed at match time, I realized I could make a small
2680 improvement to matching performance by compiling OP_ALLANY instead of
2681 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2682 on the OP_ANY path.
2683
2684 17. Compiling pcretest on Windows with readline support failed without the
2685 following two fixes: (1) Make the unistd.h include conditional on
2686 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2687
2688 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2689 ncurses library to be included for pcretest when ReadLine support is
2690 requested, but also to allow for it to be overridden. This patch came from
2691 Daniel Bergström.
2692
2693 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2694 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2695 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2696 spotting this.
2697
2698
2699 Version 7.6 28-Jan-08
2700 ---------------------
2701
2702 1. A character class containing a very large number of characters with
2703 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2704 overflow.
2705
2706 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2707 HAVE_LONG_LONG is not defined.
2708
2709 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2710 bring them up to date and include new features. This patch includes:
2711
2712 - Fixed PH's badly added libz and libbz2 support.
2713 - Fixed a problem with static linking.
2714 - Added pcredemo. [But later removed - see 7 below.]
2715 - Fixed dftables problem and added an option.
2716 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2717 HAVE_LONG_LONG.
2718 - Added readline support for pcretest.
2719 - Added an listing of the option settings after cmake has run.
2720
2721 4. A user submitted a patch to Makefile that makes it easy to create
2722 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2723 Makefile.am that cause it to include this special target, without
2724 affecting anything else. Note that the same mingw target plus all
2725 the other distribution libraries and programs are now supported
2726 when configuring with CMake (see 6 below) instead of with
2727 Configure/Make.
2728
2729 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2730 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2731 exported in the Windows port". It has not yet been confirmed that the patch
2732 solves the problem, but it does no harm.
2733
2734 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2735 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2736 with CMake, and also correct the comment about stack recursion.
2737
2738 7. Remove the automatic building of pcredemo from the ./configure system and
2739 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2740 of a program that users should build themselves after PCRE is installed, so
2741 building it automatically is not really right. What is more, it gave
2742 trouble in some build environments.
2743
2744 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2745
2746
2747 Version 7.5 10-Jan-08
2748 ---------------------
2749
2750 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2751 values in parens when parsing an RE using the C++ wrapper."
2752
2753 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2754 Characters greater than 255 were excluded from the class instead of being
2755 included.
2756
2757 3. The same bug as (2) above applied to negated POSIX classes such as
2758 [:^space:].
2759
2760 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2761 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2762 I have changed it.
2763
2764 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2765 first named subpattern) and a construct such as (?&a) would reference the
2766 first named subpattern whose name started with "a" (in other words, the
2767 length check was missing). Both these problems are fixed. "Subpattern name
2768 expected" is now given for (?&) (a zero-length name), and this patch also
2769 makes it give the same error for \k'' (previously it complained that that
2770 was a reference to a non-existent subpattern).
2771
2772 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2773 this is right because (?- can be followed by option settings as well as by
2774 digits. I have, however, made the messages clearer.
2775
2776 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2777 than the number used in the conditional) now cause a compile-time error.
2778 This is actually not compatible with Perl, which accepts such patterns, but
2779 treats the conditional as always being FALSE (as PCRE used to), but it
2780 seems to me that giving a diagnostic is better.
2781
2782 8. Change "alphameric" to the more common word "alphanumeric" in comments
2783 and messages.
2784
2785 9. Fix two occurrences of "backslash" in comments that should have been
2786 "backspace".
2787
2788 10. Remove two redundant lines of code that can never be obeyed (their function
2789 was moved elsewhere).
2790
2791 11. The program that makes PCRE's Unicode character property table had a bug
2792 which caused it to generate incorrect table entries for sequences of
2793 characters that have the same character type, but are in different scripts.
2794 It amalgamated them into a single range, with the script of the first of
2795 them. In other words, some characters were in the wrong script. There were
2796 thirteen such cases, affecting characters in the following ranges:
2797
2798 U+002b0 - U+002c1
2799 U+0060c - U+0060d
2800 U+0061e - U+00612
2801 U+0064b - U+0065e
2802 U+0074d - U+0076d
2803 U+01800 - U+01805
2804 U+01d00 - U+01d77
2805 U+01d9b - U+01dbf
2806 U+0200b - U+0200f
2807 U+030fc - U+030fe
2808 U+03260 - U+0327f
2809 U+0fb46 - U+0fbb1
2810 U+10450 - U+1049d
2811
2812 12. The -o option (show only the matching part of a line) for pcregrep was not
2813 compatible with GNU grep in that, if there was more than one match in a
2814 line, it showed only the first of them. It now behaves in the same way as
2815 GNU grep.
2816
2817 13. If the -o and -v options were combined for pcregrep, it printed a blank
2818 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2819 does the same. The return code can be used to tell if there were any
2820 non-matching lines.
2821
2822 14. Added --file-offsets and --line-offsets to pcregrep.
2823
2824 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2825 infinitely looping recursion. The bug was that positive lookaheads were not
2826 being skipped when checking for a possible empty match (negative lookaheads
2827 and both kinds of lookbehind were skipped).
2828
2829 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2830 inclusion of <windows.h> to before rather than after the definition of
2831 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2832
2833 17. Specifying a possessive quantifier with a specific limit for a Unicode
2834 character property caused pcre_compile() to compile bad code, which led at
2835 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2836 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2837 caused the error; without that there was no problem.
2838
2839 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2840
2841 19. Added --enable-pcretest-libreadline.
2842
2843 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2844 RE::GlobalReplace(). As a result, the number of replacements returned was
2845 double what it should be. I removed one of the increments, but Craig sent a
2846 later patch that removed the other one (the right fix) and added unit tests
2847 that check the return values (which was not done before).
2848
2849 21. Several CMake things:
2850
2851 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2852 the names libpcre and libpcreposix, not just pcre and pcreposix.
2853
2854 (2) The above change means that pcretest and pcregrep are now correctly
2855 linked with the newly-built libraries, not previously installed ones.
2856
2857 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2858
2859 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2860 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2861 UTF-8 newline character). The key issue is that the pattern starts .*;
2862 this means that the match must be either at the beginning, or after a
2863 newline. The bug was in the code for advancing after a failed match and
2864 checking that the new position followed a newline. It was not taking
2865 account of UTF-8 characters correctly.
2866
2867 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2868 character classes. PCRE was not treating the sequence [:...:] as a
2869 character class unless the ... were all letters. Perl, however, seems to
2870 allow any characters between [: and :], though of course it rejects as
2871 unknown any "names" that contain non-letters, because all the known class
2872 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2873 for example, whereas PCRE did not - it did not recognize a POSIX character
2874 class. This seemed a bit dangerous, so the code has been changed to be
2875 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2876 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2877 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2878 Perl does, and where it didn't before.
2879
2880 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2881 Windows environments %n is disabled by default.
2882
2883
2884 Version 7.4 21-Sep-07
2885 ---------------------
2886
2887 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2888 means that a class such as [\s] counted as "explicit reference to CR or
2889 LF". That isn't really right - the whole point of the change was to try to
2890 help when there was an actual mention of one of the two characters. So now
2891 the change happens only if \r or \n (or a literal CR or LF) character is
2892 encountered.
2893
2894 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2895 of both had grown to the point where there were only 3 bits left.
2896 Fortunately, there was spare space in the data structure, and so I have
2897 moved the internal flags into a new 16-bit field to free up more option
2898 bits.
2899
2900 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2901 but did not set the internal JCHANGED flag - either of these is enough to
2902 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2903 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2904 start sets both bits.
2905
2906 4. Added options (at build time, compile time, exec time) to change \R from
2907 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2908
2909 5. doc/pcresyntax.html was missing from the distribution.
2910
2911 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2912 compatibility, even though it is no longer used.
2913
2914 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2915 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2916 windows.h file is present (where different names are used). [This was
2917 reversed later after testing - see 16 below.]
2918
2919 8. Changed all #include <config.h> to #include "config.h". There were also
2920 some further <pcre.h> cases that I changed to "pcre.h".
2921
2922 9. When pcregrep was used with the --colour option, it missed the line ending
2923 sequence off the lines that it output.
2924
2925 10. It was pointed out to me that arrays of string pointers cause lots of
2926 relocations when a shared library is dynamically loaded. A technique of
2927 using a single long string with a table of offsets can drastically reduce
2928 these. I have refactored PCRE in four places to do this. The result is
2929 dramatic:
2930
2931 Originally: 290
2932 After changing UCP table: 187
2933 After changing error message table: 43
2934 After changing table of "verbs" 36
2935 After changing table of Posix names 22
2936
2937 Thanks to the folks working on Gregex for glib for this insight.
2938
2939 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2940 unicode-properties was also set.
2941
2942 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2943
2944 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2945 checked only for CRLF.
2946
2947 14. Added casts to pcretest.c to avoid compiler warnings.
2948
2949 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2950
2951 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2952 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2953 entirely. This removes changes made in 7 above.
2954
2955 17. The CMake files have been updated, and there is now more information about
2956 building with CMake in the NON-UNIX-USE document.
2957
2958
2959 Version 7.3 28-Aug-07
2960 ---------------------
2961
2962 1. In the rejigging of the build system that eventually resulted in 7.1, the
2963 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2964 brackets there is not right, since it causes compilers to look for an
2965 installed pcre.h, not the version that is in the source that is being
2966 compiled (which of course may be different). I have changed it back to:
2967
2968 #include "pcre.h"
2969
2970 I have a vague recollection that the change was concerned with compiling in
2971 different directories, but in the new build system, that is taken care of
2972 by the VPATH setting the Makefile.
2973
2974 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2975 when the subject happened to end in the byte 0x85 (e.g. if the last
2976 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2977 characters but of course it shouldn't be taken as a newline when it is part
2978 of another character. The bug was that, for an unlimited repeat of . in
2979 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2980 characters when looking for a newline.
2981
2982 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2983
2984 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2985 in debug output.
2986
2987 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2988 long printing in the pcrecpp unittest when running under MinGW.
2989
2990 6. ESC_K was left out of the EBCDIC table.
2991
2992 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2993 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2994 limit also applies to "virtual nesting" when a pattern is recursive, and in
2995 this case 1000 isn't so big. I have been able to remove this limit at the
2996 expense of backing off one optimization in certain circumstances. Normally,
2997 when pcre_exec() would call its internal match() function recursively and
2998 immediately return the result unconditionally, it uses a "tail recursion"
2999 feature to save stack. However, when a subpattern that can match an empty
3000 string has an unlimited repetition quantifier, it no longer makes this
3001 optimization. That gives it a stack frame in which to save the data for
3002 checking that an empty string has been matched. Previously this was taken
3003 from the 1000-entry workspace that had been reserved. So now there is no
3004 explicit limit, but more stack is used.
3005
3006 8. Applied Daniel's patches to solve problems with the import/export magic
3007 syntax that is required for Windows, and which was going wrong for the
3008 pcreposix and pcrecpp parts of the library. These were overlooked when this
3009 problem was solved for the main library.
3010
3011 9. There were some crude static tests to avoid integer overflow when computing
3012 the size of patterns that contain repeated groups with explicit upper
3013 limits. As the maximum quantifier is 65535, the maximum group length was
3014 set at 30,000 so that the product of these two numbers did not overflow a
3015 32-bit integer. However, it turns out that people want to use groups that
3016 are longer than 30,000 bytes (though not repeat them that many times).
3017 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3018 made it possible to implement the integer overflow checks in a much more
3019 dynamic way, which I have now done. The artificial limitation on group
3020 length has been removed - we now have only the limit on the total length of
3021 the compiled pattern, which depends on the LINK_SIZE setting.
3022
3023 10. Fixed a bug in the documentation for get/copy named substring when
3024 duplicate names are permitted. If none of the named substrings are set, the
3025 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3026 empty string.
3027
3028 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3029 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3030 because the ] is interpreted as the first data character and the
3031 terminating ] is not found. PCRE has been made compatible with Perl in this
3032 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3033 cause memory overwriting.
3034
3035 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3036 string has been matched (to stop an infinite loop). It was not recognizing
3037 a conditional subpattern that could match an empty string if that
3038 subpattern was within another subpattern. For example, it looped when
3039 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3040 condition was not nested. This bug has been fixed.
3041
3042 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3043 past the start of the subject in the presence of bytes with the top bit
3044 set, for example "\x8aBCD".
3045
3046 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3047 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3048
3049 14. Optimized (?!) to (*FAIL).
3050
3051 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3052 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3053 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3054 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3055 does: it's just the validity check that is more restrictive.
3056
3057 16. Inserted checks for integer overflows during escape sequence (backslash)
3058 processing, and also fixed erroneous offset values for syntax errors during
3059 backslash processing.
3060
3061 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3062 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3063
3064 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3065 caused an overrun.
3066
3067 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3068 something other than just ASCII characters) inside a group that had an
3069 unlimited repeat caused a loop at compile time (while checking to see
3070 whether the group could match an empty string).
3071
3072 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3073 [\P{Any}] did so. (Error in the code for printing property names.)
3074
3075 21. An orphan \E inside a character class could cause a crash.
3076
3077 22. A repeated capturing bracket such as (A)? could cause a wild memory
3078 reference during compilation.
3079
3080 23. There are several functions in pcre_compile() that scan along a compiled
3081 expression for various reasons (e.g. to see if it's fixed length for look
3082 behind). There were bugs in these functions when a repeated \p or \P was
3083 present in the pattern. These operators have additional parameters compared
3084 with \d, etc, and these were not being taken into account when moving along
3085 the compiled data. Specifically:
3086
3087 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3088 length.
3089
3090 (b) An item such as \pL+ within a repeated group could cause crashes or
3091 loops.
3092
3093 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3094 "reference to non-existent subpattern" error.
3095
3096 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3097
3098 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3099 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3100
3101 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3102
3103 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3104 character were causing crashes (broken optimization).
3105
3106 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3107 \p or \P) caused a compile-time loop.
3108
3109 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3110 break. For example, the unstudied pattern [\r\n]A does not match the string
3111 "\r\nA" because change 7.0/46 below moves the current point on by two
3112 characters after failing to match at the start. However, the pattern \nA
3113 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3114 the same is true. There doesn't seem any very clean way out of this, but
3115 what I have chosen to do makes the common cases work: PCRE now takes note
3116 of whether there can be an explicit match for \r or \n anywhere in the
3117 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3118 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3119 pattern has explicit CR or LF references.
3120
3121 29. Added (*CR) etc for changing newline setting at start of pattern.
3122
3123
3124 Version 7.2 19-Jun-07
3125 ---------------------
3126
3127 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3128 which is apparently normally available under Windows.
3129
3130 2. Re-jig the pcregrep tests with different newline settings in an attempt
3131 to make them independent of the local environment's newline setting.
3132
3133 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3134
3135 4. Some of the "internals" tests were previously cut out when the link size
3136 was not 2, because the output contained actual offsets. The recent new
3137 "Z" feature of pcretest means that these can be cut out, making the tests
3138 usable with all link sizes.
3139
3140 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3141 stack recursion. This gives a massive performance boost under BSD, but just
3142 a small improvement under Linux. However, it saves one field in the frame
3143 in all cases.
3144
3145 6. Added more features from the forthcoming Perl 5.10:
3146
3147 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3148 recursion call. It refers to the nth most recently opened parentheses.
3149
3150 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3151 to be opened parentheses.
3152
3153 (c) Conditions that refer to capturing parentheses can be specified
3154 relatively, for example, (?(-2)... or (?(+3)...
3155
3156 (d) \K resets the start of the current match so that everything before
3157 is not part of it.
3158
3159 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3160
3161 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3162 reference syntax.
3163
3164 (g) (?| introduces a group in which the numbering of parentheses in each
3165 alternative starts with the same number.
3166
3167 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3168
3169 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3170 PCRE_INFO_JCHANGED.
3171
3172 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3173 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3174 for detecting groups that can match an empty string.
3175
3176 9. A pattern with a very large number of alternatives (more than several
3177 hundred) was running out of internal workspace during the pre-compile
3178 phase, where pcre_compile() figures out how much memory will be needed. A
3179 bit of new cunning has reduced the workspace needed for groups with
3180 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3181 workspace instead of running out of the 4096 that are available.
3182
3183 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3184
3185 11. Applied patch from Google to remove an optimization that didn't quite work.
3186 The report of the bug said:
3187
3188 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3189 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3190 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3191
3192 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3193 it matched the wrong number of bytes.
3194
3195
3196 Version 7.1 24-Apr-07
3197 ---------------------
3198
3199 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3200 that is more "standard", making use of automake and other Autotools. There
3201 is some re-arrangement of the files and adjustment of comments consequent
3202 on this.
3203
3204 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3205 for recursive directory scanning broke on some systems because the files
3206 are not scanned in any specific order and on different systems the order
3207 was different. A call to "sort" has been inserted into RunGrepTest for the
3208 approprate test as a short-term fix. In the longer term there may be an
3209 alternative.
3210
3211 3. I had an email from Eric Raymond about problems translating some of PCRE's
3212 man pages to HTML (despite the fact that I distribute HTML pages, some
3213 people do their own conversions for various reasons). The problems
3214 concerned the use of low-level troff macros .br and .in. I have therefore
3215 removed all such uses from the man pages (some were redundant, some could
3216 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3217 HTML has been updated to handle .nf/.fi and to complain if it encounters
3218 .br or .in.
3219
3220 4. Updated comments in configure.ac that get placed in config.h.in and also
3221 arranged for config.h to be included in the distribution, with the name
3222 config.h.generic, for the benefit of those who have to compile without
3223 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3224
3225 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3226 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3227 makevp.bat for the current PCRE, using the additional files
3228 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3229
3230 6. A Windows user reported a minor discrepancy with test 2, which turned out
3231 to be caused by a trailing space on an input line that had got lost in his
3232 copy. The trailing space was an accident, so I've just removed it.
3233
3234 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3235 that is needed.
3236
3237 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3238 as "const" (a) because they are and (b) because it helps the PHP
3239 maintainers who have recently made a script to detect big data structures
3240 in the php code that should be moved to the .rodata section. I remembered
3241 to update Builducptable as well, so it won't revert if ucptable.h is ever
3242 re-created.
3243
3244 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3245 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3246 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3247 support is not required. This saves 1.5-2K of code, which is important in
3248 some applications.
3249
3250 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3251 so as not to refer to the tables, even though these functions will never be
3252 called when UTF-8 support is disabled. Otherwise there are problems with a
3253 shared library.
3254
3255 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3256
3257 (a) It was defining its arguments as char * instead of void *.
3258
3259 (b) It was assuming that all moves were upwards in memory; this was true
3260 a long time ago when I wrote it, but is no longer the case.
3261
3262 The emulated memove() is provided for those environments that have neither
3263 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3264 is clearly not the case, as these two bugs were recently reported.
3265
3266 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3267 and Detrail to create the HTML documentation, the .txt form of the man
3268 pages, and it removes trailing spaces from listed files. It also creates
3269 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3270 case, it wraps all the #defines with #ifndefs. This script should be run
3271 before "make dist".
3272
3273 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3274 with Unicode property support.
3275
3276 (a) For a maximizing quantifier, if the two different cases of the
3277 character were of different lengths in their UTF-8 codings (there are
3278 some cases like this - I found 11), and the matching function had to
3279 back up over a mixture of the two cases, it incorrectly assumed they
3280 were both the same length.
3281
3282 (b) When PCRE was configured to use the heap rather than the stack for
3283 recursion during matching, it was not correctly preserving the data for
3284 the other case of a UTF-8 character when checking ahead for a match
3285 while processing a minimizing repeat. If the check also involved
3286 matching a wide character, but failed, corruption could cause an
3287 erroneous result when trying to check for a repeat of the original
3288 character.
3289
3290 13. Some tidying changes to the testing mechanism:
3291
3292 (a) The RunTest script now detects the internal link size and whether there
3293 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3294 values substituted by "configure". (The RunGrepTest script already did
3295 this for UTF-8.) The configure.ac script no longer substitutes the
3296 relevant variables.
3297
3298 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3299 with length and offset values. This means that the output is different
3300 for different internal link sizes. Test 2 is skipped for link sizes
3301 other than 2 because of this, bypassing the problem. Unfortunately,
3302 there was also a test in test 3 (the locale tests) that used /B and
3303 failed for link sizes other than 2. Rather than cut the whole test out,
3304 I have added a new /Z option to pcretest that replaces the length and
3305 offset values with spaces. This is now used to make test 3 independent
3306 of link size. (Test 2 will be tidied up later.)
3307
3308 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3309 segmentation fault instead of returning the appropriate error message.
3310
3311 15. In multiline mode when the newline sequence was set to "any", the pattern
3312 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3313 This doesn't seem right; it now treats the CRLF combination as the line
3314 ending, and so does not match in that case. It's only a pattern such as ^$
3315 that would hit this one: something like ^ABC$ would have failed after \r
3316 and then tried again after \r\n.
3317
3318 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3319 in an attempt to make files that differ only in their line terminators
3320 compare equal. This works on Linux.
3321
3322 17. Under certain error circumstances pcregrep might try to free random memory
3323 as it exited. This is now fixed, thanks to valgrind.
3324
3325 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3326 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3327 was because its rules for how to advance for /g after matching an empty
3328 string at the end of a line did not allow for this case. They now check for
3329 it specially.
3330
3331 20. pcretest is supposed to handle patterns and data of any length, by
3332 extending its buffers when necessary. It was getting this wrong when the
3333 buffer for a data line had to be extended.
3334
3335 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3336 CRLF as a newline sequence.
3337
3338 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3339 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3340 I have nevertheless tidied it up.
3341
3342 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3343
3344 24. Added a man page for pcre-config.
3345
3346
3347 Version 7.0 19-Dec-06
3348 ---------------------
3349
3350 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3351 moving to gcc 4.1.1.
3352
3353 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3354 sys/time.h, which is documented as needed for this function. It doesn't
3355 seem to matter on Linux, but it showed up on some releases of OS X.
3356
3357 3. It seems that there are systems where bytes whose values are greater than
3358 127 match isprint() in the "C" locale. The "C" locale should be the
3359 default when a C program starts up. In most systems, only ASCII printing
3360 characters match isprint(). This difference caused the output from pcretest
3361 to vary, making some of the tests fail. I have changed pcretest so that:
3362
3363 (a) When it is outputting text in the compiled version of a pattern, bytes
3364 other than 32-126 are always shown as hex escapes.
3365
3366 (b) When it is outputting text that is a matched part of a subject string,
3367 it does the same, unless a different locale has been set for the match
3368 (using the /L modifier). In this case, it uses isprint() to decide.
3369
3370 4. Fixed a major bug that caused incorrect computation of the amount of memory
3371 required for a compiled pattern when options that changed within the
3372 pattern affected the logic of the preliminary scan that determines the
3373 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3374 that the computed length was too small. The symptoms of this bug were
3375 either the PCRE error "internal error: code overflow" from pcre_compile(),
3376 or a glibc crash with a message such as "pcretest: free(): invalid next
3377 size (fast)". Examples of patterns that provoked this bug (shown in
3378 pcretest format) are:
3379
3380 /(?-x: )/x
3381 /(?x)(?-x: \s*#\s*)/
3382 /((?i)[\x{c0}])/8
3383 /(?i:[\x{c0}])/8
3384
3385 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3386 is now done differently.
3387
3388 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3389 wrapper classes; (b) implement a new function in the C++ scanner that is
3390 more efficient than the old way of doing things because it avoids levels of
3391 recursion in the regex matching; (c) add a paragraph to the documentation
3392 for the FullMatch() function.
3393
3394 6. The escape sequence \n was being treated as whatever was defined as
3395 "newline". Not only was this contrary to the documentation, which states
3396 that \n is character 10 (hex 0A), but it also went horribly wrong when
3397 "newline" was defined as CRLF. This has been fixed.
3398
3399 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3400 was being set to -1 for the "end of line" case (supposedly a value that no
3401 character can have). Though this value is never used (the check for end of
3402 line is "zero bytes in current character"), it caused compiler complaints.
3403 I've changed it to 0xffffffff.
3404
3405 8. In pcre_version.c, the version string was being built by a sequence of
3406 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3407 string (as it is for production releases) called a macro with an empty
3408 argument. The C standard says the result of this is undefined. The gcc
3409 compiler treats it as an empty string (which was what was wanted) but it is
3410 reported that Visual C gives an error. The source has been hacked around to
3411 avoid this problem.
3412
3413 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3414 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3415 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3416 of them did).
3417
3418 10. Originally, pcretest opened its input and output without "b"; then I was
3419 told that "b" was needed in some environments, so it was added for release
3420 5.0 to both the input and output. (It makes no difference on Unix-like
3421 systems.) Later I was told that it is wrong for the input on Windows. I've
3422 now abstracted the modes into two macros, to make it easier to fiddle with
3423 them, and removed "b" from the input mode under Windows.
3424
3425 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3426
3427 12. Added -help and --help to pcretest as an official way of being reminded
3428 of the options.
3429
3430 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3431 and pcrecpp.cc because they annoy compilers at high warning levels.
3432
3433 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3434
3435 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3436 scripts are not C programs :-) and which was not noticed because it works
3437 on Linux.
3438
3439 16. pcretest is supposed to handle any length of pattern and data line (as one
3440 line or as a continued sequence of lines) by extending its input buffer if
3441 necessary. This feature was broken for very long pattern lines, leading to
3442 a string of junk being passed to pcre_compile() if the pattern was longer
3443 than about 50K.
3444
3445 17. I have done a major re-factoring of the way pcre_compile() computes the
3446 amount of memory needed for a compiled pattern. Previously, there was code
3447 that made a preliminary scan of the pattern in order to do this. That was
3448 OK when PCRE was new, but as the facilities have expanded, it has become
3449 harder and harder to keep it in step with the real compile phase, and there
3450 have been a number of bugs (see for example, 4 above). I have now found a
3451 cunning way of running the real compile function in a "fake" mode that
3452 enables it to compute how much memory it would need, while actually only
3453 ever using a few hundred bytes of working memory and without too many
3454 tests of the mode. This should make future maintenance and development
3455 easier. A side effect of this work is that the limit of 200 on the nesting
3456 depth of parentheses has been removed (though this was never a serious
3457 limitation, I suspect). However, there is a downside: pcre_compile() now
3458 runs more slowly than before (30% or more, depending on the pattern). I
3459 hope this isn't a big issue. There is no effect on runtime performance.
3460
3461 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3462 newline (only possible for the last line of a file) and it was a
3463 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3464
3465 19. Added additional timing features to pcretest. (1) The -tm option now times
3466 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3467 separate command line item, by a number that specifies the number of
3468 repeats to use when timing. The default is 50000; this gives better
3469 precision, but takes uncomfortably long for very large patterns.
3470
3471 20. Extended pcre_study() to be more clever in cases where a branch of a
3472 subpattern has no definite first character. For example, (a*|b*)[cd] would
3473 previously give no result from pcre_study(). Now it recognizes that the
3474 first character must be a, b, c, or d.
3475
3476 21. There was an incorrect error "recursive call could loop indefinitely" if
3477 a subpattern (or the entire pattern) that was being tested for matching an
3478 empty string contained only one non-empty item after a nested subpattern.
3479 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3480 incorrectly, because the \d was being skipped in the check.
3481
3482 22. The pcretest program now has a new pattern option /B and a command line
3483 option -b, which is equivalent to adding /B to every pattern. This causes
3484 it to show the compiled bytecode, without the additional information that
3485 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3486 is the same as /B/I).
3487
3488 23. A new optimization is now able automatically to treat some sequences such
3489 as a*b as a*+b. More specifically, if something simple (such as a character
3490 or a simple class like \d) has an unlimited quantifier, and is followed by
3491 something that cannot possibly match the quantified thing, the quantifier
3492 is automatically "possessified".
3493
3494 24. A recursive reference to a subpattern whose number was greater than 39
3495 went wrong under certain circumstances in UTF-8 mode. This bug could also
3496 have affected the operation of pcre_study().
3497
3498 25. Realized that a little bit of performance could be had by replacing
3499 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3500
3501 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3502
3503 27. Possessive quantifiers such as a++ were previously implemented by turning
3504 them into atomic groups such as ($>a+). Now they have their own opcodes,
3505 which improves performance. This includes the automatically created ones
3506 from 23 above.
3507
3508 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3509 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3510 the first matched character to be a colon. This applied both to named and
3511 numbered groups.
3512
3513 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3514
3515 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3516 building PCRE on Windows easier, so I have included it in the distribution.
3517
3518 31. There is now a check in pcretest against a ridiculously large number being
3519 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3520 loop, the loop is abandoned.
3521
3522 32. Forward references to subpatterns in conditions such as (?(2)...) where
3523 subpattern 2 is defined later cause pcre_compile() to search forwards in
3524 the pattern for the relevant set of parentheses. This search went wrong
3525 when there were unescaped parentheses in a character class, parentheses
3526 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3527
3528 33. "Subroutine" calls and backreferences were previously restricted to
3529 referencing subpatterns earlier in the regex. This restriction has now
3530 been removed.
3531
3532 34. Added a number of extra features that are going to be in Perl 5.10. On the
3533 whole, these are just syntactic alternatives for features that PCRE had
3534 previously implemented using the Python syntax or my own invention. The
3535 other formats are all retained for compatibility.
3536
3537 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3538 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3539 also .NET compatible.
3540
3541 (b) A recursion or subroutine call to a named group can now be defined as
3542 (?&name) as well as (?P>name).
3543
3544 (c) A backreference to a named group can now be defined as \k<name> or
3545 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3546 5.10, are also .NET compatible.
3547
3548 (d) A conditional reference to a named group can now use the syntax
3549 (?(<name>) or (?('name') as well as (?(name).
3550
3551 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3552 groups (named and numbered) that are never evaluated inline, but can be
3553 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3554 is always false. There may be only one alternative in such a group.
3555
3556 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3557 as the simple (?(R). The condition is true only if the most recent
3558 recursion is that of the given number or name. It does not search out
3559 through the entire recursion stack.
3560
3561 (g) The escape \gN or \g{N} has been added, where N is a positive or
3562 negative number, specifying an absolute or relative reference.
3563
3564 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3565 some "unreachable code" warnings.
3566
3567 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3568 things, this adds five new scripts.
3569
3570 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3571 There were also incompatibilities regarding the handling of \Q..\E inside
3572 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3573 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3574
3575 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3576 matches an empty string, and forcibly breaks the loop. There were bugs in
3577 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3578 against aaaa the result was just "a" rather than "aaaa", for example. Two
3579 separate and independent bugs (that affected different cases) have been
3580 fixed.
3581
3582 39. Refactored the code to abolish the use of different opcodes for small
3583 capturing bracket numbers. This is a tidy that I avoided doing when I
3584 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3585 The new approach is not only tidier, it makes it possible to reduce the
3586 memory needed to fix the previous bug (38).
3587
3588 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3589 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3590 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3591 mode.
3592
3593 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3594 report.
3595
3596 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3597 copy construction and assignment in the C++ wrapper.
3598
3599 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3600 couple of bugs that could have given wrong results in the "--newline=crlf"
3601 case.
3602
3603 44. Added a number of casts and did some reorganization of signed/unsigned int
3604 variables following suggestions from Dair Grant. Also renamed the variable
3605 "this" as "item" because it is a C++ keyword.
3606
3607 45. Arranged for dftables to add
3608
3609 #include "pcre_internal.h"
3610
3611 to pcre_chartables.c because without it, gcc 4.x may remove the array
3612 definition from the final binary if PCRE is built into a static library and
3613 dead code stripping is activated.
3614
3615 46. For an unanchored pattern, if a match attempt fails at the start of a
3616 newline sequence, and the newline setting is CRLF or ANY, and the next two
3617 characters are CRLF, advance by two characters instead of one.
3618
3619
3620 Version 6.7 04-Jul-06
3621 ---------------------
3622
3623 1. In order to handle tests when input lines are enormously long, pcretest has
3624 been re-factored so that it automatically extends its buffers when
3625 necessary. The code is crude, but this _is_ just a test program. The
3626 default size has been increased from 32K to 50K.
3627
3628 2. The code in pcre_study() was using the value of the re argument before
3629 testing it for NULL. (Of course, in any sensible call of the function, it
3630 won't be NULL.)
3631
3632 3. The memmove() emulation function in pcre_internal.h, which is used on
3633 systems that lack both memmove() and bcopy() - that is, hardly ever -
3634 was missing a "static" storage class specifier.
3635
3636 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3637 containing an extended class (one that cannot be represented by a bitmap
3638 because it contains high-valued characters or Unicode property items, e.g.
3639 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3640 pattern, but PCRE should not loop if you do not (it no longer does).
3641 [Detail: two cases were found: (a) a repeated subpattern containing an
3642 extended class; (b) a recursive reference to a subpattern that followed a
3643 previous extended class. It wasn't skipping over the extended class
3644 correctly when UTF-8 mode was not set.]
3645
3646 5. A negated single-character class was not being recognized as fixed-length
3647 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3648 compile error "lookbehind assertion is not fixed length".
3649
3650 6. The RunPerlTest auxiliary script was showing an unexpected difference
3651 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3652 write a Perl script that can interpret lines of an input file either as
3653 byte characters or as UTF-8, which is what "perltest" was being required to
3654 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3655 can't do is switch easily at run time between having the "use utf8;" pragma
3656 or not. In the end, I fudged it by using the RunPerlTest script to insert
3657 "use utf8;" explicitly for the UTF-8 tests.
3658
3659 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3660 the end of the subject string, contrary to the documentation and to what
3661 Perl does. This was true of both matching functions. Now it matches only at
3662 the start of the subject and immediately after *internal* newlines.
3663
3664 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3665 a pointer to an int instead of a pointer to an unsigned long int. This
3666 caused problems on 64-bit systems.
3667
3668 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3669 instance of the 'standard' template library not being so standard".
3670
3671 10. There was no check on the number of named subpatterns nor the maximum
3672 length of a subpattern name. The product of these values is used to compute
3673 the size of the memory block for a compiled pattern. By supplying a very
3674 long subpattern name and a large number of named subpatterns, the size
3675 computation could be caused to overflow. This is now prevented by limiting
3676 the length of names to 32 characters, and the number of named subpatterns
3677 to 10,000.
3678
3679 11. Subpatterns that are repeated with specific counts have to be replicated in
3680 the compiled pattern. The size of memory for this was computed from the
3681 length of the subpattern and the repeat count. The latter is limited to
3682 65535, but there was no limit on the former, meaning that integer overflow
3683 could in principle occur. The compiled length of a repeated subpattern is
3684 now limited to 30,000 bytes in order to prevent this.
3685
3686 12. Added the optional facility to have named substrings with the same name.
3687
3688 13. Added the ability to use a named substring as a condition, using the
3689 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3690 are numbers (not recommended). Forward references are permitted.
3691
3692 14. Added forward references in named backreferences (if you see what I mean).
3693
3694 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3695 pattern could run off the end of the subject. For example, the pattern
3696 "(?s)(.{1,5})"8 did this with the subject "ab".
3697
3698 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3699 PCRE_CASELESS was set when matching characters that were quantified with ?
3700 or *.
3701
3702 17. A character class other than a single negated character that had a minimum
3703 but no maximum quantifier - for example [ab]{6,} - was not handled
3704 correctly by pce_dfa_exec(). It would match only one character.
3705
3706 18. A valid (though odd) pattern that looked like a POSIX character
3707 class but used an invalid character after [ (for example [[,abc,]]) caused
3708 pcre_compile() to give the error "Failed: internal error: code overflow" or
3709 in some cases to crash with a glibc free() error. This could even happen if
3710 the pattern terminated after [[ but there just happened to be a sequence of
3711 letters, a binary zero, and a closing ] in the memory that followed.
3712
3713 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3714 over the years. Originally (before any Unicode support), just the bottom 8
3715 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3716 output from "man perlunicode" includes this:
3717
3718 The regular expression compiler produces polymorphic opcodes. That
3719 is, the pattern adapts to the data and automatically switches to
3720 the Unicode character scheme when presented with Unicode data--or
3721 instead uses a traditional byte scheme when presented with byte
3722 data.
3723
3724 Sadly, a wide octal escape does not cause a switch, and in a string with
3725 no other multibyte characters, these octal escapes are treated as before.
3726 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3727 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3728 Unicode string.
3729
3730 I have not perpetrated such confusion in PCRE. Up till now, it took just
3731 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3732 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3733 translate to the appropriate multibyte character.
3734
3735 29. Applied some refactoring to reduce the number of warnings from Microsoft
3736 and Borland compilers. This has included removing the fudge introduced
3737 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3738 a warning about an unused variable.
3739
3740 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3741 characters since release 4.0, because Perl (from release 5.004) does not.
3742 [Or at least, is documented not to: some releases seem to be in conflict
3743 with the documentation.] However, when a pattern was studied with
3744 pcre_study() and all its branches started with \s, PCRE still included VT
3745 as a possible starting character. Of course, this did no harm; it just
3746 caused an unnecessary match attempt.
3747
3748 22. Removed a now-redundant internal flag bit that recorded the fact that case
3749 dependency changed within the pattern. This was once needed for "required
3750 byte" processing, but is no longer used. This recovers a now-scarce options
3751 bit. Also moved the least significant internal flag bit to the most-
3752 significant bit of the word, which was not previously used (hangover from
3753 the days when it was an int rather than a uint) to free up another bit for
3754 the future.
3755
3756 23. Added support for CRLF line endings as well as CR and LF. As well as the
3757 default being selectable at build time, it can now be changed at runtime
3758 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3759 specify that it is scanning data with non-default line endings.
3760
3761 24. Changed the definition of CXXLINK to make it agree with the definition of
3762 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3763
3764 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3765 recursions. This makes a big different to stack usage for some patterns.
3766
3767 26. If a subpattern containing a named recursion or subroutine reference such
3768 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3769 the space required for the compiled pattern went wrong and gave too small a
3770 value. Depending on the environment, this could lead to "Failed: internal
3771 error: code overflow at offset 49" or "glibc detected double free or
3772 corruption" errors.
3773
3774 27. Applied patches from Google (a) to support the new newline modes and (b) to
3775 advance over multibyte UTF-8 characters in GlobalReplace.
3776
3777 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3778 difference for some implementation of PCRE in some Windows version.
3779
3780 29. Added some extra testing facilities to pcretest:
3781
3782 \q<number> in a data line sets the "match limit" value
3783 \Q<number> in a data line sets the "match recursion limt" value
3784 -S <number> sets the stack size, where <number> is in megabytes
3785
3786 The -S option isn't available for Windows.
3787
3788
3789 Version 6.6 06-Feb-06
3790 ---------------------
3791
3792 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3793 in pcreposix.h. I have copied the definition from pcre.h.
3794
3795 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3796 because pcre.h is no longer a built file.
3797
3798 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3799 not normally included in the compiled code.
3800
3801
3802 Version 6.5 01-Feb-06
3803 ---------------------
3804
3805 1. When using the partial match feature with pcre_dfa_exec(), it was not
3806 anchoring the second and subsequent partial matches at the new starting
3807 point. This could lead to incorrect results. For example, with the pattern
3808 /1234/, partially matching against "123" and then "a4" gave a match.
3809
3810 2. Changes to pcregrep:
3811
3812 (a) All non-match returns from pcre_exec() were being treated as failures
3813 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3814 error message is output. Some extra information is given for the
3815 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3816 probably the only errors that are likely to be caused by users (by
3817 specifying a regex that has nested indefinite repeats, for instance).
3818 If there are more than 20 of these errors, pcregrep is abandoned.
3819
3820 (b) A binary zero was treated as data while matching, but terminated the
3821 output line if it was written out. This has been fixed: binary zeroes
3822 are now no different to any other data bytes.
3823
3824 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3825 used to set a locale for matching. The --locale=xxxx long option has
3826 been added (no short equivalent) to specify a locale explicitly on the
3827 pcregrep command, overriding the environment variables.
3828
3829 (d) When -B was used with -n, some line numbers in the output were one less
3830 than they should have been.
3831
3832 (e) Added the -o (--only-matching) option.
3833
3834 (f) If -A or -C was used with -c (count only), some lines of context were
3835 accidentally printed for the final match.
3836
3837 (g) Added the -H (--with-filename) option.
3838
3839 (h) The combination of options -rh failed to suppress file names for files
3840 that were found from directory arguments.
3841
3842 (i) Added the -D (--devices) and -d (--directories) options.
3843
3844 (j) Added the -F (--fixed-strings) option.
3845
3846 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3847
3848 (l) Added the --colo(u)r option.
3849
3850 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3851 is not present by default.
3852
3853 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3854 items such as (?R) or (?1), when the recursion could match a number of
3855 alternatives. If it matched one of the alternatives, but subsequently,
3856 outside the recursion, there was a failure, the code tried to back up into
3857 the recursion. However, because of the way PCRE is implemented, this is not
3858 possible, and the result was an incorrect result from the match.
3859
3860 In order to prevent this happening, the specification of recursion has
3861 been changed so that all such subpatterns are automatically treated as
3862 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3863
3864 4. I had overlooked the fact that, in some locales, there are characters for
3865 which isalpha() is true but neither isupper() nor islower() are true. In
3866 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3867 and ordfeminine) are like this. This affected the treatment of \w and \W
3868 when they appeared in character classes, but not when they appeared outside
3869 a character class. The bit map for "word" characters is now created
3870 separately from the results of isalnum() instead of just taking it from the
3871 upper, lower, and digit maps. (Plus the underscore character, of course.)
3872
3873 5. The above bug also affected the handling of POSIX character classes such as
3874 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3875 permanent tables. Instead, the bit maps for such a class were previously
3876 created as the appropriate unions of the upper, lower, and digit bitmaps.
3877 Now they are created by subtraction from the [[:word:]] class, which has
3878 its own bitmap.
3879
3880 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3881 It is created by subtracting the vertical space characters (\x09, \x0a,
3882 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3883 subtraction was done in the overall bitmap for a character class, meaning
3884 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3885 be recognized. This bug has been fixed.
3886
3887 7. Patches from the folks at Google:
3888
3889 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3890 real life, but is still worth protecting against".
3891
3892 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3893 regular expressions".
3894
3895 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3896 have it.
3897
3898 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3899 "configure" and the latter not, in order to fix a problem somebody had
3900 with compiling the Arg class on HP-UX.
3901
3902 (e) Improve the error-handling of the C++ wrapper a little bit.
3903
3904 (f) New tests for checking recursion limiting.
3905
3906 8. The pcre_memmove() function, which is used only if the environment does not
3907 have a standard memmove() function (and is therefore rarely compiled),
3908 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3909 returning a result (though PCRE never actually uses the result).
3910
3911 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3912 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3913 returned instead of calling malloc() with an overflowing number that would
3914 most likely cause subsequent chaos.
3915
3916 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3917
3918 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3919 with this option is matched, the nmatch and pmatch options of regexec() are
3920 ignored.
3921
3922 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3923 provided in case anyone wants to the the POSIX interface with UTF-8
3924 strings.
3925
3926 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3927 C++ linking (needed for some HP-UX environments).
3928
3929 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3930 (unused parameter) and in the pcre_printint() function (omitted "default"
3931 switch label when the default is to do nothing).
3932
3933 15. Added some code to make it possible, when PCRE is compiled as a C++
3934 library, to replace subject pointers for pcre_exec() with a smart pointer
3935 class, thus making it possible to process discontinuous strings.
3936
3937 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3938 much the same function. They were added by different people who were trying
3939 to make PCRE easy to compile on non-Unix systems. It has been suggested
3940 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3941 for compiling on Windows systems. I have therefore replaced it with
3942 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3943 defaults to "extern" for C or "extern C" for C++, which works fine on
3944 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3945 SCOPE with something explicit in config.h. In addition:
3946
3947 (a) pcreposix.h still had just "extern" instead of either of these macros;
3948 I have replaced it with PCRE_DATA_SCOPE.
3949
3950 (b) Functions such as _pcre_xclass(), which are internal to the library,
3951 but external in the C sense, all had PCRE_EXPORT in their definitions.
3952 This is apparently wrong for the Windows case, so I have removed it.
3953 (It makes no difference on Unix-like systems.)
3954
3955 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3956 of recursive calls to match(). This is different to MATCH_LIMIT because
3957 that limits the total number of calls to match(), not all of which increase
3958 the depth of recursion. Limiting the recursion depth limits the amount of
3959 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3960 when PCRE is compiled, and changed at run time. A patch from Google adds
3961 this functionality to the C++ interface.
3962
3963 18. Changes to the handling of Unicode character properties:
3964
3965 (a) Updated the table to Unicode 4.1.0.
3966
3967 (b) Recognize characters that are not in the table as "Cn" (undefined).
3968
3969 (c) I revised the way the table is implemented to a much improved format
3970 which includes recognition of ranges. It now supports the ranges that
3971 are defined in UnicodeData.txt, and it also amalgamates other
3972 characters into ranges. This has reduced the number of entries in the
3973 table from around 16,000 to around 3,000, thus reducing its size
3974 considerably. I realized I did not need to use a tree structure after
3975 all - a binary chop search is just as efficient. Having reduced the
3976 number of entries, I extended their size from 6 bytes to 8 bytes to
3977 allow for more data.
3978
3979 (d) Added support for Unicode script names via properties such as \p{Han}.
3980
3981 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3982 matching that character.
3983
3984 20. When matching a repeated Unicode property with a minimum greater than zero,
3985 (for example \pL{2,}), PCRE could look past the end of the subject if it
3986 reached it while seeking the minimum number of characters. This could
3987 happen only if some of the characters were more than one byte long, because
3988 there is a check for at least the minimum number of bytes.
3989
3990 21. Refactored the implementation of \p and \P so as to be more general, to
3991 allow for more different types of property in future. This has changed the
3992 compiled form incompatibly. Anybody with saved compiled patterns that use
3993 \p or \P will have to recompile them.
3994
3995 22. Added "Any" and "L&" to the supported property types.
3996
3997 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3998 but give a compile time error if the value is greater than 0xff.
3999
4000 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4001 accidentally not being installed or uninstalled.
4002
4003 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4004 made were to insert the current release number. This seemed silly, because
4005 it made things harder for people building PCRE on systems that don't run
4006 "configure". I have turned pcre.h into a distributed file, no longer built
4007 by "configure", with the version identification directly included. There is
4008 no longer a pcre.h.in file.
4009
4010 However, this change necessitated a change to the pcre-config script as
4011 well. It is built from pcre-config.in, and one of the substitutions was the
4012 release number. I have updated configure.ac so that ./configure now finds
4013 the release number by grepping pcre.h.
4014
4015 26. Added the ability to run the tests under valgrind.
4016
4017
4018 Version 6.4 05-Sep-05
4019 ---------------------
4020
4021 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4022 "--" to be printed when multiple files were scanned, even when none of the
4023 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4024 consider it to be a bug, and have restored the previous behaviour.
4025
4026 2. A couple of code tidies to get rid of compiler warnings.
4027
4028 3. The pcretest program used to cheat by referring to symbols in the library
4029 whose names begin with _pcre_. These are internal symbols that are not
4030 really supposed to be visible externally, and in some environments it is
4031 possible to suppress them. The cheating is now confined to including
4032 certain files from the library's source, which is a bit cleaner.
4033
4034 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4035 file's purpose clearer.
4036
4037 5. Reorganized pcre_ucp_findchar().
4038
4039
4040 Version 6.3 15-Aug-05
4041 ---------------------
4042
4043 1. The file libpcre.pc.in did not have general read permission in the tarball.
4044
4045 2. There were some problems when building without C++ support:
4046
4047 (a) If C++ support was not built, "make install" and "make test" still
4048 tried to test it.
4049
4050 (b) There were problems when the value of CXX was explicitly set. Some
4051 changes have been made to try to fix these, and ...
4052
4053 (c) --disable-cpp can now be used to explicitly disable C++ support.
4054
4055 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4056 backslash in a target when C++ was disabled. This confuses some
4057 versions of "make", apparently. Using an intermediate variable solves
4058 this. (Same for CPP_LOBJ.)
4059
4060 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4061 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4062 necessary on certain architectures.
4063
4064 4. Added a setting of -export-symbols-regex to the link command to remove
4065 those symbols that are exported in the C sense, but actually are local
4066 within the library, and not documented. Their names all begin with
4067 "_pcre_". This is not a perfect job, because (a) we have to except some
4068 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4069 available (and never for static libraries). I have made a note to try to
4070 find a way round (a) in the future.
4071
4072
4073 Version 6.2 01-Aug-05
4074 ---------------------
4075
4076 1. There was no test for integer overflow of quantifier values. A construction
4077 such as {1111111111111111} would give undefined results. What is worse, if
4078 a minimum quantifier for a parenthesized subpattern overflowed and became
4079 negative, the calculation of the memory size went wrong. This could have
4080 led to memory overwriting.
4081
4082 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4083
4084 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4085 operating environments where this matters.
4086
4087 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4088 PCRE options from within the C++ wrapper.
4089
4090 5. Named capturing subpatterns were not being correctly counted when a pattern
4091 was compiled. This caused two problems: (a) If there were more than 100
4092 such subpatterns, the calculation of the memory needed for the whole
4093 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4094 back references of the form \12, where the number was greater than 9, were
4095 not recognized as back references, even though there were sufficient
4096 previous subpatterns.
4097
4098 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4099 versions of gcc, e.g. 2.95.4.
4100
4101
4102 Version 6.1 21-Jun-05
4103 ---------------------
4104
4105 1. There was one reference to the variable "posix" in pcretest.c that was not
4106 surrounded by "#if !defined NOPOSIX".
4107
4108 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4109 the cross-check on the old pcre_info() function, for the benefit of the
4110 cut-down version of PCRE that is currently imported into Exim.
4111
4112 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4113 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4114 patterns that start (?i) but I don't think that matters. The use of (?i) is
4115 just an example; this all applies to the other options as well.
4116
4117 4. Since libtool seems to echo the compile commands it is issuing, the output
4118 from "make" can be reduced a bit by putting "@" in front of each libtool
4119 compile command.
4120
4121 5. Patch from the folks at Google for configure.in to be a bit more thorough
4122 in checking for a suitable C++ installation before trying to compile the
4123 C++ stuff. This should fix a reported problem when a compiler was present,
4124 but no suitable headers.
4125
4126 6. The man pages all had just "PCRE" as their title. I have changed them to
4127 be the relevant file name. I have also arranged that these names are
4128 retained in the file doc/pcre.txt, which is a concatenation in text format
4129 of all the man pages except the little individual ones for each function.
4130
4131 7. The NON-UNIX-USE file had not been updated for the different set of source
4132 files that come with release 6. I also added a few comments about the C++
4133 wrapper.
4134
4135
4136 Version 6.0 07-Jun-05
4137 ---------------------
4138
4139 1. Some minor internal re-organization to help with my DFA experiments.
4140
4141 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4142 didn't matter for the library itself when fully configured, but did matter
4143 when compiling without UCP support, or within Exim, where the ucp files are
4144 not imported.
4145
4146 3. Refactoring of the library code to split up the various functions into
4147 different source modules. The addition of the new DFA matching code (see
4148 below) to a single monolithic source would have made it really too
4149 unwieldy, quite apart from causing all the code to be include in a
4150 statically linked application, when only some functions are used. This is
4151 relevant even without the DFA addition now that patterns can be compiled in
4152 one application and matched in another.
4153
4154 The downside of splitting up is that there have to be some external
4155 functions and data tables that are used internally in different modules of
4156 the library but which are not part of the API. These have all had their
4157 names changed to start with "_pcre_" so that they are unlikely to clash
4158 with other external names.
4159
4160 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4161 a different (DFA) algorithm. Although it is slower than the original
4162 function, it does have some advantages for certain types of matching
4163 problem.
4164
4165 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4166 including restarting after a partial match.
4167
4168 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4169 defined when compiling for Windows was sent to me. I have put it into the
4170 code, though I have no means of testing or verifying it.
4171
4172 7. Added the pcre_refcount() auxiliary function.
4173
4174 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4175 match before or at the first newline in the subject string. In pcretest,
4176 the /f option on a pattern can be used to set this.
4177
4178 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4179 would behave wrongly. This has been present in PCRE since release 4.0.
4180
4181 10. A number of changes to the pcregrep command:
4182
4183 (a) Refactored how -x works; insert ^(...)$ instead of setting
4184 PCRE_ANCHORED and checking the length, in preparation for adding
4185 something similar for -w.
4186
4187 (b) Added the -w (match as a word) option.
4188
4189 (c) Refactored the way lines are read and buffered so as to have more
4190 than one at a time available.
4191
4192 (d) Implemented a pcregrep test script.
4193
4194 (e) Added the -M (multiline match) option. This allows patterns to match
4195 over several lines of the subject. The buffering ensures that at least
4196 8K, or the rest of the document (whichever is the shorter) is available
4197 for matching (and similarly the previous 8K for lookbehind assertions).
4198
4199 (f) Changed the --help output so that it now says
4200
4201 -w, --word-regex(p)
4202
4203 instead of two lines, one with "regex" and the other with "regexp"
4204 because that confused at least one person since the short forms are the
4205 same. (This required a bit of code, as the output is generated
4206 automatically from a table. It wasn't just a text change.)
4207
4208 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4209 option but starts with a hyphen. Could be a pattern or a path name
4210 starting with a hyphen, for instance.
4211
4212 (h) "-" can be given as a file name to represent stdin.
4213
4214 (i) When file names are being printed, "(standard input)" is used for
4215 the standard input, for compatibility with GNU grep. Previously
4216 "<stdin>" was used.
4217
4218 (j) The option --label=xxx can be used to supply a name to be used for
4219 stdin when file names are being printed. There is no short form.
4220
4221 (k) Re-factored the options decoding logic because we are going to add
4222 two more options that take data. Such options can now be given in four
4223 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4224
4225 (l) Added the -A, -B, and -C options for requesting that lines of context
4226 around matches be printed.
4227
4228 (m) Added the -L option to print the names of files that do not contain
4229 any matching lines, that is, the complement of -l.
4230
4231 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4232 continue to scan other files.
4233
4234 (o) The -s option was incorrectly implemented. For compatibility with other
4235 greps, it now suppresses the error message for a non-existent or non-
4236 accessible file (but not the return code). There is a new option called
4237 -q that suppresses the output of matching lines, which was what -s was
4238 previously doing.
4239
4240 (p) Added --include and --exclude options to specify files for inclusion
4241 and exclusion when recursing.
4242
4243 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4244 Hopefully, it now does.
4245
4246 12. Missing cast in pcre_study().
4247
4248 13. Added an "uninstall" target to the makefile.
4249
4250 14. Replaced "extern" in the function prototypes in Makefile.in with
4251 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4252 world, but is set differently for Windows.
4253
4254 15. Added a second compiling function called pcre_compile2(). The only
4255 difference is that it has an extra argument, which is a pointer to an
4256 integer error code. When there is a compile-time failure, this is set
4257 non-zero, in addition to the error test pointer being set to point to an
4258 error message. The new argument may be NULL if no error number is required
4259 (but then you may as well call pcre_compile(), which is now just a
4260 wrapper). This facility is provided because some applications need a
4261 numeric error indication, but it has also enabled me to tidy up the way
4262 compile-time errors are handled in the POSIX wrapper.
4263
4264 16. Added VPATH=.libs to the makefile; this should help when building with one
4265 prefix path and installing with another. (Or so I'm told by someone who
4266 knows more about this stuff than I do.)
4267
4268 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4269 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4270 match everything, including newlines. This is not POSIX-compatible, but
4271 somebody wanted the feature. From pcretest it can be activated by using
4272 both the P and the s flags.
4273
4274 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4275
4276 19. libpcre.pc was being incorrectly installed as executable.
4277
4278 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4279 it now also looks for '\r' so that it will work unmodified on Windows.
4280
4281 21. Added Google's contributed C++ wrapper to the distribution.
4282
4283 22. Added some untidy missing memory free() calls in pcretest, to keep
4284 Electric Fence happy when testing.
4285
4286
4287
4288 Version 5.0 13-Sep-04
4289 ---------------------
4290
4291 1. Internal change: literal characters are no longer packed up into items
4292 containing multiple characters in a single byte-string. Each character
4293 is now matched using a separate opcode. However, there may be more than one
4294 byte in the character in UTF-8 mode.
4295
4296 2. The pcre_callout_block structure has two new fields: pattern_position and
4297 next_item_length. These contain the offset in the pattern to the next match
4298 item, and its length, respectively.
4299
4300 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4301 insertion of callouts before each pattern item. Added the /C option to
4302 pcretest to make use of this.
4303
4304 4. On the advice of a Windows user, the lines
4305
4306 #if defined(_WIN32) || defined(WIN32)
4307 _setmode( _fileno( stdout ), 0x8000 );
4308 #endif /* defined(_WIN32) || defined(WIN32) */
4309
4310 have been added to the source of pcretest. This apparently does useful
4311 magic in relation to line terminators.
4312
4313 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4314 for the benefit of those environments where the "b" makes a difference.
4315
4316 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4317 to know about it. I have put a hack into configure.in that adds in code
4318 to set GCC=yes if CC=icc. This seems to end up at a point in the
4319 generated configure script that is early enough to affect the setting of
4320 compiler options, which is what is needed, but I have no means of testing
4321 whether it really works. (The user who reported this had patched the
4322 generated configure script, which of course I cannot do.)
4323
4324 LATER: After change 22 below (new libtool files), the configure script
4325 seems to know about icc (and also ecc). Therefore, I have commented out
4326 this hack in configure.in.
4327
4328 7. Added support for pkg-config (2 patches were sent in).
4329
4330 8. Negated POSIX character classes that used a combination of internal tables
4331 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4332 [[:^ascii]]. Typically, they would match almost any characters. The other
4333 POSIX classes were not broken in this way.
4334
4335 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4336 to find the match, as PCRE was deluded into thinking that the match had to
4337 start at the start point or following a newline. The same bug applied to
4338 patterns with negative forward assertions or any backward assertions
4339 preceding ".*" at the start, unless the pattern required a fixed first
4340 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4341
4342 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4343 starting at the last subject character, bytes beyond the end of the subject
4344 string were read.
4345
4346 11. Renamed the variable "class" as "classbits" to make life easier for C++
4347 users. (Previously there was a macro definition, but it apparently wasn't
4348 enough.)
4349
4350 12. Added the new field "tables" to the extra data so that tables can be passed
4351 in at exec time, or the internal tables can be re-selected. This allows
4352 a compiled regex to be saved and re-used at a later time by a different
4353 program that might have everything at different addresses.
4354
4355 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4356 -R library as well as a -L library.
4357
4358 14. The debugging options of pcretest (-d on the command line or D on a
4359 pattern) showed incorrect output for anything following an extended class
4360 that contained multibyte characters and which was followed by a quantifier.
4361
4362 15. Added optional support for general category Unicode character properties
4363 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4364 support. It adds about 90K to the size of the library. The meanings of the
4365 inbuilt class escapes such as \d and \s have NOT been changed.
4366
4367 16. Updated pcredemo.c to include calls to free() to release the memory for the
4368 compiled pattern.
4369
4370 17. The generated file chartables.c was being created in the source directory
4371 instead of in the building directory. This caused the build to fail if the
4372 source directory was different from the building directory, and was
4373 read-only.
4374
4375 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4376 file. No doubt somebody will tell me if they don't make sense... Also added
4377 Dan Mooney's comments about building on OpenVMS.
4378
4379 19. Added support for partial matching via the PCRE_PARTIAL option for
4380 pcre_exec() and the \P data escape in pcretest.
4381
4382 20. Extended pcretest with 3 new pattern features:
4383
4384 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4385 write the compiled pattern to the file whose name is "rest-of-line".
4386 This is a straight binary dump of the data, with the saved pointer to
4387 the character tables forced to be NULL. The study data, if any, is
4388 written too. After writing, pcretest reads a new pattern.
4389
4390 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4391 compiled pattern from the given file. There must not be any
4392 occurrences of "<" in the file name (pretty unlikely); if there are,
4393 pcretest will instead treat the initial "<" as a pattern delimiter.
4394 After reading in the pattern, pcretest goes on to read data lines as
4395 usual.
4396
4397 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4398 and 16-bit fields in a compiled pattern, to simulate a pattern that
4399 was compiled on a host of opposite endianness.
4400
4401 21. The pcre-exec() function can now cope with patterns that were compiled on
4402 hosts of opposite endianness, with this restriction:
4403
4404 As for any compiled expression that is saved and used later, the tables
4405 pointer field cannot be preserved; the extra_data field in the arguments
4406 to pcre_exec() should be used to pass in a tables address if a value
4407 other than the default internal tables were used at compile time.
4408
4409 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4410 now diagnosed as an error. Previously, most of the time, a negative number
4411 would have been treated as zero, but if in addition "ovector" was passed as
4412 NULL, a crash could occur.
4413
4414 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4415 new versions from the libtool 1.5 distribution (the last one is a copy of
4416 a file called libtool.m4). This seems to have fixed the need to patch
4417 "configure" to support Darwin 1.3 (which I used to do). However, I still
4418 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4419 workstation).
4420
4421 24. Changed the PCRE licence to be the more standard "BSD" licence.
4422
4423
4424 Version 4.5 01-Dec-03
4425 ---------------------
4426
4427 1. There has been some re-arrangement of the code for the match() function so
4428 that it can be compiled in a version that does not call itself recursively.
4429 Instead, it keeps those local variables that need separate instances for
4430 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4431 needs to "recurse". Keeping track of where control must go is done by means
4432 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4433 hide most of the details from the main code, and operates only if
4434 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4435 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4436 operating.
4437
4438 To make it easier for callers to provide specially tailored get/free
4439 functions for this usage, two new functions, pcre_stack_malloc, and
4440 pcre_stack_free, are used. They are always called in strict stacking order,
4441 and the size of block requested is always the same.
4442
4443 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4444 PCRE has been compiled to use the stack or the heap for recursion. The
4445 -C option of pcretest uses this to show which version is compiled.
4446
4447 A new data escape \S, is added to pcretest; it causes the amounts of store
4448 obtained and freed by both kinds of malloc/free at match time to be added
4449 to the output.
4450
4451 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4452 what's available on my current Linux desktop machine.
4453
4454 3. When matching a UTF-8 string, the test for a valid string at the start has
4455 been extended. If start_offset is not zero, PCRE now checks that it points
4456 to a byte that is the start of a UTF-8 character. If not, it returns
4457 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4458 this is necessary because there may be backward assertions in the pattern.
4459 When matching the same subject several times, it may save resources to use
4460 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4461
4462 4. The code for checking the validity of UTF-8 strings has been tightened so
4463 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4464 containing "overlong sequences".
4465
4466 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4467 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4468 should have been "&", but it just so happened that all the cases this let
4469 through by mistake were picked up later in the function.
4470
4471 6. I had used a variable called "isblank" - this is a C99 function, causing
4472 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4473
4474 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4475 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4476 so the version is shown; (c) stop "make test" echoing "./RunTest".
4477
4478 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4479
4480 9. The prototype for memmove() for systems that don't have it was using
4481 size_t, but the inclusion of the header that defines size_t was later. I've
4482 moved the #includes for the C headers earlier to avoid this.
4483
4484 10. Added some adjustments to the code to make it easier to compiler on certain
4485 special systems:
4486
4487 (a) Some "const" qualifiers were missing.
4488 (b) Added the macro EXPORT before all exported functions; by default this
4489 is defined to be empty.
4490 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4491 that it reads its output file name as an argument instead of writing
4492 to the standard output and assuming this can be redirected.
4493
4494 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4495 class containing characters with values greater than 255, PCRE compilation
4496 went into a loop.
4497
4498 12. A recursive reference to a subpattern that was within another subpattern
4499 that had a minimum quantifier of zero caused PCRE to crash. For example,
4500 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4501 recursion. If the recursively-called subpattern itself had a zero repeat,
4502 that was OK.
4503
4504 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4505 buffer into which it was copied (for escape processing) was still set at
4506 1024, so long lines caused crashes.
4507
4508 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4509 "internal error: code overflow...". This applied to any character class
4510 that was followed by a possessive quantifier.
4511
4512 15. Modified the Makefile to add libpcre.la as a prerequisite for
4513 libpcreposix.la because I was told this is needed for a parallel build to
4514 work.
4515
4516 16. If a pattern that contained .* following optional items at the start was
4517 studied, the wrong optimizing data was generated, leading to matching
4518 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4519 matching string must start with a or b or c. The correct conclusion for
4520 this pattern is that a match can start with any character.
4521
4522
4523 Version 4.4 13-Aug-03
4524 ---------------------
4525
4526 1. In UTF-8 mode, a character class containing characters with values between
4527 127 and 255 was not handled correctly if the compiled pattern was studied.
4528 In fixing this, I have also improved the studying algorithm for such
4529 classes (slightly).
4530
4531 2. Three internal functions had redundant arguments passed to them. Removal
4532 might give a very teeny performance improvement.
4533
4534 3. Documentation bug: the value of the capture_top field in a callout is *one
4535 more than* the number of the hightest numbered captured substring.
4536
4537 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4538 in incorrectly linking with a previously installed version. They now link
4539 explicitly with libpcre.la.
4540
4541 5. configure.in no longer needs to recognize Cygwin specially.
4542
4543 6. A problem in pcre.in for Windows platforms is fixed.
4544
4545 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4546 pcretest, it used to include the size of the study block as part of its
4547 output. Unfortunately, the structure contains a field that has a different
4548 size on different hardware architectures. This meant that the tests that
4549 showed this size failed. As the block is currently always of a fixed size,
4550 this information isn't actually particularly useful in pcretest output, so
4551 I have just removed it.
4552
4553 8. Three pre-processor statements accidentally did not start in column 1.
4554 Sadly, there are *still* compilers around that complain, even though
4555 standard C has not required this for well over a decade. Sigh.
4556
4557 9. In pcretest, the code for checking callouts passed small integers in the
4558 callout_data field, which is a void * field. However, some picky compilers
4559 complained about the casts involved for this on 64-bit systems. Now
4560 pcretest passes the address of the small integer instead, which should get
4561 rid of the warnings.
4562
4563 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4564 both compile and run time, and gives an error if an invalid UTF-8 sequence
4565 is found. There is a option for disabling this check in cases where the
4566 string is known to be correct and/or the maximum performance is wanted.
4567
4568 11. In response to a bug report, I changed one line in Makefile.in from
4569
4570 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4571 to
4572 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4573
4574 to look similar to other lines, but I have no way of telling whether this
4575 is the right thing to do, as I do not use Windows. No doubt I'll get told
4576 if it's wrong...
4577
4578
4579 Version 4.3 21-May-03
4580 ---------------------
4581
4582 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4583 Makefile.
4584
4585 2. Some refactoring to improve the quality of the code:
4586
4587 (i) The utf8_table... variables are now declared "const".
4588
4589 (ii) The code for \cx, which used the "case flipping" table to upper case
4590 lower case letters, now just substracts 32. This is ASCII-specific,
4591 but the whole concept of \cx is ASCII-specific, so it seems
4592 reasonable.
4593
4594 (iii) PCRE was using its character types table to recognize decimal and
4595 hexadecimal digits in the pattern. This is silly, because it handles
4596 only 0-9, a-f, and A-F, but the character types table is locale-
4597 specific, which means strange things might happen. A private
4598 table is now used for this - though it costs 256 bytes, a table is
4599 much faster than multiple explicit tests. Of course, the standard
4600 character types table is still used for matching digits in subject
4601 strings against \d.
4602
4603 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4604 ending in _t are). So I've renamed it as ESC_tee.
4605
4606 3. The first argument for regexec() in the POSIX wrapper should have been
4607 defined as "const".
4608
4609 4. Changed pcretest to use malloc() for its buffers so that they can be
4610 Electric Fenced for debugging.
4611
4612 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4613 to read one or more bytes before the start of the subject string. Often this
4614 had no effect on PCRE's behaviour, but in some circumstances it could
4615 provoke a segmentation fault.
4616
4617 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4618 to try to read one or more bytes before the start of the subject string.
4619
4620 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4621 UTF-8 support could misbehave in various ways if the subject string
4622 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4623 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4624 back over UTF-8 characters.)
4625
4626
4627 Version 4.2 14-Apr-03
4628 ---------------------
4629
4630 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4631
4632 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4633 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4634 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4635 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4636 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4637 and BUILD_EXEEXT
4638 Note: automatic setting of the BUILD variables is not yet working
4639 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4640 compile-time but not at link-time
4641 [LINK]: use for linking executables only
4642 make different versions for Windows and non-Windows
4643 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4644 libraries
4645 [LINK_FOR_BUILD]: new variable
4646 [OBJEXT]: use throughout
4647 [EXEEXT]: use throughout
4648 <winshared>: new target
4649 <wininstall>: new target
4650 <dftables.o>: use native compiler
4651 <dftables>: use native linker
4652 <install>: handle Windows platform correctly
4653 <clean>: ditto
4654 <check>: ditto
4655 copy DLL to top builddir before testing
4656
4657 As part of these changes, -no-undefined was removed again. This was reported
4658 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4659 in any case.
4660
4661 3. Some tidies to get rid of compiler warnings:
4662
4663 . In the match_data structure, match_limit was an unsigned long int, whereas
4664 match_call_count was an int. I've made them both unsigned long ints.
4665
4666 . In pcretest the fact that a const uschar * doesn't automatically cast to
4667 a void * provoked a warning.
4668
4669 . Turning on some more compiler warnings threw up some "shadow" variables
4670 and a few more missing casts.
4671
4672 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4673 option, a class that contained a single character with a value between 128
4674 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4675
4676 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4677 option, a class that contained several characters, but with at least one
4678 whose value was between 128 and 255 caused PCRE to crash.
4679
4680
4681 Version 4.1 12-Mar-03
4682 ---------------------
4683
4684 1. Compiling with gcc -pedantic found a couple of places where casts were
4685 needed, and a string in dftables.c that was longer than standard compilers are
4686 required to support.
4687
4688 2. Compiling with Sun's compiler found a few more places where the code could
4689 be tidied up in order to avoid warnings.
4690
4691 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4692 first of these names is deprecated in the latest Autoconf in favour of the name
4693 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4694 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4695 analogy I have changed it to CFLAGS_FOR_BUILD.
4696
4697 4. Added -no-undefined to the linking command in the Makefile, because this is
4698 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4699 linking step for the pcreposix library.
4700
4701 5. PCRE was failing to diagnose the case of two named groups with the same
4702 name.
4703
4704 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4705 literal character that is needed in the subject for a match, and scans along to
4706 ensure that it is present before embarking on the full matching process. This
4707 saves time in cases of nested unlimited repeats that are never going to match.
4708 Problem: the scan can take a lot of time if the subject is very long (e.g.
4709 megabytes), thus penalizing straightforward matches. It is now done only if the
4710 amount of subject to be scanned is less than 1000 bytes.
4711
4712 7. A lesser problem with the same optimization is that it was recording the
4713 first character of an anchored pattern as "needed", thus provoking a search
4714 right along the subject, even when the first match of the pattern was going to
4715 fail. The "needed" character is now not set for anchored patterns, unless it
4716 follows something in the pattern that is of non-fixed length. Thus, it still
4717 fulfils its original purpose of finding quick non-matches in cases of nested
4718 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4719
4720
4721 Version 4.0 17-Feb-03
4722 ---------------------
4723
4724 1. If a comment in an extended regex that started immediately after a meta-item
4725 extended to the end of string, PCRE compiled incorrect data. This could lead to
4726 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4727
4728 2. Moved to autoconf 2.53 and libtool 1.4.2.
4729
4730 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4731 the special perltest8 script is no longer needed - all the tests can be run
4732 from a single perltest script.
4733
4734 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4735 by \s. It has now been removed in PCRE. This means it isn't recognized as
4736 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4737 class [:space:] *does* include VT, thereby creating a mess.
4738
4739 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4740 space and tab.
4741
4742 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4743 its new features into the main test script, reducing the number of scripts.
4744
4745 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4746 were backward compatible, and made the (?i) apply to the whole pattern, as if
4747 /i were given. Now it behaves more logically, and applies the option setting
4748 only to what follows. PCRE has been changed to follow suit. However, if it
4749 finds options settings right at the start of the pattern, it extracts them into
4750 the global options, as before. Thus, they show up in the info data.
4751
4752 8. Added support for the \Q...\E escape sequence. Characters in between are
4753 treated as literals. This is slightly different from Perl in that $ and @ are
4754 also handled as literals inside the quotes. In Perl, they will cause variable
4755 interpolation. Note the following examples:
4756
4757 Pattern PCRE matches Perl matches
4758
4759 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4760 \Qabc\$xyz\E abc\$xyz abc\$xyz
4761 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4762
4763 For compatibility with Perl, \Q...\E sequences are recognized inside character
4764 classes as well as outside them.
4765
4766 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4767 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4768 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4769 signed/unsigned warnings.
4770
4771 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4772 option for pcretest, so I've replaced it by a simple function that does just
4773 that job.
4774
4775 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4776 "pcregrep -".
4777
4778 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4779 Java package. This provides some syntactic sugar for simple cases of what my
4780 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4781 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4782 item, you can use this simplified notation. Note that only makes sense with
4783 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4784 greediness, whatever the setting of the PCRE_UNGREEDY option.
4785
4786 13. A change of greediness default within a pattern was not taking effect at
4787 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4788 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4789 was abstracted outside.
4790
4791 14. PCRE now supports the \G assertion. It is true when the current matching
4792 position is at the start point of the match. This differs from \A when the
4793 starting offset is non-zero. Used with the /g option of pcretest (or similar
4794 code), it works in the same way as it does for Perl's /g option. If all
4795 alternatives of a regex begin with \G, the expression is anchored to the start
4796 match position, and the "anchored" flag is set in the compiled expression.
4797
4798 15. Some bugs concerning the handling of certain option changes within patterns
4799 have been fixed. These applied to options other than (?ims). For example,
4800 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4801 been the other way round. Some of this was related to change 7 above.
4802
4803 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4804 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4805 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4806 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4807
4808 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4809 mode. Unlike ".", it always matches newline, whatever the setting of
4810 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4811 assertions. Perl allows it, but it doesn't (in general) work because it can't
4812 calculate the length of the lookbehind. At least, that's the case for Perl
4813 5.8.0 - I've been told they are going to document that it doesn't work in
4814 future.
4815
4816 18. Added an error diagnosis for escapes that PCRE does not support: these are
4817 \L, \l, \N, \P, \p, \U, \u, and \X.
4818
4819 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4820 reading past the end of the pattern in cases such as /[abcd/.
4821
4822 20. PCRE was getting more memory than necessary for patterns with classes that
4823 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4824
4825 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4826 compiling PCRE for use with Virtual Pascal.
4827
4828 22. Small fix to the Makefile to make it work properly if the build is done
4829 outside the source tree.
4830
4831 23. Added a new extension: a condition to go with recursion. If a conditional
4832 subpattern starts with (?(R) the "true" branch is used if recursion has
4833 happened, whereas the "false" branch is used only at the top level.
4834
4835 24. When there was a very long string of literal characters (over 255 bytes
4836 without UTF support, over 250 bytes with UTF support), the computation of how
4837 much memory was required could be incorrect, leading to segfaults or other
4838 strange effects.
4839
4840 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4841 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4842 there was a subsequent back reference to those brackets. This meant that, for
4843 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4844 possible to check for precisely this case. All we can do is abandon the
4845 optimization if .* occurs inside capturing brackets when there are any back
4846 references whatsoever. (See below for a better fix that came later.)
4847
4848 26. The handling of the optimization for finding the first character of a
4849 non-anchored pattern, and for finding a character that is required later in the
4850 match were failing in some cases. This didn't break the matching; it just
4851 failed to optimize when it could. The way this is done has been re-implemented.
4852
4853 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4854
4855 28. Added a new feature that provides some of the functionality that Perl
4856 provides with (?{...}). The facility is termed a "callout". The way it is done
4857 in PCRE is for the caller to provide an optional function, by setting
4858 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4859 global variable. By default it is unset, which disables all calling out. To get
4860 the function called, the regex must include (?C) at appropriate points. This
4861 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4862 This provides a means of identifying different callout points. When PCRE
4863 reaches such a point in the regex, if pcre_callout has been set, the external
4864 function is called. It is provided with data in a structure called
4865 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4866 matching continues; if it returns a non-zero value, the match at the current
4867 point fails. However, backtracking will occur if possible. [This was changed
4868 later and other features added - see item 49 below.]
4869
4870 29. pcretest is upgraded to test the callout functionality. It provides a
4871 callout function that displays information. By default, it shows the start of
4872 the match and the current position in the text. There are some new data escapes
4873 to vary what happens:
4874
4875 \C+ in addition, show current contents of captured substrings
4876 \C- do not supply a callout function
4877 \C!n return 1 when callout number n is reached
4878 \C!n!m return 1 when callout number n is reached for the mth time
4879
4880 30. If pcregrep was called with the -l option and just a single file name, it
4881 output "<stdin>" if a match was found, instead of the file name.
4882
4883 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4884 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4885 pcre_exec(). This saves a malloc/free per call. The default value of
4886 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4887 when configuring.
4888
4889 32. The default maximum size of a compiled pattern is 64K. There have been a
4890 few cases of people hitting this limit. The code now uses macros to handle the
4891 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4892 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4893 configuring. Tests 2 and 5 work only with 2-byte links because they output
4894 debugging information about compiled patterns.
4895
4896 33. Internal code re-arrangements:
4897
4898 (a) Moved the debugging function for printing out a compiled regex into
4899 its own source file (printint.c) and used #include to pull it into
4900 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4901 separate copies.
4902
4903 (b) Defined the list of op-code names for debugging as a macro in
4904 internal.h so that it is next to the definition of the opcodes.
4905
4906 (c) Defined a table of op-code lengths for simpler skipping along compiled
4907 code. This is again a macro in internal.h so that it is next to the
4908 definition of the opcodes.
4909
4910 34. Added support for recursive calls to individual subpatterns, along the
4911 lines of Robin Houston's patch (but implemented somewhat differently).
4912
4913 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4914 allow it to read and process whole directories in Win32. This code was
4915 contributed by Lionel Fourquaux; it has not been tested by me.
4916
4917 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4918 used to name a group. Names consist of alphanumerics and underscores, and must
4919 be unique. Back references use the syntax (?P=name) and recursive calls use
4920 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4921 numbers. The function pcre_fullinfo() can be used after compilation to extract
4922 a name/number map. There are three relevant calls:
4923
4924 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4925 PCRE_INFO_NAMECOUNT yields the number of entries
4926 PCRE_INFO_NAMETABLE yields a pointer to the map.
4927
4928 The map is a vector of fixed-size entries. The size of each entry depends on
4929 the length of the longest name used. The first two bytes of each entry are the
4930 group number, most significant byte first. There follows the corresponding
4931 name, zero terminated. The names are in alphabetical order.
4932
4933 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4934 case instead of 255. Making it the same both with and without UTF-8 support
4935 means that the same test output works with both.
4936
4937 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4938 calling malloc() with a zero argument.
4939
4940 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4941 optimization. I've improved things by keeping a bitmap of backreferences with
4942 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4943 fact referenced, the optimization can be applied. It is unlikely that a
4944 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4945 the match to follow \n) will appear inside brackets with a number greater than
4946 31, but if it does, any back reference > 31 suppresses the optimization.
4947
4948 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4949 of disabling numbered capturing parentheses. Any opening parenthesis that is
4950 not followed by ? behaves as if it were followed by ?: but named parentheses
4951 can still be used for capturing (and they will acquire numbers in the usual
4952 way).
4953
4954 41. Redesigned the return codes from the match() function into yes/no/error so
4955 that errors can be passed back from deep inside the nested calls. A malloc
4956 failure while inside a recursive subpattern call now causes the
4957 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4958
4959 42. It is now possible to set a limit on the number of times the match()
4960 function is called in a call to pcre_exec(). This facility makes it possible to
4961 limit the amount of recursion and backtracking, though not in a directly
4962 obvious way, because the match() function is used in a number of different
4963 circumstances. The count starts from zero for each position in the subject
4964 string (for non-anchored patterns). The default limit is, for compatibility, a
4965 large number, namely 10 000 000. You can change this in two ways:
4966
4967 (a) When configuring PCRE before making, you can use --with-match-limit=n
4968 to set a default value for the compiled library.
4969
4970 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4971 a different value is set. See 45 below.
4972
4973 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4974
4975 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4976 of things that can be changed at compile time. The first argument specifies
4977 what is wanted and the second points to where the information is to be placed.
4978 The current list of available information is:
4979
4980 PCRE_CONFIG_UTF8
4981
4982 The output is an integer that is set to one if UTF-8 support is available;
4983 otherwise it is set to zero.
4984
4985 PCRE_CONFIG_NEWLINE
4986
4987 The output is an integer that it set to the value of the code that is used for
4988 newline. It is either LF (10) or CR (13).
4989
4990 PCRE_CONFIG_LINK_SIZE
4991
4992 The output is an integer that contains the number of bytes used for internal
4993 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4994
4995 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4996
4997 The output is an integer that contains the threshold above which the POSIX
4998 interface uses malloc() for output vectors. See item 31 above.
4999
5000 PCRE_CONFIG_MATCH_LIMIT
5001
5002 The output is an unsigned integer that contains the default limit of the number
5003 of match() calls in a pcre_exec() execution. See 42 above.
5004
5005 44. pcretest has been upgraded by the addition of the -C option. This causes it
5006 to extract all the available output from the new pcre_config() function, and to
5007 output it. The program then exits immediately.
5008
5009 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5010 order to support additional features. One way would have been to define
5011 pcre_exec2() (for example) with extra arguments, but this would not have been
5012 extensible, and would also have required all calls to the original function to
5013 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5014 is used for passing in "extra" data from pcre_study().
5015
5016 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5017 contains the following fields:
5018
5019 flags a bitmap indicating which of the following fields are set
5020 study_data opaque data from pcre_study()
5021 match_limit a way of specifying a limit on match() calls for a specific
5022 call to pcre_exec()
5023 callout_data data for callouts (see 49 below)
5024
5025 The flag bits are also defined in pcre.h, and are
5026
5027 PCRE_EXTRA_STUDY_DATA
5028 PCRE_EXTRA_MATCH_LIMIT
5029 PCRE_EXTRA_CALLOUT_DATA
5030
5031 The pcre_study() function now returns one of these new pcre_extra blocks, with
5032 the actual study data pointed to by the study_data field, and the
5033 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5034 before. That is, this change is entirely upwards-compatible and requires no
5035 change to existing code.
5036
5037 If you want to pass in additional data to pcre_exec(), you can either place it
5038 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5039 block.
5040
5041 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5042 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5043 times with different match limits, until it finds the minimum value needed for
5044 pcre_exec() to complete. The value is then output. This can be instructive; for
5045 most simple matches the number is quite small, but for pathological cases it
5046 gets very large very quickly.
5047
5048 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5049 returns the size of the data block pointed to by the study_data field in a
5050 pcre_extra block, that is, the value that was passed as the argument to
5051 pcre_malloc() when PCRE was getting memory in which to place the information
5052 created by pcre_study(). The fourth argument should point to a size_t variable.
5053 pcretest has been extended so that this information is shown after a successful
5054 pcre_study() call when information about the compiled regex is being displayed.
5055
5056 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5057 because what follows is always an absolute path. (Later: it turns out that this
5058 is more than cosmetic for MinGW, because it doesn't like empty path
5059 components.)
5060
5061 49. Some changes have been made to the callout feature (see 28 above):
5062
5063 (i) A callout function now has three choices for what it returns:
5064
5065 0 => success, carry on matching
5066 > 0 => failure at this point, but backtrack if possible
5067 < 0 => serious error, return this value from pcre_exec()
5068
5069 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5070 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5071 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5072 use by callout functions. It will never be used by PCRE itself.
5073
5074 (ii) The pcre_extra structure (see 45 above) has a void * field called
5075 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5076 pcre_callout_block structure has a field of the same name. The contents of
5077 the field passed in the pcre_extra structure are passed to the callout
5078 function in the corresponding field in the callout block. This makes it
5079 easier to use the same callout-containing regex from multiple threads. For
5080 testing, the pcretest program has a new data escape
5081
5082 \C*n pass the number n (may be negative) as callout_data
5083
5084 If the callout function in pcretest receives a non-zero value as
5085 callout_data, it returns that value.
5086
5087 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5088 there were some redundant $(CFLAGS) in commands that are now specified as
5089 $(LINK), which already includes $(CFLAGS).
5090
5091 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5092 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5093 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5094 one-byte characters throughout. Note that case-insensitive matching applies
5095 only to characters whose values are less than 256. PCRE doesn't support the
5096 notion of cases for higher-valued characters.
5097
5098 (i) A character class whose characters are all within 0-255 is handled as
5099 a bit map, and the map is inverted for negative classes. Previously, a
5100 character > 255 always failed to match such a class; however it should
5101 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5102
5103 (ii) A negated character class with a single character < 255 is coded as
5104 "not this character" (OP_NOT). This wasn't working properly when the test
5105 character was multibyte, either singly or repeated.
5106
5107 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5108 mode, for example: \x{100}{2,3}.
5109
5110 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5111 singly or repeated) now correctly test multibyte characters. However,
5112 PCRE doesn't recognize any characters with values greater than 255 as
5113 digits, spaces, or word characters. Such characters always match \D, \S,
5114 and \W, and never match \d, \s, or \w.
5115
5116 (v) Classes may now contain characters and character ranges with values
5117 greater than 255. For example: [ab\x{100}-\x{400}].
5118
5119 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5120 PCRE in UTF-8 mode.
5121
5122 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5123 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5124 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5125 value.)
5126
5127 53. The single man page has become too large. I have therefore split it up into
5128 a number of separate man pages. These also give rise to individual HTML pages;
5129 these are now put in a separate directory, and there is an index.html page that
5130 lists them all. Some hyperlinking between the pages has been installed.
5131
5132 54. Added convenience functions for handling named capturing parentheses.
5133
5134 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5135 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5136 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5137 are faulted.
5138
5139 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5140 calling configure. These values are used when compiling the dftables.c program
5141 which is run to generate the source of the default character tables. They
5142 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5143 you will need to set these values.
5144
5145 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5146
5147
5148 Version 3.9 02-Jan-02
5149 ---------------------
5150
5151 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5152
5153 2. If --disable-static was given, the building process failed when trying to
5154 build pcretest and pcregrep. (For some reason it was using libtool to compile
5155 them, which is not right, as they aren't part of the library.)
5156
5157
5158 Version 3.8 18-Dec-01
5159 ---------------------
5160
5161 1. The experimental UTF-8 code was completely screwed up. It was packing the
5162 bytes in the wrong order. How dumb can you get?
5163
5164
5165 Version 3.7 29-Oct-01
5166 ---------------------
5167
5168 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5169 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5170 this didn't happen under Solaris 8, where I normally test things.
5171
5172 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5173 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5174 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5175 this fix an hour or so after the initial 3.7 release.)
5176
5177
5178 Version 3.6 23-Oct-01
5179 ---------------------
5180
5181 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5182 offsets passed as NULL with zero offset count.
5183
5184 2. The config.guess and config.sub files had not been updated when I moved to
5185 the latest autoconf.
5186
5187
5188 Version 3.5 15-Aug-01
5189 ---------------------
5190
5191 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5192 had been forgotten.
5193
5194 2. By using declared but undefined structures, we can avoid using "void"
5195 definitions in pcre.h while keeping the internal definitions of the structures
5196 private.
5197
5198 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5199 user point of view, this means that both static and shared libraries are built
5200 by default, but this can be individually controlled. More of the work of
5201 handling this static/shared cases is now inside libtool instead of PCRE's make
5202 file.
5203
5204 4. The pcretest utility is now installed along with pcregrep because it is
5205 useful for users (to test regexs) and by doing this, it automatically gets
5206 relinked by libtool. The documentation has been turned into a man page, so
5207 there are now .1, .txt, and .html versions in /doc.
5208
5209 5. Upgrades to pcregrep:
5210 (i) Added long-form option names like gnu grep.
5211 (ii) Added --help to list all options with an explanatory phrase.
5212 (iii) Added -r, --recursive to recurse into sub-directories.
5213 (iv) Added -f, --file to read patterns from a file.
5214
5215 6. pcre_exec() was referring to its "code" argument before testing that
5216 argument for NULL (and giving an error if it was NULL).
5217
5218 7. Upgraded Makefile.in to allow for compiling in a different directory from
5219 the source directory.
5220
5221 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5222 options bits, the pointer it was passed was to an int instead of to an unsigned
5223 long int. This mattered only on 64-bit systems.
5224
5225 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5226 generated) instead of pcre.in, which it its source. Also made the same change
5227 in several of the .c files.
5228
5229 10. A new release of gcc defines printf() as a macro, which broke pcretest
5230 because it had an ifdef in the middle of a string argument for printf(). Fixed
5231 by using separate calls to printf().
5232
5233 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5234 script, to force use of CR or LF instead of \n in the source. On non-Unix
5235 systems, the value can be set in config.h.
5236
5237 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5238 absolute limit. Changed the text of the error message to make this clear, and
5239 likewise updated the man page.
5240
5241 13. The limit of 99 on the number of capturing subpatterns has been removed.
5242 The new limit is 65535, which I hope will not be a "real" limit.
5243
5244
5245 Version 3.4 22-Aug-00
5246 ---------------------
5247
5248 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5249
5250 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5251
5252
5253 Version 3.3 01-Aug-00
5254 ---------------------
5255
5256 1. If an octal character was given, but the value was greater than \377, it
5257 was not getting masked to the least significant bits, as documented. This could
5258 lead to crashes in some systems.
5259
5260 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5261 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5262
5263 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5264 These just pass their arguments on to (pcre_free)(), but they are provided
5265 because some uses of PCRE bind it to non-C systems that can call its functions,
5266 but cannot call free() or pcre_free() directly.
5267
5268 4. Add "make test" as a synonym for "make check". Corrected some comments in
5269 the Makefile.
5270
5271 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5272 Makefile.
5273
5274 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5275 command called pgrep for grepping around the active processes.
5276
5277 7. Added the beginnings of support for UTF-8 character strings.
5278
5279 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5280 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5281 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5282 out for the ar command.)
5283
5284
5285 Version 3.2 12-May-00
5286 ---------------------
5287
5288 This is purely a bug fixing release.
5289
5290 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5291 of ZA. This was just one example of several cases that could provoke this bug,
5292 which was introduced by change 9 of version 2.00. The code for breaking
5293 infinite loops after an iteration that matches an empty string was't working
5294 correctly.
5295
5296 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5297 when matched against abbab (for example). After matching an empty string, it
5298 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5299 caused it to match further down the string than it should.
5300
5301 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5302 was there because it doesn't seem to be needed, and it causes trouble on some
5303 systems, as it is not a Standard C header. It has been removed.
5304
5305 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5306 were reported on the Macintosh. The changes were from
5307
5308 while ((c = *(++ptr)) != 0 && c != '\n');
5309 to
5310 while ((c = *(++ptr)) != 0 && c != '\n') ;
5311
5312 Totally extraordinary, but if that's what it takes...
5313
5314 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5315 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5316 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5317 assumes the way PCRE uses memmove() (always moving upwards).
5318
5319 6. PCRE is being used in one environment where strchr() is not available. There
5320 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5321 faster code anyway.
5322
5323
5324 Version 3.1 09-Feb-00
5325 ---------------------
5326
5327 The only change in this release is the fixing of some bugs in Makefile.in for
5328 the "install" target:
5329
5330 (1) It was failing to install pcreposix.h.
5331
5332 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5333
5334
5335 Version 3.0 01-Feb-00
5336 ---------------------
5337
5338 1. Add support for the /+ modifier to perltest (to output $` like it does in
5339 pcretest).
5340
5341 2. Add support for the /g modifier to perltest.
5342
5343 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5344 matches null strings.
5345
5346 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5347 pattern. Perl treats empty patterns specially - it reuses the most recent
5348 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5349 effect.
5350
5351 5. The POSIX interface was broken in that it was just handing over the POSIX
5352 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5353 required a bigger vector, with some working space on the end. This means that
5354 the POSIX wrapper now has to get and free some memory, and copy the results.
5355
5356 6. Added some simple autoconf support, placing the test data and the
5357 documentation in separate directories, re-organizing some of the
5358 information files, and making it build pcre-config (a GNU standard). Also added
5359 libtool support for building PCRE as a shared library, which is now the
5360 default.
5361
5362 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5363 09 are not valid octal constants. Single digits will be used for minor values
5364 less than 10.
5365
5366 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5367 existing programs that set these in the POSIX interface can use PCRE without
5368 modification.
5369
5370 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5371 return all that pcre_info() returns, plus additional data. The pcre_info()
5372 function is retained for compatibility, but is considered to be obsolete.
5373
5374 10. Added experimental recursion feature (?R) to handle one common case that
5375 Perl 5.6 will be able to do with (?p{...}).
5376
5377 11. Added support for POSIX character classes like [:alpha:], which Perl is
5378 adopting.
5379
5380
5381 Version 2.08 31-Aug-99
5382 ----------------------
5383
5384 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5385 trying to match at the startoffset position, but instead was moving forward to
5386 the next newline as if a previous match had failed.
5387
5388 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5389 and could get into a loop if a null string was matched other than at the start
5390 of the subject.
5391
5392 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5393 be distinguished at compile time, and for completeness also added PCRE_DATE.
5394
5395 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5396 in GnuWin32 environments.
5397
5398
5399 Version 2.07 29-Jul-99
5400 ----------------------
5401
5402 1. The documentation is now supplied in plain text form and HTML as well as in
5403 the form of man page sources.
5404
5405 2. C++ compilers don't like assigning (void *) values to other pointer types.
5406 In particular this affects malloc(). Although there is no problem in Standard
5407 C, I've put in casts to keep C++ compilers happy.
5408
5409 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5410 should be (const char *).
5411
5412 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5413 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5414 However, I haven't made this a standard facility. The documentation doesn't
5415 mention it, and the Makefile doesn't support it.
5416
5417 5. The Makefile now contains an "install" target, with editable destinations at
5418 the top of the file. The pcretest program is not installed.
5419
5420 6. pgrep -V now gives the PCRE version number and date.
5421
5422 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5423 causing the entire string to be ignored, instead of just the last character.
5424
5425 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5426 non-matching string, it can take a very, very long time, even for strings of
5427 quite modest length, because of the nested recursion. PCRE now does better in
5428 some of these cases. It does this by remembering the last required literal
5429 character in the pattern, and pre-searching the subject to ensure it is present
5430 before running the real match. In other words, it applies a heuristic to detect
5431 some types of certain failure quickly, and in the above example, if presented
5432 with a string that has no trailing " it gives "no match" very quickly.
5433
5434 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5435 other alternatives are tried instead.
5436
5437
5438 Version 2.06 09-Jun-99
5439 ----------------------
5440
5441 1. Change pcretest's output for amount of store used to show just the code
5442 space, because the remainder (the data block) varies in size between 32-bit and
5443 64-bit systems.
5444
5445 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5446 start matching at. This allows lookbehinds to work when searching for multiple
5447 occurrences in a string.
5448
5449 3. Added additional options to pcretest for testing multiple occurrences:
5450
5451 /+ outputs the rest of the string that follows a match
5452 /g loops for multiple occurrences, using the new startoffset argument
5453 /G loops for multiple occurrences by passing an incremented pointer
5454
5455 4. PCRE wasn't doing the "first character" optimization for patterns starting
5456 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5457 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5458 the letter 'x'. On long subject strings, this gives a significant speed-up.
5459
5460
5461 Version 2.05 21-Apr-99
5462 ----------------------
5463
5464 1. Changed the type of magic_number from int to long int so that it works
5465 properly on 16-bit systems.
5466
5467 2. Fixed a bug which caused patterns starting with .* not to work correctly
5468 when the subject string contained newline characters. PCRE was assuming
5469 anchoring for such patterns in all cases, which is not correct because .* will
5470 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5471 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5472 must be retried after every newline in the subject.
5473
5474
5475 Version 2.04 18-Feb-99
5476 ----------------------
5477
5478 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5479 computation of the store needed to hold the pattern was incorrect (too large).
5480 If such patterns were nested a few deep, this could multiply and become a real
5481 problem.
5482
5483 2. Added /M option to pcretest to show the memory requirement of a specific
5484 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5485
5486 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5487 compiled in such a way that the backtracking after subsequent failure was
5488 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5489 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5490
5491
5492 Version 2.03 02-Feb-99
5493 ----------------------
5494
5495 1. Fixed typo and small mistake in man page.
5496
5497 2. Added 4th condition (GPL supersedes if conflict) and created separate
5498 LICENCE file containing the conditions.
5499
5500 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5501 Perl, that is the internal \ allows the delimiter to be included in the
5502 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5503 the final delimiter, add \ to the end of the pattern (to test the error).
5504
5505 4. Added the convenience functions for extracting substrings after a successful
5506 match. Updated pcretest to make it able to test these functions.
5507
5508
5509 Version 2.02 14-Jan-99
5510 ----------------------
5511
5512 1. Initialized the working variables associated with each extraction so that
5513 their saving and restoring doesn't refer to uninitialized store.
5514
5515 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5516 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5517 fix the problem.
5518
5519 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5520 calls, and wasn't printing the correct value for compiling calls. Increased the
5521 default value of LOOPREPEAT, and the number of significant figures in the
5522 times.
5523
5524 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5525
5526 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5527 a building problem on Windows NT with a FAT file system.
5528
5529
5530 Version 2.01 21-Oct-98
5531 ----------------------
5532
5533 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5534 to character tables built by pcre_maketables() in the current locale. If NULL
5535 is passed, the default tables are used.
5536
5537
5538 Version 2.00 24-Sep-98
5539 ----------------------
5540
5541 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5542 it any more.
5543
5544 2. Allow quantification of (?>) groups, and make it work correctly.
5545
5546 3. The first character computation wasn't working for (?>) groups.
5547
5548 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5549 end of the subject) and add 5.005's \z, which really does match only at the
5550 very end of the subject.
5551
5552 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5553
5554 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5555 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5556 localized options. All options to pcre_study() were also removed.
5557
5558 7. Add other new features from 5.005:
5559
5560 $(?<= positive lookbehind
5561 $(?<! negative lookbehind
5562 (?imsx-imsx) added the unsetting capability
5563 such a setting is global if at outer level; local otherwise
5564 (?imsx-imsx:) non-capturing groups with option setting
5565 (?(cond)re|re) conditional pattern matching
5566
5567 A backreference to itself in a repeated group matches the previous
5568 captured string.
5569
5570 8. General tidying up of studying (both automatic and via "study")
5571 consequential on the addition of new assertions.
5572
5573 9. As in 5.005, unlimited repeated groups that could match an empty substring
5574 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5575 runtime if any iteration does actually match an empty substring.
5576
5577 10. Include the RunTest script in the distribution.
5578
5579 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5580 discrepancies, some of which were old and were also with respect to 5.004. They
5581 have now been fixed.
5582
5583
5584 Version 1.09 28-Apr-98
5585 ----------------------
5586
5587 1. A negated single character class followed by a quantifier with a minimum
5588 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5589 program crashes, or just wrong answers. This did not apply to negated classes
5590 containing more than one character, or to minima other than one.
5591
5592
5593 Version 1.08 27-Mar-98
5594 ----------------------
5595
5596 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5597
5598 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5599 latter must appear before anything that relies on it in the pattern.
5600
5601
5602 Version 1.07 16-Feb-98
5603 ----------------------
5604
5605 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5606 repeat of a potentially empty string).
5607
5608
5609 Version 1.06 23-Jan-98
5610 ----------------------
5611
5612 1. Added Markus Oberhumer's little patches for C++.
5613
5614 2. Literal strings longer than 255 characters were broken.
5615
5616
5617 Version 1.05 23-Dec-97
5618 ----------------------
5619
5620 1. Negated character classes containing more than one character were failing if
5621 PCRE_CASELESS was set at run time.
5622
5623
5624 Version 1.04 19-Dec-97
5625 ----------------------
5626
5627 1. Corrected the man page, where some "const" qualifiers had been omitted.
5628
5629 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5630 input syntax.
5631
5632 3. Fixed memory leak which occurred when a regex with back references was
5633 matched with an offsets vector that wasn't big enough. The temporary memory
5634 that is used in this case wasn't being freed if the match failed.
5635
5636 4. Tidied pcretest to ensure it frees memory that it gets.
5637
5638 5. Temporary memory was being obtained in the case where the passed offsets
5639 vector was exactly big enough.
5640
5641 6. Corrected definition of offsetof() from change 5 below.
5642
5643 7. I had screwed up change 6 below and broken the rules for the use of
5644 setjmp(). Now fixed.
5645
5646
5647 Version 1.03 18-Dec-97
5648 ----------------------
5649
5650 1. A erroneous regex with a missing opening parenthesis was correctly
5651 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5652 on some systems.
5653
5654 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5655 it was reported that one broken compiler failed on the former because "code" is
5656 also an independent variable.
5657
5658 3. The erroneous regex a[]b caused an array overrun reference.
5659
5660 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5661 fail on data ending with that character. (It was going on too far, and checking
5662 the next character, typically a binary zero.) This was specific to the
5663 optimized code for single-character negative classes.
5664
5665 5. Added a contributed patch from the TIN world which does the following:
5666
5667 + Add an undef for memmove, in case the the system defines a macro for it.
5668
5669 + Add a definition of offsetof(), in case there isn't one. (I don't know
5670 the reason behind this - offsetof() is part of the ANSI standard - but
5671 it does no harm).
5672
5673 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5674 most of the places where whitespace preceded '#'. I have given up and
5675 allowed the remaining 2 cases to be at the margin.
5676
5677 + Rename some variables in pcre to eliminate shadowing. This seems very
5678 pedantic, but does no harm, of course.
5679
5680 6. Moved the call to setjmp() into its own function, to get rid of warnings
5681 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5682
5683 7. Constructs such as \d{8,} were compiling into the equivalent of
5684 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5685 outcome, but in this particular case used more store than had been allocated,
5686 which caused the bug to be discovered because it threw up an internal error.
5687
5688 8. The debugging code in both pcre and pcretest for outputting the compiled
5689 form of a regex was going wrong in the case of back references followed by
5690 curly-bracketed repeats.
5691
5692
5693 Version 1.02 12-Dec-97
5694 ----------------------
5695
5696 1. Typos in pcre.3 and comments in the source fixed.
5697
5698 2. Applied a contributed patch to get rid of places where it used to remove
5699 'const' from variables, and fixed some signed/unsigned and uninitialized
5700 variable warnings.
5701
5702 3. Added the "runtest" target to Makefile.
5703
5704 4. Set default compiler flag to -O2 rather than just -O.
5705
5706
5707 Version 1.01 19-Nov-97
5708 ----------------------
5709
5710 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5711 like /([ab]*)*/, that is, for classes with more than one character in them.
5712
5713 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5714 as /((?>a*))*/ (a PCRE_EXTRA facility).
5715
5716
5717 Version 1.00 18-Nov-97
5718 ----------------------
5719
5720 1. Added compile-time macros to support systems such as SunOS4 which don't have
5721 memmove() or strerror() but have other things that can be used instead.
5722
5723 2. Arranged that "make clean" removes the executables.
5724
5725
5726 Version 0.99 27-Oct-97
5727 ----------------------
5728
5729 1. Fixed bug in code for optimizing classes with only one character. It was
5730 initializing a 32-byte map regardless, which could cause it to run off the end
5731 of the memory it had got.
5732
5733 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5734
5735
5736 Version 0.98 22-Oct-97
5737 ----------------------
5738
5739 1. Fixed bug in code for handling temporary memory usage when there are more
5740 back references than supplied space in the ovector. This could cause segfaults.
5741
5742
5743 Version 0.97 21-Oct-97
5744 ----------------------
5745
5746 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5747
5748 2. Optimized negated single characters not to use a bit map.
5749
5750 3. Brought error texts together as macro definitions; clarified some of them;
5751 fixed one that was wrong - it said "range out of order" when it meant "invalid
5752 escape sequence".
5753
5754 4. Changed some char * arguments to const char *.
5755
5756 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5757
5758 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5759 pcretest.
5760
5761
5762 Version 0.96 16-Oct-97
5763 ----------------------
5764
5765 1. Added a simple "pgrep" utility to the distribution.
5766
5767 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5768 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5769 where "ddd" means "one or more decimal digits".
5770
5771 3. Fixed serious bug. If a pattern had a back reference, but the call to
5772 pcre_exec() didn't supply a large enough ovector to record the related
5773 identifying subpattern, the match always failed. PCRE now remembers the number
5774 of the largest back reference, and gets some temporary memory in which to save
5775 the offsets during matching if necessary, in order to ensure that
5776 backreferences always work.
5777
5778 4. Increased the compatibility with Perl in a number of ways:
5779
5780 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5781 to request this handling. The option can be set at compile or exec time.
5782
5783 (b) $ matches before a terminating newline by default; an option
5784 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5785 mode). The option can be set at compile or exec time.
5786
5787 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5788 the same as Perl's. If the decimal number it represents is less than 10
5789 or there aren't that many previous left capturing parentheses, an octal
5790 escape is read. Inside a character class, it's always an octal escape,
5791 even if it is a single digit.
5792
5793 (d) An escaped but undefined alphabetic character is taken as a literal,
5794 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5795 escapes.
5796
5797 (e) {0} is now permitted. (The previous item is removed from the compiled
5798 pattern).
5799
5800 5. Changed all the names of code files so that the basic parts are no longer
5801 than 10 characters, and abolished the teeny "globals.c" file.
5802
5803 6. Changed the handling of character classes; they are now done with a 32-byte
5804 bit map always.
5805
5806 7. Added the -d and /D options to pcretest to make it possible to look at the
5807 internals of compilation without having to recompile pcre.
5808
5809
5810 Version 0.95 23-Sep-97
5811 ----------------------
5812
5813 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5814 \x20 at the start of a run of normal characters. These were being treated as
5815 real characters, instead of the source characters being re-checked.
5816
5817
5818 Version 0.94 18-Sep-97
5819 ----------------------
5820
5821 1. The functions are now thread-safe, with the caveat that the global variables
5822 containing pointers to malloc() and free() or alternative functions are the
5823 same for all threads.
5824
5825 2. Get pcre_study() to generate a bitmap of initial characters for non-
5826 anchored patterns when this is possible, and use it if passed to pcre_exec().
5827
5828
5829 Version 0.93 15-Sep-97
5830 ----------------------
5831
5832 1. /(b)|(:+)/ was computing an incorrect first character.
5833
5834 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5835 but not actually doing anything yet.
5836
5837 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5838 as Perl does (e.g. [-az] or [az-]).
5839
5840 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5841 all possible positions.
5842
5843 5. Split up into different modules to avoid including unneeded functions in a
5844 compiled binary. However, compile and exec are still in one module. The "study"
5845 function is split off.
5846
5847 6. The character tables are now in a separate module whose source is generated
5848 by an auxiliary program - but can then be edited by hand if required. There are
5849 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5850 toupper() in the code.
5851
5852 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5853 make them global. Abolish the function for setting them, as the caller can now
5854 set them directly.
5855
5856
5857 Version 0.92 11-Sep-97
5858 ----------------------
5859
5860 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5861 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5862
5863 2. Caseless matching was not working in character classes if the characters in
5864 the pattern were in upper case.
5865
5866 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5867
5868 4. Make PCRE_ANCHORED public and accept as a compile option.
5869
5870 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5871 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5872 pass them.
5873
5874 6. Give an error if bad option bits passed at compile or run time.
5875
5876 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5877 pcretest to cause it to pass that flag.
5878
5879 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5880 options, and the first character, if set.
5881
5882 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5883
5884
5885 Version 0.91 10-Sep-97
5886 ----------------------
5887
5888 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5889 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5890
5891 2. PCRE was looping on encountering an indefinitely repeated back reference to
5892 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5893 Perl does - treats the match as successful.
5894
5895 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5