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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1593 - (show annotations)
Fri Aug 14 09:34:32 2015 UTC (4 years, 4 months ago) by zherczeg
File size: 283584 byte(s)
Error occurred while calculating annotation data.
Match limit check added to recursion.
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 recursion. This issue was found by Karl Skomski
136 with a custom LLVM fuzzer.
137
138
139 Version 8.37 28-April-2015
140 --------------------------
141
142 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
143 for those parentheses to be closed with whatever has been captured so far.
144 However, it was failing to mark any other groups between the hightest
145 capture so far and the currrent group as "unset". Thus, the ovector for
146 those groups contained whatever was previously there. An example is the
147 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
148
149 2. If an assertion condition was quantified with a minimum of zero (an odd
150 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
151
152 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
153 unrecognized modifier, a crash could occur.
154
155 4. An attempt to do global matching in pcretest with a zero-length ovector
156 caused a crash.
157
158 5. Fixed a memory leak during matching that could occur for a subpattern
159 subroutine call (recursive or otherwise) if the number of captured groups
160 that had to be saved was greater than ten.
161
162 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
163 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
164 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
165
166 7. A UTF pattern containing a "not" match of a non-ASCII character and a
167 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
168
169 8. When a pattern is compiled, it remembers the highest back reference so that
170 when matching, if the ovector is too small, extra memory can be obtained to
171 use instead. A conditional subpattern whose condition is a check on a
172 capture having happened, such as, for example in the pattern
173 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
174 setting the highest backreference number. This mattered only if pcre_exec()
175 was called with an ovector that was too small to hold the capture, and there
176 was no other kind of back reference (a situation which is probably quite
177 rare). The effect of the bug was that the condition was always treated as
178 FALSE when the capture could not be consulted, leading to a incorrect
179 behaviour by pcre_exec(). This bug has been fixed.
180
181 9. A reference to a duplicated named group (either a back reference or a test
182 for being set in a conditional) that occurred in a part of the pattern where
183 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
184 to be incorrectly calculated, leading to overwriting.
185
186 10. A mutually recursive set of back references such as (\2)(\1) caused a
187 segfault at study time (while trying to find the minimum matching length).
188 The infinite loop is now broken (with the minimum length unset, that is,
189 zero).
190
191 11. If an assertion that was used as a condition was quantified with a minimum
192 of zero, matching went wrong. In particular, if the whole group had
193 unlimited repetition and could match an empty string, a segfault was
194 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
195 assertions to be quantified, but not if they are being used as conditions,
196 so the above pattern is faulted by Perl. PCRE has now been changed so that
197 it also rejects such patterns.
198
199 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
200 failed to allow the zero-repeat case if pcre2_exec() was called with an
201 ovector too small to capture the group.
202
203 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
204 Red Hat Product Security:
205
206 (a) A crash if /K and /F were both set with the option to save the compiled
207 pattern.
208
209 (b) Another crash if the option to print captured substrings in a callout
210 was combined with setting a null ovector, for example \O\C+ as a subject
211 string.
212
213 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
214 forward reference repeated a large (but limited) number of times within a
215 repeated outer group that has a zero minimum quantifier, caused incorrect
216 code to be compiled, leading to the error "internal error:
217 previously-checked referenced subpattern not found" when an incorrect
218 memory address was read. This bug was reported as "heap overflow",
219 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
220 CVE-2015-2325.
221
222 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
223 call within a group that also contained a recursive back reference caused
224 incorrect code to be compiled. This bug was reported as "heap overflow",
225 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
226 number CVE-2015-2326.
227
228 24. Computing the size of the JIT read-only data in advance has been a source
229 of various issues, and new ones are still appear unfortunately. To fix
230 existing and future issues, size computation is eliminated from the code,
231 and replaced by on-demand memory allocation.
232
233 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
234 adjacent to the end of the range, and the range contained characters with
235 more than one other case, caused incorrect behaviour when compiled in UTF
236 mode. In that example, the range a-j was left out of the class.
237
238 26. Fix JIT compilation of conditional blocks, which assertion
239 is converted to (*FAIL). E.g: /(?(?!))/.
240
241 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
242 discovered by the LLVM fuzzer.
243
244 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
245 when this assertion was used as a condition, for example (?(?!)a|b). In
246 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
247 error about an unsupported item.
248
249 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
250 possessification code could take exponential time to complete. A recursion
251 depth limit of 1000 has been imposed to limit the resources used by this
252 optimization.
253
254 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
255 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
256 because \S ensures they are all in the class. The code for doing this was
257 interacting badly with the code for computing the amount of space needed to
258 compile the pattern, leading to a buffer overflow. This bug was discovered
259 by the LLVM fuzzer.
260
261 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
262 other kinds of group caused stack overflow at compile time. This bug was
263 discovered by the LLVM fuzzer.
264
265 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
266 between a subroutine call and its quantifier was incorrectly compiled,
267 leading to buffer overflow or other errors. This bug was discovered by the
268 LLVM fuzzer.
269
270 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
271 assertion after (?(. The code was failing to check the character after
272 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
273 was discovered by the LLVM fuzzer.
274
275 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
276 a fixed maximum following a group that contains a subroutine reference was
277 incorrectly compiled and could trigger buffer overflow. This bug was
278 discovered by the LLVM fuzzer.
279
280 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
281 caused a stack overflow instead of the diagnosis of a non-fixed length
282 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
283
284 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
285 (e.g. /(?<=\Ka)/) could make pcregrep loop.
286
287 37. There was a similar problem to 36 in pcretest for global matches.
288
289 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
290 and a subsequent item in the pattern caused a non-match, backtracking over
291 the repeated \X did not stop, but carried on past the start of the subject,
292 causing reference to random memory and/or a segfault. There were also some
293 other cases where backtracking after \C could crash. This set of bugs was
294 discovered by the LLVM fuzzer.
295
296 39. The function for finding the minimum length of a matching string could take
297 a very long time if mutual recursion was present many times in a pattern,
298 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
299 method has been implemented. This infelicity was discovered by the LLVM
300 fuzzer.
301
302 40. Static linking against the PCRE library using the pkg-config module was
303 failing on missing pthread symbols.
304
305
306 Version 8.36 26-September-2014
307 ------------------------------
308
309 1. Got rid of some compiler warnings in the C++ modules that were shown up by
310 -Wmissing-field-initializers and -Wunused-parameter.
311
312 2. The tests for quantifiers being too big (greater than 65535) were being
313 applied after reading the number, and stupidly assuming that integer
314 overflow would give a negative number. The tests are now applied as the
315 numbers are read.
316
317 3. Tidy code in pcre_exec.c where two branches that used to be different are
318 now the same.
319
320 4. The JIT compiler did not generate match limit checks for certain
321 bracketed expressions with quantifiers. This may lead to exponential
322 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
323 issue should be resolved now.
324
325 5. Fixed an issue, which occures when nested alternatives are optimized
326 with table jumps.
327
328 6. Inserted two casts and changed some ints to size_t in the light of some
329 reported 64-bit compiler warnings (Bugzilla 1477).
330
331 7. Fixed a bug concerned with zero-minimum possessive groups that could match
332 an empty string, which sometimes were behaving incorrectly in the
333 interpreter (though correctly in the JIT matcher). This pcretest input is
334 an example:
335
336 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
337 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
338
339 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
340 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
341 for an empty string was breaking the inner loop and carrying on at a lower
342 level, when possessive repeated groups should always return to a higher
343 level as they have no backtrack points in them. The empty string test now
344 occurs at the outer level.
345
346 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
347 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
348
349 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
350 doesn't).
351
352 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
353 to hold a pattern to be incorrectly computed (too small) when there were
354 named back references to duplicated names. This could cause "internal
355 error: code overflow" or "double free or corruption" or other memory
356 handling errors.
357
358 11. When named subpatterns had the same prefixes, back references could be
359 confused. For example, in this pattern:
360
361 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
362
363 the reference to 'Name' was incorrectly treated as a reference to a
364 duplicate name.
365
366 12. A pattern such as /^s?c/mi8 where the optional character has more than
367 one "other case" was incorrectly compiled such that it would only try to
368 match starting at "c".
369
370 13. When a pattern starting with \s was studied, VT was not included in the
371 list of possible starting characters; this should have been part of the
372 8.34/18 patch.
373
374 14. If a character class started [\Qx]... where x is any character, the class
375 was incorrectly terminated at the ].
376
377 15. If a pattern that started with a caseless match for a character with more
378 than one "other case" was studied, PCRE did not set up the starting code
379 unit bit map for the list of possible characters. Now it does. This is an
380 optimization improvement, not a bug fix.
381
382 16. The Unicode data tables have been updated to Unicode 7.0.0.
383
384 17. Fixed a number of memory leaks in pcregrep.
385
386 18. Avoid a compiler warning (from some compilers) for a function call with
387 a cast that removes "const" from an lvalue by using an intermediate
388 variable (to which the compiler does not object).
389
390 19. Incorrect code was compiled if a group that contained an internal recursive
391 back reference was optional (had quantifier with a minimum of zero). This
392 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
393 caused segmentation faults because of stack overflows at compile time.
394
395 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
396 group that is quantified with an indefinite repeat, caused a compile-time
397 loop which used up all the system stack and provoked a segmentation fault.
398 This was not the same bug as 19 above.
399
400 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
401 Patch by Mike Frysinger.
402
403
404 Version 8.35 04-April-2014
405 --------------------------
406
407 1. A new flag is set, when property checks are present in an XCLASS.
408 When this flag is not set, PCRE can perform certain optimizations
409 such as studying these XCLASS-es.
410
411 2. The auto-possessification of character sets were improved: a normal
412 and an extended character set can be compared now. Furthermore
413 the JIT compiler optimizes more character set checks.
414
415 3. Got rid of some compiler warnings for potentially uninitialized variables
416 that show up only when compiled with -O2.
417
418 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
419 of a match later then the end of the match. The pcretest program was not
420 handling the case sensibly - it was outputting from the start to the next
421 binary zero. It now reports this situation in a message, and outputs the
422 text from the end to the start.
423
424 5. Fast forward search is improved in JIT. Instead of the first three
425 characters, any three characters with fixed position can be searched.
426 Search order: first, last, middle.
427
428 6. Improve character range checks in JIT. Characters are read by an inprecise
429 function now, which returns with an unknown value if the character code is
430 above a certain threshold (e.g: 256). The only limitation is that the value
431 must be bigger than the threshold as well. This function is useful when
432 the characters above the threshold are handled in the same way.
433
434 7. The macros whose names start with RAWUCHAR are placeholders for a future
435 mode in which only the bottom 21 bits of 32-bit data items are used. To
436 make this more memorable for those maintaining the code, the names have
437 been changed to start with UCHAR21, and an extensive comment has been added
438 to their definition.
439
440 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
441 to the export list in Makefile.am (they were accidentally omitted from the
442 8.34 tarball).
443
444 9. The informational output from pcretest used the phrase "starting byte set"
445 which is inappropriate for the 16-bit and 32-bit libraries. As the output
446 for "first char" and "need char" really means "non-UTF-char", I've changed
447 "byte" to "char", and slightly reworded the output. The documentation about
448 these values has also been (I hope) clarified.
449
450 10. Another JIT related optimization: use table jumps for selecting the correct
451 backtracking path, when more than four alternatives are present inside a
452 bracket.
453
454 11. Empty match is not possible, when the minimum length is greater than zero,
455 and there is no \K in the pattern. JIT should avoid empty match checks in
456 such cases.
457
458 12. In a caseless character class with UCP support, when a character with more
459 than one alternative case was not the first character of a range, not all
460 the alternative cases were added to the class. For example, s and \x{17f}
461 are both alternative cases for S: the class [RST] was handled correctly,
462 but [R-T] was not.
463
464 13. The configure.ac file always checked for pthread support when JIT was
465 enabled. This is not used in Windows, so I have put this test inside a
466 check for the presence of windows.h (which was already tested for).
467
468 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
469 The algorithm provides a way to skip certain starting offsets, and usually
470 faster than linear prefix searches.
471
472 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
473 as for 'fr_FR' and 'french'. For some reason, however, it then used the
474 Windows-specific input and output files, which have 'french' screwed in.
475 So this could never have worked. One of the problems with locales is that
476 they aren't always the same. I have now updated RunTest so that it checks
477 the output of the locale test (test 3) against three different output
478 files, and it allows the test to pass if any one of them matches. With luck
479 this should make the test pass on some versions of Solaris where it was
480 failing. Because of the uncertainty, the script did not used to stop if
481 test 3 failed; it now does. If further versions of a French locale ever
482 come to light, they can now easily be added.
483
484 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
485 there was a message during ./configure, but it did not stop. This now
486 provokes an error. The invalid example in README has been corrected.
487 If a value less than the minimum is given, the minimum value has always
488 been used, but now a warning is given.
489
490 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
491 was a bug in the test system, which is now fixed. Also, the list of various
492 configurations that are tested for each release did not have one with both
493 16/32 bits and --enable-bar-anycrlf. It now does.
494
495 18. pcretest was missing "-C bsr" for displaying the \R default setting.
496
497 19. Little endian PowerPC systems are supported now by the JIT compiler.
498
499 20. The fast forward newline mechanism could enter to an infinite loop on
500 certain invalid UTF-8 input. Although we don't support these cases
501 this issue can be fixed by a performance optimization.
502
503 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
504 not take account if existing stack usage. There is now a new global
505 variable called pcre_stack_guard that can be set to point to an external
506 function to check stack availability. It is called at the start of
507 processing every parenthesized group.
508
509 22. A typo in the code meant that in ungreedy mode the max/min qualifier
510 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
511 not match "ab".
512
513 23. When UTF was disabled, the JIT program reported some incorrect compile
514 errors. These messages are silenced now.
515
516 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
517 compiler.
518
519 25. Change all the temporary files used in RunGrepTest to be different to those
520 used by RunTest so that the tests can be run simultaneously, for example by
521 "make -j check".
522
523
524 Version 8.34 15-December-2013
525 -----------------------------
526
527 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
528 executable memory. Patch inspired by Carsten Klein.
529
530 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
531 this macro is never tested and has no effect, because the work to support
532 coverage involves only compiling and linking options and special targets in
533 the Makefile. The comment in config.h implied that defining the macro would
534 enable coverage support, which is totally false. There was also support for
535 setting this macro in the CMake files (my fault, I just copied it from
536 configure). SUPPORT_GCOV has now been removed.
537
538 3. Make a small performance improvement in strlen16() and strlen32() in
539 pcretest.
540
541 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
542 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
543 these cases). There was also one in pcretest.c.
544
545 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
546
547 6. In UTF mode, the code for checking whether a group could match an empty
548 string (which is used for indefinitely repeated groups to allow for
549 breaking an infinite loop) was broken when the group contained a repeated
550 negated single-character class with a character that occupied more than one
551 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
552 UTF-8 mode). The effect was undefined: the group might or might not be
553 deemed as matching an empty string, or the program might have crashed.
554
555 7. The code for checking whether a group could match an empty string was not
556 recognizing that \h, \H, \v, \V, and \R must match a character.
557
558 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
559 an empty string. If it can, pcretest shows this in its information output.
560
561 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
562 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
563 matched by pcre_exec() without using JIT:
564
565 (a) If the rest of the pattern did not match after a maximal run of
566 grapheme clusters, the code for backing up to try with fewer of them
567 did not always back up over a full grapheme when characters that do not
568 have the modifier quality were involved, e.g. Hangul syllables.
569
570 (b) If the match point in a subject started with modifier character, and
571 there was no match, the code could incorrectly back up beyond the match
572 point, and potentially beyond the first character in the subject,
573 leading to a segfault or an incorrect match result.
574
575 10. A conditional group with an assertion condition could lead to PCRE
576 recording an incorrect first data item for a match if no other first data
577 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
578 first data item, and therefore matched "ca" after "c" instead of at the
579 start.
580
581 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
582 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
583
584 12. The source of pcregrep now includes z/OS-specific code so that it can be
585 compiled for z/OS as part of the special z/OS distribution.
586
587 13. Added the -T and -TM options to pcretest.
588
589 14. The code in pcre_compile.c for creating the table of named capturing groups
590 has been refactored. Instead of creating the table dynamically during the
591 actual compiling pass, the information is remembered during the pre-compile
592 pass (on the stack unless there are more than 20 named groups, in which
593 case malloc() is used) and the whole table is created before the actual
594 compile happens. This has simplified the code (it is now nearly 150 lines
595 shorter) and prepared the way for better handling of references to groups
596 with duplicate names.
597
598 15. A back reference to a named subpattern when there is more than one of the
599 same name now checks them in the order in which they appear in the pattern.
600 The first one that is set is used for the reference. Previously only the
601 first one was inspected. This change makes PCRE more compatible with Perl.
602
603 16. Unicode character properties were updated from Unicode 6.3.0.
604
605 17. The compile-time code for auto-possessification has been refactored, based
606 on a patch by Zoltan Herczeg. It now happens after instead of during
607 compilation. The code is cleaner, and more cases are handled. The option
608 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
609 options in pcretest are provided to set it. It can also be set by
610 (*NO_AUTO_POSSESS) at the start of a pattern.
611
612 18. The character VT has been added to the default ("C" locale) set of
613 characters that match \s and are generally treated as white space,
614 following this same change in Perl 5.18. There is now no difference between
615 "Perl space" and "POSIX space". Whether VT is treated as white space in
616 other locales depends on the locale.
617
618 19. The code for checking named groups as conditions, either for being set or
619 for being recursed, has been refactored (this is related to 14 and 15
620 above). Processing unduplicated named groups should now be as fast at
621 numerical groups, and processing duplicated groups should be faster than
622 before.
623
624 20. Two patches to the CMake build system, by Alexander Barkov:
625
626 (1) Replace the "source" command by "." in CMakeLists.txt because
627 "source" is a bash-ism.
628
629 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
630 without these the CMake build does not work on Solaris.
631
632 21. Perl has changed its handling of \8 and \9. If there is no previously
633 encountered capturing group of those numbers, they are treated as the
634 literal characters 8 and 9 instead of a binary zero followed by the
635 literals. PCRE now does the same.
636
637 22. Following Perl, added \o{} to specify codepoints in octal, making it
638 possible to specify values greater than 0777 and also making them
639 unambiguous.
640
641 23. Perl now gives an error for missing closing braces after \x{... instead of
642 treating the string as literal. PCRE now does the same.
643
644 24. RunTest used to grumble if an inappropriate test was selected explicitly,
645 but just skip it when running all tests. This make it awkward to run ranges
646 of tests when one of them was inappropriate. Now it just skips any
647 inappropriate tests, as it always did when running all tests.
648
649 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
650 character types such as \d or \w, too many callouts were inserted, and the
651 data that they returned was rubbish.
652
653 26. In UCP mode, \s was not matching two of the characters that Perl matches,
654 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
655 were matched by \h. The code has now been refactored so that the lists of
656 the horizontal and vertical whitespace characters used for \h and \v (which
657 are defined only in one place) are now also used for \s.
658
659 27. Add JIT support for the 64 bit TileGX architecture.
660 Patch by Jiong Wang (Tilera Corporation).
661
662 28. Possessive quantifiers for classes (both explicit and automatically
663 generated) now use special opcodes instead of wrapping in ONCE brackets.
664
665 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
666 (because it's meaningless), this was not happening when PCRE_CASELESS was
667 set. Not wrong, but inefficient.
668
669 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
670 properties for \w, \d, etc) is present in a test regex. Otherwise if the
671 test contains no characters greater than 255, Perl doesn't realise it
672 should be using Unicode semantics.
673
674 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
675 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
676 does in Unicode mode.
677
678 32. Added the "forbid" facility to pcretest so that putting tests into the
679 wrong test files can sometimes be quickly detected.
680
681 33. There is now a limit (default 250) on the depth of nesting of parentheses.
682 This limit is imposed to control the amount of system stack used at compile
683 time. It can be changed at build time by --with-parens-nest-limit=xxx or
684 the equivalent in CMake.
685
686 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
687 errors. Perl warns for these when in warning mode, but PCRE has no facility
688 for giving warnings.
689
690 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
691 However, this was not working for (?!) because it is optimized to (*FAIL),
692 for which PCRE does not allow quantifiers. The optimization is now disabled
693 when a quantifier follows (?!). I can't see any use for this, but it makes
694 things uniform.
695
696 36. Perl no longer allows group names to start with digits, so I have made this
697 change also in PCRE. It simplifies the code a bit.
698
699 37. In extended mode, Perl ignores spaces before a + that indicates a
700 possessive quantifier. PCRE allowed a space before the quantifier, but not
701 before the possessive +. It now does.
702
703 38. The use of \K (reset reported match start) within a repeated possessive
704 group such as (a\Kb)*+ was not working.
705
706 40. Document that the same character tables must be used at compile time and
707 run time, and that the facility to pass tables to pcre_exec() and
708 pcre_dfa_exec() is for use only with saved/restored patterns.
709
710 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
711 features for Builds with MSVC:
712
713 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
714 builds). The .rc files can be used to set FileDescription and many other
715 attributes.
716
717 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
718 This allows higher-level build scripts which want .pdb files to avoid
719 hard-coding the exact files needed."
720
721 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
722 mean "start of word" and "end of word", respectively, as a transition aid.
723
724 43. A minimizing repeat of a class containing codepoints greater than 255 in
725 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
726 compiled to use the heap for recursion.
727
728 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
729 is configured.
730
731
732 Version 8.33 28-May-2013
733 ------------------------
734
735 1. Added 'U' to some constants that are compared to unsigned integers, to
736 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
737 variables that are added to signed variables, to ensure the result is
738 signed and can be negated.
739
740 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
741 CMake config files.
742
743 3. Revise the creation of config.h.generic so that all boolean macros are
744 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
745 overriding via -D on the command line possible.
746
747 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
748 to unsigned int is reported to make a quite noticeable speed difference in
749 a specific Windows environment. Testing on Linux did also appear to show
750 some benefit (and it is clearly not harmful). Also fixed the definition of
751 Xop which should be unsigned.
752
753 5. Related to (4), changing the definition of the intermediate variable cc
754 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
755 improvements.
756
757 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
758 unnecessary spaces.
759
760 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
761 and later.
762
763 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
764
765 9. Optimizing fast_forward_start_bits in JIT.
766
767 10. Adding support for callouts in JIT, and fixing some issues revealed
768 during this work. Namely:
769
770 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
771
772 (b) Minimum length was not checked before the matching is started.
773
774 11. The value of capture_last that is passed to callouts was incorrect in some
775 cases when there was a capture on one path that was subsequently abandoned
776 after a backtrack. Also, the capture_last value is now reset after a
777 recursion, since all captures are also reset in this case.
778
779 12. The interpreter no longer returns the "too many substrings" error in the
780 case when an overflowing capture is in a branch that is subsequently
781 abandoned after a backtrack.
782
783 13. In the pathological case when an offset vector of size 2 is used, pcretest
784 now prints out the matched string after a yield of 0 or 1.
785
786 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
787 Only supported by the JIT compiler at the moment.
788
789 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
790
791 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
792 offset of the starting point of the matching process, provided the offsets
793 vector is large enough.
794
795 17. The \A escape now records a lookbehind value of 1, though its execution
796 does not actually inspect the previous character. This is to ensure that,
797 in partial multi-segment matching, at least one character from the old
798 segment is retained when a new segment is processed. Otherwise, if there
799 are no lookbehinds in the pattern, \A might match incorrectly at the start
800 of a new segment.
801
802 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
803
804 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
805 gives some modest performance improvement in 8-bit mode.
806
807 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
808 be expressed in certain programming languages using Universal Character
809 Names.
810
811 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
812 which points out that "non characters" are not "characters that may not
813 appear in Unicode strings" but rather "characters that are reserved for
814 internal use and have only local meaning".
815
816 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
817 there was a conditional group that depended on an assertion, if the
818 assertion was false, the callout that immediately followed the alternation
819 in the condition was skipped when pcre_exec() was used for matching.
820
821 23. Allow an explicit callout to be inserted before an assertion that is the
822 condition for a conditional group, for compatibility with automatic
823 callouts, which always insert a callout at this point.
824
825 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
826 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
827
828 25. (*PRUNE) is now supported by the JIT compiler.
829
830 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
831
832 27. Fix the case where there are two or more SKIPs with arguments that may be
833 ignored.
834
835 28. (*SKIP) is now supported by the JIT compiler.
836
837 29. (*THEN) is now supported by the JIT compiler.
838
839 30. Update RunTest with additional test selector options.
840
841 31. The way PCRE handles backtracking verbs has been changed in two ways.
842
843 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
844 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
845 backtracking. In some cases this makes it more Perl-compatible, but Perl's
846 rather obscure rules do not always do the same thing.
847
848 (2) Previously, backtracking verbs were confined within assertions. This is
849 no longer the case for positive assertions, except for (*ACCEPT). Again,
850 this sometimes improves Perl compatibility, and sometimes does not.
851
852 32. A number of tests that were in test 2 because Perl did things differently
853 have been moved to test 1, because either Perl or PCRE has changed, and
854 these tests are now compatible.
855
856 32. Backtracking control verbs are now handled in the same way in JIT and
857 interpreter.
858
859 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
860 contained a forward subroutine reference caused a compile error.
861
862 34. Auto-detect and optimize limited repetitions in JIT.
863
864 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
865 blocking (*UTF) etc.
866
867 36. In the interpreter, maximizing pattern repetitions for characters and
868 character types now use tail recursion, which reduces stack usage.
869
870 37. The value of the max lookbehind was not correctly preserved if a compiled
871 and saved regex was reloaded on a host of different endianness.
872
873 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
874 of the compiled pattern block, expand the flags field from 16 to 32 bits
875 because it was almost full.
876
877 39. Try madvise first before posix_madvise.
878
879 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
880 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
881 original change locked out all matches of empty strings. This has been
882 changed so that one match of an empty string per line is recognized.
883 Subsequent searches on the same line (for colouring or for --only-matching,
884 for example) do not recognize empty strings.
885
886 41. Applied a user patch to fix a number of spelling mistakes in comments.
887
888 42. Data lines longer than 65536 caused pcretest to crash.
889
890 43. Clarified the data type for length and startoffset arguments for pcre_exec
891 and pcre_dfa_exec in the function-specific man pages, where they were
892 explicitly stated to be in bytes, never having been updated. I also added
893 some clarification to the pcreapi man page.
894
895 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
896 a segmentation fault.
897
898
899 Version 8.32 30-November-2012
900 -----------------------------
901
902 1. Improved JIT compiler optimizations for first character search and single
903 character iterators.
904
905 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
906 Patch by Daniel Richard G.
907
908 3. Single character iterator optimizations in the JIT compiler.
909
910 4. Improved JIT compiler optimizations for character ranges.
911
912 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
913 Reported by Giuseppe D'Angelo.
914
915 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
916 of a line, was being set incorrectly in cases where .* appeared inside
917 atomic brackets at the start of a pattern, or where there was a subsequent
918 *PRUNE or *SKIP.
919
920 7. Improved instruction cache flush for POWER/PowerPC.
921 Patch by Daniel Richard G.
922
923 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
924 grep:
925
926 (a) There is now no limit to the number of patterns to be matched.
927
928 (b) An error is given if a pattern is too long.
929
930 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
931 are now supported.
932
933 (d) --exclude-from and --include-from (multiple use) have been added.
934
935 (e) Exclusions and inclusions now apply to all files and directories, not
936 just to those obtained from scanning a directory recursively.
937
938 (f) Multiple uses of -f and --file-list are now supported.
939
940 (g) In a Windows environment, the default for -d has been changed from
941 "read" (the GNU grep default) to "skip", because otherwise the presence
942 of a directory in the file list provokes an error.
943
944 (h) The documentation has been revised and clarified in places.
945
946 9. Improve the matching speed of capturing brackets.
947
948 10. Changed the meaning of \X so that it now matches a Unicode extended
949 grapheme cluster.
950
951 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
952 out POSIX threads when JIT support is configured.
953
954 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
955
956 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
957 structure could go wrong in environments where size_t is not the same size
958 as int.
959
960 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
961 set.
962
963 15. The EBCDIC support had decayed; later updates to the code had included
964 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
965 general tidy up of EBCDIC-related issues, and the documentation was also
966 not quite right. There is now a test that can be run on ASCII systems to
967 check some of the EBCDIC-related things (but is it not a full test).
968
969 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
970 in a small tidy to the code.
971
972 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
973
974 18. If the --only-matching (-o) option in pcregrep is specified multiple
975 times, each one causes appropriate output. For example, -o1 -o2 outputs the
976 substrings matched by the 1st and 2nd capturing parentheses. A separating
977 string can be specified by --om-separator (default empty).
978
979 19. Improving the first n character searches.
980
981 20. Turn case lists for horizontal and vertical white space into macros so that
982 they are defined only once.
983
984 21. This set of changes together give more compatible Unicode case-folding
985 behaviour for characters that have more than one other case when UCP
986 support is available.
987
988 (a) The Unicode property table now has offsets into a new table of sets of
989 three or more characters that are case-equivalent. The MultiStage2.py
990 script that generates these tables (the pcre_ucd.c file) now scans
991 CaseFolding.txt instead of UnicodeData.txt for character case
992 information.
993
994 (b) The code for adding characters or ranges of characters to a character
995 class has been abstracted into a generalized function that also handles
996 case-independence. In UTF-mode with UCP support, this uses the new data
997 to handle characters with more than one other case.
998
999 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1000 whose other case is greater than 256 are now correctly matched
1001 caselessly. Previously, the high codepoint matched the low one, but not
1002 vice versa.
1003
1004 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1005 of the new class addition function, using character lists defined as
1006 macros alongside the case definitions of 20 above.
1007
1008 (e) Caseless back references now work with characters that have more than
1009 one other case.
1010
1011 (f) General caseless matching of characters with more than one other case
1012 is supported.
1013
1014 22. Unicode character properties were updated from Unicode 6.2.0
1015
1016 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1017
1018 24. Add support for 32-bit character strings, and UTF-32
1019
1020 25. Major JIT compiler update (code refactoring and bugfixing).
1021 Experimental Sparc 32 support is added.
1022
1023 26. Applied a modified version of Daniel Richard G's patch to create
1024 pcre.h.generic and config.h.generic by "make" instead of in the
1025 PrepareRelease script.
1026
1027 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1028 pcre_compile.c when checking for a zero character.
1029
1030 28. Introducing a native interface for JIT. Through this interface, the compiled
1031 machine code can be directly executed. The purpose of this interface is to
1032 provide fast pattern matching, so several sanity checks are not performed.
1033 However, feature tests are still performed. The new interface provides
1034 1.4x speedup compared to the old one.
1035
1036 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1037 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1038 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1039
1040 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1041 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1042 the "old" RFC 2279). Instead, it ended up passing a negative length to
1043 pcre_exec().
1044
1045 31. Add support for GCC's visibility feature to hide internal functions.
1046
1047 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1048 "unknown -C option" after outputting 0 or 1.
1049
1050 33. There is now support for generating a code coverage report for the test
1051 suite in environments where gcc is the compiler and lcov is installed. This
1052 is mainly for the benefit of the developers.
1053
1054 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1055 unaddressable using valgrind annotations, allowing valgrind to detect
1056 invalid memory accesses. This is mainly for the benefit of the developers.
1057
1058 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1059
1060 26. Give configure error if --enable-cpp but no C++ compiler found.
1061
1062
1063 Version 8.31 06-July-2012
1064 -------------------------
1065
1066 1. Fixing a wrong JIT test case and some compiler warnings.
1067
1068 2. Removed a bashism from the RunTest script.
1069
1070 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1071 to unsigned type, result still unsigned" that was given by an MS compiler
1072 on encountering the code "-sizeof(xxx)".
1073
1074 4. Partial matching support is added to the JIT compiler.
1075
1076 5. Fixed several bugs concerned with partial matching of items that consist
1077 of more than one character:
1078
1079 (a) /^(..)\1/ did not partially match "aba" because checking references was
1080 done on an "all or nothing" basis. This also applied to repeated
1081 references.
1082
1083 (b) \R did not give a hard partial match if \r was found at the end of the
1084 subject.
1085
1086 (c) \X did not give a hard partial match after matching one or more
1087 characters at the end of the subject.
1088
1089 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1090 a partial match for the string "\r".
1091
1092 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1093 a partial match for a CR character at the end of the subject string.
1094
1095 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1096 running pcretest, the text "(JIT)" added to the output whenever JIT is
1097 actually used to run the match.
1098
1099 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1100 or /S+[+] with a digit between 1 and 7.
1101
1102 8. OP_NOT now supports any UTF character not just single-byte ones.
1103
1104 9. (*MARK) control verb is now supported by the JIT compiler.
1105
1106 10. The command "./RunTest list" lists the available tests without actually
1107 running any of them. (Because I keep forgetting what they all are.)
1108
1109 11. Add PCRE_INFO_MAXLOOKBEHIND.
1110
1111 12. Applied a (slightly modified) user-supplied patch that improves performance
1112 when the heap is used for recursion (compiled with --disable-stack-for-
1113 recursion). Instead of malloc and free for each heap frame each time a
1114 logical recursion happens, frames are retained on a chain and re-used where
1115 possible. This sometimes gives as much as 30% improvement.
1116
1117 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1118 call.
1119
1120 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1121
1122 15. It is now possible to link pcretest with libedit as an alternative to
1123 libreadline.
1124
1125 16. (*COMMIT) control verb is now supported by the JIT compiler.
1126
1127 17. The Unicode data tables have been updated to Unicode 6.1.0.
1128
1129 18. Added --file-list option to pcregrep.
1130
1131 19. Added binary file support to pcregrep, including the -a, --binary-files,
1132 -I, and --text options.
1133
1134 20. The madvise function is renamed for posix_madvise for QNX compatibility
1135 reasons. Fixed by Giuseppe D'Angelo.
1136
1137 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1138
1139 22. Changed the option for creating symbolic links for 16-bit man pages from
1140 -s to -sf so that re-installing does not cause issues.
1141
1142 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1143
1144 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1145 matches in certain environments (the workspace was not being correctly
1146 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1147 some of the workspace data at the beginning of a restart.
1148
1149 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1150 was not doing so when it should - probably a typo introduced by SVN 528
1151 (change 8.10/14).
1152
1153 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1154 \w+ when the character tables indicated that \x{c4} was a word character.
1155 There were several related cases, all because the tests for doing a table
1156 lookup were testing for characters less than 127 instead of 255.
1157
1158 27. If a pattern contains capturing parentheses that are not used in a match,
1159 their slots in the ovector are set to -1. For those that are higher than
1160 any matched groups, this happens at the end of processing. In the case when
1161 there were back references that the ovector was too small to contain
1162 (causing temporary malloc'd memory to be used during matching), and the
1163 highest capturing number was not used, memory off the end of the ovector
1164 was incorrectly being set to -1. (It was using the size of the temporary
1165 memory instead of the true size.)
1166
1167 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1168 ovector size, it uses memory at the end of the block that it has got.
1169
1170 29. Check for an overlong MARK name and give an error at compile time. The
1171 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1172
1173 30. JIT compiler update.
1174
1175 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1176 Rill for the patch.
1177
1178 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1179
1180 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1181
1182 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1183 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1184 was enabled.
1185
1186 35. Improve JIT code generation for greedy plus quantifier.
1187
1188 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1189 1 to "aa" instead of to an empty string. The bug affected repeated groups
1190 that could potentially match an empty string.
1191
1192 37. Optimizing single character iterators in JIT.
1193
1194 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1195 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1196 codepoints that are too big for the mode are faulted, and in a UTF mode,
1197 disallowed codepoints are also faulted.
1198
1199 39. If PCRE was compiled with UTF support, in three places in the DFA
1200 matcher there was code that should only have been obeyed in UTF mode, but
1201 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1202 processing when bytes with values greater than 127 were present. In 16-bit
1203 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1204 both cases the values are those that cannot be the first data item in a UTF
1205 character. The three items that might have provoked this were recursions,
1206 possessively repeated groups, and atomic groups.
1207
1208 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1209 and pcregrep, because some OS require shared objects to be explicitly
1210 passed to ld, causing the link step to fail if they are not.
1211
1212 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1213 mode, patterns that started with \h* or \R* might be incorrectly matched.
1214
1215
1216 Version 8.30 04-February-2012
1217 -----------------------------
1218
1219 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1220 name is defined in ctype.h.
1221
1222 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1223 only in quite long subpatterns.
1224
1225 3. Removed the function pcre_info(), which has been obsolete and deprecated
1226 since it was replaced by pcre_fullinfo() in February 2000.
1227
1228 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1229 match a (*MARK), and the match failed at the start of the subject, a
1230 reference to memory before the start of the subject could occur. This bug
1231 was introduced by fix 17 of release 8.21.
1232
1233 5. A reference to an unset group with zero minimum repetition was giving
1234 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1235 /(another)?(\1?)test/ matched against "hello world test". This bug was
1236 introduced in release 8.13.
1237
1238 6. Add support for 16-bit character strings (a large amount of work involving
1239 many changes and refactorings).
1240
1241 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1242 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1243 from a file.
1244
1245 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1246 rounding is not applied in this particular case).
1247
1248 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1249 if they appear, or are escaped, in patterns.
1250
1251 10. Get rid of a number of -Wunused-but-set-variable warnings.
1252
1253 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1254 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1255 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1256 also returns the mark "x". This bug applied to capturing parentheses,
1257 non-capturing parentheses, and atomic parentheses. It also applied to some
1258 assertions.
1259
1260 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1261 information out of configure.ac instead of relying on pcre.h.generic, which
1262 is not stored in the repository.
1263
1264 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1265 -lreadline.
1266
1267 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1268
1269 15. Applied Graycode's patch to put the top-level frame on the stack rather
1270 than the heap when not using the stack for recursion. This gives a
1271 performance improvement in many cases when recursion is not deep.
1272
1273 16. Experimental code added to "pcretest -C" to output the stack frame size.
1274
1275
1276 Version 8.21 12-Dec-2011
1277 ------------------------
1278
1279 1. Updating the JIT compiler.
1280
1281 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1282 are added as well.
1283
1284 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1285 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1286 calling _pcre_jit_exec. Some extra comments are added.
1287
1288 4. (*MARK) settings inside atomic groups that do not contain any capturing
1289 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1290 was introduced by change 18 for 8.20.
1291
1292 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1293 ECMA-262 standard.
1294
1295 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1296 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1297 This bug was probably introduced by change 9 of 8.13.
1298
1299 7. While fixing 6 above, I noticed that a number of other items were being
1300 incorrectly rejected as "not fixed length". This arose partly because newer
1301 opcodes had not been added to the fixed-length checking code. I have (a)
1302 corrected the bug and added tests for these items, and (b) arranged for an
1303 error to occur if an unknown opcode is encountered while checking for fixed
1304 length instead of just assuming "not fixed length". The items that were
1305 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1306 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1307 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1308
1309 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1310 being incorrectly compiled and would have given unpredicatble results.
1311
1312 9. A possessively repeated subpattern with minimum repeat count greater than
1313 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1314 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1315 the first (A) could occur when it should not.
1316
1317 10. Add a cast and remove a redundant test from the code.
1318
1319 11. JIT should use pcre_malloc/pcre_free for allocation.
1320
1321 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1322 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1323 is anything other than /usr, -L is still shown).
1324
1325 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1326
1327 14. Perl does not support \N without a following name in a [] class; PCRE now
1328 also gives an error.
1329
1330 15. If a forward reference was repeated with an upper limit of around 2000,
1331 it caused the error "internal error: overran compiling workspace". The
1332 maximum number of forward references (including repeats) was limited by the
1333 internal workspace, and dependent on the LINK_SIZE. The code has been
1334 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1335 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1336 of around 200,000 forward references. While doing this, I also speeded up
1337 the filling in of repeated forward references.
1338
1339 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1340 incorrectly expecting the subject to contain another "a" after the start.
1341
1342 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1343 in the match, the SKIP should be ignored. This was not happening; instead
1344 the SKIP was being treated as NOMATCH. For patterns such as
1345 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1346 tested.
1347
1348 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1349 now much more compatible with Perl, in particular in cases where the result
1350 is a non-match for a non-anchored pattern. For example, if
1351 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1352 "m", where previously it did not return a name. A side effect of this
1353 change is that for partial matches, the last encountered mark name is
1354 returned, as for non matches. A number of tests that were previously not
1355 Perl-compatible have been moved into the Perl-compatible test files. The
1356 refactoring has had the pleasing side effect of removing one argument from
1357 the match() function, thus reducing its stack requirements.
1358
1359 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1360 subsequent uses of /S (without +) incorrectly behaved like /S+.
1361
1362 21. Retrieve executable code size support for the JIT compiler and fixing
1363 some warnings.
1364
1365 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1366 not work when the shorter character appeared right at the end of the
1367 subject string.
1368
1369 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1370 systems.
1371
1372 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1373 output it when the /M option is used in pcretest.
1374
1375 25. The CheckMan script was not being included in the distribution. Also, added
1376 an explicit "perl" to run Perl scripts from the PrepareRelease script
1377 because this is reportedly needed in Windows.
1378
1379 26. If study data was being save in a file and studying had not found a set of
1380 "starts with" bytes for the pattern, the data written to the file (though
1381 never used) was taken from uninitialized memory and so caused valgrind to
1382 complain.
1383
1384 27. Updated RunTest.bat as provided by Sheri Pierce.
1385
1386 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1387
1388 29. Computation of memory usage for the table of capturing group names was
1389 giving an unnecessarily large value.
1390
1391
1392 Version 8.20 21-Oct-2011
1393 ------------------------
1394
1395 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1396 a POSIX class. After further experiments with Perl, which convinced me that
1397 Perl has bugs and confusions, a closing square bracket is no longer allowed
1398 in a POSIX name. This bug also affected patterns with classes that started
1399 with full stops.
1400
1401 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1402 captured substring, but while checking the failing first alternative,
1403 substring 1 is temporarily captured. If the output vector supplied to
1404 pcre_exec() was not big enough for this capture, the yield of the function
1405 was still zero ("insufficient space for captured substrings"). This cannot
1406 be totally fixed without adding another stack variable, which seems a lot
1407 of expense for a edge case. However, I have improved the situation in cases
1408 such as /(a)(b)x|abc/ matched against "abc", where the return code
1409 indicates that fewer than the maximum number of slots in the ovector have
1410 been set.
1411
1412 3. Related to (2) above: when there are more back references in a pattern than
1413 slots in the output vector, pcre_exec() uses temporary memory during
1414 matching, and copies in the captures as far as possible afterwards. It was
1415 using the entire output vector, but this conflicts with the specification
1416 that only 2/3 is used for passing back captured substrings. Now it uses
1417 only the first 2/3, for compatibility. This is, of course, another edge
1418 case.
1419
1420 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1421 main code base, and can be used by building with --enable-jit. When this is
1422 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1423 runtime --no-jit option is given.
1424
1425 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1426 ovector, the return from the function was zero, implying that there were
1427 other matches that did not fit. The correct "exactly full" value is now
1428 returned.
1429
1430 6. If a subpattern that was called recursively or as a subroutine contained
1431 (*PRUNE) or any other control that caused it to give a non-standard return,
1432 invalid errors such as "Error -26 (nested recursion at the same subject
1433 position)" or even infinite loops could occur.
1434
1435 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1436 computing the minimum length on reaching *ACCEPT, and so ended up with the
1437 wrong value of 1 rather than 0. Further investigation indicates that
1438 computing a minimum subject length in the presence of *ACCEPT is difficult
1439 (think back references, subroutine calls), and so I have changed the code
1440 so that no minimum is registered for a pattern that contains *ACCEPT.
1441
1442 8. If (*THEN) was present in the first (true) branch of a conditional group,
1443 it was not handled as intended. [But see 16 below.]
1444
1445 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1446 Sheri Pierce.
1447
1448 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1449 the first byte in a match must be "a".
1450
1451 11. Change 17 for 8.13 increased the recursion depth for patterns like
1452 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1453 pattern contains any instances of (*THEN). If it does not, the old
1454 optimizations are restored. It would be nice to do this on a per-group
1455 basis, but at the moment that is not feasible.
1456
1457 12. In some environments, the output of pcretest -C is CRLF terminated. This
1458 broke RunTest's code that checks for the link size. A single white space
1459 character after the value is now allowed for.
1460
1461 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1462 For "fr", it uses the Windows-specific input and output files.
1463
1464 14. If (*THEN) appeared in a group that was called recursively or as a
1465 subroutine, it did not work as intended. [But see next item.]
1466
1467 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1468 pattern fragments (but not containing any | characters). If A and B are
1469 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1470 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1471 D. In other words, Perl considers parentheses that do not contain any |
1472 characters to be part of a surrounding alternative, whereas PCRE was
1473 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1474 differently. PCRE now behaves in the same way as Perl, except in the case
1475 of subroutine/recursion calls such as (?1) which have in any case always
1476 been different (but PCRE had them first :-).
1477
1478 16. Related to 15 above: Perl does not treat the | in a conditional group as
1479 creating alternatives. Such a group is treated in the same way as an
1480 ordinary group without any | characters when processing (*THEN). PCRE has
1481 been changed to match Perl's behaviour.
1482
1483 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1484 RunGrepTest script failed.
1485
1486 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1487 inevitable for groups that contain captures, but it can lead to a lot of
1488 stack use in large patterns. The old behaviour has been restored for atomic
1489 groups that do not contain any capturing parentheses.
1490
1491 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1492 suppress the check for a minimum subject length at run time. (If it was
1493 given to pcre_exec() or pcre_dfa_exec() it did work.)
1494
1495 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1496 fail to work when decoding hex characters in data strings in EBCDIC
1497 environments.
1498
1499 21. It appears that in at least one Mac OS environment, the isxdigit() function
1500 is implemented as a macro that evaluates to its argument more than once,
1501 contravening the C 90 Standard (I haven't checked a later standard). There
1502 was an instance in pcretest which caused it to go wrong when processing
1503 \x{...} escapes in subject strings. The has been rewritten to avoid using
1504 things like p++ in the argument of isxdigit().
1505
1506
1507 Version 8.13 16-Aug-2011
1508 ------------------------
1509
1510 1. The Unicode data tables have been updated to Unicode 6.0.0.
1511
1512 2. Two minor typos in pcre_internal.h have been fixed.
1513
1514 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1515 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1516 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1517
1518 4. There were a number of related bugs in the code for matching backrefences
1519 caselessly in UTF-8 mode when codes for the characters concerned were
1520 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1521 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1522 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1523 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1524 2-byte code at the end of the subject (it thought there wasn't enough data
1525 left).
1526
1527 5. Comprehensive information about what went wrong is now returned by
1528 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1529 as the output vector has at least 2 elements. The offset of the start of
1530 the failing character and a reason code are placed in the vector.
1531
1532 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1533 now returned is for the first byte of the failing character, instead of the
1534 last byte inspected. This is an incompatible change, but I hope it is small
1535 enough not to be a problem. It makes the returned offset consistent with
1536 pcre_exec() and pcre_dfa_exec().
1537
1538 7. pcretest now gives a text phrase as well as the error number when
1539 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1540 failure, the offset and reason code are output.
1541
1542 8. When \R was used with a maximizing quantifier it failed to skip backwards
1543 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1544 back over a single character (\n). This seems wrong (because it treated the
1545 two characters as a single entity when going forwards), conflicts with the
1546 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1547 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1548 has been changed.
1549
1550 9. Some internal refactoring has changed the processing so that the handling
1551 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1552 time (the PCRE_DOTALL option was changed this way some time ago: version
1553 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1554 which was always a bit of a fudge. It also means that there is one less
1555 argument for the match() function, which reduces its stack requirements
1556 slightly. This change also fixes an incompatibility with Perl: the pattern
1557 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1558
1559 10. More internal refactoring has drastically reduced the number of recursive
1560 calls to match() for possessively repeated groups such as (abc)++ when
1561 using pcre_exec().
1562
1563 11. While implementing 10, a number of bugs in the handling of groups were
1564 discovered and fixed:
1565
1566 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1567 (a|)*(?1) gave a compile-time internal error.
1568 ((a|)+)+ did not notice that the outer group could match an empty string.
1569 (^a|^)+ was not marked as anchored.
1570 (.*a|.*)+ was not marked as matching at start or after a newline.
1571
1572 12. Yet more internal refactoring has removed another argument from the match()
1573 function. Special calls to this function are now indicated by setting a
1574 value in a variable in the "match data" data block.
1575
1576 13. Be more explicit in pcre_study() instead of relying on "default" for
1577 opcodes that mean there is no starting character; this means that when new
1578 ones are added and accidentally left out of pcre_study(), testing should
1579 pick them up.
1580
1581 14. The -s option of pcretest has been documented for ages as being an old
1582 synonym of -m (show memory usage). I have changed it to mean "force study
1583 for every regex", that is, assume /S for every regex. This is similar to -i
1584 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1585 using it. It makes it easier to run collections of tests with and without
1586 study enabled, and thereby test pcre_study() more easily. All the standard
1587 tests are now run with and without -s (but some patterns can be marked as
1588 "never study" - see 20 below).
1589
1590 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1591 restoration of the capturing data to the outer values was not happening
1592 correctly.
1593
1594 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1595 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1596 pattern had matched an empty string, and so incorrectly returned a no
1597 match.
1598
1599 17. There was optimizing code for the last branch of non-capturing parentheses,
1600 and also for the obeyed branch of a conditional subexpression, which used
1601 tail recursion to cut down on stack usage. Unfortunately, now that there is
1602 the possibility of (*THEN) occurring in these branches, tail recursion is
1603 no longer possible because the return has to be checked for (*THEN). These
1604 two optimizations have therefore been removed. [But see 8.20/11 above.]
1605
1606 18. If a pattern containing \R was studied, it was assumed that \R always
1607 matched two bytes, thus causing the minimum subject length to be
1608 incorrectly computed because \R can also match just one byte.
1609
1610 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1611 was incorrectly computed.
1612
1613 20. If /S is present twice on a test pattern in pcretest input, it now
1614 *disables* studying, thereby overriding the use of -s on the command line
1615 (see 14 above). This is necessary for one or two tests to keep the output
1616 identical in both cases.
1617
1618 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1619 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1620
1621 22. When an atomic group that contained a capturing parenthesis was
1622 successfully matched, but the branch in which it appeared failed, the
1623 capturing was not being forgotten if a higher numbered group was later
1624 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1625 group 1 to "a", when in fact it should be unset. This applied to multi-
1626 branched capturing and non-capturing groups, repeated or not, and also to
1627 positive assertions (capturing in negative assertions does not happen
1628 in PCRE) and also to nested atomic groups.
1629
1630 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1631 subject after a captured substring, to make it easier to tell which of a
1632 number of identical substrings has been captured.
1633
1634 24. The way atomic groups are processed by pcre_exec() has been changed so that
1635 if they are repeated, backtracking one repetition now resets captured
1636 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1637 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1638 "aaa". Previously, it would have been "a". As part of this code
1639 refactoring, the way recursive calls are handled has also been changed.
1640
1641 25. If an assertion condition captured any substrings, they were not passed
1642 back unless some other capturing happened later. For example, if
1643 (?(?=(a))a) was matched against "a", no capturing was returned.
1644
1645 26. When studying a pattern that contained subroutine calls or assertions,
1646 the code for finding the minimum length of a possible match was handling
1647 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1648 group 1 called group 2 while simultaneously a separate group 2 called group
1649 1). A stack overflow occurred in this case. I have fixed this by limiting
1650 the recursion depth to 10.
1651
1652 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1653 Fortmann. This supports explicit test numbers on the command line, and has
1654 argument validation and error reporting.
1655
1656 28. An instance of \X with an unlimited repeat could fail if at any point the
1657 first character it looked at was a mark character.
1658
1659 29. Some minor code refactoring concerning Unicode properties and scripts
1660 should reduce the stack requirement of match() slightly.
1661
1662 30. Added the '=' option to pcretest to check the setting of unused capturing
1663 slots at the end of the pattern, which are documented as being -1, but are
1664 not included in the return count.
1665
1666 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1667 compiled something random. Now it gives a compile-time error (as does
1668 Perl).
1669
1670 32. A *MARK encountered during the processing of a positive assertion is now
1671 recorded and passed back (compatible with Perl).
1672
1673 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1674 had alternative anchored branches, the search for a second match in a line
1675 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1676 matched the line "0102" twice. The same bug affected patterns that started
1677 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1678 twice.
1679
1680 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1681 does, and because of capturing effects, quantifying parenthesized
1682 assertions may at times be useful. Quantifiers are now allowed for
1683 parenthesized assertions.
1684
1685 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1686
1687 36. \g was being checked for fancy things in a character class, when it should
1688 just be a literal "g".
1689
1690 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1691 appearance of a nested POSIX class supersedes an apparent external class.
1692 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1693 unescaped square brackets may also appear as part of class names. For
1694 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1695 more like Perl. (But see 8.20/1 above.)
1696
1697 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1698 was because it thought it was \N{name}, which is not supported).
1699
1700 39. Add minix to OS list not supporting the -S option in pcretest.
1701
1702 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1703 cannot analyze patterns in sufficient detail to catch mutual recursions
1704 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1705 subgroup is called recursively as a subpattern for a second time at the
1706 same position in the subject string. In previous releases this might have
1707 been caught by the recursion limit, or it might have run out of stack.
1708
1709 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1710 happen only once. PCRE was, however incorrectly giving a compile time error
1711 "recursive call could loop indefinitely" because it cannot analyze the
1712 pattern in sufficient detail. The compile time test no longer happens when
1713 PCRE is compiling a conditional subpattern, but actual runaway loops are
1714 now caught at runtime (see 40 above).
1715
1716 42. It seems that Perl allows any characters other than a closing parenthesis
1717 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1718 has been changed to be the same.
1719
1720 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1721 as not to get warnings when autogen.sh is called. Also changed
1722 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1723
1724 44. To help people who use pcregrep to scan files containing exceedingly long
1725 lines, the following changes have been made:
1726
1727 (a) The default value of the buffer size parameter has been increased from
1728 8K to 20K. (The actual buffer used is three times this size.)
1729
1730 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1731 PCRE is built.
1732
1733 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1734 to be set at run time.
1735
1736 (d) Numerical values in pcregrep options can be followed by K or M, for
1737 example --buffer-size=50K.
1738
1739 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1740 given and the return code is set to 2.
1741
1742 45. Add a pointer to the latest mark to the callout data block.
1743
1744 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1745 partial match of an empty string instead of no match. This was specific to
1746 the use of ".".
1747
1748 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1749 complete match instead of a partial match. This bug was dependent on both
1750 the PCRE_UTF8 and PCRE_DOTALL options being set.
1751
1752 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1753 starting byte set, because \b was not being ignored.
1754
1755
1756 Version 8.12 15-Jan-2011
1757 ------------------------
1758
1759 1. Fixed some typos in the markup of the man pages, and wrote a script that
1760 checks for such things as part of the documentation building process.
1761
1762 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1763 --match-limit and --recursion-limit options (added for 8.11). In
1764 particular, this made one of the standard tests fail. (The integer value
1765 went into the wrong half of a long int.)
1766
1767 3. If the --colour option was given to pcregrep with -v (invert match), it
1768 did strange things, either producing crazy output, or crashing. It should,
1769 of course, ignore a request for colour when reporting lines that do not
1770 match.
1771
1772 4. Another pcregrep bug caused similar problems if --colour was specified with
1773 -M (multiline) and the pattern match finished with a line ending.
1774
1775 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1776 matched in multiline mode, the following line was shown as part of the
1777 match. This seems wrong, so I have changed it.
1778
1779 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1780 the check for further matches in the same line (so they could be coloured)
1781 to overrun the end of the current line. If another match was found, it was
1782 incorrectly shown (and then shown again when found in the next line).
1783
1784 7. If pcregrep was compiled under Windows, there was a reference to the
1785 function pcregrep_exit() before it was defined. I am assuming this was
1786 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1787 reported by a user. I've moved the definition above the reference.
1788
1789
1790 Version 8.11 10-Dec-2010
1791 ------------------------
1792
1793 1. (*THEN) was not working properly if there were untried alternatives prior
1794 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1795 backtracked to try for "b" instead of moving to the next alternative branch
1796 at the same level (in this case, to look for "c"). The Perl documentation
1797 is clear that when (*THEN) is backtracked onto, it goes to the "next
1798 alternative in the innermost enclosing group".
1799
1800 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1801 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1802 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1803 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1804 (*THEN).
1805
1806 3. If \s appeared in a character class, it removed the VT character from
1807 the class, even if it had been included by some previous item, for example
1808 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1809 of \s, but is part of the POSIX "space" class.)
1810
1811 4. A partial match never returns an empty string (because you can always
1812 match an empty string at the end of the subject); however the checking for
1813 an empty string was starting at the "start of match" point. This has been
1814 changed to the "earliest inspected character" point, because the returned
1815 data for a partial match starts at this character. This means that, for
1816 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1817 (previously it gave "no match").
1818
1819 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1820 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1821 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1822 has an implication that the given string is incomplete (because a partial
1823 match is preferred over a full match). For this reason, these items now
1824 give a partial match in this situation. [Aside: previously, the one case
1825 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1826 match rather than a full match, which was wrong by the old rules, but is
1827 now correct.]
1828
1829 6. There was a bug in the handling of #-introduced comments, recognized when
1830 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1831 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1832 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1833 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1834 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1835 places in pcre_compile().
1836
1837 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1838 comments when looking ahead for named forward references to subpatterns,
1839 the only newline sequence it recognized was NL. It now handles newlines
1840 according to the set newline convention.
1841
1842 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1843 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1844 cater for a lack of strerror(). These oversights have been fixed.
1845
1846 9. Added --match-limit and --recursion-limit to pcregrep.
1847
1848 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1849
1850 11. When the -o option was used, pcregrep was setting a return code of 1, even
1851 when matches were found, and --line-buffered was not being honoured.
1852
1853 12. Added an optional parentheses number to the -o and --only-matching options
1854 of pcregrep.
1855
1856 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1857 can match an empty string. The code to do it in pcretest and pcredemo
1858 needed fixing:
1859
1860 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1861 only one byte after an empty string match just before CRLF (this case
1862 just got forgotten; "any" and "anycrlf" were OK).
1863
1864 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1865 mode when an empty string match preceded an ASCII character followed by
1866 a non-ASCII character. (The code for advancing by one character rather
1867 than one byte was nonsense.)
1868
1869 (c) The pcredemo.c sample program did not have any code at all to handle
1870 the cases when CRLF is a valid newline sequence.
1871
1872 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1873 as a starting offset was within the subject string. There is now a new
1874 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1875 negative or greater than the length of the string. In order to test this,
1876 pcretest is extended to allow the setting of negative starting offsets.
1877
1878 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1879 starting offset points to the beginning of a UTF-8 character was
1880 unnecessarily clumsy. I tidied it up.
1881
1882 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1883 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1884
1885 17. Nobody had reported that the --include_dir option, which was added in
1886 release 7.7 should have been called --include-dir (hyphen, not underscore)
1887 for compatibility with GNU grep. I have changed it to --include-dir, but
1888 left --include_dir as an undocumented synonym, and the same for
1889 --exclude-dir, though that is not available in GNU grep, at least as of
1890 release 2.5.4.
1891
1892 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1893 characters from a string of bytes) have been redefined so as not to use
1894 loops, in order to improve performance in some environments. At the same
1895 time, I abstracted some of the common code into auxiliary macros to save
1896 repetition (this should not affect the compiled code).
1897
1898 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1899 compile-time error is now given if \c is not followed by an ASCII
1900 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1901 different, and any byte value is allowed.)
1902
1903 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1904 START_OPTIMIZE option, which is now allowed at compile time - but just
1905 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1906 to pcregrep and other applications that have no direct access to PCRE
1907 options. The new /Y option in pcretest sets this option when calling
1908 pcre_compile().
1909
1910 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1911 back references. Groups containing recursive back references were forced to
1912 be atomic by that change, but in the case of named groups, the amount of
1913 memory required was incorrectly computed, leading to "Failed: internal
1914 error: code overflow". This has been fixed.
1915
1916 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1917 pcretest.c, to avoid build problems in some Borland environments.
1918
1919
1920 Version 8.10 25-Jun-2010
1921 ------------------------
1922
1923 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1924 THEN.
1925
1926 2. (*ACCEPT) was not working when inside an atomic group.
1927
1928 3. Inside a character class, \B is treated as a literal by default, but
1929 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1930 causes the error). The code is unchanged, but I tidied the documentation.
1931
1932 4. Inside a character class, PCRE always treated \R and \X as literals,
1933 whereas Perl faults them if its -w option is set. I have changed PCRE so
1934 that it faults them when PCRE_EXTRA is set.
1935
1936 5. Added support for \N, which always matches any character other than
1937 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1938
1939 6. When compiling pcregrep with newer versions of gcc which may have
1940 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1941 declared with attribute warn_unused_result" were given. Just casting the
1942 result to (void) does not stop the warnings; a more elaborate fudge is
1943 needed. I've used a macro to implement this.
1944
1945 7. Minor change to pcretest.c to avoid a compiler warning.
1946
1947 8. Added four artifical Unicode properties to help with an option to make
1948 \s etc use properties (see next item). The new properties are: Xan
1949 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1950
1951 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1952 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1953 this option. Modified pcretest to add /W to test this facility. Added
1954 REG_UCP to make it available via the POSIX interface.
1955
1956 10. Added --line-buffered to pcregrep.
1957
1958 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1959 studied, and the match started with a letter with a code point greater than
1960 127 whose first byte was different to the first byte of the other case of
1961 the letter, the other case of this starting letter was not recognized
1962 (#976).
1963
1964 12. If a pattern that was studied started with a repeated Unicode property
1965 test, for example, \p{Nd}+, there was the theoretical possibility of
1966 setting up an incorrect bitmap of starting bytes, but fortunately it could
1967 not have actually happened in practice until change 8 above was made (it
1968 added property types that matched character-matching opcodes).
1969
1970 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1971 possible starting bytes for non-anchored patterns.
1972
1973 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1974 \R, and also a number of cases that involve Unicode properties, both
1975 explicit and implicit when PCRE_UCP is set.
1976
1977 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1978 input, it could crash or give wrong results if characters with values
1979 greater than 0xc0 were present in the subject string. (Detail: it assumed
1980 UTF-8 input when processing these items.)
1981
1982 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1983 size_t is 64-bit (#991).
1984
1985 17. Added a check for running out of memory when PCRE is compiled with
1986 --disable-stack-for-recursion (#990).
1987
1988 18. If the last data line in a file for pcretest does not have a newline on
1989 the end, a newline was missing in the output.
1990
1991 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1992 less than 128) in its various bitmaps. However, there is a facility for
1993 generating tables according to the current locale when PCRE is compiled. It
1994 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1995 characters, are recognized by isspace() and therefore were getting set in
1996 these tables, and indeed these tables seem to approximate to ISO 8859. This
1997 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1998 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1999 which of course cannot start UTF-8 characters. I have changed the code so
2000 that only real ASCII characters (less than 128) and the correct starting
2001 bytes for UTF-8 encodings are set for characters greater than 127 when in
2002 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2003 altogether.)
2004
2005 20. Added the /T option to pcretest so as to be able to run tests with non-
2006 standard character tables, thus making it possible to include the tests
2007 used for 19 above in the standard set of tests.
2008
2009 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2010 reference to a subpattern the other side of a comment that contains an
2011 opening parenthesis caused either an internal compiling error, or a
2012 reference to the wrong subpattern.
2013
2014
2015 Version 8.02 19-Mar-2010
2016 ------------------------
2017
2018 1. The Unicode data tables have been updated to Unicode 5.2.0.
2019
2020 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2021 configured.
2022
2023 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2024 original author of that file, following a query about its status.
2025
2026 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2027 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2028
2029 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2030 quantifier applied to a forward-referencing subroutine call, could compile
2031 incorrect code or give the error "internal error: previously-checked
2032 referenced subpattern not found".
2033
2034 6. Both MS Visual Studio and Symbian OS have problems with initializing
2035 variables to point to external functions. For these systems, therefore,
2036 pcre_malloc etc. are now initialized to local functions that call the
2037 relevant global functions.
2038
2039 7. There were two entries missing in the vectors called coptable and poptable
2040 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2041 I've fixed the data, and added a kludgy way of testing at compile time that
2042 the lengths are correct (equal to the number of opcodes).
2043
2044 8. Following on from 7, I added a similar kludge to check the length of the
2045 eint vector in pcreposix.c.
2046
2047 9. Error texts for pcre_compile() are held as one long string to avoid too
2048 much relocation at load time. To find a text, the string is searched,
2049 counting zeros. There was no check for running off the end of the string,
2050 which could happen if a new error number was added without updating the
2051 string.
2052
2053 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2054
2055 11. \K was not working if it appeared in an atomic group or in a group that
2056 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2057 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2058 the assertion is positive, but not if it is negative.
2059
2060 12. Change 11 fortuitously reduced the size of the stack frame used in the
2061 "match()" function of pcre_exec.c by one pointer. Forthcoming
2062 implementation of support for (*MARK) will need an extra pointer on the
2063 stack; I have reserved it now, so that the stack frame size does not
2064 decrease.
2065
2066 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2067 item in branch that calls a recursion is a subroutine call - as in the
2068 second branch in the above example - was incorrectly given the compile-
2069 time error "recursive call could loop indefinitely" because pcre_compile()
2070 was not correctly checking the subroutine for matching a non-empty string.
2071
2072 14. The checks for overrunning compiling workspace could trigger after an
2073 overrun had occurred. This is a "should never occur" error, but it can be
2074 triggered by pathological patterns such as hundreds of nested parentheses.
2075 The checks now trigger 100 bytes before the end of the workspace.
2076
2077 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2078
2079
2080 Version 8.01 19-Jan-2010
2081 ------------------------
2082
2083 1. If a pattern contained a conditional subpattern with only one branch (in
2084 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2085 computed the wrong minimum data length (which is of course zero for such
2086 subpatterns). This could cause incorrect "no match" results.
2087
2088 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2089 the pattern is reset in the first branch, pcre_compile() failed with
2090 "internal error: code overflow at offset...". This happened only when
2091 the reset was to the original external option setting. (An optimization
2092 abstracts leading options settings into an external setting, which was the
2093 cause of this.)
2094
2095 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2096 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2097 assertion pattern did not match (meaning that the assertion was true), it
2098 was incorrectly treated as false if the SKIP had been reached during the
2099 matching. This also applied to assertions used as conditions.
2100
2101 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2102 assertion subpattern, including such a pattern used as a condition,
2103 unpredictable results occurred, instead of the error return
2104 PCRE_ERROR_DFA_UITEM.
2105
2106 5. The C++ GlobalReplace function was not working like Perl for the special
2107 situation when an empty string is matched. It now does the fancy magic
2108 stuff that is necessary.
2109
2110 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2111 removed. (These were left over from very, very early versions of PCRE.)
2112
2113 7. Some cosmetic changes to the code to make life easier when compiling it
2114 as part of something else:
2115
2116 (a) Change DEBUG to PCRE_DEBUG.
2117
2118 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2119 called "current" as "current_branch", to prevent a collision with the
2120 Linux macro when compiled as a kernel module.
2121
2122 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2123 prevent a collision with the Linux macro when compiled as a kernel
2124 module.
2125
2126 8. In pcre_compile() there are some checks for integer overflows that used to
2127 cast potentially large values to (double). This has been changed to that
2128 when building, a check for int64_t is made, and if it is found, it is used
2129 instead, thus avoiding the use of floating point arithmetic. (There is no
2130 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2131 double.
2132
2133 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2134 2005 (difference between two addresses compared to an unsigned value).
2135
2136 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2137 custom one, because of the following reported problem in Windows:
2138
2139 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2140 under Win32.
2141 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2142 therefore missing the function definition.
2143 - The compiler thus generates a "C" signature for the test function.
2144 - The linker fails to find the "C" function.
2145 - PCRE fails to configure if asked to do so against libbz2.
2146
2147 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2148 messages were output:
2149
2150 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2151 rerunning libtoolize, to keep the correct libtool macros in-tree.
2152 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2153
2154 I have done both of these things.
2155
2156 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2157 most of the time, it *can* run out if it is given a pattern that contains a
2158 runaway infinite recursion. I updated the discussion in the pcrestack man
2159 page.
2160
2161 13. Now that we have gone to the x.xx style of version numbers, the minor
2162 version may start with zero. Using 08 or 09 is a bad idea because users
2163 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2164 interpreted as invalid octal numbers. I've updated the previous comment in
2165 configure.ac, and also added a check that gives an error if 08 or 09 are
2166 used.
2167
2168 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2169 causing partial matching to fail when the end of the subject matched \W
2170 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2171
2172 15. There were some discrepancies between the declarations in pcre_internal.h
2173 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2174 their definitions. The declarations used "const uschar *" and the
2175 definitions used USPTR. Even though USPTR is normally defined as "const
2176 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2177 reported that: "This difference in casting confuses some C++ compilers, for
2178 example, SunCC recognizes above declarations as different functions and
2179 generates broken code for hbpcre." I have changed the declarations to use
2180 USPTR.
2181
2182 16. GNU libtool is named differently on some systems. The autogen.sh script now
2183 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2184 (FreeBSD).
2185
2186 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2187 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2188 comment: "Figure out how to create a longlong from a string: strtoll and
2189 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2190 instance, but it only takes 2 args instead of 3!"
2191
2192 18. A subtle bug concerned with back references has been fixed by a change of
2193 specification, with a corresponding code fix. A pattern such as
2194 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2195 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2196 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2197 same bug. Such groups have to be quantified to be useful, or contained
2198 inside another quantified group. (If there's no repetition, the reference
2199 can never match.) The problem arises because, having left the group and
2200 moved on to the rest of the pattern, a later failure that backtracks into
2201 the group uses the captured value from the final iteration of the group
2202 rather than the correct earlier one. I have fixed this in PCRE by forcing
2203 any group that contains a reference to itself to be an atomic group; that
2204 is, there cannot be any backtracking into it once it has completed. This is
2205 similar to recursive and subroutine calls.
2206
2207
2208 Version 8.00 19-Oct-09
2209 ----------------------
2210
2211 1. The table for translating pcre_compile() error codes into POSIX error codes
2212 was out-of-date, and there was no check on the pcre_compile() error code
2213 being within the table. This could lead to an OK return being given in
2214 error.
2215
2216 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2217 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2218 in a Windows environment.
2219
2220 3. The pcregrep --count option prints the count for each file even when it is
2221 zero, as does GNU grep. However, pcregrep was also printing all files when
2222 --files-with-matches was added. Now, when both options are given, it prints
2223 counts only for those files that have at least one match. (GNU grep just
2224 prints the file name in this circumstance, but including the count seems
2225 more useful - otherwise, why use --count?) Also ensured that the
2226 combination -clh just lists non-zero counts, with no names.
2227
2228 4. The long form of the pcregrep -F option was incorrectly implemented as
2229 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2230 but it seems right to fix it, and I didn't think it was worth preserving
2231 the old behaviour.
2232
2233 5. The command line items --regex=pattern and --regexp=pattern were not
2234 recognized by pcregrep, which required --regex pattern or --regexp pattern
2235 (with a space rather than an '='). The man page documented the '=' forms,
2236 which are compatible with GNU grep; these now work.
2237
2238 6. No libpcreposix.pc file was created for pkg-config; there was just
2239 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2240
2241 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2242 when UCP support is not needed, by modifying the Python script that
2243 generates it from Unicode data files. This should not matter if the module
2244 is correctly used as a library, but I received one complaint about 50K of
2245 unwanted data. My guess is that the person linked everything into his
2246 program rather than using a library. Anyway, it does no harm.
2247
2248 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2249 was a minimum greater than 1 for a wide character in a possessive
2250 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2251 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2252 character. Chaos in the form of incorrect output or a compiling loop could
2253 result.
2254
2255 9. The restrictions on what a pattern can contain when partial matching is
2256 requested for pcre_exec() have been removed. All patterns can now be
2257 partially matched by this function. In addition, if there are at least two
2258 slots in the offset vector, the offset of the earliest inspected character
2259 for the match and the offset of the end of the subject are set in them when
2260 PCRE_ERROR_PARTIAL is returned.
2261
2262 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2263 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2264 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2265 and may be more useful for multi-segment matching.
2266
2267 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2268 used to be given if ever the end of the subject was reached; now it is
2269 given only if matching could not proceed because another character was
2270 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2271 string "Z", which now yields "no match" instead of "partial match". In the
2272 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2273 final character ended with (*FAIL).
2274
2275 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2276 if the pattern had a "must contain" character that was already found in the
2277 earlier partial match, unless partial matching was again requested. For
2278 example, with the pattern /dog.(body)?/, the "must contain" character is
2279 "g". If the first part-match was for the string "dog", restarting with
2280 "sbody" failed. This bug has been fixed.
2281
2282 13. The string returned by pcre_dfa_exec() after a partial match has been
2283 changed so that it starts at the first inspected character rather than the
2284 first character of the match. This makes a difference only if the pattern
2285 starts with a lookbehind assertion or \b or \B (\K is not supported by
2286 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2287 matching functions compatible, and I think it's the right thing to do.
2288
2289 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2290 so that the demonstration program is easily available in environments where
2291 PCRE has not been installed from source.
2292
2293 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2294 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2295 library.
2296
2297 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2298 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2299 is not the first non-POSIX option to be added. Clearly some people find
2300 these options useful.
2301
2302 17. If a caller to the POSIX matching function regexec() passes a non-zero
2303 value for nmatch with a NULL value for pmatch, the value of
2304 nmatch is forced to zero.
2305
2306 18. RunGrepTest did not have a test for the availability of the -u option of
2307 the diff command, as RunTest does. It now checks in the same way as
2308 RunTest, and also checks for the -b option.
2309
2310 19. If an odd number of negated classes containing just a single character
2311 interposed, within parentheses, between a forward reference to a named
2312 subpattern and the definition of the subpattern, compilation crashed with
2313 an internal error, complaining that it could not find the referenced
2314 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2315 [The bug was that it was starting one character too far in when skipping
2316 over the character class, thus treating the ] as data rather than
2317 terminating the class. This meant it could skip too much.]
2318
2319 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2320 /g option in pcretest when the pattern contains \K, which makes it possible
2321 to have an empty string match not at the start, even when the pattern is
2322 anchored. Updated pcretest and pcredemo to use this option.
2323
2324 21. If the maximum number of capturing subpatterns in a recursion was greater
2325 than the maximum at the outer level, the higher number was returned, but
2326 with unset values at the outer level. The correct (outer level) value is
2327 now given.
2328
2329 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2330 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2331 make it do so. The string so far is captured, making this feature
2332 compatible with Perl.
2333
2334 23. The tests have been re-organized, adding tests 11 and 12, to make it
2335 possible to check the Perl 5.10 features against Perl 5.10.
2336
2337 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2338 pattern matches a fixed length string. PCRE did not allow this; now it
2339 does. Neither allows recursion.
2340
2341 25. I finally figured out how to implement a request to provide the minimum
2342 length of subject string that was needed in order to match a given pattern.
2343 (It was back references and recursion that I had previously got hung up
2344 on.) This code has now been added to pcre_study(); it finds a lower bound
2345 to the length of subject needed. It is not necessarily the greatest lower
2346 bound, but using it to avoid searching strings that are too short does give
2347 some useful speed-ups. The value is available to calling programs via
2348 pcre_fullinfo().
2349
2350 26. While implementing 25, I discovered to my embarrassment that pcretest had
2351 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2352 study optimizations had never been tested with that matching function.
2353 Oops. What is worse, even when it was passed study data, there was a bug in
2354 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2355 were also very few tests of studied patterns with pcre_dfa_exec().
2356
2357 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2358 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2359 on the other side of the coin, they are no longer allowed to have different
2360 names, because these cannot be distinguished in PCRE, and this has caused
2361 confusion. (This is a difference from Perl.)
2362
2363 28. When duplicate subpattern names are present (necessarily with different
2364 numbers, as required by 27 above), and a test is made by name in a
2365 conditional pattern, either for a subpattern having been matched, or for
2366 recursion in such a pattern, all the associated numbered subpatterns are
2367 tested, and the overall condition is true if the condition is true for any
2368 one of them. This is the way Perl works, and is also more like the way
2369 testing by number works.
2370
2371
2372 Version 7.9 11-Apr-09
2373 ---------------------
2374
2375 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2376 (pcretest), all targets were linked against these libraries. This included
2377 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2378 libraries. This caused unwanted dependencies to be created. This problem
2379 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2380 pcretest is linked with readline.
2381
2382 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2383 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2384 moved outside it again, because FALSE and TRUE are already defined in AIX,
2385 but BOOL is not.
2386
2387 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2388 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2389
2390 4. The pcregrep documentation said spaces were inserted as well as colons (or
2391 hyphens) following file names and line numbers when outputting matching
2392 lines. This is not true; no spaces are inserted. I have also clarified the
2393 wording for the --colour (or --color) option.
2394
2395 5. In pcregrep, when --colour was used with -o, the list of matching strings
2396 was not coloured; this is different to GNU grep, so I have changed it to be
2397 the same.
2398
2399 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2400 each matching line was coloured. Now it goes on to look for further matches
2401 of any of the test patterns, which is the same behaviour as GNU grep.
2402
2403 7. A pattern that could match an empty string could cause pcregrep to loop; it
2404 doesn't make sense to accept an empty string match in pcregrep, so I have
2405 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2406 seems to be how GNU grep behaves. [But see later change 40 for release
2407 8.33.]
2408
2409 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2410 start or after a newline", because the conditional assertion was not being
2411 correctly handled. The rule now is that both the assertion and what follows
2412 in the first alternative must satisfy the test.
2413
2414 9. If auto-callout was enabled in a pattern with a conditional group whose
2415 condition was an assertion, PCRE could crash during matching, both with
2416 pcre_exec() and pcre_dfa_exec().
2417
2418 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2419 used for matching.
2420
2421 11. Unicode property support in character classes was not working for
2422 characters (bytes) greater than 127 when not in UTF-8 mode.
2423
2424 12. Added the -M command line option to pcretest.
2425
2426 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2427
2428 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2429
2430 16. Added comments and documentation about mis-use of no_arg in the C++
2431 wrapper.
2432
2433 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2434 from Martin Jerabek that uses macro names for all relevant character and
2435 string constants.
2436
2437 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2438 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2439 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2440 these, but not everybody uses configure.
2441
2442 19. A conditional group that had only one branch was not being correctly
2443 recognized as an item that could match an empty string. This meant that an
2444 enclosing group might also not be so recognized, causing infinite looping
2445 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2446 with the subject "ab", where knowledge that the repeated group can match
2447 nothing is needed in order to break the loop.
2448
2449 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2450 exec(), but without supplying a callout function, matching went wrong.
2451
2452 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2453 leak if the size of the offset vector was greater than 30. When the vector
2454 is smaller, the saved offsets during recursion go onto a local stack
2455 vector, but for larger vectors malloc() is used. It was failing to free
2456 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2457 error, in fact).
2458
2459 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2460 heapframe that is used only when UTF-8 support is enabled. This caused no
2461 problem, but was untidy.
2462
2463 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2464 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2465 included within another project.
2466
2467 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2468 slightly modified by me:
2469
2470 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2471 not building pcregrep.
2472
2473 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2474 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2475
2476 25. Forward references, both numeric and by name, in patterns that made use of
2477 duplicate group numbers, could behave incorrectly or give incorrect errors,
2478 because when scanning forward to find the reference group, PCRE was not
2479 taking into account the duplicate group numbers. A pattern such as
2480 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2481
2482 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2483 the feature of a custom pointer more persuasive (as requested by a user).
2484
2485 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2486 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2487 pre-defined.
2488
2489 28. Added support for (*UTF8) at the start of a pattern.
2490
2491 29. Arrange for flags added by the "release type" setting in CMake to be shown
2492 in the configuration summary.
2493
2494
2495 Version 7.8 05-Sep-08
2496 ---------------------
2497
2498 1. Replaced UCP searching code with optimized version as implemented for Ad
2499 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2500 stage table and inline lookup instead of a function, giving speed ups of 2
2501 to 5 times on some simple patterns that I tested. Permission was given to
2502 distribute the MultiStage2.py script that generates the tables (it's not in
2503 the tarball, but is in the Subversion repository).
2504
2505 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2506 scripts.
2507
2508 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2509 a group with a zero qualifier. The result of the study could be incorrect,
2510 or the function might crash, depending on the pattern.
2511
2512 4. Caseless matching was not working for non-ASCII characters in back
2513 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2514 It now works when Unicode Property Support is available.
2515
2516 5. In pcretest, an escape such as \x{de} in the data was always generating
2517 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2518 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2519 truncation.
2520
2521 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2522
2523 7. Added two (int) casts to pcregrep when printing the difference of two
2524 pointers, in case they are 64-bit values.
2525
2526 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2527 test 2 if it fails.
2528
2529 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2530 and a #define of that name to empty if it is not externally set. This is to
2531 allow users of MSVC to set it if necessary.
2532
2533 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2534 the convenience functions in the pcre_get.c source file.
2535
2536 11. An option change at the start of a pattern that had top-level alternatives
2537 could cause overwriting and/or a crash. This command provoked a crash in
2538 some environments:
2539
2540 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2541
2542 This potential security problem was recorded as CVE-2008-2371.
2543
2544 12. For a pattern where the match had to start at the beginning or immediately
2545 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2546 pcre_dfa_exec() could read past the end of the passed subject if there was
2547 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2548 pcretest so that it places the subject at the end of its malloc-ed buffer.
2549
2550 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2551 exec() might read past the end of the data buffer in UTF-8 mode.
2552
2553 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2554 the data contained the byte 0x85 as part of a UTF-8 character within its
2555 first line. This applied both to normal and DFA matching.
2556
2557 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2558 /^[^d]*?$/8 failed to match "abc".
2559
2560 16. Added a missing copyright notice to pcrecpp_internal.h.
2561
2562 17. Make it more clear in the documentation that values returned from
2563 pcre_exec() in ovector are byte offsets, not character counts.
2564
2565 18. Tidied a few places to stop certain compilers from issuing warnings.
2566
2567 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2568 supplied by Stefan Weber. I made a further small update for 7.8 because
2569 there is a change of source arrangements: the pcre_searchfuncs.c module is
2570 replaced by pcre_ucd.c.
2571
2572
2573 Version 7.7 07-May-08
2574 ---------------------
2575
2576 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2577 a string to a long long, pretend we don't even have a long long." This is
2578 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2579
2580 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2581 pre-7.6 versions, which defined a global no_arg variable instead of putting
2582 it in the RE class. (See also #8 below.)
2583
2584 3. Remove a line of dead code, identified by coverity and reported by Nuno
2585 Lopes.
2586
2587 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2588
2589 (1) The include/exclude patterns were being applied to the whole pathnames
2590 of files, instead of just to the final components.
2591
2592 (2) If there was more than one level of directory, the subdirectories were
2593 skipped unless they satisfied the include/exclude conditions. This is
2594 inconsistent with GNU grep (and could even be seen as contrary to the
2595 pcregrep specification - which I improved to make it absolutely clear).
2596 The action now is always to scan all levels of directory, and just
2597 apply the include/exclude patterns to regular files.
2598
2599 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2600 --exclude_dir in the tests to avoid scanning .svn directories.
2601
2602 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2603 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2604 doesn't support NULs in patterns.
2605
2606 7. Added some missing "const"s to declarations of static tables in
2607 pcre_compile.c and pcre_dfa_exec.c.
2608
2609 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2610 caused by fix #2 above. (Subsequently also a second patch to fix the
2611 first patch. And a third patch - this was a messy problem.)
2612
2613 9. Applied Craig's patch to remove the use of push_back().
2614
2615 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2616 matching function regexec().
2617
2618 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2619 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2620 references. PCRE supports relative numbers with this syntax (I don't think
2621 Oniguruma does).
2622
2623 12. Previously, a group with a zero repeat such as (...){0} was completely
2624 omitted from the compiled regex. However, this means that if the group
2625 was called as a subroutine from elsewhere in the pattern, things went wrong
2626 (an internal error was given). Such groups are now left in the compiled
2627 pattern, with a new opcode that causes them to be skipped at execution
2628 time.
2629
2630 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2631 to the way PCRE behaves:
2632
2633 (a) A lone ] character is dis-allowed (Perl treats it as data).
2634
2635 (b) A back reference to an unmatched subpattern matches an empty string
2636 (Perl fails the current match path).
2637
2638 (c) A data ] in a character class must be notated as \] because if the
2639 first data character in a class is ], it defines an empty class. (In
2640 Perl it is not possible to have an empty class.) The empty class []
2641 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2642 The negative empty class [^] matches any one character, independently
2643 of the DOTALL setting.
2644
2645 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2646 non-existent subpattern following a character class starting with ']' and
2647 containing () gave an internal compiling error instead of "reference to
2648 non-existent subpattern". Fortunately, when the pattern did exist, the
2649 compiled code was correct. (When scanning forwards to check for the
2650 existence of the subpattern, it was treating the data ']' as terminating
2651 the class, so got the count wrong. When actually compiling, the reference
2652 was subsequently set up correctly.)
2653
2654 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2655 it was being rejected as not supported by pcre_dfa_exec(), even though
2656 other assertions are supported. I have made pcre_dfa_exec() support
2657 (*FAIL).
2658
2659 16. The implementation of 13c above involved the invention of a new opcode,
2660 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2661 cannot be changed at match time, I realized I could make a small
2662 improvement to matching performance by compiling OP_ALLANY instead of
2663 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2664 on the OP_ANY path.
2665
2666 17. Compiling pcretest on Windows with readline support failed without the
2667 following two fixes: (1) Make the unistd.h include conditional on
2668 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2669
2670 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2671 ncurses library to be included for pcretest when ReadLine support is
2672 requested, but also to allow for it to be overridden. This patch came from
2673 Daniel Bergström.
2674
2675 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2676 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2677 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2678 spotting this.
2679
2680
2681 Version 7.6 28-Jan-08
2682 ---------------------
2683
2684 1. A character class containing a very large number of characters with
2685 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2686 overflow.
2687
2688 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2689 HAVE_LONG_LONG is not defined.
2690
2691 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2692 bring them up to date and include new features. This patch includes:
2693
2694 - Fixed PH's badly added libz and libbz2 support.
2695 - Fixed a problem with static linking.
2696 - Added pcredemo. [But later removed - see 7 below.]
2697 - Fixed dftables problem and added an option.
2698 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2699 HAVE_LONG_LONG.
2700 - Added readline support for pcretest.
2701 - Added an listing of the option settings after cmake has run.
2702
2703 4. A user submitted a patch to Makefile that makes it easy to create
2704 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2705 Makefile.am that cause it to include this special target, without
2706 affecting anything else. Note that the same mingw target plus all
2707 the other distribution libraries and programs are now supported
2708 when configuring with CMake (see 6 below) instead of with
2709 Configure/Make.
2710
2711 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2712 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2713 exported in the Windows port". It has not yet been confirmed that the patch
2714 solves the problem, but it does no harm.
2715
2716 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2717 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2718 with CMake, and also correct the comment about stack recursion.
2719
2720 7. Remove the automatic building of pcredemo from the ./configure system and
2721 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2722 of a program that users should build themselves after PCRE is installed, so
2723 building it automatically is not really right. What is more, it gave
2724 trouble in some build environments.
2725
2726 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2727
2728
2729 Version 7.5 10-Jan-08
2730 ---------------------
2731
2732 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2733 values in parens when parsing an RE using the C++ wrapper."
2734
2735 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2736 Characters greater than 255 were excluded from the class instead of being
2737 included.
2738
2739 3. The same bug as (2) above applied to negated POSIX classes such as
2740 [:^space:].
2741
2742 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2743 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2744 I have changed it.
2745
2746 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2747 first named subpattern) and a construct such as (?&a) would reference the
2748 first named subpattern whose name started with "a" (in other words, the
2749 length check was missing). Both these problems are fixed. "Subpattern name
2750 expected" is now given for (?&) (a zero-length name), and this patch also
2751 makes it give the same error for \k'' (previously it complained that that
2752 was a reference to a non-existent subpattern).
2753
2754 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2755 this is right because (?- can be followed by option settings as well as by
2756 digits. I have, however, made the messages clearer.
2757
2758 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2759 than the number used in the conditional) now cause a compile-time error.
2760 This is actually not compatible with Perl, which accepts such patterns, but
2761 treats the conditional as always being FALSE (as PCRE used to), but it
2762 seems to me that giving a diagnostic is better.
2763
2764 8. Change "alphameric" to the more common word "alphanumeric" in comments
2765 and messages.
2766
2767 9. Fix two occurrences of "backslash" in comments that should have been
2768 "backspace".
2769
2770 10. Remove two redundant lines of code that can never be obeyed (their function
2771 was moved elsewhere).
2772
2773 11. The program that makes PCRE's Unicode character property table had a bug
2774 which caused it to generate incorrect table entries for sequences of
2775 characters that have the same character type, but are in different scripts.
2776 It amalgamated them into a single range, with the script of the first of
2777 them. In other words, some characters were in the wrong script. There were
2778 thirteen such cases, affecting characters in the following ranges:
2779
2780 U+002b0 - U+002c1
2781 U+0060c - U+0060d
2782 U+0061e - U+00612
2783 U+0064b - U+0065e
2784 U+0074d - U+0076d
2785 U+01800 - U+01805
2786 U+01d00 - U+01d77
2787 U+01d9b - U+01dbf
2788 U+0200b - U+0200f
2789 U+030fc - U+030fe
2790 U+03260 - U+0327f
2791 U+0fb46 - U+0fbb1
2792 U+10450 - U+1049d
2793
2794 12. The -o option (show only the matching part of a line) for pcregrep was not
2795 compatible with GNU grep in that, if there was more than one match in a
2796 line, it showed only the first of them. It now behaves in the same way as
2797 GNU grep.
2798
2799 13. If the -o and -v options were combined for pcregrep, it printed a blank
2800 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2801 does the same. The return code can be used to tell if there were any
2802 non-matching lines.
2803
2804 14. Added --file-offsets and --line-offsets to pcregrep.
2805
2806 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2807 infinitely looping recursion. The bug was that positive lookaheads were not
2808 being skipped when checking for a possible empty match (negative lookaheads
2809 and both kinds of lookbehind were skipped).
2810
2811 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2812 inclusion of <windows.h> to before rather than after the definition of
2813 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2814
2815 17. Specifying a possessive quantifier with a specific limit for a Unicode
2816 character property caused pcre_compile() to compile bad code, which led at
2817 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2818 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2819 caused the error; without that there was no problem.
2820
2821 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2822
2823 19. Added --enable-pcretest-libreadline.
2824
2825 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2826 RE::GlobalReplace(). As a result, the number of replacements returned was
2827 double what it should be. I removed one of the increments, but Craig sent a
2828 later patch that removed the other one (the right fix) and added unit tests
2829 that check the return values (which was not done before).
2830
2831 21. Several CMake things:
2832
2833 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2834 the names libpcre and libpcreposix, not just pcre and pcreposix.
2835
2836 (2) The above change means that pcretest and pcregrep are now correctly
2837 linked with the newly-built libraries, not previously installed ones.
2838
2839 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2840
2841 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2842 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2843 UTF-8 newline character). The key issue is that the pattern starts .*;
2844 this means that the match must be either at the beginning, or after a
2845 newline. The bug was in the code for advancing after a failed match and
2846 checking that the new position followed a newline. It was not taking
2847 account of UTF-8 characters correctly.
2848
2849 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2850 character classes. PCRE was not treating the sequence [:...:] as a
2851 character class unless the ... were all letters. Perl, however, seems to
2852 allow any characters between [: and :], though of course it rejects as
2853 unknown any "names" that contain non-letters, because all the known class
2854 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2855 for example, whereas PCRE did not - it did not recognize a POSIX character
2856 class. This seemed a bit dangerous, so the code has been changed to be
2857 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2858 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2859 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2860 Perl does, and where it didn't before.
2861
2862 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2863 Windows environments %n is disabled by default.
2864
2865
2866 Version 7.4 21-Sep-07
2867 ---------------------
2868
2869 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2870 means that a class such as [\s] counted as "explicit reference to CR or
2871 LF". That isn't really right - the whole point of the change was to try to
2872 help when there was an actual mention of one of the two characters. So now
2873 the change happens only if \r or \n (or a literal CR or LF) character is
2874 encountered.
2875
2876 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2877 of both had grown to the point where there were only 3 bits left.
2878 Fortunately, there was spare space in the data structure, and so I have
2879 moved the internal flags into a new 16-bit field to free up more option
2880 bits.
2881
2882 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2883 but did not set the internal JCHANGED flag - either of these is enough to
2884 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2885 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2886 start sets both bits.
2887
2888 4. Added options (at build time, compile time, exec time) to change \R from
2889 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2890
2891 5. doc/pcresyntax.html was missing from the distribution.
2892
2893 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2894 compatibility, even though it is no longer used.
2895
2896 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2897 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2898 windows.h file is present (where different names are used). [This was
2899 reversed later after testing - see 16 below.]
2900
2901 8. Changed all #include <config.h> to #include "config.h". There were also
2902 some further <pcre.h> cases that I changed to "pcre.h".
2903
2904 9. When pcregrep was used with the --colour option, it missed the line ending
2905 sequence off the lines that it output.
2906
2907 10. It was pointed out to me that arrays of string pointers cause lots of
2908 relocations when a shared library is dynamically loaded. A technique of
2909 using a single long string with a table of offsets can drastically reduce
2910 these. I have refactored PCRE in four places to do this. The result is
2911 dramatic:
2912
2913 Originally: 290
2914 After changing UCP table: 187
2915 After changing error message table: 43
2916 After changing table of "verbs" 36
2917 After changing table of Posix names 22
2918
2919 Thanks to the folks working on Gregex for glib for this insight.
2920
2921 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2922 unicode-properties was also set.
2923
2924 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2925
2926 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2927 checked only for CRLF.
2928
2929 14. Added casts to pcretest.c to avoid compiler warnings.
2930
2931 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2932
2933 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2934 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2935 entirely. This removes changes made in 7 above.
2936
2937 17. The CMake files have been updated, and there is now more information about
2938 building with CMake in the NON-UNIX-USE document.
2939
2940
2941 Version 7.3 28-Aug-07
2942 ---------------------
2943
2944 1. In the rejigging of the build system that eventually resulted in 7.1, the
2945 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2946 brackets there is not right, since it causes compilers to look for an
2947 installed pcre.h, not the version that is in the source that is being
2948 compiled (which of course may be different). I have changed it back to:
2949
2950 #include "pcre.h"
2951
2952 I have a vague recollection that the change was concerned with compiling in
2953 different directories, but in the new build system, that is taken care of
2954 by the VPATH setting the Makefile.
2955
2956 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2957 when the subject happened to end in the byte 0x85 (e.g. if the last
2958 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2959 characters but of course it shouldn't be taken as a newline when it is part
2960 of another character. The bug was that, for an unlimited repeat of . in
2961 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2962 characters when looking for a newline.
2963
2964 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2965
2966 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2967 in debug output.
2968
2969 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2970 long printing in the pcrecpp unittest when running under MinGW.
2971
2972 6. ESC_K was left out of the EBCDIC table.
2973
2974 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2975 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2976 limit also applies to "virtual nesting" when a pattern is recursive, and in
2977 this case 1000 isn't so big. I have been able to remove this limit at the
2978 expense of backing off one optimization in certain circumstances. Normally,
2979 when pcre_exec() would call its internal match() function recursively and
2980 immediately return the result unconditionally, it uses a "tail recursion"
2981 feature to save stack. However, when a subpattern that can match an empty
2982 string has an unlimited repetition quantifier, it no longer makes this
2983 optimization. That gives it a stack frame in which to save the data for
2984 checking that an empty string has been matched. Previously this was taken
2985 from the 1000-entry workspace that had been reserved. So now there is no
2986 explicit limit, but more stack is used.
2987
2988 8. Applied Daniel's patches to solve problems with the import/export magic
2989 syntax that is required for Windows, and which was going wrong for the
2990 pcreposix and pcrecpp parts of the library. These were overlooked when this
2991 problem was solved for the main library.
2992
2993 9. There were some crude static tests to avoid integer overflow when computing
2994 the size of patterns that contain repeated groups with explicit upper
2995 limits. As the maximum quantifier is 65535, the maximum group length was
2996 set at 30,000 so that the product of these two numbers did not overflow a
2997 32-bit integer. However, it turns out that people want to use groups that
2998 are longer than 30,000 bytes (though not repeat them that many times).
2999 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3000 made it possible to implement the integer overflow checks in a much more
3001 dynamic way, which I have now done. The artificial limitation on group
3002 length has been removed - we now have only the limit on the total length of
3003 the compiled pattern, which depends on the LINK_SIZE setting.
3004
3005 10. Fixed a bug in the documentation for get/copy named substring when
3006 duplicate names are permitted. If none of the named substrings are set, the
3007 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3008 empty string.
3009
3010 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3011 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3012 because the ] is interpreted as the first data character and the
3013 terminating ] is not found. PCRE has been made compatible with Perl in this
3014 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3015 cause memory overwriting.
3016
3017 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3018 string has been matched (to stop an infinite loop). It was not recognizing
3019 a conditional subpattern that could match an empty string if that
3020 subpattern was within another subpattern. For example, it looped when
3021 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3022 condition was not nested. This bug has been fixed.
3023
3024 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3025 past the start of the subject in the presence of bytes with the top bit
3026 set, for example "\x8aBCD".
3027
3028 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3029 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3030
3031 14. Optimized (?!) to (*FAIL).
3032
3033 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3034 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3035 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3036 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3037 does: it's just the validity check that is more restrictive.
3038
3039 16. Inserted checks for integer overflows during escape sequence (backslash)
3040 processing, and also fixed erroneous offset values for syntax errors during
3041 backslash processing.
3042
3043 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3044 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3045
3046 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3047 caused an overrun.
3048
3049 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3050 something other than just ASCII characters) inside a group that had an
3051 unlimited repeat caused a loop at compile time (while checking to see
3052 whether the group could match an empty string).
3053
3054 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3055 [\P{Any}] did so. (Error in the code for printing property names.)
3056
3057 21. An orphan \E inside a character class could cause a crash.
3058
3059 22. A repeated capturing bracket such as (A)? could cause a wild memory
3060 reference during compilation.
3061
3062 23. There are several functions in pcre_compile() that scan along a compiled
3063 expression for various reasons (e.g. to see if it's fixed length for look
3064 behind). There were bugs in these functions when a repeated \p or \P was
3065 present in the pattern. These operators have additional parameters compared
3066 with \d, etc, and these were not being taken into account when moving along
3067 the compiled data. Specifically:
3068
3069 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3070 length.
3071
3072 (b) An item such as \pL+ within a repeated group could cause crashes or
3073 loops.
3074
3075 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3076 "reference to non-existent subpattern" error.
3077
3078 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3079
3080 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3081 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3082
3083 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3084
3085 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3086 character were causing crashes (broken optimization).
3087
3088 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3089 \p or \P) caused a compile-time loop.
3090
3091 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3092 break. For example, the unstudied pattern [\r\n]A does not match the string
3093 "\r\nA" because change 7.0/46 below moves the current point on by two
3094 characters after failing to match at the start. However, the pattern \nA
3095 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3096 the same is true. There doesn't seem any very clean way out of this, but
3097 what I have chosen to do makes the common cases work: PCRE now takes note
3098 of whether there can be an explicit match for \r or \n anywhere in the
3099 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3100 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3101 pattern has explicit CR or LF references.
3102
3103 29. Added (*CR) etc for changing newline setting at start of pattern.
3104
3105
3106 Version 7.2 19-Jun-07
3107 ---------------------
3108
3109 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3110 which is apparently normally available under Windows.
3111
3112 2. Re-jig the pcregrep tests with different newline settings in an attempt
3113 to make them independent of the local environment's newline setting.
3114
3115 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3116
3117 4. Some of the "internals" tests were previously cut out when the link size
3118 was not 2, because the output contained actual offsets. The recent new
3119 "Z" feature of pcretest means that these can be cut out, making the tests
3120 usable with all link sizes.
3121
3122 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3123 stack recursion. This gives a massive performance boost under BSD, but just
3124 a small improvement under Linux. However, it saves one field in the frame
3125 in all cases.
3126
3127 6. Added more features from the forthcoming Perl 5.10:
3128
3129 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3130 recursion call. It refers to the nth most recently opened parentheses.
3131
3132 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3133 to be opened parentheses.
3134
3135 (c) Conditions that refer to capturing parentheses can be specified
3136 relatively, for example, (?(-2)... or (?(+3)...
3137
3138 (d) \K resets the start of the current match so that everything before
3139 is not part of it.
3140
3141 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3142
3143 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3144 reference syntax.
3145
3146 (g) (?| introduces a group in which the numbering of parentheses in each
3147 alternative starts with the same number.
3148
3149 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3150
3151 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3152 PCRE_INFO_JCHANGED.
3153
3154 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3155 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3156 for detecting groups that can match an empty string.
3157
3158 9. A pattern with a very large number of alternatives (more than several
3159 hundred) was running out of internal workspace during the pre-compile
3160 phase, where pcre_compile() figures out how much memory will be needed. A
3161 bit of new cunning has reduced the workspace needed for groups with
3162 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3163 workspace instead of running out of the 4096 that are available.
3164
3165 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3166
3167 11. Applied patch from Google to remove an optimization that didn't quite work.
3168 The report of the bug said:
3169
3170 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3171 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3172 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3173
3174 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3175 it matched the wrong number of bytes.
3176
3177
3178 Version 7.1 24-Apr-07
3179 ---------------------
3180
3181 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3182 that is more "standard", making use of automake and other Autotools. There
3183 is some re-arrangement of the files and adjustment of comments consequent
3184 on this.
3185
3186 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3187 for recursive directory scanning broke on some systems because the files
3188 are not scanned in any specific order and on different systems the order
3189 was different. A call to "sort" has been inserted into RunGrepTest for the
3190 approprate test as a short-term fix. In the longer term there may be an
3191 alternative.
3192
3193 3. I had an email from Eric Raymond about problems translating some of PCRE's
3194 man pages to HTML (despite the fact that I distribute HTML pages, some
3195 people do their own conversions for various reasons). The problems
3196 concerned the use of low-level troff macros .br and .in. I have therefore
3197 removed all such uses from the man pages (some were redundant, some could
3198 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3199 HTML has been updated to handle .nf/.fi and to complain if it encounters
3200 .br or .in.
3201
3202 4. Updated comments in configure.ac that get placed in config.h.in and also
3203 arranged for config.h to be included in the distribution, with the name
3204 config.h.generic, for the benefit of those who have to compile without
3205 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3206
3207 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3208 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3209 makevp.bat for the current PCRE, using the additional files
3210 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3211
3212 6. A Windows user reported a minor discrepancy with test 2, which turned out
3213 to be caused by a trailing space on an input line that had got lost in his
3214 copy. The trailing space was an accident, so I've just removed it.
3215
3216 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3217 that is needed.
3218
3219 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3220 as "const" (a) because they are and (b) because it helps the PHP
3221 maintainers who have recently made a script to detect big data structures
3222 in the php code that should be moved to the .rodata section. I remembered
3223 to update Builducptable as well, so it won't revert if ucptable.h is ever
3224 re-created.
3225
3226 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3227 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3228 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3229 support is not required. This saves 1.5-2K of code, which is important in
3230 some applications.
3231
3232 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3233 so as not to refer to the tables, even though these functions will never be
3234 called when UTF-8 support is disabled. Otherwise there are problems with a
3235 shared library.
3236
3237 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3238
3239 (a) It was defining its arguments as char * instead of void *.
3240
3241 (b) It was assuming that all moves were upwards in memory; this was true
3242 a long time ago when I wrote it, but is no longer the case.
3243
3244 The emulated memove() is provided for those environments that have neither
3245 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3246 is clearly not the case, as these two bugs were recently reported.
3247
3248 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3249 and Detrail to create the HTML documentation, the .txt form of the man
3250 pages, and it removes trailing spaces from listed files. It also creates
3251 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3252 case, it wraps all the #defines with #ifndefs. This script should be run
3253 before "make dist".
3254
3255 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3256 with Unicode property support.
3257
3258 (a) For a maximizing quantifier, if the two different cases of the
3259 character were of different lengths in their UTF-8 codings (there are
3260 some cases like this - I found 11), and the matching function had to
3261 back up over a mixture of the two cases, it incorrectly assumed they
3262 were both the same length.
3263
3264 (b) When PCRE was configured to use the heap rather than the stack for
3265 recursion during matching, it was not correctly preserving the data for
3266 the other case of a UTF-8 character when checking ahead for a match
3267 while processing a minimizing repeat. If the check also involved
3268 matching a wide character, but failed, corruption could cause an
3269 erroneous result when trying to check for a repeat of the original
3270 character.
3271
3272 13. Some tidying changes to the testing mechanism:
3273
3274 (a) The RunTest script now detects the internal link size and whether there
3275 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3276 values substituted by "configure". (The RunGrepTest script already did
3277 this for UTF-8.) The configure.ac script no longer substitutes the
3278 relevant variables.
3279
3280 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3281 with length and offset values. This means that the output is different
3282 for different internal link sizes. Test 2 is skipped for link sizes
3283 other than 2 because of this, bypassing the problem. Unfortunately,
3284 there was also a test in test 3 (the locale tests) that used /B and
3285 failed for link sizes other than 2. Rather than cut the whole test out,
3286 I have added a new /Z option to pcretest that replaces the length and
3287 offset values with spaces. This is now used to make test 3 independent
3288 of link size. (Test 2 will be tidied up later.)
3289
3290 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3291 segmentation fault instead of returning the appropriate error message.
3292
3293 15. In multiline mode when the newline sequence was set to "any", the pattern
3294 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3295 This doesn't seem right; it now treats the CRLF combination as the line
3296 ending, and so does not match in that case. It's only a pattern such as ^$
3297 that would hit this one: something like ^ABC$ would have failed after \r
3298 and then tried again after \r\n.
3299
3300 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3301 in an attempt to make files that differ only in their line terminators
3302 compare equal. This works on Linux.
3303
3304 17. Under certain error circumstances pcregrep might try to free random memory
3305 as it exited. This is now fixed, thanks to valgrind.
3306
3307 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3308 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3309 was because its rules for how to advance for /g after matching an empty
3310 string at the end of a line did not allow for this case. They now check for
3311 it specially.
3312
3313 20. pcretest is supposed to handle patterns and data of any length, by
3314 extending its buffers when necessary. It was getting this wrong when the
3315 buffer for a data line had to be extended.
3316
3317 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3318 CRLF as a newline sequence.
3319
3320 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3321 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3322 I have nevertheless tidied it up.
3323
3324 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3325
3326 24. Added a man page for pcre-config.
3327
3328
3329 Version 7.0 19-Dec-06
3330 ---------------------
3331
3332 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3333 moving to gcc 4.1.1.
3334
3335 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3336 sys/time.h, which is documented as needed for this function. It doesn't
3337 seem to matter on Linux, but it showed up on some releases of OS X.
3338
3339 3. It seems that there are systems where bytes whose values are greater than
3340 127 match isprint() in the "C" locale. The "C" locale should be the
3341 default when a C program starts up. In most systems, only ASCII printing
3342 characters match isprint(). This difference caused the output from pcretest
3343 to vary, making some of the tests fail. I have changed pcretest so that:
3344
3345 (a) When it is outputting text in the compiled version of a pattern, bytes
3346 other than 32-126 are always shown as hex escapes.
3347
3348 (b) When it is outputting text that is a matched part of a subject string,
3349 it does the same, unless a different locale has been set for the match
3350 (using the /L modifier). In this case, it uses isprint() to decide.
3351
3352 4. Fixed a major bug that caused incorrect computation of the amount of memory
3353 required for a compiled pattern when options that changed within the
3354 pattern affected the logic of the preliminary scan that determines the
3355 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3356 that the computed length was too small. The symptoms of this bug were
3357 either the PCRE error "internal error: code overflow" from pcre_compile(),
3358 or a glibc crash with a message such as "pcretest: free(): invalid next
3359 size (fast)". Examples of patterns that provoked this bug (shown in
3360 pcretest format) are:
3361
3362 /(?-x: )/x
3363 /(?x)(?-x: \s*#\s*)/
3364 /((?i)[\x{c0}])/8
3365 /(?i:[\x{c0}])/8
3366
3367 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3368 is now done differently.
3369
3370 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3371 wrapper classes; (b) implement a new function in the C++ scanner that is
3372 more efficient than the old way of doing things because it avoids levels of
3373 recursion in the regex matching; (c) add a paragraph to the documentation
3374 for the FullMatch() function.
3375
3376 6. The escape sequence \n was being treated as whatever was defined as
3377 "newline". Not only was this contrary to the documentation, which states
3378 that \n is character 10 (hex 0A), but it also went horribly wrong when
3379 "newline" was defined as CRLF. This has been fixed.
3380
3381 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3382 was being set to -1 for the "end of line" case (supposedly a value that no
3383 character can have). Though this value is never used (the check for end of
3384 line is "zero bytes in current character"), it caused compiler complaints.
3385 I've changed it to 0xffffffff.
3386
3387 8. In pcre_version.c, the version string was being built by a sequence of
3388 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3389 string (as it is for production releases) called a macro with an empty
3390 argument. The C standard says the result of this is undefined. The gcc
3391 compiler treats it as an empty string (which was what was wanted) but it is
3392 reported that Visual C gives an error. The source has been hacked around to
3393 avoid this problem.
3394
3395 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3396 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3397 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3398 of them did).
3399
3400 10. Originally, pcretest opened its input and output without "b"; then I was
3401 told that "b" was needed in some environments, so it was added for release
3402 5.0 to both the input and output. (It makes no difference on Unix-like
3403 systems.) Later I was told that it is wrong for the input on Windows. I've
3404 now abstracted the modes into two macros, to make it easier to fiddle with
3405 them, and removed "b" from the input mode under Windows.
3406
3407 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3408
3409 12. Added -help and --help to pcretest as an official way of being reminded
3410 of the options.
3411
3412 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3413 and pcrecpp.cc because they annoy compilers at high warning levels.
3414
3415 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3416
3417 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3418 scripts are not C programs :-) and which was not noticed because it works
3419 on Linux.
3420
3421 16. pcretest is supposed to handle any length of pattern and data line (as one
3422 line or as a continued sequence of lines) by extending its input buffer if
3423 necessary. This feature was broken for very long pattern lines, leading to
3424 a string of junk being passed to pcre_compile() if the pattern was longer
3425 than about 50K.
3426
3427 17. I have done a major re-factoring of the way pcre_compile() computes the
3428 amount of memory needed for a compiled pattern. Previously, there was code
3429 that made a preliminary scan of the pattern in order to do this. That was
3430 OK when PCRE was new, but as the facilities have expanded, it has become
3431 harder and harder to keep it in step with the real compile phase, and there
3432 have been a number of bugs (see for example, 4 above). I have now found a
3433 cunning way of running the real compile function in a "fake" mode that
3434 enables it to compute how much memory it would need, while actually only
3435 ever using a few hundred bytes of working memory and without too many
3436 tests of the mode. This should make future maintenance and development
3437 easier. A side effect of this work is that the limit of 200 on the nesting
3438 depth of parentheses has been removed (though this was never a serious
3439 limitation, I suspect). However, there is a downside: pcre_compile() now
3440 runs more slowly than before (30% or more, depending on the pattern). I
3441 hope this isn't a big issue. There is no effect on runtime performance.
3442
3443 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3444 newline (only possible for the last line of a file) and it was a
3445 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3446
3447 19. Added additional timing features to pcretest. (1) The -tm option now times
3448 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3449 separate command line item, by a number that specifies the number of
3450 repeats to use when timing. The default is 50000; this gives better
3451 precision, but takes uncomfortably long for very large patterns.
3452
3453 20. Extended pcre_study() to be more clever in cases where a branch of a
3454 subpattern has no definite first character. For example, (a*|b*)[cd] would
3455 previously give no result from pcre_study(). Now it recognizes that the
3456 first character must be a, b, c, or d.
3457
3458 21. There was an incorrect error "recursive call could loop indefinitely" if
3459 a subpattern (or the entire pattern) that was being tested for matching an
3460 empty string contained only one non-empty item after a nested subpattern.
3461 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3462 incorrectly, because the \d was being skipped in the check.
3463
3464 22. The pcretest program now has a new pattern option /B and a command line
3465 option -b, which is equivalent to adding /B to every pattern. This causes
3466 it to show the compiled bytecode, without the additional information that
3467 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3468 is the same as /B/I).
3469
3470 23. A new optimization is now able automatically to treat some sequences such
3471 as a*b as a*+b. More specifically, if something simple (such as a character
3472 or a simple class like \d) has an unlimited quantifier, and is followed by
3473 something that cannot possibly match the quantified thing, the quantifier
3474 is automatically "possessified".
3475
3476 24. A recursive reference to a subpattern whose number was greater than 39
3477 went wrong under certain circumstances in UTF-8 mode. This bug could also
3478 have affected the operation of pcre_study().
3479
3480 25. Realized that a little bit of performance could be had by replacing
3481 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3482
3483 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3484
3485 27. Possessive quantifiers such as a++ were previously implemented by turning
3486 them into atomic groups such as ($>a+). Now they have their own opcodes,
3487 which improves performance. This includes the automatically created ones
3488 from 23 above.
3489
3490 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3491 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3492 the first matched character to be a colon. This applied both to named and
3493 numbered groups.
3494
3495 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3496
3497 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3498 building PCRE on Windows easier, so I have included it in the distribution.
3499
3500 31. There is now a check in pcretest against a ridiculously large number being
3501 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3502 loop, the loop is abandoned.
3503
3504 32. Forward references to subpatterns in conditions such as (?(2)...) where
3505 subpattern 2 is defined later cause pcre_compile() to search forwards in
3506 the pattern for the relevant set of parentheses. This search went wrong
3507 when there were unescaped parentheses in a character class, parentheses
3508 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3509
3510 33. "Subroutine" calls and backreferences were previously restricted to
3511 referencing subpatterns earlier in the regex. This restriction has now
3512 been removed.
3513
3514 34. Added a number of extra features that are going to be in Perl 5.10. On the
3515 whole, these are just syntactic alternatives for features that PCRE had
3516 previously implemented using the Python syntax or my own invention. The
3517 other formats are all retained for compatibility.
3518
3519 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3520 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3521 also .NET compatible.
3522
3523 (b) A recursion or subroutine call to a named group can now be defined as
3524 (?&name) as well as (?P>name).
3525
3526 (c) A backreference to a named group can now be defined as \k<name> or
3527 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3528 5.10, are also .NET compatible.
3529
3530 (d) A conditional reference to a named group can now use the syntax
3531 (?(<name>) or (?('name') as well as (?(name).
3532
3533 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3534 groups (named and numbered) that are never evaluated inline, but can be
3535 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3536 is always false. There may be only one alternative in such a group.
3537
3538 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3539 as the simple (?(R). The condition is true only if the most recent
3540 recursion is that of the given number or name. It does not search out
3541 through the entire recursion stack.
3542
3543 (g) The escape \gN or \g{N} has been added, where N is a positive or
3544 negative number, specifying an absolute or relative reference.
3545
3546 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3547 some "unreachable code" warnings.
3548
3549 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3550 things, this adds five new scripts.
3551
3552 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3553 There were also incompatibilities regarding the handling of \Q..\E inside
3554 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3555 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3556
3557 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3558 matches an empty string, and forcibly breaks the loop. There were bugs in
3559 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3560 against aaaa the result was just "a" rather than "aaaa", for example. Two
3561 separate and independent bugs (that affected different cases) have been
3562 fixed.
3563
3564 39. Refactored the code to abolish the use of different opcodes for small
3565 capturing bracket numbers. This is a tidy that I avoided doing when I
3566 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3567 The new approach is not only tidier, it makes it possible to reduce the
3568 memory needed to fix the previous bug (38).
3569
3570 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3571 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3572 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3573 mode.
3574
3575 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3576 report.
3577
3578 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3579 copy construction and assignment in the C++ wrapper.
3580
3581 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3582 couple of bugs that could have given wrong results in the "--newline=crlf"
3583 case.
3584
3585 44. Added a number of casts and did some reorganization of signed/unsigned int
3586 variables following suggestions from Dair Grant. Also renamed the variable
3587 "this" as "item" because it is a C++ keyword.
3588
3589 45. Arranged for dftables to add
3590
3591 #include "pcre_internal.h"
3592
3593 to pcre_chartables.c because without it, gcc 4.x may remove the array
3594 definition from the final binary if PCRE is built into a static library and
3595 dead code stripping is activated.
3596
3597 46. For an unanchored pattern, if a match attempt fails at the start of a
3598 newline sequence, and the newline setting is CRLF or ANY, and the next two
3599 characters are CRLF, advance by two characters instead of one.
3600
3601
3602 Version 6.7 04-Jul-06
3603 ---------------------
3604
3605 1. In order to handle tests when input lines are enormously long, pcretest has
3606 been re-factored so that it automatically extends its buffers when
3607 necessary. The code is crude, but this _is_ just a test program. The
3608 default size has been increased from 32K to 50K.
3609
3610 2. The code in pcre_study() was using the value of the re argument before
3611 testing it for NULL. (Of course, in any sensible call of the function, it
3612 won't be NULL.)
3613
3614 3. The memmove() emulation function in pcre_internal.h, which is used on
3615 systems that lack both memmove() and bcopy() - that is, hardly ever -
3616 was missing a "static" storage class specifier.
3617
3618 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3619 containing an extended class (one that cannot be represented by a bitmap
3620 because it contains high-valued characters or Unicode property items, e.g.
3621 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3622 pattern, but PCRE should not loop if you do not (it no longer does).
3623 [Detail: two cases were found: (a) a repeated subpattern containing an
3624 extended class; (b) a recursive reference to a subpattern that followed a
3625 previous extended class. It wasn't skipping over the extended class
3626 correctly when UTF-8 mode was not set.]
3627
3628 5. A negated single-character class was not being recognized as fixed-length
3629 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3630 compile error "lookbehind assertion is not fixed length".
3631
3632 6. The RunPerlTest auxiliary script was showing an unexpected difference
3633 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3634 write a Perl script that can interpret lines of an input file either as
3635 byte characters or as UTF-8, which is what "perltest" was being required to
3636 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3637 can't do is switch easily at run time between having the "use utf8;" pragma
3638 or not. In the end, I fudged it by using the RunPerlTest script to insert
3639 "use utf8;" explicitly for the UTF-8 tests.
3640
3641 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3642 the end of the subject string, contrary to the documentation and to what
3643 Perl does. This was true of both matching functions. Now it matches only at
3644 the start of the subject and immediately after *internal* newlines.
3645
3646 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3647 a pointer to an int instead of a pointer to an unsigned long int. This
3648 caused problems on 64-bit systems.
3649
3650 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3651 instance of the 'standard' template library not being so standard".
3652
3653 10. There was no check on the number of named subpatterns nor the maximum
3654 length of a subpattern name. The product of these values is used to compute
3655 the size of the memory block for a compiled pattern. By supplying a very
3656 long subpattern name and a large number of named subpatterns, the size
3657 computation could be caused to overflow. This is now prevented by limiting
3658 the length of names to 32 characters, and the number of named subpatterns
3659 to 10,000.
3660
3661 11. Subpatterns that are repeated with specific counts have to be replicated in
3662 the compiled pattern. The size of memory for this was computed from the
3663 length of the subpattern and the repeat count. The latter is limited to
3664 65535, but there was no limit on the former, meaning that integer overflow
3665 could in principle occur. The compiled length of a repeated subpattern is
3666 now limited to 30,000 bytes in order to prevent this.
3667
3668 12. Added the optional facility to have named substrings with the same name.
3669
3670 13. Added the ability to use a named substring as a condition, using the
3671 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3672 are numbers (not recommended). Forward references are permitted.
3673
3674 14. Added forward references in named backreferences (if you see what I mean).
3675
3676 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3677 pattern could run off the end of the subject. For example, the pattern
3678 "(?s)(.{1,5})"8 did this with the subject "ab".
3679
3680 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3681 PCRE_CASELESS was set when matching characters that were quantified with ?
3682 or *.
3683
3684 17. A character class other than a single negated character that had a minimum
3685 but no maximum quantifier - for example [ab]{6,} - was not handled
3686 correctly by pce_dfa_exec(). It would match only one character.
3687
3688 18. A valid (though odd) pattern that looked like a POSIX character
3689 class but used an invalid character after [ (for example [[,abc,]]) caused
3690 pcre_compile() to give the error "Failed: internal error: code overflow" or
3691 in some cases to crash with a glibc free() error. This could even happen if
3692 the pattern terminated after [[ but there just happened to be a sequence of
3693 letters, a binary zero, and a closing ] in the memory that followed.
3694
3695 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3696 over the years. Originally (before any Unicode support), just the bottom 8
3697 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3698 output from "man perlunicode" includes this:
3699
3700 The regular expression compiler produces polymorphic opcodes. That
3701 is, the pattern adapts to the data and automatically switches to
3702 the Unicode character scheme when presented with Unicode data--or
3703 instead uses a traditional byte scheme when presented with byte
3704 data.
3705
3706 Sadly, a wide octal escape does not cause a switch, and in a string with
3707 no other multibyte characters, these octal escapes are treated as before.
3708 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3709 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3710 Unicode string.
3711
3712 I have not perpetrated such confusion in PCRE. Up till now, it took just
3713 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3714 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3715 translate to the appropriate multibyte character.
3716
3717 29. Applied some refactoring to reduce the number of warnings from Microsoft
3718 and Borland compilers. This has included removing the fudge introduced
3719 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3720 a warning about an unused variable.
3721
3722 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3723 characters since release 4.0, because Perl (from release 5.004) does not.
3724 [Or at least, is documented not to: some releases seem to be in conflict
3725 with the documentation.] However, when a pattern was studied with
3726 pcre_study() and all its branches started with \s, PCRE still included VT
3727 as a possible starting character. Of course, this did no harm; it just
3728 caused an unnecessary match attempt.
3729
3730 22. Removed a now-redundant internal flag bit that recorded the fact that case
3731 dependency changed within the pattern. This was once needed for "required
3732 byte" processing, but is no longer used. This recovers a now-scarce options
3733 bit. Also moved the least significant internal flag bit to the most-
3734 significant bit of the word, which was not previously used (hangover from
3735 the days when it was an int rather than a uint) to free up another bit for
3736 the future.
3737
3738 23. Added support for CRLF line endings as well as CR and LF. As well as the
3739 default being selectable at build time, it can now be changed at runtime
3740 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3741 specify that it is scanning data with non-default line endings.
3742
3743 24. Changed the definition of CXXLINK to make it agree with the definition of
3744 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3745
3746 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3747 recursions. This makes a big different to stack usage for some patterns.
3748
3749 26. If a subpattern containing a named recursion or subroutine reference such
3750 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3751 the space required for the compiled pattern went wrong and gave too small a
3752 value. Depending on the environment, this could lead to "Failed: internal
3753 error: code overflow at offset 49" or "glibc detected double free or
3754 corruption" errors.
3755
3756 27. Applied patches from Google (a) to support the new newline modes and (b) to
3757 advance over multibyte UTF-8 characters in GlobalReplace.
3758
3759 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3760 difference for some implementation of PCRE in some Windows version.
3761
3762 29. Added some extra testing facilities to pcretest:
3763
3764 \q<number> in a data line sets the "match limit" value
3765 \Q<number> in a data line sets the "match recursion limt" value
3766 -S <number> sets the stack size, where <number> is in megabytes
3767
3768 The -S option isn't available for Windows.
3769
3770
3771 Version 6.6 06-Feb-06
3772 ---------------------
3773
3774 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3775 in pcreposix.h. I have copied the definition from pcre.h.
3776
3777 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3778 because pcre.h is no longer a built file.
3779
3780 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3781 not normally included in the compiled code.
3782
3783
3784 Version 6.5 01-Feb-06
3785 ---------------------
3786
3787 1. When using the partial match feature with pcre_dfa_exec(), it was not
3788 anchoring the second and subsequent partial matches at the new starting
3789 point. This could lead to incorrect results. For example, with the pattern
3790 /1234/, partially matching against "123" and then "a4" gave a match.
3791
3792 2. Changes to pcregrep:
3793
3794 (a) All non-match returns from pcre_exec() were being treated as failures
3795 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3796 error message is output. Some extra information is given for the
3797 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3798 probably the only errors that are likely to be caused by users (by
3799 specifying a regex that has nested indefinite repeats, for instance).
3800 If there are more than 20 of these errors, pcregrep is abandoned.
3801
3802 (b) A binary zero was treated as data while matching, but terminated the
3803 output line if it was written out. This has been fixed: binary zeroes
3804 are now no different to any other data bytes.
3805
3806 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3807 used to set a locale for matching. The --locale=xxxx long option has
3808 been added (no short equivalent) to specify a locale explicitly on the
3809 pcregrep command, overriding the environment variables.
3810
3811 (d) When -B was used with -n, some line numbers in the output were one less
3812 than they should have been.
3813
3814 (e) Added the -o (--only-matching) option.
3815
3816 (f) If -A or -C was used with -c (count only), some lines of context were
3817 accidentally printed for the final match.
3818
3819 (g) Added the -H (--with-filename) option.
3820
3821 (h) The combination of options -rh failed to suppress file names for files
3822 that were found from directory arguments.
3823
3824 (i) Added the -D (--devices) and -d (--directories) options.
3825
3826 (j) Added the -F (--fixed-strings) option.
3827
3828 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3829
3830 (l) Added the --colo(u)r option.
3831
3832 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3833 is not present by default.
3834
3835 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3836 items such as (?R) or (?1), when the recursion could match a number of
3837 alternatives. If it matched one of the alternatives, but subsequently,
3838 outside the recursion, there was a failure, the code tried to back up into
3839 the recursion. However, because of the way PCRE is implemented, this is not
3840 possible, and the result was an incorrect result from the match.
3841
3842 In order to prevent this happening, the specification of recursion has
3843 been changed so that all such subpatterns are automatically treated as
3844 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3845
3846 4. I had overlooked the fact that, in some locales, there are characters for
3847 which isalpha() is true but neither isupper() nor islower() are true. In
3848 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3849 and ordfeminine) are like this. This affected the treatment of \w and \W
3850 when they appeared in character classes, but not when they appeared outside
3851 a character class. The bit map for "word" characters is now created
3852 separately from the results of isalnum() instead of just taking it from the
3853 upper, lower, and digit maps. (Plus the underscore character, of course.)
3854
3855 5. The above bug also affected the handling of POSIX character classes such as
3856 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3857 permanent tables. Instead, the bit maps for such a class were previously
3858 created as the appropriate unions of the upper, lower, and digit bitmaps.
3859 Now they are created by subtraction from the [[:word:]] class, which has
3860 its own bitmap.
3861
3862 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3863 It is created by subtracting the vertical space characters (\x09, \x0a,
3864 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3865 subtraction was done in the overall bitmap for a character class, meaning
3866 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3867 be recognized. This bug has been fixed.
3868
3869 7. Patches from the folks at Google:
3870
3871 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3872 real life, but is still worth protecting against".
3873
3874 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3875 regular expressions".
3876
3877 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3878 have it.
3879
3880 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3881 "configure" and the latter not, in order to fix a problem somebody had
3882 with compiling the Arg class on HP-UX.
3883
3884 (e) Improve the error-handling of the C++ wrapper a little bit.
3885
3886 (f) New tests for checking recursion limiting.
3887
3888 8. The pcre_memmove() function, which is used only if the environment does not
3889 have a standard memmove() function (and is therefore rarely compiled),
3890 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3891 returning a result (though PCRE never actually uses the result).
3892
3893 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3894 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3895 returned instead of calling malloc() with an overflowing number that would
3896 most likely cause subsequent chaos.
3897
3898 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3899
3900 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3901 with this option is matched, the nmatch and pmatch options of regexec() are
3902 ignored.
3903
3904 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3905 provided in case anyone wants to the the POSIX interface with UTF-8
3906 strings.
3907
3908 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3909 C++ linking (needed for some HP-UX environments).
3910
3911 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3912 (unused parameter) and in the pcre_printint() function (omitted "default"
3913 switch label when the default is to do nothing).
3914
3915 15. Added some code to make it possible, when PCRE is compiled as a C++
3916 library, to replace subject pointers for pcre_exec() with a smart pointer
3917 class, thus making it possible to process discontinuous strings.
3918
3919 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3920 much the same function. They were added by different people who were trying
3921 to make PCRE easy to compile on non-Unix systems. It has been suggested
3922 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3923 for compiling on Windows systems. I have therefore replaced it with
3924 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3925 defaults to "extern" for C or "extern C" for C++, which works fine on
3926 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3927 SCOPE with something explicit in config.h. In addition:
3928
3929 (a) pcreposix.h still had just "extern" instead of either of these macros;
3930 I have replaced it with PCRE_DATA_SCOPE.
3931
3932 (b) Functions such as _pcre_xclass(), which are internal to the library,
3933 but external in the C sense, all had PCRE_EXPORT in their definitions.
3934 This is apparently wrong for the Windows case, so I have removed it.
3935 (It makes no difference on Unix-like systems.)
3936
3937 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3938 of recursive calls to match(). This is different to MATCH_LIMIT because
3939 that limits the total number of calls to match(), not all of which increase
3940 the depth of recursion. Limiting the recursion depth limits the amount of
3941 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3942 when PCRE is compiled, and changed at run time. A patch from Google adds
3943 this functionality to the C++ interface.
3944
3945 18. Changes to the handling of Unicode character properties:
3946
3947 (a) Updated the table to Unicode 4.1.0.
3948
3949 (b) Recognize characters that are not in the table as "Cn" (undefined).
3950
3951 (c) I revised the way the table is implemented to a much improved format
3952 which includes recognition of ranges. It now supports the ranges that
3953 are defined in UnicodeData.txt, and it also amalgamates other
3954 characters into ranges. This has reduced the number of entries in the
3955 table from around 16,000 to around 3,000, thus reducing its size
3956 considerably. I realized I did not need to use a tree structure after
3957 all - a binary chop search is just as efficient. Having reduced the
3958 number of entries, I extended their size from 6 bytes to 8 bytes to
3959 allow for more data.
3960
3961 (d) Added support for Unicode script names via properties such as \p{Han}.
3962
3963 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3964 matching that character.
3965
3966 20. When matching a repeated Unicode property with a minimum greater than zero,
3967 (for example \pL{2,}), PCRE could look past the end of the subject if it
3968 reached it while seeking the minimum number of characters. This could
3969 happen only if some of the characters were more than one byte long, because
3970 there is a check for at least the minimum number of bytes.
3971
3972 21. Refactored the implementation of \p and \P so as to be more general, to
3973 allow for more different types of property in future. This has changed the
3974 compiled form incompatibly. Anybody with saved compiled patterns that use
3975 \p or \P will have to recompile them.
3976
3977 22. Added "Any" and "L&" to the supported property types.
3978
3979 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3980 but give a compile time error if the value is greater than 0xff.
3981
3982 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3983 accidentally not being installed or uninstalled.
3984
3985 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3986 made were to insert the current release number. This seemed silly, because
3987 it made things harder for people building PCRE on systems that don't run
3988 "configure". I have turned pcre.h into a distributed file, no longer built
3989 by "configure", with the version identification directly included. There is
3990 no longer a pcre.h.in file.
3991
3992 However, this change necessitated a change to the pcre-config script as
3993 well. It is built from pcre-config.in, and one of the substitutions was the
3994 release number. I have updated configure.ac so that ./configure now finds
3995 the release number by grepping pcre.h.
3996
3997 26. Added the ability to run the tests under valgrind.
3998
3999
4000 Version 6.4 05-Sep-05
4001 ---------------------
4002
4003 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4004 "--" to be printed when multiple files were scanned, even when none of the
4005 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4006 consider it to be a bug, and have restored the previous behaviour.
4007
4008 2. A couple of code tidies to get rid of compiler warnings.
4009
4010 3. The pcretest program used to cheat by referring to symbols in the library
4011 whose names begin with _pcre_. These are internal symbols that are not
4012 really supposed to be visible externally, and in some environments it is
4013 possible to suppress them. The cheating is now confined to including
4014 certain files from the library's source, which is a bit cleaner.
4015
4016 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4017 file's purpose clearer.
4018
4019 5. Reorganized pcre_ucp_findchar().
4020
4021
4022 Version 6.3 15-Aug-05
4023 ---------------------
4024
4025 1. The file libpcre.pc.in did not have general read permission in the tarball.
4026
4027 2. There were some problems when building without C++ support:
4028
4029 (a) If C++ support was not built, "make install" and "make test" still
4030 tried to test it.
4031
4032 (b) There were problems when the value of CXX was explicitly set. Some
4033 changes have been made to try to fix these, and ...
4034
4035 (c) --disable-cpp can now be used to explicitly disable C++ support.
4036
4037 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4038 backslash in a target when C++ was disabled. This confuses some
4039 versions of "make", apparently. Using an intermediate variable solves
4040 this. (Same for CPP_LOBJ.)
4041
4042 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4043 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4044 necessary on certain architectures.
4045
4046 4. Added a setting of -export-symbols-regex to the link command to remove
4047 those symbols that are exported in the C sense, but actually are local
4048 within the library, and not documented. Their names all begin with
4049 "_pcre_". This is not a perfect job, because (a) we have to except some
4050 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4051 available (and never for static libraries). I have made a note to try to
4052 find a way round (a) in the future.
4053
4054
4055 Version 6.2 01-Aug-05
4056 ---------------------
4057
4058 1. There was no test for integer overflow of quantifier values. A construction
4059 such as {1111111111111111} would give undefined results. What is worse, if
4060 a minimum quantifier for a parenthesized subpattern overflowed and became
4061 negative, the calculation of the memory size went wrong. This could have
4062 led to memory overwriting.
4063
4064 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4065
4066 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4067 operating environments where this matters.
4068
4069 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4070 PCRE options from within the C++ wrapper.
4071
4072 5. Named capturing subpatterns were not being correctly counted when a pattern
4073 was compiled. This caused two problems: (a) If there were more than 100
4074 such subpatterns, the calculation of the memory needed for the whole
4075 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4076 back references of the form \12, where the number was greater than 9, were
4077 not recognized as back references, even though there were sufficient
4078 previous subpatterns.
4079
4080 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4081 versions of gcc, e.g. 2.95.4.
4082
4083
4084 Version 6.1 21-Jun-05
4085 ---------------------
4086
4087 1. There was one reference to the variable "posix" in pcretest.c that was not
4088 surrounded by "#if !defined NOPOSIX".
4089
4090 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4091 the cross-check on the old pcre_info() function, for the benefit of the
4092 cut-down version of PCRE that is currently imported into Exim.
4093
4094 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4095 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4096 patterns that start (?i) but I don't think that matters. The use of (?i) is
4097 just an example; this all applies to the other options as well.
4098
4099 4. Since libtool seems to echo the compile commands it is issuing, the output
4100 from "make" can be reduced a bit by putting "@" in front of each libtool
4101 compile command.
4102
4103 5. Patch from the folks at Google for configure.in to be a bit more thorough
4104 in checking for a suitable C++ installation before trying to compile the
4105 C++ stuff. This should fix a reported problem when a compiler was present,
4106 but no suitable headers.
4107
4108 6. The man pages all had just "PCRE" as their title. I have changed them to
4109 be the relevant file name. I have also arranged that these names are
4110 retained in the file doc/pcre.txt, which is a concatenation in text format
4111 of all the man pages except the little individual ones for each function.
4112
4113 7. The NON-UNIX-USE file had not been updated for the different set of source
4114 files that come with release 6. I also added a few comments about the C++
4115 wrapper.
4116
4117
4118 Version 6.0 07-Jun-05
4119 ---------------------
4120
4121 1. Some minor internal re-organization to help with my DFA experiments.
4122
4123 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4124 didn't matter for the library itself when fully configured, but did matter
4125 when compiling without UCP support, or within Exim, where the ucp files are
4126 not imported.
4127
4128 3. Refactoring of the library code to split up the various functions into
4129 different source modules. The addition of the new DFA matching code (see
4130 below) to a single monolithic source would have made it really too
4131 unwieldy, quite apart from causing all the code to be include in a
4132 statically linked application, when only some functions are used. This is
4133 relevant even without the DFA addition now that patterns can be compiled in
4134 one application and matched in another.
4135
4136 The downside of splitting up is that there have to be some external
4137 functions and data tables that are used internally in different modules of
4138 the library but which are not part of the API. These have all had their
4139 names changed to start with "_pcre_" so that they are unlikely to clash
4140 with other external names.
4141
4142 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4143 a different (DFA) algorithm. Although it is slower than the original
4144 function, it does have some advantages for certain types of matching
4145 problem.
4146
4147 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4148 including restarting after a partial match.
4149
4150 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4151 defined when compiling for Windows was sent to me. I have put it into the
4152 code, though I have no means of testing or verifying it.
4153
4154 7. Added the pcre_refcount() auxiliary function.
4155
4156 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4157 match before or at the first newline in the subject string. In pcretest,
4158 the /f option on a pattern can be used to set this.
4159
4160 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4161 would behave wrongly. This has been present in PCRE since release 4.0.
4162
4163 10. A number of changes to the pcregrep command:
4164
4165 (a) Refactored how -x works; insert ^(...)$ instead of setting
4166 PCRE_ANCHORED and checking the length, in preparation for adding
4167 something similar for -w.
4168
4169 (b) Added the -w (match as a word) option.
4170
4171 (c) Refactored the way lines are read and buffered so as to have more
4172 than one at a time available.
4173
4174 (d) Implemented a pcregrep test script.
4175
4176 (e) Added the -M (multiline match) option. This allows patterns to match
4177 over several lines of the subject. The buffering ensures that at least
4178 8K, or the rest of the document (whichever is the shorter) is available
4179 for matching (and similarly the previous 8K for lookbehind assertions).
4180
4181 (f) Changed the --help output so that it now says
4182
4183 -w, --word-regex(p)
4184
4185 instead of two lines, one with "regex" and the other with "regexp"
4186 because that confused at least one person since the short forms are the
4187 same. (This required a bit of code, as the output is generated
4188 automatically from a table. It wasn't just a text change.)
4189
4190 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4191 option but starts with a hyphen. Could be a pattern or a path name
4192 starting with a hyphen, for instance.
4193
4194 (h) "-" can be given as a file name to represent stdin.
4195
4196 (i) When file names are being printed, "(standard input)" is used for
4197 the standard input, for compatibility with GNU grep. Previously
4198 "<stdin>" was used.
4199
4200 (j) The option --label=xxx can be used to supply a name to be used for
4201 stdin when file names are being printed. There is no short form.
4202
4203 (k) Re-factored the options decoding logic because we are going to add
4204 two more options that take data. Such options can now be given in four
4205 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4206
4207 (l) Added the -A, -B, and -C options for requesting that lines of context
4208 around matches be printed.
4209
4210 (m) Added the -L option to print the names of files that do not contain
4211 any matching lines, that is, the complement of -l.
4212
4213 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4214 continue to scan other files.
4215
4216 (o) The -s option was incorrectly implemented. For compatibility with other
4217 greps, it now suppresses the error message for a non-existent or non-
4218 accessible file (but not the return code). There is a new option called
4219 -q that suppresses the output of matching lines, which was what -s was
4220 previously doing.
4221
4222 (p) Added --include and --exclude options to specify files for inclusion
4223 and exclusion when recursing.
4224
4225 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4226 Hopefully, it now does.
4227
4228 12. Missing cast in pcre_study().
4229
4230 13. Added an "uninstall" target to the makefile.
4231
4232 14. Replaced "extern" in the function prototypes in Makefile.in with
4233 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4234 world, but is set differently for Windows.
4235
4236 15. Added a second compiling function called pcre_compile2(). The only
4237 difference is that it has an extra argument, which is a pointer to an
4238 integer error code. When there is a compile-time failure, this is set
4239 non-zero, in addition to the error test pointer being set to point to an
4240 error message. The new argument may be NULL if no error number is required
4241 (but then you may as well call pcre_compile(), which is now just a
4242 wrapper). This facility is provided because some applications need a
4243 numeric error indication, but it has also enabled me to tidy up the way
4244 compile-time errors are handled in the POSIX wrapper.
4245
4246 16. Added VPATH=.libs to the makefile; this should help when building with one
4247 prefix path and installing with another. (Or so I'm told by someone who
4248 knows more about this stuff than I do.)
4249
4250 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4251 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4252 match everything, including newlines. This is not POSIX-compatible, but
4253 somebody wanted the feature. From pcretest it can be activated by using
4254 both the P and the s flags.
4255
4256 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4257
4258 19. libpcre.pc was being incorrectly installed as executable.
4259
4260 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4261 it now also looks for '\r' so that it will work unmodified on Windows.
4262
4263 21. Added Google's contributed C++ wrapper to the distribution.
4264
4265 22. Added some untidy missing memory free() calls in pcretest, to keep
4266 Electric Fence happy when testing.
4267
4268
4269
4270 Version 5.0 13-Sep-04
4271 ---------------------
4272
4273 1. Internal change: literal characters are no longer packed up into items
4274 containing multiple characters in a single byte-string. Each character
4275 is now matched using a separate opcode. However, there may be more than one
4276 byte in the character in UTF-8 mode.
4277
4278 2. The pcre_callout_block structure has two new fields: pattern_position and
4279 next_item_length. These contain the offset in the pattern to the next match
4280 item, and its length, respectively.
4281
4282 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4283 insertion of callouts before each pattern item. Added the /C option to
4284 pcretest to make use of this.
4285
4286 4. On the advice of a Windows user, the lines
4287
4288 #if defined(_WIN32) || defined(WIN32)
4289 _setmode( _fileno( stdout ), 0x8000 );
4290 #endif /* defined(_WIN32) || defined(WIN32) */
4291
4292 have been added to the source of pcretest. This apparently does useful
4293 magic in relation to line terminators.
4294
4295 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4296 for the benefit of those environments where the "b" makes a difference.
4297
4298 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4299 to know about it. I have put a hack into configure.in that adds in code
4300 to set GCC=yes if CC=icc. This seems to end up at a point in the
4301 generated configure script that is early enough to affect the setting of
4302 compiler options, which is what is needed, but I have no means of testing
4303 whether it really works. (The user who reported this had patched the
4304 generated configure script, which of course I cannot do.)
4305
4306 LATER: After change 22 below (new libtool files), the configure script
4307 seems to know about icc (and also ecc). Therefore, I have commented out
4308 this hack in configure.in.
4309
4310 7. Added support for pkg-config (2 patches were sent in).
4311
4312 8. Negated POSIX character classes that used a combination of internal tables
4313 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4314 [[:^ascii]]. Typically, they would match almost any characters. The other
4315 POSIX classes were not broken in this way.
4316
4317 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4318 to find the match, as PCRE was deluded into thinking that the match had to
4319 start at the start point or following a newline. The same bug applied to
4320 patterns with negative forward assertions or any backward assertions
4321 preceding ".*" at the start, unless the pattern required a fixed first
4322 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4323
4324 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4325 starting at the last subject character, bytes beyond the end of the subject
4326 string were read.
4327
4328 11. Renamed the variable "class" as "classbits" to make life easier for C++
4329 users. (Previously there was a macro definition, but it apparently wasn't
4330 enough.)
4331
4332 12. Added the new field "tables" to the extra data so that tables can be passed
4333 in at exec time, or the internal tables can be re-selected. This allows
4334 a compiled regex to be saved and re-used at a later time by a different
4335 program that might have everything at different addresses.
4336
4337 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4338 -R library as well as a -L library.
4339
4340 14. The debugging options of pcretest (-d on the command line or D on a
4341 pattern) showed incorrect output for anything following an extended class
4342 that contained multibyte characters and which was followed by a quantifier.
4343
4344 15. Added optional support for general category Unicode character properties
4345 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4346 support. It adds about 90K to the size of the library. The meanings of the
4347 inbuilt class escapes such as \d and \s have NOT been changed.
4348
4349 16. Updated pcredemo.c to include calls to free() to release the memory for the
4350 compiled pattern.
4351
4352 17. The generated file chartables.c was being created in the source directory
4353 instead of in the building directory. This caused the build to fail if the
4354 source directory was different from the building directory, and was
4355 read-only.
4356
4357 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4358 file. No doubt somebody will tell me if they don't make sense... Also added
4359 Dan Mooney's comments about building on OpenVMS.
4360
4361 19. Added support for partial matching via the PCRE_PARTIAL option for
4362 pcre_exec() and the \P data escape in pcretest.
4363
4364 20. Extended pcretest with 3 new pattern features:
4365
4366 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4367 write the compiled pattern to the file whose name is "rest-of-line".
4368 This is a straight binary dump of the data, with the saved pointer to
4369 the character tables forced to be NULL. The study data, if any, is
4370 written too. After writing, pcretest reads a new pattern.
4371
4372 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4373 compiled pattern from the given file. There must not be any
4374 occurrences of "<" in the file name (pretty unlikely); if there are,
4375 pcretest will instead treat the initial "<" as a pattern delimiter.
4376 After reading in the pattern, pcretest goes on to read data lines as
4377 usual.
4378
4379 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4380 and 16-bit fields in a compiled pattern, to simulate a pattern that
4381 was compiled on a host of opposite endianness.
4382
4383 21. The pcre-exec() function can now cope with patterns that were compiled on
4384 hosts of opposite endianness, with this restriction:
4385
4386 As for any compiled expression that is saved and used later, the tables
4387 pointer field cannot be preserved; the extra_data field in the arguments
4388 to pcre_exec() should be used to pass in a tables address if a value
4389 other than the default internal tables were used at compile time.
4390
4391 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4392 now diagnosed as an error. Previously, most of the time, a negative number
4393 would have been treated as zero, but if in addition "ovector" was passed as
4394 NULL, a crash could occur.
4395
4396 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4397 new versions from the libtool 1.5 distribution (the last one is a copy of
4398 a file called libtool.m4). This seems to have fixed the need to patch
4399 "configure" to support Darwin 1.3 (which I used to do). However, I still
4400 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4401 workstation).
4402
4403 24. Changed the PCRE licence to be the more standard "BSD" licence.
4404
4405
4406 Version 4.5 01-Dec-03
4407 ---------------------
4408
4409 1. There has been some re-arrangement of the code for the match() function so
4410 that it can be compiled in a version that does not call itself recursively.
4411 Instead, it keeps those local variables that need separate instances for
4412 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4413 needs to "recurse". Keeping track of where control must go is done by means
4414 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4415 hide most of the details from the main code, and operates only if
4416 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4417 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4418 operating.
4419
4420 To make it easier for callers to provide specially tailored get/free
4421 functions for this usage, two new functions, pcre_stack_malloc, and
4422 pcre_stack_free, are used. They are always called in strict stacking order,
4423 and the size of block requested is always the same.
4424
4425 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4426 PCRE has been compiled to use the stack or the heap for recursion. The
4427 -C option of pcretest uses this to show which version is compiled.
4428
4429 A new data escape \S, is added to pcretest; it causes the amounts of store
4430 obtained and freed by both kinds of malloc/free at match time to be added
4431 to the output.
4432
4433 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4434 what's available on my current Linux desktop machine.
4435
4436 3. When matching a UTF-8 string, the test for a valid string at the start has
4437 been extended. If start_offset is not zero, PCRE now checks that it points
4438 to a byte that is the start of a UTF-8 character. If not, it returns
4439 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4440 this is necessary because there may be backward assertions in the pattern.
4441 When matching the same subject several times, it may save resources to use
4442 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4443
4444 4. The code for checking the validity of UTF-8 strings has been tightened so
4445 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4446 containing "overlong sequences".
4447
4448 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4449 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4450 should have been "&", but it just so happened that all the cases this let
4451 through by mistake were picked up later in the function.
4452
4453 6. I had used a variable called "isblank" - this is a C99 function, causing
4454 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4455
4456 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4457 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4458 so the version is shown; (c) stop "make test" echoing "./RunTest".
4459
4460 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4461
4462 9. The prototype for memmove() for systems that don't have it was using
4463 size_t, but the inclusion of the header that defines size_t was later. I've
4464 moved the #includes for the C headers earlier to avoid this.
4465
4466 10. Added some adjustments to the code to make it easier to compiler on certain
4467 special systems:
4468
4469 (a) Some "const" qualifiers were missing.
4470 (b) Added the macro EXPORT before all exported functions; by default this
4471 is defined to be empty.
4472 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4473 that it reads its output file name as an argument instead of writing
4474 to the standard output and assuming this can be redirected.
4475
4476 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4477 class containing characters with values greater than 255, PCRE compilation
4478 went into a loop.
4479
4480 12. A recursive reference to a subpattern that was within another subpattern
4481 that had a minimum quantifier of zero caused PCRE to crash. For example,
4482 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4483 recursion. If the recursively-called subpattern itself had a zero repeat,
4484 that was OK.
4485
4486 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4487 buffer into which it was copied (for escape processing) was still set at
4488 1024, so long lines caused crashes.
4489
4490 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4491 "internal error: code overflow...". This applied to any character class
4492 that was followed by a possessive quantifier.
4493
4494 15. Modified the Makefile to add libpcre.la as a prerequisite for
4495 libpcreposix.la because I was told this is needed for a parallel build to
4496 work.
4497
4498 16. If a pattern that contained .* following optional items at the start was
4499 studied, the wrong optimizing data was generated, leading to matching
4500 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4501 matching string must start with a or b or c. The correct conclusion for
4502 this pattern is that a match can start with any character.
4503
4504
4505 Version 4.4 13-Aug-03
4506 ---------------------
4507
4508 1. In UTF-8 mode, a character class containing characters with values between
4509 127 and 255 was not handled correctly if the compiled pattern was studied.
4510 In fixing this, I have also improved the studying algorithm for such
4511 classes (slightly).
4512
4513 2. Three internal functions had redundant arguments passed to them. Removal
4514 might give a very teeny performance improvement.
4515
4516 3. Documentation bug: the value of the capture_top field in a callout is *one
4517 more than* the number of the hightest numbered captured substring.
4518
4519 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4520 in incorrectly linking with a previously installed version. They now link
4521 explicitly with libpcre.la.
4522
4523 5. configure.in no longer needs to recognize Cygwin specially.
4524
4525 6. A problem in pcre.in for Windows platforms is fixed.
4526
4527 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4528 pcretest, it used to include the size of the study block as part of its
4529 output. Unfortunately, the structure contains a field that has a different
4530 size on different hardware architectures. This meant that the tests that
4531 showed this size failed. As the block is currently always of a fixed size,
4532 this information isn't actually particularly useful in pcretest output, so
4533 I have just removed it.
4534
4535 8. Three pre-processor statements accidentally did not start in column 1.
4536 Sadly, there are *still* compilers around that complain, even though
4537 standard C has not required this for well over a decade. Sigh.
4538
4539 9. In pcretest, the code for checking callouts passed small integers in the
4540 callout_data field, which is a void * field. However, some picky compilers
4541 complained about the casts involved for this on 64-bit systems. Now
4542 pcretest passes the address of the small integer instead, which should get
4543 rid of the warnings.
4544
4545 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4546 both compile and run time, and gives an error if an invalid UTF-8 sequence
4547 is found. There is a option for disabling this check in cases where the
4548 string is known to be correct and/or the maximum performance is wanted.
4549
4550 11. In response to a bug report, I changed one line in Makefile.in from
4551
4552 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4553 to
4554 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4555
4556 to look similar to other lines, but I have no way of telling whether this
4557 is the right thing to do, as I do not use Windows. No doubt I'll get told
4558 if it's wrong...
4559
4560
4561 Version 4.3 21-May-03
4562 ---------------------
4563
4564 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4565 Makefile.
4566
4567 2. Some refactoring to improve the quality of the code:
4568
4569 (i) The utf8_table... variables are now declared "const".
4570
4571 (ii) The code for \cx, which used the "case flipping" table to upper case
4572 lower case letters, now just substracts 32. This is ASCII-specific,
4573 but the whole concept of \cx is ASCII-specific, so it seems
4574 reasonable.
4575
4576 (iii) PCRE was using its character types table to recognize decimal and
4577 hexadecimal digits in the pattern. This is silly, because it handles
4578 only 0-9, a-f, and A-F, but the character types table is locale-
4579 specific, which means strange things might happen. A private
4580 table is now used for this - though it costs 256 bytes, a table is
4581 much faster than multiple explicit tests. Of course, the standard
4582 character types table is still used for matching digits in subject
4583 strings against \d.
4584
4585 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4586 ending in _t are). So I've renamed it as ESC_tee.
4587
4588 3. The first argument for regexec() in the POSIX wrapper should have been
4589 defined as "const".
4590
4591 4. Changed pcretest to use malloc() for its buffers so that they can be
4592 Electric Fenced for debugging.
4593
4594 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4595 to read one or more bytes before the start of the subject string. Often this
4596 had no effect on PCRE's behaviour, but in some circumstances it could
4597 provoke a segmentation fault.
4598
4599 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4600 to try to read one or more bytes before the start of the subject string.
4601
4602 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4603 UTF-8 support could misbehave in various ways if the subject string
4604 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4605 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4606 back over UTF-8 characters.)
4607
4608
4609 Version 4.2 14-Apr-03
4610 ---------------------
4611
4612 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4613
4614 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4615 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4616 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4617 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4618 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4619 and BUILD_EXEEXT
4620 Note: automatic setting of the BUILD variables is not yet working
4621 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4622 compile-time but not at link-time
4623 [LINK]: use for linking executables only
4624 make different versions for Windows and non-Windows
4625 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4626 libraries
4627 [LINK_FOR_BUILD]: new variable
4628 [OBJEXT]: use throughout
4629 [EXEEXT]: use throughout
4630 <winshared>: new target
4631 <wininstall>: new target
4632 <dftables.o>: use native compiler
4633 <dftables>: use native linker
4634 <install>: handle Windows platform correctly
4635 <clean>: ditto
4636 <check>: ditto
4637 copy DLL to top builddir before testing
4638
4639 As part of these changes, -no-undefined was removed again. This was reported
4640 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4641 in any case.
4642
4643 3. Some tidies to get rid of compiler warnings:
4644
4645 . In the match_data structure, match_limit was an unsigned long int, whereas
4646 match_call_count was an int. I've made them both unsigned long ints.
4647
4648 . In pcretest the fact that a const uschar * doesn't automatically cast to
4649 a void * provoked a warning.
4650
4651 . Turning on some more compiler warnings threw up some "shadow" variables
4652 and a few more missing casts.
4653
4654 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4655 option, a class that contained a single character with a value between 128
4656 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4657
4658 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4659 option, a class that contained several characters, but with at least one
4660 whose value was between 128 and 255 caused PCRE to crash.
4661
4662
4663 Version 4.1 12-Mar-03
4664 ---------------------
4665
4666 1. Compiling with gcc -pedantic found a couple of places where casts were
4667 needed, and a string in dftables.c that was longer than standard compilers are
4668 required to support.
4669
4670 2. Compiling with Sun's compiler found a few more places where the code could
4671 be tidied up in order to avoid warnings.
4672
4673 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4674 first of these names is deprecated in the latest Autoconf in favour of the name
4675 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4676 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4677 analogy I have changed it to CFLAGS_FOR_BUILD.
4678
4679 4. Added -no-undefined to the linking command in the Makefile, because this is
4680 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4681 linking step for the pcreposix library.
4682
4683 5. PCRE was failing to diagnose the case of two named groups with the same
4684 name.
4685
4686 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4687 literal character that is needed in the subject for a match, and scans along to
4688 ensure that it is present before embarking on the full matching process. This
4689 saves time in cases of nested unlimited repeats that are never going to match.
4690 Problem: the scan can take a lot of time if the subject is very long (e.g.
4691 megabytes), thus penalizing straightforward matches. It is now done only if the
4692 amount of subject to be scanned is less than 1000 bytes.
4693
4694 7. A lesser problem with the same optimization is that it was recording the
4695 first character of an anchored pattern as "needed", thus provoking a search
4696 right along the subject, even when the first match of the pattern was going to
4697 fail. The "needed" character is now not set for anchored patterns, unless it
4698 follows something in the pattern that is of non-fixed length. Thus, it still
4699 fulfils its original purpose of finding quick non-matches in cases of nested
4700 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4701
4702
4703 Version 4.0 17-Feb-03
4704 ---------------------
4705
4706 1. If a comment in an extended regex that started immediately after a meta-item
4707 extended to the end of string, PCRE compiled incorrect data. This could lead to
4708 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4709
4710 2. Moved to autoconf 2.53 and libtool 1.4.2.
4711
4712 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4713 the special perltest8 script is no longer needed - all the tests can be run
4714 from a single perltest script.
4715
4716 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4717 by \s. It has now been removed in PCRE. This means it isn't recognized as
4718 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4719 class [:space:] *does* include VT, thereby creating a mess.
4720
4721 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4722 space and tab.
4723
4724 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4725 its new features into the main test script, reducing the number of scripts.
4726
4727 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4728 were backward compatible, and made the (?i) apply to the whole pattern, as if
4729 /i were given. Now it behaves more logically, and applies the option setting
4730 only to what follows. PCRE has been changed to follow suit. However, if it
4731 finds options settings right at the start of the pattern, it extracts them into
4732 the global options, as before. Thus, they show up in the info data.
4733
4734 8. Added support for the \Q...\E escape sequence. Characters in between are
4735 treated as literals. This is slightly different from Perl in that $ and @ are
4736 also handled as literals inside the quotes. In Perl, they will cause variable
4737 interpolation. Note the following examples:
4738
4739 Pattern PCRE matches Perl matches
4740
4741 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4742 \Qabc\$xyz\E abc\$xyz abc\$xyz
4743 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4744
4745 For compatibility with Perl, \Q...\E sequences are recognized inside character
4746 classes as well as outside them.
4747
4748 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4749 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4750 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4751 signed/unsigned warnings.
4752
4753 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4754 option for pcretest, so I've replaced it by a simple function that does just
4755 that job.
4756
4757 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4758 "pcregrep -".
4759
4760 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4761 Java package. This provides some syntactic sugar for simple cases of what my
4762 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4763 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4764 item, you can use this simplified notation. Note that only makes sense with
4765 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4766 greediness, whatever the setting of the PCRE_UNGREEDY option.
4767
4768 13. A change of greediness default within a pattern was not taking effect at
4769 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4770 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4771 was abstracted outside.
4772
4773 14. PCRE now supports the \G assertion. It is true when the current matching
4774 position is at the start point of the match. This differs from \A when the
4775 starting offset is non-zero. Used with the /g option of pcretest (or similar
4776 code), it works in the same way as it does for Perl's /g option. If all
4777 alternatives of a regex begin with \G, the expression is anchored to the start
4778 match position, and the "anchored" flag is set in the compiled expression.
4779
4780 15. Some bugs concerning the handling of certain option changes within patterns
4781 have been fixed. These applied to options other than (?ims). For example,
4782 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4783 been the other way round. Some of this was related to change 7 above.
4784
4785 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4786 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4787 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4788 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4789
4790 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4791 mode. Unlike ".", it always matches newline, whatever the setting of
4792 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4793 assertions. Perl allows it, but it doesn't (in general) work because it can't
4794 calculate the length of the lookbehind. At least, that's the case for Perl
4795 5.8.0 - I've been told they are going to document that it doesn't work in
4796 future.
4797
4798 18. Added an error diagnosis for escapes that PCRE does not support: these are
4799 \L, \l, \N, \P, \p, \U, \u, and \X.
4800
4801 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4802 reading past the end of the pattern in cases such as /[abcd/.
4803
4804 20. PCRE was getting more memory than necessary for patterns with classes that
4805 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4806
4807 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4808 compiling PCRE for use with Virtual Pascal.
4809
4810 22. Small fix to the Makefile to make it work properly if the build is done
4811 outside the source tree.
4812
4813 23. Added a new extension: a condition to go with recursion. If a conditional
4814 subpattern starts with (?(R) the "true" branch is used if recursion has
4815 happened, whereas the "false" branch is used only at the top level.
4816
4817 24. When there was a very long string of literal characters (over 255 bytes
4818 without UTF support, over 250 bytes with UTF support), the computation of how
4819 much memory was required could be incorrect, leading to segfaults or other
4820 strange effects.
4821
4822 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4823 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4824 there was a subsequent back reference to those brackets. This meant that, for
4825 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4826 possible to check for precisely this case. All we can do is abandon the
4827 optimization if .* occurs inside capturing brackets when there are any back
4828 references whatsoever. (See below for a better fix that came later.)
4829
4830 26. The handling of the optimization for finding the first character of a
4831 non-anchored pattern, and for finding a character that is required later in the
4832 match were failing in some cases. This didn't break the matching; it just
4833 failed to optimize when it could. The way this is done has been re-implemented.
4834
4835 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4836
4837 28. Added a new feature that provides some of the functionality that Perl
4838 provides with (?{...}). The facility is termed a "callout". The way it is done
4839 in PCRE is for the caller to provide an optional function, by setting
4840 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4841 global variable. By default it is unset, which disables all calling out. To get
4842 the function called, the regex must include (?C) at appropriate points. This
4843 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4844 This provides a means of identifying different callout points. When PCRE
4845 reaches such a point in the regex, if pcre_callout has been set, the external
4846 function is called. It is provided with data in a structure called
4847 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4848 matching continues; if it returns a non-zero value, the match at the current
4849 point fails. However, backtracking will occur if possible. [This was changed
4850 later and other features added - see item 49 below.]
4851
4852 29. pcretest is upgraded to test the callout functionality. It provides a
4853 callout function that displays information. By default, it shows the start of
4854 the match and the current position in the text. There are some new data escapes
4855 to vary what happens:
4856
4857 \C+ in addition, show current contents of captured substrings
4858 \C- do not supply a callout function
4859 \C!n return 1 when callout number n is reached
4860 \C!n!m return 1 when callout number n is reached for the mth time
4861
4862 30. If pcregrep was called with the -l option and just a single file name, it
4863 output "<stdin>" if a match was found, instead of the file name.
4864
4865 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4866 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4867 pcre_exec(). This saves a malloc/free per call. The default value of
4868 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4869 when configuring.
4870
4871 32. The default maximum size of a compiled pattern is 64K. There have been a
4872 few cases of people hitting this limit. The code now uses macros to handle the
4873 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4874 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4875 configuring. Tests 2 and 5 work only with 2-byte links because they output
4876 debugging information about compiled patterns.
4877
4878 33. Internal code re-arrangements:
4879
4880 (a) Moved the debugging function for printing out a compiled regex into
4881 its own source file (printint.c) and used #include to pull it into
4882 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4883 separate copies.
4884
4885 (b) Defined the list of op-code names for debugging as a macro in
4886 internal.h so that it is next to the definition of the opcodes.
4887
4888 (c) Defined a table of op-code lengths for simpler skipping along compiled
4889 code. This is again a macro in internal.h so that it is next to the
4890 definition of the opcodes.
4891
4892 34. Added support for recursive calls to individual subpatterns, along the
4893 lines of Robin Houston's patch (but implemented somewhat differently).
4894
4895 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4896 allow it to read and process whole directories in Win32. This code was
4897 contributed by Lionel Fourquaux; it has not been tested by me.
4898
4899 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4900 used to name a group. Names consist of alphanumerics and underscores, and must
4901 be unique. Back references use the syntax (?P=name) and recursive calls use
4902 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4903 numbers. The function pcre_fullinfo() can be used after compilation to extract
4904 a name/number map. There are three relevant calls:
4905
4906 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4907 PCRE_INFO_NAMECOUNT yields the number of entries
4908 PCRE_INFO_NAMETABLE yields a pointer to the map.
4909
4910 The map is a vector of fixed-size entries. The size of each entry depends on
4911 the length of the longest name used. The first two bytes of each entry are the
4912 group number, most significant byte first. There follows the corresponding
4913 name, zero terminated. The names are in alphabetical order.
4914
4915 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4916 case instead of 255. Making it the same both with and without UTF-8 support
4917 means that the same test output works with both.
4918
4919 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4920 calling malloc() with a zero argument.
4921
4922 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4923 optimization. I've improved things by keeping a bitmap of backreferences with
4924 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4925 fact referenced, the optimization can be applied. It is unlikely that a
4926 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4927 the match to follow \n) will appear inside brackets with a number greater than
4928 31, but if it does, any back reference > 31 suppresses the optimization.
4929
4930 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4931 of disabling numbered capturing parentheses. Any opening parenthesis that is
4932 not followed by ? behaves as if it were followed by ?: but named parentheses
4933 can still be used for capturing (and they will acquire numbers in the usual
4934 way).
4935
4936 41. Redesigned the return codes from the match() function into yes/no/error so
4937 that errors can be passed back from deep inside the nested calls. A malloc
4938 failure while inside a recursive subpattern call now causes the
4939 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4940
4941 42. It is now possible to set a limit on the number of times the match()
4942 function is called in a call to pcre_exec(). This facility makes it possible to
4943 limit the amount of recursion and backtracking, though not in a directly
4944 obvious way, because the match() function is used in a number of different
4945 circumstances. The count starts from zero for each position in the subject
4946 string (for non-anchored patterns). The default limit is, for compatibility, a
4947 large number, namely 10 000 000. You can change this in two ways:
4948
4949 (a) When configuring PCRE before making, you can use --with-match-limit=n
4950 to set a default value for the compiled library.
4951
4952 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4953 a different value is set. See 45 below.
4954
4955 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4956
4957 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4958 of things that can be changed at compile time. The first argument specifies
4959 what is wanted and the second points to where the information is to be placed.
4960 The current list of available information is:
4961
4962 PCRE_CONFIG_UTF8
4963
4964 The output is an integer that is set to one if UTF-8 support is available;
4965 otherwise it is set to zero.
4966
4967 PCRE_CONFIG_NEWLINE
4968
4969 The output is an integer that it set to the value of the code that is used for
4970 newline. It is either LF (10) or CR (13).
4971
4972 PCRE_CONFIG_LINK_SIZE
4973
4974 The output is an integer that contains the number of bytes used for internal
4975 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4976
4977 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4978
4979 The output is an integer that contains the threshold above which the POSIX
4980 interface uses malloc() for output vectors. See item 31 above.
4981
4982 PCRE_CONFIG_MATCH_LIMIT
4983
4984 The output is an unsigned integer that contains the default limit of the number
4985 of match() calls in a pcre_exec() execution. See 42 above.
4986
4987 44. pcretest has been upgraded by the addition of the -C option. This causes it
4988 to extract all the available output from the new pcre_config() function, and to
4989 output it. The program then exits immediately.
4990
4991 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4992 order to support additional features. One way would have been to define
4993 pcre_exec2() (for example) with extra arguments, but this would not have been
4994 extensible, and would also have required all calls to the original function to
4995 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4996 is used for passing in "extra" data from pcre_study().
4997
4998 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4999 contains the following fields:
5000
5001 flags a bitmap indicating which of the following fields are set
5002 study_data opaque data from pcre_study()
5003 match_limit a way of specifying a limit on match() calls for a specific
5004 call to pcre_exec()
5005 callout_data data for callouts (see 49 below)
5006
5007 The flag bits are also defined in pcre.h, and are
5008
5009 PCRE_EXTRA_STUDY_DATA
5010 PCRE_EXTRA_MATCH_LIMIT
5011 PCRE_EXTRA_CALLOUT_DATA
5012
5013 The pcre_study() function now returns one of these new pcre_extra blocks, with
5014 the actual study data pointed to by the study_data field, and the
5015 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5016 before. That is, this change is entirely upwards-compatible and requires no
5017 change to existing code.
5018
5019 If you want to pass in additional data to pcre_exec(), you can either place it
5020 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5021 block.
5022
5023 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5024 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5025 times with different match limits, until it finds the minimum value needed for
5026 pcre_exec() to complete. The value is then output. This can be instructive; for
5027 most simple matches the number is quite small, but for pathological cases it
5028 gets very large very quickly.
5029
5030 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5031 returns the size of the data block pointed to by the study_data field in a
5032 pcre_extra block, that is, the value that was passed as the argument to
5033 pcre_malloc() when PCRE was getting memory in which to place the information
5034 created by pcre_study(). The fourth argument should point to a size_t variable.
5035 pcretest has been extended so that this information is shown after a successful
5036 pcre_study() call when information about the compiled regex is being displayed.
5037
5038 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5039 because what follows is always an absolute path. (Later: it turns out that this
5040 is more than cosmetic for MinGW, because it doesn't like empty path
5041 components.)
5042
5043 49. Some changes have been made to the callout feature (see 28 above):
5044
5045 (i) A callout function now has three choices for what it returns:
5046
5047 0 => success, carry on matching
5048 > 0 => failure at this point, but backtrack if possible
5049 < 0 => serious error, return this value from pcre_exec()
5050
5051 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5052 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5053 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5054 use by callout functions. It will never be used by PCRE itself.
5055
5056 (ii) The pcre_extra structure (see 45 above) has a void * field called
5057 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5058 pcre_callout_block structure has a field of the same name. The contents of
5059 the field passed in the pcre_extra structure are passed to the callout
5060 function in the corresponding field in the callout block. This makes it
5061 easier to use the same callout-containing regex from multiple threads. For
5062 testing, the pcretest program has a new data escape
5063
5064 \C*n pass the number n (may be negative) as callout_data
5065
5066 If the callout function in pcretest receives a non-zero value as
5067 callout_data, it returns that value.
5068
5069 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5070 there were some redundant $(CFLAGS) in commands that are now specified as
5071 $(LINK), which already includes $(CFLAGS).
5072
5073 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5074 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5075 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5076 one-byte characters throughout. Note that case-insensitive matching applies
5077 only to characters whose values are less than 256. PCRE doesn't support the
5078 notion of cases for higher-valued characters.
5079
5080 (i) A character class whose characters are all within 0-255 is handled as
5081 a bit map, and the map is inverted for negative classes. Previously, a
5082 character > 255 always failed to match such a class; however it should
5083 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5084
5085 (ii) A negated character class with a single character < 255 is coded as
5086 "not this character" (OP_NOT). This wasn't working properly when the test
5087 character was multibyte, either singly or repeated.
5088
5089 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5090 mode, for example: \x{100}{2,3}.
5091
5092 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5093 singly or repeated) now correctly test multibyte characters. However,
5094 PCRE doesn't recognize any characters with values greater than 255 as
5095 digits, spaces, or word characters. Such characters always match \D, \S,
5096 and \W, and never match \d, \s, or \w.
5097
5098 (v) Classes may now contain characters and character ranges with values
5099 greater than 255. For example: [ab\x{100}-\x{400}].
5100
5101 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5102 PCRE in UTF-8 mode.
5103
5104 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5105 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5106 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5107 value.)
5108
5109 53. The single man page has become too large. I have therefore split it up into
5110 a number of separate man pages. These also give rise to individual HTML pages;
5111 these are now put in a separate directory, and there is an index.html page that
5112 lists them all. Some hyperlinking between the pages has been installed.
5113
5114 54. Added convenience functions for handling named capturing parentheses.
5115
5116 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5117 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5118 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5119 are faulted.
5120
5121 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5122 calling configure. These values are used when compiling the dftables.c program
5123 which is run to generate the source of the default character tables. They
5124 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5125 you will need to set these values.
5126
5127 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5128
5129
5130 Version 3.9 02-Jan-02
5131 ---------------------
5132
5133 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5134
5135 2. If --disable-static was given, the building process failed when trying to
5136 build pcretest and pcregrep. (For some reason it was using libtool to compile
5137 them, which is not right, as they aren't part of the library.)
5138
5139
5140 Version 3.8 18-Dec-01
5141 ---------------------
5142
5143 1. The experimental UTF-8 code was completely screwed up. It was packing the
5144 bytes in the wrong order. How dumb can you get?
5145
5146
5147 Version 3.7 29-Oct-01
5148 ---------------------
5149
5150 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5151 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5152 this didn't happen under Solaris 8, where I normally test things.
5153
5154 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5155 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5156 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5157 this fix an hour or so after the initial 3.7 release.)
5158
5159
5160 Version 3.6 23-Oct-01
5161 ---------------------
5162
5163 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5164 offsets passed as NULL with zero offset count.
5165
5166 2. The config.guess and config.sub files had not been updated when I moved to
5167 the latest autoconf.
5168
5169
5170 Version 3.5 15-Aug-01
5171 ---------------------
5172
5173 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5174 had been forgotten.
5175
5176 2. By using declared but undefined structures, we can avoid using "void"
5177 definitions in pcre.h while keeping the internal definitions of the structures
5178 private.
5179
5180 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5181 user point of view, this means that both static and shared libraries are built
5182 by default, but this can be individually controlled. More of the work of
5183 handling this static/shared cases is now inside libtool instead of PCRE's make
5184 file.
5185
5186 4. The pcretest utility is now installed along with pcregrep because it is
5187 useful for users (to test regexs) and by doing this, it automatically gets
5188 relinked by libtool. The documentation has been turned into a man page, so
5189 there are now .1, .txt, and .html versions in /doc.
5190
5191 5. Upgrades to pcregrep:
5192 (i) Added long-form option names like gnu grep.
5193 (ii) Added --help to list all options with an explanatory phrase.
5194 (iii) Added -r, --recursive to recurse into sub-directories.
5195 (iv) Added -f, --file to read patterns from a file.
5196
5197 6. pcre_exec() was referring to its "code" argument before testing that
5198 argument for NULL (and giving an error if it was NULL).
5199
5200 7. Upgraded Makefile.in to allow for compiling in a different directory from
5201 the source directory.
5202
5203 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5204 options bits, the pointer it was passed was to an int instead of to an unsigned
5205 long int. This mattered only on 64-bit systems.
5206
5207 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5208 generated) instead of pcre.in, which it its source. Also made the same change
5209 in several of the .c files.
5210
5211 10. A new release of gcc defines printf() as a macro, which broke pcretest
5212 because it had an ifdef in the middle of a string argument for printf(). Fixed
5213 by using separate calls to printf().
5214
5215 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5216 script, to force use of CR or LF instead of \n in the source. On non-Unix
5217 systems, the value can be set in config.h.
5218
5219 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5220 absolute limit. Changed the text of the error message to make this clear, and
5221 likewise updated the man page.
5222
5223 13. The limit of 99 on the number of capturing subpatterns has been removed.
5224 The new limit is 65535, which I hope will not be a "real" limit.
5225
5226
5227 Version 3.4 22-Aug-00
5228 ---------------------
5229
5230 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5231
5232 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5233
5234
5235 Version 3.3 01-Aug-00
5236 ---------------------
5237
5238 1. If an octal character was given, but the value was greater than \377, it
5239 was not getting masked to the least significant bits, as documented. This could
5240 lead to crashes in some systems.
5241
5242 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5243 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5244
5245 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5246 These just pass their arguments on to (pcre_free)(), but they are provided
5247 because some uses of PCRE bind it to non-C systems that can call its functions,
5248 but cannot call free() or pcre_free() directly.
5249
5250 4. Add "make test" as a synonym for "make check". Corrected some comments in
5251 the Makefile.
5252
5253 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5254 Makefile.
5255
5256 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5257 command called pgrep for grepping around the active processes.
5258
5259 7. Added the beginnings of support for UTF-8 character strings.
5260
5261 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5262 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5263 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5264 out for the ar command.)
5265
5266
5267 Version 3.2 12-May-00
5268 ---------------------
5269
5270 This is purely a bug fixing release.
5271
5272 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5273 of ZA. This was just one example of several cases that could provoke this bug,
5274 which was introduced by change 9 of version 2.00. The code for breaking
5275 infinite loops after an iteration that matches an empty string was't working
5276 correctly.
5277
5278 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5279 when matched against abbab (for example). After matching an empty string, it
5280 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5281 caused it to match further down the string than it should.
5282
5283 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5284 was there because it doesn't seem to be needed, and it causes trouble on some
5285 systems, as it is not a Standard C header. It has been removed.
5286
5287 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5288 were reported on the Macintosh. The changes were from
5289
5290 while ((c = *(++ptr)) != 0 && c != '\n');
5291 to
5292 while ((c = *(++ptr)) != 0 && c != '\n') ;
5293
5294 Totally extraordinary, but if that's what it takes...
5295
5296 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5297 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5298 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5299 assumes the way PCRE uses memmove() (always moving upwards).
5300
5301 6. PCRE is being used in one environment where strchr() is not available. There
5302 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5303 faster code anyway.
5304
5305
5306 Version 3.1 09-Feb-00
5307 ---------------------
5308
5309 The only change in this release is the fixing of some bugs in Makefile.in for
5310 the "install" target:
5311
5312 (1) It was failing to install pcreposix.h.
5313
5314 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5315
5316
5317 Version 3.0 01-Feb-00
5318 ---------------------
5319
5320 1. Add support for the /+ modifier to perltest (to output $` like it does in
5321 pcretest).
5322
5323 2. Add support for the /g modifier to perltest.
5324
5325 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5326 matches null strings.
5327
5328 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5329 pattern. Perl treats empty patterns specially - it reuses the most recent
5330 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5331 effect.
5332
5333 5. The POSIX interface was broken in that it was just handing over the POSIX
5334 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5335 required a bigger vector, with some working space on the end. This means that
5336 the POSIX wrapper now has to get and free some memory, and copy the results.
5337
5338 6. Added some simple autoconf support, placing the test data and the
5339 documentation in separate directories, re-organizing some of the
5340 information files, and making it build pcre-config (a GNU standard). Also added
5341 libtool support for building PCRE as a shared library, which is now the
5342 default.
5343
5344 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5345 09 are not valid octal constants. Single digits will be used for minor values
5346 less than 10.
5347
5348 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5349 existing programs that set these in the POSIX interface can use PCRE without
5350 modification.
5351
5352 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5353 return all that pcre_info() returns, plus additional data. The pcre_info()
5354 function is retained for compatibility, but is considered to be obsolete.
5355
5356 10. Added experimental recursion feature (?R) to handle one common case that
5357 Perl 5.6 will be able to do with (?p{...}).
5358
5359 11. Added support for POSIX character classes like [:alpha:], which Perl is
5360 adopting.
5361
5362
5363 Version 2.08 31-Aug-99
5364 ----------------------
5365
5366 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5367 trying to match at the startoffset position, but instead was moving forward to
5368 the next newline as if a previous match had failed.
5369
5370 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5371 and could get into a loop if a null string was matched other than at the start
5372 of the subject.
5373
5374 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5375 be distinguished at compile time, and for completeness also added PCRE_DATE.
5376
5377 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5378 in GnuWin32 environments.
5379
5380
5381 Version 2.07 29-Jul-99
5382 ----------------------
5383
5384 1. The documentation is now supplied in plain text form and HTML as well as in
5385 the form of man page sources.
5386
5387 2. C++ compilers don't like assigning (void *) values to other pointer types.
5388 In particular this affects malloc(). Although there is no problem in Standard
5389 C, I've put in casts to keep C++ compilers happy.
5390
5391 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5392 should be (const char *).
5393
5394 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5395 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5396 However, I haven't made this a standard facility. The documentation doesn't
5397 mention it, and the Makefile doesn't support it.
5398
5399 5. The Makefile now contains an "install" target, with editable destinations at
5400 the top of the file. The pcretest program is not installed.
5401
5402 6. pgrep -V now gives the PCRE version number and date.
5403
5404 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5405 causing the entire string to be ignored, instead of just the last character.
5406
5407 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5408 non-matching string, it can take a very, very long time, even for strings of
5409 quite modest length, because of the nested recursion. PCRE now does better in
5410 some of these cases. It does this by remembering the last required literal
5411 character in the pattern, and pre-searching the subject to ensure it is present
5412 before running the real match. In other words, it applies a heuristic to detect
5413 some types of certain failure quickly, and in the above example, if presented
5414 with a string that has no trailing " it gives "no match" very quickly.
5415
5416 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5417 other alternatives are tried instead.
5418
5419
5420 Version 2.06 09-Jun-99
5421 ----------------------
5422
5423 1. Change pcretest's output for amount of store used to show just the code
5424 space, because the remainder (the data block) varies in size between 32-bit and
5425 64-bit systems.
5426
5427 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5428 start matching at. This allows lookbehinds to work when searching for multiple
5429 occurrences in a string.
5430
5431 3. Added additional options to pcretest for testing multiple occurrences:
5432
5433 /+ outputs the rest of the string that follows a match
5434 /g loops for multiple occurrences, using the new startoffset argument
5435 /G loops for multiple occurrences by passing an incremented pointer
5436
5437 4. PCRE wasn't doing the "first character" optimization for patterns starting
5438 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5439 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5440 the letter 'x'. On long subject strings, this gives a significant speed-up.
5441
5442
5443 Version 2.05 21-Apr-99
5444 ----------------------
5445
5446 1. Changed the type of magic_number from int to long int so that it works
5447 properly on 16-bit systems.
5448
5449 2. Fixed a bug which caused patterns starting with .* not to work correctly
5450 when the subject string contained newline characters. PCRE was assuming
5451 anchoring for such patterns in all cases, which is not correct because .* will
5452 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5453 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5454 must be retried after every newline in the subject.
5455
5456
5457 Version 2.04 18-Feb-99
5458 ----------------------
5459
5460 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5461 computation of the store needed to hold the pattern was incorrect (too large).
5462 If such patterns were nested a few deep, this could multiply and become a real
5463 problem.
5464
5465 2. Added /M option to pcretest to show the memory requirement of a specific
5466 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5467
5468 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5469 compiled in such a way that the backtracking after subsequent failure was
5470 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5471 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5472
5473
5474 Version 2.03 02-Feb-99
5475 ----------------------
5476
5477 1. Fixed typo and small mistake in man page.
5478
5479 2. Added 4th condition (GPL supersedes if conflict) and created separate
5480 LICENCE file containing the conditions.
5481
5482 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5483 Perl, that is the internal \ allows the delimiter to be included in the
5484 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5485 the final delimiter, add \ to the end of the pattern (to test the error).
5486
5487 4. Added the convenience functions for extracting substrings after a successful
5488 match. Updated pcretest to make it able to test these functions.
5489
5490
5491 Version 2.02 14-Jan-99
5492 ----------------------
5493
5494 1. Initialized the working variables associated with each extraction so that
5495 their saving and restoring doesn't refer to uninitialized store.
5496
5497 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5498 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5499 fix the problem.
5500
5501 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5502 calls, and wasn't printing the correct value for compiling calls. Increased the
5503 default value of LOOPREPEAT, and the number of significant figures in the
5504 times.
5505
5506 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5507
5508 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5509 a building problem on Windows NT with a FAT file system.
5510
5511
5512 Version 2.01 21-Oct-98
5513 ----------------------
5514
5515 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5516 to character tables built by pcre_maketables() in the current locale. If NULL
5517 is passed, the default tables are used.
5518
5519
5520 Version 2.00 24-Sep-98
5521 ----------------------
5522
5523 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5524 it any more.
5525
5526 2. Allow quantification of (?>) groups, and make it work correctly.
5527
5528 3. The first character computation wasn't working for (?>) groups.
5529
5530 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5531 end of the subject) and add 5.005's \z, which really does match only at the
5532 very end of the subject.
5533
5534 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5535
5536 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5537 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5538 localized options. All options to pcre_study() were also removed.
5539
5540 7. Add other new features from 5.005:
5541
5542 $(?<= positive lookbehind
5543 $(?<! negative lookbehind
5544 (?imsx-imsx) added the unsetting capability
5545 such a setting is global if at outer level; local otherwise
5546 (?imsx-imsx:) non-capturing groups with option setting
5547 (?(cond)re|re) conditional pattern matching
5548
5549 A backreference to itself in a repeated group matches the previous
5550 captured string.
5551
5552 8. General tidying up of studying (both automatic and via "study")
5553 consequential on the addition of new assertions.
5554
5555 9. As in 5.005, unlimited repeated groups that could match an empty substring
5556 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5557 runtime if any iteration does actually match an empty substring.
5558
5559 10. Include the RunTest script in the distribution.
5560
5561 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5562 discrepancies, some of which were old and were also with respect to 5.004. They
5563 have now been fixed.
5564
5565
5566 Version 1.09 28-Apr-98
5567 ----------------------
5568
5569 1. A negated single character class followed by a quantifier with a minimum
5570 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5571 program crashes, or just wrong answers. This did not apply to negated classes
5572 containing more than one character, or to minima other than one.
5573
5574
5575 Version 1.08 27-Mar-98
5576 ----------------------
5577
5578 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5579
5580 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5581 latter must appear before anything that relies on it in the pattern.
5582
5583
5584 Version 1.07 16-Feb-98
5585 ----------------------
5586
5587 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5588 repeat of a potentially empty string).
5589
5590
5591 Version 1.06 23-Jan-98
5592 ----------------------
5593
5594 1. Added Markus Oberhumer's little patches for C++.
5595
5596 2. Literal strings longer than 255 characters were broken.
5597
5598
5599 Version 1.05 23-Dec-97
5600 ----------------------
5601
5602 1. Negated character classes containing more than one character were failing if
5603 PCRE_CASELESS was set at run time.
5604
5605
5606 Version 1.04 19-Dec-97
5607 ----------------------
5608
5609 1. Corrected the man page, where some "const" qualifiers had been omitted.
5610
5611 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5612 input syntax.
5613
5614 3. Fixed memory leak which occurred when a regex with back references was
5615 matched with an offsets vector that wasn't big enough. The temporary memory
5616 that is used in this case wasn't being freed if the match failed.
5617
5618 4. Tidied pcretest to ensure it frees memory that it gets.
5619
5620 5. Temporary memory was being obtained in the case where the passed offsets
5621 vector was exactly big enough.
5622
5623 6. Corrected definition of offsetof() from change 5 below.
5624
5625 7. I had screwed up change 6 below and broken the rules for the use of
5626 setjmp(). Now fixed.
5627
5628
5629 Version 1.03 18-Dec-97
5630 ----------------------
5631
5632 1. A erroneous regex with a missing opening parenthesis was correctly
5633 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5634 on some systems.
5635
5636 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5637 it was reported that one broken compiler failed on the former because "code" is
5638 also an independent variable.
5639
5640 3. The erroneous regex a[]b caused an array overrun reference.
5641
5642 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5643 fail on data ending with that character. (It was going on too far, and checking
5644 the next character, typically a binary zero.) This was specific to the
5645 optimized code for single-character negative classes.
5646
5647 5. Added a contributed patch from the TIN world which does the following:
5648
5649 + Add an undef for memmove, in case the the system defines a macro for it.
5650
5651 + Add a definition of offsetof(), in case there isn't one. (I don't know
5652 the reason behind this - offsetof() is part of the ANSI standard - but
5653 it does no harm).
5654
5655 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5656 most of the places where whitespace preceded '#'. I have given up and
5657 allowed the remaining 2 cases to be at the margin.
5658
5659 + Rename some variables in pcre to eliminate shadowing. This seems very
5660 pedantic, but does no harm, of course.
5661
5662 6. Moved the call to setjmp() into its own function, to get rid of warnings
5663 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5664
5665 7. Constructs such as \d{8,} were compiling into the equivalent of
5666 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5667 outcome, but in this particular case used more store than had been allocated,
5668 which caused the bug to be discovered because it threw up an internal error.
5669
5670 8. The debugging code in both pcre and pcretest for outputting the compiled
5671 form of a regex was going wrong in the case of back references followed by
5672 curly-bracketed repeats.
5673
5674
5675 Version 1.02 12-Dec-97
5676 ----------------------
5677
5678 1. Typos in pcre.3 and comments in the source fixed.
5679
5680 2. Applied a contributed patch to get rid of places where it used to remove
5681 'const' from variables, and fixed some signed/unsigned and uninitialized
5682 variable warnings.
5683
5684 3. Added the "runtest" target to Makefile.
5685
5686 4. Set default compiler flag to -O2 rather than just -O.
5687
5688
5689 Version 1.01 19-Nov-97
5690 ----------------------
5691
5692 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5693 like /([ab]*)*/, that is, for classes with more than one character in them.
5694
5695 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5696 as /((?>a*))*/ (a PCRE_EXTRA facility).
5697
5698
5699 Version 1.00 18-Nov-97
5700 ----------------------
5701
5702 1. Added compile-time macros to support systems such as SunOS4 which don't have
5703 memmove() or strerror() but have other things that can be used instead.
5704
5705 2. Arranged that "make clean" removes the executables.
5706
5707
5708 Version 0.99 27-Oct-97
5709 ----------------------
5710
5711 1. Fixed bug in code for optimizing classes with only one character. It was
5712 initializing a 32-byte map regardless, which could cause it to run off the end
5713 of the memory it had got.
5714
5715 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5716
5717
5718 Version 0.98 22-Oct-97
5719 ----------------------
5720
5721 1. Fixed bug in code for handling temporary memory usage when there are more
5722 back references than supplied space in the ovector. This could cause segfaults.
5723
5724
5725 Version 0.97 21-Oct-97
5726 ----------------------
5727
5728 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5729
5730 2. Optimized negated single characters not to use a bit map.
5731
5732 3. Brought error texts together as macro definitions; clarified some of them;
5733 fixed one that was wrong - it said "range out of order" when it meant "invalid
5734 escape sequence".
5735
5736 4. Changed some char * arguments to const char *.
5737
5738 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5739
5740 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5741 pcretest.
5742
5743
5744 Version 0.96 16-Oct-97
5745 ----------------------
5746
5747 1. Added a simple "pgrep" utility to the distribution.
5748
5749 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5750 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5751 where "ddd" means "one or more decimal digits".
5752
5753 3. Fixed serious bug. If a pattern had a back reference, but the call to
5754 pcre_exec() didn't supply a large enough ovector to record the related
5755 identifying subpattern, the match always failed. PCRE now remembers the number
5756 of the largest back reference, and gets some temporary memory in which to save
5757 the offsets during matching if necessary, in order to ensure that
5758 backreferences always work.
5759
5760 4. Increased the compatibility with Perl in a number of ways:
5761
5762 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5763 to request this handling. The option can be set at compile or exec time.
5764
5765 (b) $ matches before a terminating newline by default; an option
5766 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5767 mode). The option can be set at compile or exec time.
5768
5769 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5770 the same as Perl's. If the decimal number it represents is less than 10
5771 or there aren't that many previous left capturing parentheses, an octal
5772 escape is read. Inside a character class, it's always an octal escape,
5773 even if it is a single digit.
5774
5775 (d) An escaped but undefined alphabetic character is taken as a literal,
5776 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5777 escapes.
5778
5779 (e) {0} is now permitted. (The previous item is removed from the compiled
5780 pattern).
5781
5782 5. Changed all the names of code files so that the basic parts are no longer
5783 than 10 characters, and abolished the teeny "globals.c" file.
5784
5785 6. Changed the handling of character classes; they are now done with a 32-byte
5786 bit map always.
5787
5788 7. Added the -d and /D options to pcretest to make it possible to look at the
5789 internals of compilation without having to recompile pcre.
5790
5791
5792 Version 0.95 23-Sep-97
5793 ----------------------
5794
5795 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5796 \x20 at the start of a run of normal characters. These were being treated as
5797 real characters, instead of the source characters being re-checked.
5798
5799
5800 Version 0.94 18-Sep-97
5801 ----------------------
5802
5803 1. The functions are now thread-safe, with the caveat that the global variables
5804 containing pointers to malloc() and free() or alternative functions are the
5805 same for all threads.
5806
5807 2. Get pcre_study() to generate a bitmap of initial characters for non-
5808 anchored patterns when this is possible, and use it if passed to pcre_exec().
5809
5810
5811 Version 0.93 15-Sep-97
5812 ----------------------
5813
5814 1. /(b)|(:+)/ was computing an incorrect first character.
5815
5816 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5817 but not actually doing anything yet.
5818
5819 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5820 as Perl does (e.g. [-az] or [az-]).
5821
5822 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5823 all possible positions.
5824
5825 5. Split up into different modules to avoid including unneeded functions in a
5826 compiled binary. However, compile and exec are still in one module. The "study"
5827 function is split off.
5828
5829 6. The character tables are now in a separate module whose source is generated
5830 by an auxiliary program - but can then be edited by hand if required. There are
5831 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5832 toupper() in the code.
5833
5834 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5835 make them global. Abolish the function for setting them, as the caller can now
5836 set them directly.
5837
5838
5839 Version 0.92 11-Sep-97
5840 ----------------------
5841
5842 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5843 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5844
5845 2. Caseless matching was not working in character classes if the characters in
5846 the pattern were in upper case.
5847
5848 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5849
5850 4. Make PCRE_ANCHORED public and accept as a compile option.
5851
5852 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5853 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5854 pass them.
5855
5856 6. Give an error if bad option bits passed at compile or run time.
5857
5858 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5859 pcretest to cause it to pass that flag.
5860
5861 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5862 options, and the first character, if set.
5863
5864 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5865
5866
5867 Version 0.91 10-Sep-97
5868 ----------------------
5869
5870 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5871 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5872
5873 2. PCRE was looping on encountering an indefinitely repeated back reference to
5874 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5875 Perl does - treats the match as successful.
5876
5877 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5