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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1562 - (show annotations)
Wed Jun 3 16:51:59 2015 UTC (5 years, 11 months ago) by ph10
File size: 279259 byte(s)
Fix another buffer overflow.
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
44 Version 8.37 28-April-2015
45 --------------------------
46
47 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
48 for those parentheses to be closed with whatever has been captured so far.
49 However, it was failing to mark any other groups between the hightest
50 capture so far and the currrent group as "unset". Thus, the ovector for
51 those groups contained whatever was previously there. An example is the
52 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
53
54 2. If an assertion condition was quantified with a minimum of zero (an odd
55 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
56
57 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
58 unrecognized modifier, a crash could occur.
59
60 4. An attempt to do global matching in pcretest with a zero-length ovector
61 caused a crash.
62
63 5. Fixed a memory leak during matching that could occur for a subpattern
64 subroutine call (recursive or otherwise) if the number of captured groups
65 that had to be saved was greater than ten.
66
67 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
68 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
69 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
70
71 7. A UTF pattern containing a "not" match of a non-ASCII character and a
72 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
73
74 8. When a pattern is compiled, it remembers the highest back reference so that
75 when matching, if the ovector is too small, extra memory can be obtained to
76 use instead. A conditional subpattern whose condition is a check on a
77 capture having happened, such as, for example in the pattern
78 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
79 setting the highest backreference number. This mattered only if pcre_exec()
80 was called with an ovector that was too small to hold the capture, and there
81 was no other kind of back reference (a situation which is probably quite
82 rare). The effect of the bug was that the condition was always treated as
83 FALSE when the capture could not be consulted, leading to a incorrect
84 behaviour by pcre_exec(). This bug has been fixed.
85
86 9. A reference to a duplicated named group (either a back reference or a test
87 for being set in a conditional) that occurred in a part of the pattern where
88 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
89 to be incorrectly calculated, leading to overwriting.
90
91 10. A mutually recursive set of back references such as (\2)(\1) caused a
92 segfault at study time (while trying to find the minimum matching length).
93 The infinite loop is now broken (with the minimum length unset, that is,
94 zero).
95
96 11. If an assertion that was used as a condition was quantified with a minimum
97 of zero, matching went wrong. In particular, if the whole group had
98 unlimited repetition and could match an empty string, a segfault was
99 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
100 assertions to be quantified, but not if they are being used as conditions,
101 so the above pattern is faulted by Perl. PCRE has now been changed so that
102 it also rejects such patterns.
103
104 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
105 failed to allow the zero-repeat case if pcre2_exec() was called with an
106 ovector too small to capture the group.
107
108 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
109 Red Hat Product Security:
110
111 (a) A crash if /K and /F were both set with the option to save the compiled
112 pattern.
113
114 (b) Another crash if the option to print captured substrings in a callout
115 was combined with setting a null ovector, for example \O\C+ as a subject
116 string.
117
118 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
119 forward reference repeated a large (but limited) number of times within a
120 repeated outer group that has a zero minimum quantifier, caused incorrect
121 code to be compiled, leading to the error "internal error:
122 previously-checked referenced subpattern not found" when an incorrect
123 memory address was read. This bug was reported as "heap overflow",
124 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
125 CVE-2015-2325.
126
127 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
128 call within a group that also contained a recursive back reference caused
129 incorrect code to be compiled. This bug was reported as "heap overflow",
130 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
131 number CVE-2015-2326.
132
133 24. Computing the size of the JIT read-only data in advance has been a source
134 of various issues, and new ones are still appear unfortunately. To fix
135 existing and future issues, size computation is eliminated from the code,
136 and replaced by on-demand memory allocation.
137
138 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
139 adjacent to the end of the range, and the range contained characters with
140 more than one other case, caused incorrect behaviour when compiled in UTF
141 mode. In that example, the range a-j was left out of the class.
142
143 26. Fix JIT compilation of conditional blocks, which assertion
144 is converted to (*FAIL). E.g: /(?(?!))/.
145
146 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
147 discovered by the LLVM fuzzer.
148
149 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
150 when this assertion was used as a condition, for example (?(?!)a|b). In
151 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
152 error about an unsupported item.
153
154 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
155 possessification code could take exponential time to complete. A recursion
156 depth limit of 1000 has been imposed to limit the resources used by this
157 optimization.
158
159 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
160 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
161 because \S ensures they are all in the class. The code for doing this was
162 interacting badly with the code for computing the amount of space needed to
163 compile the pattern, leading to a buffer overflow. This bug was discovered
164 by the LLVM fuzzer.
165
166 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
167 other kinds of group caused stack overflow at compile time. This bug was
168 discovered by the LLVM fuzzer.
169
170 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
171 between a subroutine call and its quantifier was incorrectly compiled,
172 leading to buffer overflow or other errors. This bug was discovered by the
173 LLVM fuzzer.
174
175 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
176 assertion after (?(. The code was failing to check the character after
177 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
178 was discovered by the LLVM fuzzer.
179
180 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
181 a fixed maximum following a group that contains a subroutine reference was
182 incorrectly compiled and could trigger buffer overflow. This bug was
183 discovered by the LLVM fuzzer.
184
185 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
186 caused a stack overflow instead of the diagnosis of a non-fixed length
187 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
188
189 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
190 (e.g. /(?<=\Ka)/) could make pcregrep loop.
191
192 37. There was a similar problem to 36 in pcretest for global matches.
193
194 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
195 and a subsequent item in the pattern caused a non-match, backtracking over
196 the repeated \X did not stop, but carried on past the start of the subject,
197 causing reference to random memory and/or a segfault. There were also some
198 other cases where backtracking after \C could crash. This set of bugs was
199 discovered by the LLVM fuzzer.
200
201 39. The function for finding the minimum length of a matching string could take
202 a very long time if mutual recursion was present many times in a pattern,
203 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
204 method has been implemented. This infelicity was discovered by the LLVM
205 fuzzer.
206
207 40. Static linking against the PCRE library using the pkg-config module was
208 failing on missing pthread symbols.
209
210
211 Version 8.36 26-September-2014
212 ------------------------------
213
214 1. Got rid of some compiler warnings in the C++ modules that were shown up by
215 -Wmissing-field-initializers and -Wunused-parameter.
216
217 2. The tests for quantifiers being too big (greater than 65535) were being
218 applied after reading the number, and stupidly assuming that integer
219 overflow would give a negative number. The tests are now applied as the
220 numbers are read.
221
222 3. Tidy code in pcre_exec.c where two branches that used to be different are
223 now the same.
224
225 4. The JIT compiler did not generate match limit checks for certain
226 bracketed expressions with quantifiers. This may lead to exponential
227 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
228 issue should be resolved now.
229
230 5. Fixed an issue, which occures when nested alternatives are optimized
231 with table jumps.
232
233 6. Inserted two casts and changed some ints to size_t in the light of some
234 reported 64-bit compiler warnings (Bugzilla 1477).
235
236 7. Fixed a bug concerned with zero-minimum possessive groups that could match
237 an empty string, which sometimes were behaving incorrectly in the
238 interpreter (though correctly in the JIT matcher). This pcretest input is
239 an example:
240
241 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
242 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
243
244 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
245 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
246 for an empty string was breaking the inner loop and carrying on at a lower
247 level, when possessive repeated groups should always return to a higher
248 level as they have no backtrack points in them. The empty string test now
249 occurs at the outer level.
250
251 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
252 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
253
254 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
255 doesn't).
256
257 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
258 to hold a pattern to be incorrectly computed (too small) when there were
259 named back references to duplicated names. This could cause "internal
260 error: code overflow" or "double free or corruption" or other memory
261 handling errors.
262
263 11. When named subpatterns had the same prefixes, back references could be
264 confused. For example, in this pattern:
265
266 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
267
268 the reference to 'Name' was incorrectly treated as a reference to a
269 duplicate name.
270
271 12. A pattern such as /^s?c/mi8 where the optional character has more than
272 one "other case" was incorrectly compiled such that it would only try to
273 match starting at "c".
274
275 13. When a pattern starting with \s was studied, VT was not included in the
276 list of possible starting characters; this should have been part of the
277 8.34/18 patch.
278
279 14. If a character class started [\Qx]... where x is any character, the class
280 was incorrectly terminated at the ].
281
282 15. If a pattern that started with a caseless match for a character with more
283 than one "other case" was studied, PCRE did not set up the starting code
284 unit bit map for the list of possible characters. Now it does. This is an
285 optimization improvement, not a bug fix.
286
287 16. The Unicode data tables have been updated to Unicode 7.0.0.
288
289 17. Fixed a number of memory leaks in pcregrep.
290
291 18. Avoid a compiler warning (from some compilers) for a function call with
292 a cast that removes "const" from an lvalue by using an intermediate
293 variable (to which the compiler does not object).
294
295 19. Incorrect code was compiled if a group that contained an internal recursive
296 back reference was optional (had quantifier with a minimum of zero). This
297 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
298 caused segmentation faults because of stack overflows at compile time.
299
300 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
301 group that is quantified with an indefinite repeat, caused a compile-time
302 loop which used up all the system stack and provoked a segmentation fault.
303 This was not the same bug as 19 above.
304
305 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
306 Patch by Mike Frysinger.
307
308
309 Version 8.35 04-April-2014
310 --------------------------
311
312 1. A new flag is set, when property checks are present in an XCLASS.
313 When this flag is not set, PCRE can perform certain optimizations
314 such as studying these XCLASS-es.
315
316 2. The auto-possessification of character sets were improved: a normal
317 and an extended character set can be compared now. Furthermore
318 the JIT compiler optimizes more character set checks.
319
320 3. Got rid of some compiler warnings for potentially uninitialized variables
321 that show up only when compiled with -O2.
322
323 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
324 of a match later then the end of the match. The pcretest program was not
325 handling the case sensibly - it was outputting from the start to the next
326 binary zero. It now reports this situation in a message, and outputs the
327 text from the end to the start.
328
329 5. Fast forward search is improved in JIT. Instead of the first three
330 characters, any three characters with fixed position can be searched.
331 Search order: first, last, middle.
332
333 6. Improve character range checks in JIT. Characters are read by an inprecise
334 function now, which returns with an unknown value if the character code is
335 above a certain threshold (e.g: 256). The only limitation is that the value
336 must be bigger than the threshold as well. This function is useful when
337 the characters above the threshold are handled in the same way.
338
339 7. The macros whose names start with RAWUCHAR are placeholders for a future
340 mode in which only the bottom 21 bits of 32-bit data items are used. To
341 make this more memorable for those maintaining the code, the names have
342 been changed to start with UCHAR21, and an extensive comment has been added
343 to their definition.
344
345 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
346 to the export list in Makefile.am (they were accidentally omitted from the
347 8.34 tarball).
348
349 9. The informational output from pcretest used the phrase "starting byte set"
350 which is inappropriate for the 16-bit and 32-bit libraries. As the output
351 for "first char" and "need char" really means "non-UTF-char", I've changed
352 "byte" to "char", and slightly reworded the output. The documentation about
353 these values has also been (I hope) clarified.
354
355 10. Another JIT related optimization: use table jumps for selecting the correct
356 backtracking path, when more than four alternatives are present inside a
357 bracket.
358
359 11. Empty match is not possible, when the minimum length is greater than zero,
360 and there is no \K in the pattern. JIT should avoid empty match checks in
361 such cases.
362
363 12. In a caseless character class with UCP support, when a character with more
364 than one alternative case was not the first character of a range, not all
365 the alternative cases were added to the class. For example, s and \x{17f}
366 are both alternative cases for S: the class [RST] was handled correctly,
367 but [R-T] was not.
368
369 13. The configure.ac file always checked for pthread support when JIT was
370 enabled. This is not used in Windows, so I have put this test inside a
371 check for the presence of windows.h (which was already tested for).
372
373 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
374 The algorithm provides a way to skip certain starting offsets, and usually
375 faster than linear prefix searches.
376
377 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
378 as for 'fr_FR' and 'french'. For some reason, however, it then used the
379 Windows-specific input and output files, which have 'french' screwed in.
380 So this could never have worked. One of the problems with locales is that
381 they aren't always the same. I have now updated RunTest so that it checks
382 the output of the locale test (test 3) against three different output
383 files, and it allows the test to pass if any one of them matches. With luck
384 this should make the test pass on some versions of Solaris where it was
385 failing. Because of the uncertainty, the script did not used to stop if
386 test 3 failed; it now does. If further versions of a French locale ever
387 come to light, they can now easily be added.
388
389 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
390 there was a message during ./configure, but it did not stop. This now
391 provokes an error. The invalid example in README has been corrected.
392 If a value less than the minimum is given, the minimum value has always
393 been used, but now a warning is given.
394
395 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
396 was a bug in the test system, which is now fixed. Also, the list of various
397 configurations that are tested for each release did not have one with both
398 16/32 bits and --enable-bar-anycrlf. It now does.
399
400 18. pcretest was missing "-C bsr" for displaying the \R default setting.
401
402 19. Little endian PowerPC systems are supported now by the JIT compiler.
403
404 20. The fast forward newline mechanism could enter to an infinite loop on
405 certain invalid UTF-8 input. Although we don't support these cases
406 this issue can be fixed by a performance optimization.
407
408 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
409 not take account if existing stack usage. There is now a new global
410 variable called pcre_stack_guard that can be set to point to an external
411 function to check stack availability. It is called at the start of
412 processing every parenthesized group.
413
414 22. A typo in the code meant that in ungreedy mode the max/min qualifier
415 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
416 not match "ab".
417
418 23. When UTF was disabled, the JIT program reported some incorrect compile
419 errors. These messages are silenced now.
420
421 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
422 compiler.
423
424 25. Change all the temporary files used in RunGrepTest to be different to those
425 used by RunTest so that the tests can be run simultaneously, for example by
426 "make -j check".
427
428
429 Version 8.34 15-December-2013
430 -----------------------------
431
432 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
433 executable memory. Patch inspired by Carsten Klein.
434
435 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
436 this macro is never tested and has no effect, because the work to support
437 coverage involves only compiling and linking options and special targets in
438 the Makefile. The comment in config.h implied that defining the macro would
439 enable coverage support, which is totally false. There was also support for
440 setting this macro in the CMake files (my fault, I just copied it from
441 configure). SUPPORT_GCOV has now been removed.
442
443 3. Make a small performance improvement in strlen16() and strlen32() in
444 pcretest.
445
446 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
447 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
448 these cases). There was also one in pcretest.c.
449
450 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
451
452 6. In UTF mode, the code for checking whether a group could match an empty
453 string (which is used for indefinitely repeated groups to allow for
454 breaking an infinite loop) was broken when the group contained a repeated
455 negated single-character class with a character that occupied more than one
456 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
457 UTF-8 mode). The effect was undefined: the group might or might not be
458 deemed as matching an empty string, or the program might have crashed.
459
460 7. The code for checking whether a group could match an empty string was not
461 recognizing that \h, \H, \v, \V, and \R must match a character.
462
463 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
464 an empty string. If it can, pcretest shows this in its information output.
465
466 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
467 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
468 matched by pcre_exec() without using JIT:
469
470 (a) If the rest of the pattern did not match after a maximal run of
471 grapheme clusters, the code for backing up to try with fewer of them
472 did not always back up over a full grapheme when characters that do not
473 have the modifier quality were involved, e.g. Hangul syllables.
474
475 (b) If the match point in a subject started with modifier character, and
476 there was no match, the code could incorrectly back up beyond the match
477 point, and potentially beyond the first character in the subject,
478 leading to a segfault or an incorrect match result.
479
480 10. A conditional group with an assertion condition could lead to PCRE
481 recording an incorrect first data item for a match if no other first data
482 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
483 first data item, and therefore matched "ca" after "c" instead of at the
484 start.
485
486 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
487 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
488
489 12. The source of pcregrep now includes z/OS-specific code so that it can be
490 compiled for z/OS as part of the special z/OS distribution.
491
492 13. Added the -T and -TM options to pcretest.
493
494 14. The code in pcre_compile.c for creating the table of named capturing groups
495 has been refactored. Instead of creating the table dynamically during the
496 actual compiling pass, the information is remembered during the pre-compile
497 pass (on the stack unless there are more than 20 named groups, in which
498 case malloc() is used) and the whole table is created before the actual
499 compile happens. This has simplified the code (it is now nearly 150 lines
500 shorter) and prepared the way for better handling of references to groups
501 with duplicate names.
502
503 15. A back reference to a named subpattern when there is more than one of the
504 same name now checks them in the order in which they appear in the pattern.
505 The first one that is set is used for the reference. Previously only the
506 first one was inspected. This change makes PCRE more compatible with Perl.
507
508 16. Unicode character properties were updated from Unicode 6.3.0.
509
510 17. The compile-time code for auto-possessification has been refactored, based
511 on a patch by Zoltan Herczeg. It now happens after instead of during
512 compilation. The code is cleaner, and more cases are handled. The option
513 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
514 options in pcretest are provided to set it. It can also be set by
515 (*NO_AUTO_POSSESS) at the start of a pattern.
516
517 18. The character VT has been added to the default ("C" locale) set of
518 characters that match \s and are generally treated as white space,
519 following this same change in Perl 5.18. There is now no difference between
520 "Perl space" and "POSIX space". Whether VT is treated as white space in
521 other locales depends on the locale.
522
523 19. The code for checking named groups as conditions, either for being set or
524 for being recursed, has been refactored (this is related to 14 and 15
525 above). Processing unduplicated named groups should now be as fast at
526 numerical groups, and processing duplicated groups should be faster than
527 before.
528
529 20. Two patches to the CMake build system, by Alexander Barkov:
530
531 (1) Replace the "source" command by "." in CMakeLists.txt because
532 "source" is a bash-ism.
533
534 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
535 without these the CMake build does not work on Solaris.
536
537 21. Perl has changed its handling of \8 and \9. If there is no previously
538 encountered capturing group of those numbers, they are treated as the
539 literal characters 8 and 9 instead of a binary zero followed by the
540 literals. PCRE now does the same.
541
542 22. Following Perl, added \o{} to specify codepoints in octal, making it
543 possible to specify values greater than 0777 and also making them
544 unambiguous.
545
546 23. Perl now gives an error for missing closing braces after \x{... instead of
547 treating the string as literal. PCRE now does the same.
548
549 24. RunTest used to grumble if an inappropriate test was selected explicitly,
550 but just skip it when running all tests. This make it awkward to run ranges
551 of tests when one of them was inappropriate. Now it just skips any
552 inappropriate tests, as it always did when running all tests.
553
554 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
555 character types such as \d or \w, too many callouts were inserted, and the
556 data that they returned was rubbish.
557
558 26. In UCP mode, \s was not matching two of the characters that Perl matches,
559 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
560 were matched by \h. The code has now been refactored so that the lists of
561 the horizontal and vertical whitespace characters used for \h and \v (which
562 are defined only in one place) are now also used for \s.
563
564 27. Add JIT support for the 64 bit TileGX architecture.
565 Patch by Jiong Wang (Tilera Corporation).
566
567 28. Possessive quantifiers for classes (both explicit and automatically
568 generated) now use special opcodes instead of wrapping in ONCE brackets.
569
570 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
571 (because it's meaningless), this was not happening when PCRE_CASELESS was
572 set. Not wrong, but inefficient.
573
574 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
575 properties for \w, \d, etc) is present in a test regex. Otherwise if the
576 test contains no characters greater than 255, Perl doesn't realise it
577 should be using Unicode semantics.
578
579 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
580 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
581 does in Unicode mode.
582
583 32. Added the "forbid" facility to pcretest so that putting tests into the
584 wrong test files can sometimes be quickly detected.
585
586 33. There is now a limit (default 250) on the depth of nesting of parentheses.
587 This limit is imposed to control the amount of system stack used at compile
588 time. It can be changed at build time by --with-parens-nest-limit=xxx or
589 the equivalent in CMake.
590
591 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
592 errors. Perl warns for these when in warning mode, but PCRE has no facility
593 for giving warnings.
594
595 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
596 However, this was not working for (?!) because it is optimized to (*FAIL),
597 for which PCRE does not allow quantifiers. The optimization is now disabled
598 when a quantifier follows (?!). I can't see any use for this, but it makes
599 things uniform.
600
601 36. Perl no longer allows group names to start with digits, so I have made this
602 change also in PCRE. It simplifies the code a bit.
603
604 37. In extended mode, Perl ignores spaces before a + that indicates a
605 possessive quantifier. PCRE allowed a space before the quantifier, but not
606 before the possessive +. It now does.
607
608 38. The use of \K (reset reported match start) within a repeated possessive
609 group such as (a\Kb)*+ was not working.
610
611 40. Document that the same character tables must be used at compile time and
612 run time, and that the facility to pass tables to pcre_exec() and
613 pcre_dfa_exec() is for use only with saved/restored patterns.
614
615 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
616 features for Builds with MSVC:
617
618 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
619 builds). The .rc files can be used to set FileDescription and many other
620 attributes.
621
622 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
623 This allows higher-level build scripts which want .pdb files to avoid
624 hard-coding the exact files needed."
625
626 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
627 mean "start of word" and "end of word", respectively, as a transition aid.
628
629 43. A minimizing repeat of a class containing codepoints greater than 255 in
630 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
631 compiled to use the heap for recursion.
632
633 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
634 is configured.
635
636
637 Version 8.33 28-May-2013
638 ------------------------
639
640 1. Added 'U' to some constants that are compared to unsigned integers, to
641 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
642 variables that are added to signed variables, to ensure the result is
643 signed and can be negated.
644
645 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
646 CMake config files.
647
648 3. Revise the creation of config.h.generic so that all boolean macros are
649 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
650 overriding via -D on the command line possible.
651
652 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
653 to unsigned int is reported to make a quite noticeable speed difference in
654 a specific Windows environment. Testing on Linux did also appear to show
655 some benefit (and it is clearly not harmful). Also fixed the definition of
656 Xop which should be unsigned.
657
658 5. Related to (4), changing the definition of the intermediate variable cc
659 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
660 improvements.
661
662 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
663 unnecessary spaces.
664
665 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
666 and later.
667
668 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
669
670 9. Optimizing fast_forward_start_bits in JIT.
671
672 10. Adding support for callouts in JIT, and fixing some issues revealed
673 during this work. Namely:
674
675 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
676
677 (b) Minimum length was not checked before the matching is started.
678
679 11. The value of capture_last that is passed to callouts was incorrect in some
680 cases when there was a capture on one path that was subsequently abandoned
681 after a backtrack. Also, the capture_last value is now reset after a
682 recursion, since all captures are also reset in this case.
683
684 12. The interpreter no longer returns the "too many substrings" error in the
685 case when an overflowing capture is in a branch that is subsequently
686 abandoned after a backtrack.
687
688 13. In the pathological case when an offset vector of size 2 is used, pcretest
689 now prints out the matched string after a yield of 0 or 1.
690
691 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
692 Only supported by the JIT compiler at the moment.
693
694 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
695
696 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
697 offset of the starting point of the matching process, provided the offsets
698 vector is large enough.
699
700 17. The \A escape now records a lookbehind value of 1, though its execution
701 does not actually inspect the previous character. This is to ensure that,
702 in partial multi-segment matching, at least one character from the old
703 segment is retained when a new segment is processed. Otherwise, if there
704 are no lookbehinds in the pattern, \A might match incorrectly at the start
705 of a new segment.
706
707 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
708
709 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
710 gives some modest performance improvement in 8-bit mode.
711
712 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
713 be expressed in certain programming languages using Universal Character
714 Names.
715
716 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
717 which points out that "non characters" are not "characters that may not
718 appear in Unicode strings" but rather "characters that are reserved for
719 internal use and have only local meaning".
720
721 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
722 there was a conditional group that depended on an assertion, if the
723 assertion was false, the callout that immediately followed the alternation
724 in the condition was skipped when pcre_exec() was used for matching.
725
726 23. Allow an explicit callout to be inserted before an assertion that is the
727 condition for a conditional group, for compatibility with automatic
728 callouts, which always insert a callout at this point.
729
730 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
731 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
732
733 25. (*PRUNE) is now supported by the JIT compiler.
734
735 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
736
737 27. Fix the case where there are two or more SKIPs with arguments that may be
738 ignored.
739
740 28. (*SKIP) is now supported by the JIT compiler.
741
742 29. (*THEN) is now supported by the JIT compiler.
743
744 30. Update RunTest with additional test selector options.
745
746 31. The way PCRE handles backtracking verbs has been changed in two ways.
747
748 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
749 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
750 backtracking. In some cases this makes it more Perl-compatible, but Perl's
751 rather obscure rules do not always do the same thing.
752
753 (2) Previously, backtracking verbs were confined within assertions. This is
754 no longer the case for positive assertions, except for (*ACCEPT). Again,
755 this sometimes improves Perl compatibility, and sometimes does not.
756
757 32. A number of tests that were in test 2 because Perl did things differently
758 have been moved to test 1, because either Perl or PCRE has changed, and
759 these tests are now compatible.
760
761 32. Backtracking control verbs are now handled in the same way in JIT and
762 interpreter.
763
764 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
765 contained a forward subroutine reference caused a compile error.
766
767 34. Auto-detect and optimize limited repetitions in JIT.
768
769 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
770 blocking (*UTF) etc.
771
772 36. In the interpreter, maximizing pattern repetitions for characters and
773 character types now use tail recursion, which reduces stack usage.
774
775 37. The value of the max lookbehind was not correctly preserved if a compiled
776 and saved regex was reloaded on a host of different endianness.
777
778 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
779 of the compiled pattern block, expand the flags field from 16 to 32 bits
780 because it was almost full.
781
782 39. Try madvise first before posix_madvise.
783
784 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
785 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
786 original change locked out all matches of empty strings. This has been
787 changed so that one match of an empty string per line is recognized.
788 Subsequent searches on the same line (for colouring or for --only-matching,
789 for example) do not recognize empty strings.
790
791 41. Applied a user patch to fix a number of spelling mistakes in comments.
792
793 42. Data lines longer than 65536 caused pcretest to crash.
794
795 43. Clarified the data type for length and startoffset arguments for pcre_exec
796 and pcre_dfa_exec in the function-specific man pages, where they were
797 explicitly stated to be in bytes, never having been updated. I also added
798 some clarification to the pcreapi man page.
799
800 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
801 a segmentation fault.
802
803
804 Version 8.32 30-November-2012
805 -----------------------------
806
807 1. Improved JIT compiler optimizations for first character search and single
808 character iterators.
809
810 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
811 Patch by Daniel Richard G.
812
813 3. Single character iterator optimizations in the JIT compiler.
814
815 4. Improved JIT compiler optimizations for character ranges.
816
817 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
818 Reported by Giuseppe D'Angelo.
819
820 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
821 of a line, was being set incorrectly in cases where .* appeared inside
822 atomic brackets at the start of a pattern, or where there was a subsequent
823 *PRUNE or *SKIP.
824
825 7. Improved instruction cache flush for POWER/PowerPC.
826 Patch by Daniel Richard G.
827
828 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
829 grep:
830
831 (a) There is now no limit to the number of patterns to be matched.
832
833 (b) An error is given if a pattern is too long.
834
835 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
836 are now supported.
837
838 (d) --exclude-from and --include-from (multiple use) have been added.
839
840 (e) Exclusions and inclusions now apply to all files and directories, not
841 just to those obtained from scanning a directory recursively.
842
843 (f) Multiple uses of -f and --file-list are now supported.
844
845 (g) In a Windows environment, the default for -d has been changed from
846 "read" (the GNU grep default) to "skip", because otherwise the presence
847 of a directory in the file list provokes an error.
848
849 (h) The documentation has been revised and clarified in places.
850
851 9. Improve the matching speed of capturing brackets.
852
853 10. Changed the meaning of \X so that it now matches a Unicode extended
854 grapheme cluster.
855
856 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
857 out POSIX threads when JIT support is configured.
858
859 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
860
861 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
862 structure could go wrong in environments where size_t is not the same size
863 as int.
864
865 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
866 set.
867
868 15. The EBCDIC support had decayed; later updates to the code had included
869 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
870 general tidy up of EBCDIC-related issues, and the documentation was also
871 not quite right. There is now a test that can be run on ASCII systems to
872 check some of the EBCDIC-related things (but is it not a full test).
873
874 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
875 in a small tidy to the code.
876
877 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
878
879 18. If the --only-matching (-o) option in pcregrep is specified multiple
880 times, each one causes appropriate output. For example, -o1 -o2 outputs the
881 substrings matched by the 1st and 2nd capturing parentheses. A separating
882 string can be specified by --om-separator (default empty).
883
884 19. Improving the first n character searches.
885
886 20. Turn case lists for horizontal and vertical white space into macros so that
887 they are defined only once.
888
889 21. This set of changes together give more compatible Unicode case-folding
890 behaviour for characters that have more than one other case when UCP
891 support is available.
892
893 (a) The Unicode property table now has offsets into a new table of sets of
894 three or more characters that are case-equivalent. The MultiStage2.py
895 script that generates these tables (the pcre_ucd.c file) now scans
896 CaseFolding.txt instead of UnicodeData.txt for character case
897 information.
898
899 (b) The code for adding characters or ranges of characters to a character
900 class has been abstracted into a generalized function that also handles
901 case-independence. In UTF-mode with UCP support, this uses the new data
902 to handle characters with more than one other case.
903
904 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
905 whose other case is greater than 256 are now correctly matched
906 caselessly. Previously, the high codepoint matched the low one, but not
907 vice versa.
908
909 (d) The processing of \h, \H, \v, and \ in character classes now makes use
910 of the new class addition function, using character lists defined as
911 macros alongside the case definitions of 20 above.
912
913 (e) Caseless back references now work with characters that have more than
914 one other case.
915
916 (f) General caseless matching of characters with more than one other case
917 is supported.
918
919 22. Unicode character properties were updated from Unicode 6.2.0
920
921 23. Improved CMake support under Windows. Patch by Daniel Richard G.
922
923 24. Add support for 32-bit character strings, and UTF-32
924
925 25. Major JIT compiler update (code refactoring and bugfixing).
926 Experimental Sparc 32 support is added.
927
928 26. Applied a modified version of Daniel Richard G's patch to create
929 pcre.h.generic and config.h.generic by "make" instead of in the
930 PrepareRelease script.
931
932 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
933 pcre_compile.c when checking for a zero character.
934
935 28. Introducing a native interface for JIT. Through this interface, the compiled
936 machine code can be directly executed. The purpose of this interface is to
937 provide fast pattern matching, so several sanity checks are not performed.
938 However, feature tests are still performed. The new interface provides
939 1.4x speedup compared to the old one.
940
941 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
942 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
943 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
944
945 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
946 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
947 the "old" RFC 2279). Instead, it ended up passing a negative length to
948 pcre_exec().
949
950 31. Add support for GCC's visibility feature to hide internal functions.
951
952 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
953 "unknown -C option" after outputting 0 or 1.
954
955 33. There is now support for generating a code coverage report for the test
956 suite in environments where gcc is the compiler and lcov is installed. This
957 is mainly for the benefit of the developers.
958
959 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
960 unaddressable using valgrind annotations, allowing valgrind to detect
961 invalid memory accesses. This is mainly for the benefit of the developers.
962
963 25. (*UTF) can now be used to start a pattern in any of the three libraries.
964
965 26. Give configure error if --enable-cpp but no C++ compiler found.
966
967
968 Version 8.31 06-July-2012
969 -------------------------
970
971 1. Fixing a wrong JIT test case and some compiler warnings.
972
973 2. Removed a bashism from the RunTest script.
974
975 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
976 to unsigned type, result still unsigned" that was given by an MS compiler
977 on encountering the code "-sizeof(xxx)".
978
979 4. Partial matching support is added to the JIT compiler.
980
981 5. Fixed several bugs concerned with partial matching of items that consist
982 of more than one character:
983
984 (a) /^(..)\1/ did not partially match "aba" because checking references was
985 done on an "all or nothing" basis. This also applied to repeated
986 references.
987
988 (b) \R did not give a hard partial match if \r was found at the end of the
989 subject.
990
991 (c) \X did not give a hard partial match after matching one or more
992 characters at the end of the subject.
993
994 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
995 a partial match for the string "\r".
996
997 (e) When newline was set to CRLF, the metacharacter "." did not recognize
998 a partial match for a CR character at the end of the subject string.
999
1000 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1001 running pcretest, the text "(JIT)" added to the output whenever JIT is
1002 actually used to run the match.
1003
1004 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1005 or /S+[+] with a digit between 1 and 7.
1006
1007 8. OP_NOT now supports any UTF character not just single-byte ones.
1008
1009 9. (*MARK) control verb is now supported by the JIT compiler.
1010
1011 10. The command "./RunTest list" lists the available tests without actually
1012 running any of them. (Because I keep forgetting what they all are.)
1013
1014 11. Add PCRE_INFO_MAXLOOKBEHIND.
1015
1016 12. Applied a (slightly modified) user-supplied patch that improves performance
1017 when the heap is used for recursion (compiled with --disable-stack-for-
1018 recursion). Instead of malloc and free for each heap frame each time a
1019 logical recursion happens, frames are retained on a chain and re-used where
1020 possible. This sometimes gives as much as 30% improvement.
1021
1022 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1023 call.
1024
1025 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1026
1027 15. It is now possible to link pcretest with libedit as an alternative to
1028 libreadline.
1029
1030 16. (*COMMIT) control verb is now supported by the JIT compiler.
1031
1032 17. The Unicode data tables have been updated to Unicode 6.1.0.
1033
1034 18. Added --file-list option to pcregrep.
1035
1036 19. Added binary file support to pcregrep, including the -a, --binary-files,
1037 -I, and --text options.
1038
1039 20. The madvise function is renamed for posix_madvise for QNX compatibility
1040 reasons. Fixed by Giuseppe D'Angelo.
1041
1042 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1043
1044 22. Changed the option for creating symbolic links for 16-bit man pages from
1045 -s to -sf so that re-installing does not cause issues.
1046
1047 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1048
1049 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1050 matches in certain environments (the workspace was not being correctly
1051 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1052 some of the workspace data at the beginning of a restart.
1053
1054 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1055 was not doing so when it should - probably a typo introduced by SVN 528
1056 (change 8.10/14).
1057
1058 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1059 \w+ when the character tables indicated that \x{c4} was a word character.
1060 There were several related cases, all because the tests for doing a table
1061 lookup were testing for characters less than 127 instead of 255.
1062
1063 27. If a pattern contains capturing parentheses that are not used in a match,
1064 their slots in the ovector are set to -1. For those that are higher than
1065 any matched groups, this happens at the end of processing. In the case when
1066 there were back references that the ovector was too small to contain
1067 (causing temporary malloc'd memory to be used during matching), and the
1068 highest capturing number was not used, memory off the end of the ovector
1069 was incorrectly being set to -1. (It was using the size of the temporary
1070 memory instead of the true size.)
1071
1072 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1073 ovector size, it uses memory at the end of the block that it has got.
1074
1075 29. Check for an overlong MARK name and give an error at compile time. The
1076 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1077
1078 30. JIT compiler update.
1079
1080 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1081 Rill for the patch.
1082
1083 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1084
1085 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1086
1087 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1088 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1089 was enabled.
1090
1091 35. Improve JIT code generation for greedy plus quantifier.
1092
1093 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1094 1 to "aa" instead of to an empty string. The bug affected repeated groups
1095 that could potentially match an empty string.
1096
1097 37. Optimizing single character iterators in JIT.
1098
1099 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1100 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1101 codepoints that are too big for the mode are faulted, and in a UTF mode,
1102 disallowed codepoints are also faulted.
1103
1104 39. If PCRE was compiled with UTF support, in three places in the DFA
1105 matcher there was code that should only have been obeyed in UTF mode, but
1106 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1107 processing when bytes with values greater than 127 were present. In 16-bit
1108 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1109 both cases the values are those that cannot be the first data item in a UTF
1110 character. The three items that might have provoked this were recursions,
1111 possessively repeated groups, and atomic groups.
1112
1113 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1114 and pcregrep, because some OS require shared objects to be explicitly
1115 passed to ld, causing the link step to fail if they are not.
1116
1117 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1118 mode, patterns that started with \h* or \R* might be incorrectly matched.
1119
1120
1121 Version 8.30 04-February-2012
1122 -----------------------------
1123
1124 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1125 name is defined in ctype.h.
1126
1127 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1128 only in quite long subpatterns.
1129
1130 3. Removed the function pcre_info(), which has been obsolete and deprecated
1131 since it was replaced by pcre_fullinfo() in February 2000.
1132
1133 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1134 match a (*MARK), and the match failed at the start of the subject, a
1135 reference to memory before the start of the subject could occur. This bug
1136 was introduced by fix 17 of release 8.21.
1137
1138 5. A reference to an unset group with zero minimum repetition was giving
1139 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1140 /(another)?(\1?)test/ matched against "hello world test". This bug was
1141 introduced in release 8.13.
1142
1143 6. Add support for 16-bit character strings (a large amount of work involving
1144 many changes and refactorings).
1145
1146 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1147 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1148 from a file.
1149
1150 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1151 rounding is not applied in this particular case).
1152
1153 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1154 if they appear, or are escaped, in patterns.
1155
1156 10. Get rid of a number of -Wunused-but-set-variable warnings.
1157
1158 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1159 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1160 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1161 also returns the mark "x". This bug applied to capturing parentheses,
1162 non-capturing parentheses, and atomic parentheses. It also applied to some
1163 assertions.
1164
1165 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1166 information out of configure.ac instead of relying on pcre.h.generic, which
1167 is not stored in the repository.
1168
1169 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1170 -lreadline.
1171
1172 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1173
1174 15. Applied Graycode's patch to put the top-level frame on the stack rather
1175 than the heap when not using the stack for recursion. This gives a
1176 performance improvement in many cases when recursion is not deep.
1177
1178 16. Experimental code added to "pcretest -C" to output the stack frame size.
1179
1180
1181 Version 8.21 12-Dec-2011
1182 ------------------------
1183
1184 1. Updating the JIT compiler.
1185
1186 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1187 are added as well.
1188
1189 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1190 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1191 calling _pcre_jit_exec. Some extra comments are added.
1192
1193 4. (*MARK) settings inside atomic groups that do not contain any capturing
1194 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1195 was introduced by change 18 for 8.20.
1196
1197 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1198 ECMA-262 standard.
1199
1200 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1201 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1202 This bug was probably introduced by change 9 of 8.13.
1203
1204 7. While fixing 6 above, I noticed that a number of other items were being
1205 incorrectly rejected as "not fixed length". This arose partly because newer
1206 opcodes had not been added to the fixed-length checking code. I have (a)
1207 corrected the bug and added tests for these items, and (b) arranged for an
1208 error to occur if an unknown opcode is encountered while checking for fixed
1209 length instead of just assuming "not fixed length". The items that were
1210 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1211 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1212 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1213
1214 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1215 being incorrectly compiled and would have given unpredicatble results.
1216
1217 9. A possessively repeated subpattern with minimum repeat count greater than
1218 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1219 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1220 the first (A) could occur when it should not.
1221
1222 10. Add a cast and remove a redundant test from the code.
1223
1224 11. JIT should use pcre_malloc/pcre_free for allocation.
1225
1226 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1227 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1228 is anything other than /usr, -L is still shown).
1229
1230 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1231
1232 14. Perl does not support \N without a following name in a [] class; PCRE now
1233 also gives an error.
1234
1235 15. If a forward reference was repeated with an upper limit of around 2000,
1236 it caused the error "internal error: overran compiling workspace". The
1237 maximum number of forward references (including repeats) was limited by the
1238 internal workspace, and dependent on the LINK_SIZE. The code has been
1239 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1240 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1241 of around 200,000 forward references. While doing this, I also speeded up
1242 the filling in of repeated forward references.
1243
1244 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1245 incorrectly expecting the subject to contain another "a" after the start.
1246
1247 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1248 in the match, the SKIP should be ignored. This was not happening; instead
1249 the SKIP was being treated as NOMATCH. For patterns such as
1250 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1251 tested.
1252
1253 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1254 now much more compatible with Perl, in particular in cases where the result
1255 is a non-match for a non-anchored pattern. For example, if
1256 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1257 "m", where previously it did not return a name. A side effect of this
1258 change is that for partial matches, the last encountered mark name is
1259 returned, as for non matches. A number of tests that were previously not
1260 Perl-compatible have been moved into the Perl-compatible test files. The
1261 refactoring has had the pleasing side effect of removing one argument from
1262 the match() function, thus reducing its stack requirements.
1263
1264 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1265 subsequent uses of /S (without +) incorrectly behaved like /S+.
1266
1267 21. Retrieve executable code size support for the JIT compiler and fixing
1268 some warnings.
1269
1270 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1271 not work when the shorter character appeared right at the end of the
1272 subject string.
1273
1274 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1275 systems.
1276
1277 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1278 output it when the /M option is used in pcretest.
1279
1280 25. The CheckMan script was not being included in the distribution. Also, added
1281 an explicit "perl" to run Perl scripts from the PrepareRelease script
1282 because this is reportedly needed in Windows.
1283
1284 26. If study data was being save in a file and studying had not found a set of
1285 "starts with" bytes for the pattern, the data written to the file (though
1286 never used) was taken from uninitialized memory and so caused valgrind to
1287 complain.
1288
1289 27. Updated RunTest.bat as provided by Sheri Pierce.
1290
1291 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1292
1293 29. Computation of memory usage for the table of capturing group names was
1294 giving an unnecessarily large value.
1295
1296
1297 Version 8.20 21-Oct-2011
1298 ------------------------
1299
1300 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1301 a POSIX class. After further experiments with Perl, which convinced me that
1302 Perl has bugs and confusions, a closing square bracket is no longer allowed
1303 in a POSIX name. This bug also affected patterns with classes that started
1304 with full stops.
1305
1306 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1307 captured substring, but while checking the failing first alternative,
1308 substring 1 is temporarily captured. If the output vector supplied to
1309 pcre_exec() was not big enough for this capture, the yield of the function
1310 was still zero ("insufficient space for captured substrings"). This cannot
1311 be totally fixed without adding another stack variable, which seems a lot
1312 of expense for a edge case. However, I have improved the situation in cases
1313 such as /(a)(b)x|abc/ matched against "abc", where the return code
1314 indicates that fewer than the maximum number of slots in the ovector have
1315 been set.
1316
1317 3. Related to (2) above: when there are more back references in a pattern than
1318 slots in the output vector, pcre_exec() uses temporary memory during
1319 matching, and copies in the captures as far as possible afterwards. It was
1320 using the entire output vector, but this conflicts with the specification
1321 that only 2/3 is used for passing back captured substrings. Now it uses
1322 only the first 2/3, for compatibility. This is, of course, another edge
1323 case.
1324
1325 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1326 main code base, and can be used by building with --enable-jit. When this is
1327 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1328 runtime --no-jit option is given.
1329
1330 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1331 ovector, the return from the function was zero, implying that there were
1332 other matches that did not fit. The correct "exactly full" value is now
1333 returned.
1334
1335 6. If a subpattern that was called recursively or as a subroutine contained
1336 (*PRUNE) or any other control that caused it to give a non-standard return,
1337 invalid errors such as "Error -26 (nested recursion at the same subject
1338 position)" or even infinite loops could occur.
1339
1340 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1341 computing the minimum length on reaching *ACCEPT, and so ended up with the
1342 wrong value of 1 rather than 0. Further investigation indicates that
1343 computing a minimum subject length in the presence of *ACCEPT is difficult
1344 (think back references, subroutine calls), and so I have changed the code
1345 so that no minimum is registered for a pattern that contains *ACCEPT.
1346
1347 8. If (*THEN) was present in the first (true) branch of a conditional group,
1348 it was not handled as intended. [But see 16 below.]
1349
1350 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1351 Sheri Pierce.
1352
1353 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1354 the first byte in a match must be "a".
1355
1356 11. Change 17 for 8.13 increased the recursion depth for patterns like
1357 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1358 pattern contains any instances of (*THEN). If it does not, the old
1359 optimizations are restored. It would be nice to do this on a per-group
1360 basis, but at the moment that is not feasible.
1361
1362 12. In some environments, the output of pcretest -C is CRLF terminated. This
1363 broke RunTest's code that checks for the link size. A single white space
1364 character after the value is now allowed for.
1365
1366 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1367 For "fr", it uses the Windows-specific input and output files.
1368
1369 14. If (*THEN) appeared in a group that was called recursively or as a
1370 subroutine, it did not work as intended. [But see next item.]
1371
1372 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1373 pattern fragments (but not containing any | characters). If A and B are
1374 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1375 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1376 D. In other words, Perl considers parentheses that do not contain any |
1377 characters to be part of a surrounding alternative, whereas PCRE was
1378 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1379 differently. PCRE now behaves in the same way as Perl, except in the case
1380 of subroutine/recursion calls such as (?1) which have in any case always
1381 been different (but PCRE had them first :-).
1382
1383 16. Related to 15 above: Perl does not treat the | in a conditional group as
1384 creating alternatives. Such a group is treated in the same way as an
1385 ordinary group without any | characters when processing (*THEN). PCRE has
1386 been changed to match Perl's behaviour.
1387
1388 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1389 RunGrepTest script failed.
1390
1391 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1392 inevitable for groups that contain captures, but it can lead to a lot of
1393 stack use in large patterns. The old behaviour has been restored for atomic
1394 groups that do not contain any capturing parentheses.
1395
1396 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1397 suppress the check for a minimum subject length at run time. (If it was
1398 given to pcre_exec() or pcre_dfa_exec() it did work.)
1399
1400 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1401 fail to work when decoding hex characters in data strings in EBCDIC
1402 environments.
1403
1404 21. It appears that in at least one Mac OS environment, the isxdigit() function
1405 is implemented as a macro that evaluates to its argument more than once,
1406 contravening the C 90 Standard (I haven't checked a later standard). There
1407 was an instance in pcretest which caused it to go wrong when processing
1408 \x{...} escapes in subject strings. The has been rewritten to avoid using
1409 things like p++ in the argument of isxdigit().
1410
1411
1412 Version 8.13 16-Aug-2011
1413 ------------------------
1414
1415 1. The Unicode data tables have been updated to Unicode 6.0.0.
1416
1417 2. Two minor typos in pcre_internal.h have been fixed.
1418
1419 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1420 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1421 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1422
1423 4. There were a number of related bugs in the code for matching backrefences
1424 caselessly in UTF-8 mode when codes for the characters concerned were
1425 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1426 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1427 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1428 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1429 2-byte code at the end of the subject (it thought there wasn't enough data
1430 left).
1431
1432 5. Comprehensive information about what went wrong is now returned by
1433 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1434 as the output vector has at least 2 elements. The offset of the start of
1435 the failing character and a reason code are placed in the vector.
1436
1437 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1438 now returned is for the first byte of the failing character, instead of the
1439 last byte inspected. This is an incompatible change, but I hope it is small
1440 enough not to be a problem. It makes the returned offset consistent with
1441 pcre_exec() and pcre_dfa_exec().
1442
1443 7. pcretest now gives a text phrase as well as the error number when
1444 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1445 failure, the offset and reason code are output.
1446
1447 8. When \R was used with a maximizing quantifier it failed to skip backwards
1448 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1449 back over a single character (\n). This seems wrong (because it treated the
1450 two characters as a single entity when going forwards), conflicts with the
1451 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1452 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1453 has been changed.
1454
1455 9. Some internal refactoring has changed the processing so that the handling
1456 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1457 time (the PCRE_DOTALL option was changed this way some time ago: version
1458 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1459 which was always a bit of a fudge. It also means that there is one less
1460 argument for the match() function, which reduces its stack requirements
1461 slightly. This change also fixes an incompatibility with Perl: the pattern
1462 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1463
1464 10. More internal refactoring has drastically reduced the number of recursive
1465 calls to match() for possessively repeated groups such as (abc)++ when
1466 using pcre_exec().
1467
1468 11. While implementing 10, a number of bugs in the handling of groups were
1469 discovered and fixed:
1470
1471 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1472 (a|)*(?1) gave a compile-time internal error.
1473 ((a|)+)+ did not notice that the outer group could match an empty string.
1474 (^a|^)+ was not marked as anchored.
1475 (.*a|.*)+ was not marked as matching at start or after a newline.
1476
1477 12. Yet more internal refactoring has removed another argument from the match()
1478 function. Special calls to this function are now indicated by setting a
1479 value in a variable in the "match data" data block.
1480
1481 13. Be more explicit in pcre_study() instead of relying on "default" for
1482 opcodes that mean there is no starting character; this means that when new
1483 ones are added and accidentally left out of pcre_study(), testing should
1484 pick them up.
1485
1486 14. The -s option of pcretest has been documented for ages as being an old
1487 synonym of -m (show memory usage). I have changed it to mean "force study
1488 for every regex", that is, assume /S for every regex. This is similar to -i
1489 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1490 using it. It makes it easier to run collections of tests with and without
1491 study enabled, and thereby test pcre_study() more easily. All the standard
1492 tests are now run with and without -s (but some patterns can be marked as
1493 "never study" - see 20 below).
1494
1495 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1496 restoration of the capturing data to the outer values was not happening
1497 correctly.
1498
1499 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1500 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1501 pattern had matched an empty string, and so incorrectly returned a no
1502 match.
1503
1504 17. There was optimizing code for the last branch of non-capturing parentheses,
1505 and also for the obeyed branch of a conditional subexpression, which used
1506 tail recursion to cut down on stack usage. Unfortunately, now that there is
1507 the possibility of (*THEN) occurring in these branches, tail recursion is
1508 no longer possible because the return has to be checked for (*THEN). These
1509 two optimizations have therefore been removed. [But see 8.20/11 above.]
1510
1511 18. If a pattern containing \R was studied, it was assumed that \R always
1512 matched two bytes, thus causing the minimum subject length to be
1513 incorrectly computed because \R can also match just one byte.
1514
1515 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1516 was incorrectly computed.
1517
1518 20. If /S is present twice on a test pattern in pcretest input, it now
1519 *disables* studying, thereby overriding the use of -s on the command line
1520 (see 14 above). This is necessary for one or two tests to keep the output
1521 identical in both cases.
1522
1523 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1524 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1525
1526 22. When an atomic group that contained a capturing parenthesis was
1527 successfully matched, but the branch in which it appeared failed, the
1528 capturing was not being forgotten if a higher numbered group was later
1529 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1530 group 1 to "a", when in fact it should be unset. This applied to multi-
1531 branched capturing and non-capturing groups, repeated or not, and also to
1532 positive assertions (capturing in negative assertions does not happen
1533 in PCRE) and also to nested atomic groups.
1534
1535 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1536 subject after a captured substring, to make it easier to tell which of a
1537 number of identical substrings has been captured.
1538
1539 24. The way atomic groups are processed by pcre_exec() has been changed so that
1540 if they are repeated, backtracking one repetition now resets captured
1541 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1542 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1543 "aaa". Previously, it would have been "a". As part of this code
1544 refactoring, the way recursive calls are handled has also been changed.
1545
1546 25. If an assertion condition captured any substrings, they were not passed
1547 back unless some other capturing happened later. For example, if
1548 (?(?=(a))a) was matched against "a", no capturing was returned.
1549
1550 26. When studying a pattern that contained subroutine calls or assertions,
1551 the code for finding the minimum length of a possible match was handling
1552 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1553 group 1 called group 2 while simultaneously a separate group 2 called group
1554 1). A stack overflow occurred in this case. I have fixed this by limiting
1555 the recursion depth to 10.
1556
1557 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1558 Fortmann. This supports explicit test numbers on the command line, and has
1559 argument validation and error reporting.
1560
1561 28. An instance of \X with an unlimited repeat could fail if at any point the
1562 first character it looked at was a mark character.
1563
1564 29. Some minor code refactoring concerning Unicode properties and scripts
1565 should reduce the stack requirement of match() slightly.
1566
1567 30. Added the '=' option to pcretest to check the setting of unused capturing
1568 slots at the end of the pattern, which are documented as being -1, but are
1569 not included in the return count.
1570
1571 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1572 compiled something random. Now it gives a compile-time error (as does
1573 Perl).
1574
1575 32. A *MARK encountered during the processing of a positive assertion is now
1576 recorded and passed back (compatible with Perl).
1577
1578 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1579 had alternative anchored branches, the search for a second match in a line
1580 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1581 matched the line "0102" twice. The same bug affected patterns that started
1582 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1583 twice.
1584
1585 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1586 does, and because of capturing effects, quantifying parenthesized
1587 assertions may at times be useful. Quantifiers are now allowed for
1588 parenthesized assertions.
1589
1590 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1591
1592 36. \g was being checked for fancy things in a character class, when it should
1593 just be a literal "g".
1594
1595 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1596 appearance of a nested POSIX class supersedes an apparent external class.
1597 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1598 unescaped square brackets may also appear as part of class names. For
1599 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1600 more like Perl. (But see 8.20/1 above.)
1601
1602 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1603 was because it thought it was \N{name}, which is not supported).
1604
1605 39. Add minix to OS list not supporting the -S option in pcretest.
1606
1607 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1608 cannot analyze patterns in sufficient detail to catch mutual recursions
1609 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1610 subgroup is called recursively as a subpattern for a second time at the
1611 same position in the subject string. In previous releases this might have
1612 been caught by the recursion limit, or it might have run out of stack.
1613
1614 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1615 happen only once. PCRE was, however incorrectly giving a compile time error
1616 "recursive call could loop indefinitely" because it cannot analyze the
1617 pattern in sufficient detail. The compile time test no longer happens when
1618 PCRE is compiling a conditional subpattern, but actual runaway loops are
1619 now caught at runtime (see 40 above).
1620
1621 42. It seems that Perl allows any characters other than a closing parenthesis
1622 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1623 has been changed to be the same.
1624
1625 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1626 as not to get warnings when autogen.sh is called. Also changed
1627 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1628
1629 44. To help people who use pcregrep to scan files containing exceedingly long
1630 lines, the following changes have been made:
1631
1632 (a) The default value of the buffer size parameter has been increased from
1633 8K to 20K. (The actual buffer used is three times this size.)
1634
1635 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1636 PCRE is built.
1637
1638 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1639 to be set at run time.
1640
1641 (d) Numerical values in pcregrep options can be followed by K or M, for
1642 example --buffer-size=50K.
1643
1644 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1645 given and the return code is set to 2.
1646
1647 45. Add a pointer to the latest mark to the callout data block.
1648
1649 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1650 partial match of an empty string instead of no match. This was specific to
1651 the use of ".".
1652
1653 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1654 complete match instead of a partial match. This bug was dependent on both
1655 the PCRE_UTF8 and PCRE_DOTALL options being set.
1656
1657 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1658 starting byte set, because \b was not being ignored.
1659
1660
1661 Version 8.12 15-Jan-2011
1662 ------------------------
1663
1664 1. Fixed some typos in the markup of the man pages, and wrote a script that
1665 checks for such things as part of the documentation building process.
1666
1667 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1668 --match-limit and --recursion-limit options (added for 8.11). In
1669 particular, this made one of the standard tests fail. (The integer value
1670 went into the wrong half of a long int.)
1671
1672 3. If the --colour option was given to pcregrep with -v (invert match), it
1673 did strange things, either producing crazy output, or crashing. It should,
1674 of course, ignore a request for colour when reporting lines that do not
1675 match.
1676
1677 4. Another pcregrep bug caused similar problems if --colour was specified with
1678 -M (multiline) and the pattern match finished with a line ending.
1679
1680 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1681 matched in multiline mode, the following line was shown as part of the
1682 match. This seems wrong, so I have changed it.
1683
1684 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1685 the check for further matches in the same line (so they could be coloured)
1686 to overrun the end of the current line. If another match was found, it was
1687 incorrectly shown (and then shown again when found in the next line).
1688
1689 7. If pcregrep was compiled under Windows, there was a reference to the
1690 function pcregrep_exit() before it was defined. I am assuming this was
1691 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1692 reported by a user. I've moved the definition above the reference.
1693
1694
1695 Version 8.11 10-Dec-2010
1696 ------------------------
1697
1698 1. (*THEN) was not working properly if there were untried alternatives prior
1699 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1700 backtracked to try for "b" instead of moving to the next alternative branch
1701 at the same level (in this case, to look for "c"). The Perl documentation
1702 is clear that when (*THEN) is backtracked onto, it goes to the "next
1703 alternative in the innermost enclosing group".
1704
1705 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1706 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1707 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1708 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1709 (*THEN).
1710
1711 3. If \s appeared in a character class, it removed the VT character from
1712 the class, even if it had been included by some previous item, for example
1713 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1714 of \s, but is part of the POSIX "space" class.)
1715
1716 4. A partial match never returns an empty string (because you can always
1717 match an empty string at the end of the subject); however the checking for
1718 an empty string was starting at the "start of match" point. This has been
1719 changed to the "earliest inspected character" point, because the returned
1720 data for a partial match starts at this character. This means that, for
1721 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1722 (previously it gave "no match").
1723
1724 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1725 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1726 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1727 has an implication that the given string is incomplete (because a partial
1728 match is preferred over a full match). For this reason, these items now
1729 give a partial match in this situation. [Aside: previously, the one case
1730 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1731 match rather than a full match, which was wrong by the old rules, but is
1732 now correct.]
1733
1734 6. There was a bug in the handling of #-introduced comments, recognized when
1735 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1736 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1737 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1738 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1739 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1740 places in pcre_compile().
1741
1742 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1743 comments when looking ahead for named forward references to subpatterns,
1744 the only newline sequence it recognized was NL. It now handles newlines
1745 according to the set newline convention.
1746
1747 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1748 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1749 cater for a lack of strerror(). These oversights have been fixed.
1750
1751 9. Added --match-limit and --recursion-limit to pcregrep.
1752
1753 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1754
1755 11. When the -o option was used, pcregrep was setting a return code of 1, even
1756 when matches were found, and --line-buffered was not being honoured.
1757
1758 12. Added an optional parentheses number to the -o and --only-matching options
1759 of pcregrep.
1760
1761 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1762 can match an empty string. The code to do it in pcretest and pcredemo
1763 needed fixing:
1764
1765 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1766 only one byte after an empty string match just before CRLF (this case
1767 just got forgotten; "any" and "anycrlf" were OK).
1768
1769 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1770 mode when an empty string match preceded an ASCII character followed by
1771 a non-ASCII character. (The code for advancing by one character rather
1772 than one byte was nonsense.)
1773
1774 (c) The pcredemo.c sample program did not have any code at all to handle
1775 the cases when CRLF is a valid newline sequence.
1776
1777 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1778 as a starting offset was within the subject string. There is now a new
1779 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1780 negative or greater than the length of the string. In order to test this,
1781 pcretest is extended to allow the setting of negative starting offsets.
1782
1783 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1784 starting offset points to the beginning of a UTF-8 character was
1785 unnecessarily clumsy. I tidied it up.
1786
1787 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1788 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1789
1790 17. Nobody had reported that the --include_dir option, which was added in
1791 release 7.7 should have been called --include-dir (hyphen, not underscore)
1792 for compatibility with GNU grep. I have changed it to --include-dir, but
1793 left --include_dir as an undocumented synonym, and the same for
1794 --exclude-dir, though that is not available in GNU grep, at least as of
1795 release 2.5.4.
1796
1797 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1798 characters from a string of bytes) have been redefined so as not to use
1799 loops, in order to improve performance in some environments. At the same
1800 time, I abstracted some of the common code into auxiliary macros to save
1801 repetition (this should not affect the compiled code).
1802
1803 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1804 compile-time error is now given if \c is not followed by an ASCII
1805 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1806 different, and any byte value is allowed.)
1807
1808 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1809 START_OPTIMIZE option, which is now allowed at compile time - but just
1810 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1811 to pcregrep and other applications that have no direct access to PCRE
1812 options. The new /Y option in pcretest sets this option when calling
1813 pcre_compile().
1814
1815 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1816 back references. Groups containing recursive back references were forced to
1817 be atomic by that change, but in the case of named groups, the amount of
1818 memory required was incorrectly computed, leading to "Failed: internal
1819 error: code overflow". This has been fixed.
1820
1821 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1822 pcretest.c, to avoid build problems in some Borland environments.
1823
1824
1825 Version 8.10 25-Jun-2010
1826 ------------------------
1827
1828 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1829 THEN.
1830
1831 2. (*ACCEPT) was not working when inside an atomic group.
1832
1833 3. Inside a character class, \B is treated as a literal by default, but
1834 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1835 causes the error). The code is unchanged, but I tidied the documentation.
1836
1837 4. Inside a character class, PCRE always treated \R and \X as literals,
1838 whereas Perl faults them if its -w option is set. I have changed PCRE so
1839 that it faults them when PCRE_EXTRA is set.
1840
1841 5. Added support for \N, which always matches any character other than
1842 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1843
1844 6. When compiling pcregrep with newer versions of gcc which may have
1845 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1846 declared with attribute warn_unused_result" were given. Just casting the
1847 result to (void) does not stop the warnings; a more elaborate fudge is
1848 needed. I've used a macro to implement this.
1849
1850 7. Minor change to pcretest.c to avoid a compiler warning.
1851
1852 8. Added four artifical Unicode properties to help with an option to make
1853 \s etc use properties (see next item). The new properties are: Xan
1854 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1855
1856 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1857 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1858 this option. Modified pcretest to add /W to test this facility. Added
1859 REG_UCP to make it available via the POSIX interface.
1860
1861 10. Added --line-buffered to pcregrep.
1862
1863 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1864 studied, and the match started with a letter with a code point greater than
1865 127 whose first byte was different to the first byte of the other case of
1866 the letter, the other case of this starting letter was not recognized
1867 (#976).
1868
1869 12. If a pattern that was studied started with a repeated Unicode property
1870 test, for example, \p{Nd}+, there was the theoretical possibility of
1871 setting up an incorrect bitmap of starting bytes, but fortunately it could
1872 not have actually happened in practice until change 8 above was made (it
1873 added property types that matched character-matching opcodes).
1874
1875 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1876 possible starting bytes for non-anchored patterns.
1877
1878 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1879 \R, and also a number of cases that involve Unicode properties, both
1880 explicit and implicit when PCRE_UCP is set.
1881
1882 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1883 input, it could crash or give wrong results if characters with values
1884 greater than 0xc0 were present in the subject string. (Detail: it assumed
1885 UTF-8 input when processing these items.)
1886
1887 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1888 size_t is 64-bit (#991).
1889
1890 17. Added a check for running out of memory when PCRE is compiled with
1891 --disable-stack-for-recursion (#990).
1892
1893 18. If the last data line in a file for pcretest does not have a newline on
1894 the end, a newline was missing in the output.
1895
1896 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1897 less than 128) in its various bitmaps. However, there is a facility for
1898 generating tables according to the current locale when PCRE is compiled. It
1899 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1900 characters, are recognized by isspace() and therefore were getting set in
1901 these tables, and indeed these tables seem to approximate to ISO 8859. This
1902 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1903 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1904 which of course cannot start UTF-8 characters. I have changed the code so
1905 that only real ASCII characters (less than 128) and the correct starting
1906 bytes for UTF-8 encodings are set for characters greater than 127 when in
1907 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1908 altogether.)
1909
1910 20. Added the /T option to pcretest so as to be able to run tests with non-
1911 standard character tables, thus making it possible to include the tests
1912 used for 19 above in the standard set of tests.
1913
1914 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1915 reference to a subpattern the other side of a comment that contains an
1916 opening parenthesis caused either an internal compiling error, or a
1917 reference to the wrong subpattern.
1918
1919
1920 Version 8.02 19-Mar-2010
1921 ------------------------
1922
1923 1. The Unicode data tables have been updated to Unicode 5.2.0.
1924
1925 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1926 configured.
1927
1928 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1929 original author of that file, following a query about its status.
1930
1931 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1932 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1933
1934 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1935 quantifier applied to a forward-referencing subroutine call, could compile
1936 incorrect code or give the error "internal error: previously-checked
1937 referenced subpattern not found".
1938
1939 6. Both MS Visual Studio and Symbian OS have problems with initializing
1940 variables to point to external functions. For these systems, therefore,
1941 pcre_malloc etc. are now initialized to local functions that call the
1942 relevant global functions.
1943
1944 7. There were two entries missing in the vectors called coptable and poptable
1945 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1946 I've fixed the data, and added a kludgy way of testing at compile time that
1947 the lengths are correct (equal to the number of opcodes).
1948
1949 8. Following on from 7, I added a similar kludge to check the length of the
1950 eint vector in pcreposix.c.
1951
1952 9. Error texts for pcre_compile() are held as one long string to avoid too
1953 much relocation at load time. To find a text, the string is searched,
1954 counting zeros. There was no check for running off the end of the string,
1955 which could happen if a new error number was added without updating the
1956 string.
1957
1958 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1959
1960 11. \K was not working if it appeared in an atomic group or in a group that
1961 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1962 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1963 the assertion is positive, but not if it is negative.
1964
1965 12. Change 11 fortuitously reduced the size of the stack frame used in the
1966 "match()" function of pcre_exec.c by one pointer. Forthcoming
1967 implementation of support for (*MARK) will need an extra pointer on the
1968 stack; I have reserved it now, so that the stack frame size does not
1969 decrease.
1970
1971 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1972 item in branch that calls a recursion is a subroutine call - as in the
1973 second branch in the above example - was incorrectly given the compile-
1974 time error "recursive call could loop indefinitely" because pcre_compile()
1975 was not correctly checking the subroutine for matching a non-empty string.
1976
1977 14. The checks for overrunning compiling workspace could trigger after an
1978 overrun had occurred. This is a "should never occur" error, but it can be
1979 triggered by pathological patterns such as hundreds of nested parentheses.
1980 The checks now trigger 100 bytes before the end of the workspace.
1981
1982 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1983
1984
1985 Version 8.01 19-Jan-2010
1986 ------------------------
1987
1988 1. If a pattern contained a conditional subpattern with only one branch (in
1989 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1990 computed the wrong minimum data length (which is of course zero for such
1991 subpatterns). This could cause incorrect "no match" results.
1992
1993 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1994 the pattern is reset in the first branch, pcre_compile() failed with
1995 "internal error: code overflow at offset...". This happened only when
1996 the reset was to the original external option setting. (An optimization
1997 abstracts leading options settings into an external setting, which was the
1998 cause of this.)
1999
2000 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2001 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2002 assertion pattern did not match (meaning that the assertion was true), it
2003 was incorrectly treated as false if the SKIP had been reached during the
2004 matching. This also applied to assertions used as conditions.
2005
2006 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2007 assertion subpattern, including such a pattern used as a condition,
2008 unpredictable results occurred, instead of the error return
2009 PCRE_ERROR_DFA_UITEM.
2010
2011 5. The C++ GlobalReplace function was not working like Perl for the special
2012 situation when an empty string is matched. It now does the fancy magic
2013 stuff that is necessary.
2014
2015 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2016 removed. (These were left over from very, very early versions of PCRE.)
2017
2018 7. Some cosmetic changes to the code to make life easier when compiling it
2019 as part of something else:
2020
2021 (a) Change DEBUG to PCRE_DEBUG.
2022
2023 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2024 called "current" as "current_branch", to prevent a collision with the
2025 Linux macro when compiled as a kernel module.
2026
2027 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2028 prevent a collision with the Linux macro when compiled as a kernel
2029 module.
2030
2031 8. In pcre_compile() there are some checks for integer overflows that used to
2032 cast potentially large values to (double). This has been changed to that
2033 when building, a check for int64_t is made, and if it is found, it is used
2034 instead, thus avoiding the use of floating point arithmetic. (There is no
2035 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2036 double.
2037
2038 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2039 2005 (difference between two addresses compared to an unsigned value).
2040
2041 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2042 custom one, because of the following reported problem in Windows:
2043
2044 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2045 under Win32.
2046 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2047 therefore missing the function definition.
2048 - The compiler thus generates a "C" signature for the test function.
2049 - The linker fails to find the "C" function.
2050 - PCRE fails to configure if asked to do so against libbz2.
2051
2052 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2053 messages were output:
2054
2055 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2056 rerunning libtoolize, to keep the correct libtool macros in-tree.
2057 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2058
2059 I have done both of these things.
2060
2061 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2062 most of the time, it *can* run out if it is given a pattern that contains a
2063 runaway infinite recursion. I updated the discussion in the pcrestack man
2064 page.
2065
2066 13. Now that we have gone to the x.xx style of version numbers, the minor
2067 version may start with zero. Using 08 or 09 is a bad idea because users
2068 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2069 interpreted as invalid octal numbers. I've updated the previous comment in
2070 configure.ac, and also added a check that gives an error if 08 or 09 are
2071 used.
2072
2073 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2074 causing partial matching to fail when the end of the subject matched \W
2075 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2076
2077 15. There were some discrepancies between the declarations in pcre_internal.h
2078 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2079 their definitions. The declarations used "const uschar *" and the
2080 definitions used USPTR. Even though USPTR is normally defined as "const
2081 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2082 reported that: "This difference in casting confuses some C++ compilers, for
2083 example, SunCC recognizes above declarations as different functions and
2084 generates broken code for hbpcre." I have changed the declarations to use
2085 USPTR.
2086
2087 16. GNU libtool is named differently on some systems. The autogen.sh script now
2088 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2089 (FreeBSD).
2090
2091 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2092 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2093 comment: "Figure out how to create a longlong from a string: strtoll and
2094 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2095 instance, but it only takes 2 args instead of 3!"
2096
2097 18. A subtle bug concerned with back references has been fixed by a change of
2098 specification, with a corresponding code fix. A pattern such as
2099 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2100 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2101 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2102 same bug. Such groups have to be quantified to be useful, or contained
2103 inside another quantified group. (If there's no repetition, the reference
2104 can never match.) The problem arises because, having left the group and
2105 moved on to the rest of the pattern, a later failure that backtracks into
2106 the group uses the captured value from the final iteration of the group
2107 rather than the correct earlier one. I have fixed this in PCRE by forcing
2108 any group that contains a reference to itself to be an atomic group; that
2109 is, there cannot be any backtracking into it once it has completed. This is
2110 similar to recursive and subroutine calls.
2111
2112
2113 Version 8.00 19-Oct-09
2114 ----------------------
2115
2116 1. The table for translating pcre_compile() error codes into POSIX error codes
2117 was out-of-date, and there was no check on the pcre_compile() error code
2118 being within the table. This could lead to an OK return being given in
2119 error.
2120
2121 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2122 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2123 in a Windows environment.
2124
2125 3. The pcregrep --count option prints the count for each file even when it is
2126 zero, as does GNU grep. However, pcregrep was also printing all files when
2127 --files-with-matches was added. Now, when both options are given, it prints
2128 counts only for those files that have at least one match. (GNU grep just
2129 prints the file name in this circumstance, but including the count seems
2130 more useful - otherwise, why use --count?) Also ensured that the
2131 combination -clh just lists non-zero counts, with no names.
2132
2133 4. The long form of the pcregrep -F option was incorrectly implemented as
2134 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2135 but it seems right to fix it, and I didn't think it was worth preserving
2136 the old behaviour.
2137
2138 5. The command line items --regex=pattern and --regexp=pattern were not
2139 recognized by pcregrep, which required --regex pattern or --regexp pattern
2140 (with a space rather than an '='). The man page documented the '=' forms,
2141 which are compatible with GNU grep; these now work.
2142
2143 6. No libpcreposix.pc file was created for pkg-config; there was just
2144 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2145
2146 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2147 when UCP support is not needed, by modifying the Python script that
2148 generates it from Unicode data files. This should not matter if the module
2149 is correctly used as a library, but I received one complaint about 50K of
2150 unwanted data. My guess is that the person linked everything into his
2151 program rather than using a library. Anyway, it does no harm.
2152
2153 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2154 was a minimum greater than 1 for a wide character in a possessive
2155 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2156 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2157 character. Chaos in the form of incorrect output or a compiling loop could
2158 result.
2159
2160 9. The restrictions on what a pattern can contain when partial matching is
2161 requested for pcre_exec() have been removed. All patterns can now be
2162 partially matched by this function. In addition, if there are at least two
2163 slots in the offset vector, the offset of the earliest inspected character
2164 for the match and the offset of the end of the subject are set in them when
2165 PCRE_ERROR_PARTIAL is returned.
2166
2167 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2168 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2169 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2170 and may be more useful for multi-segment matching.
2171
2172 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2173 used to be given if ever the end of the subject was reached; now it is
2174 given only if matching could not proceed because another character was
2175 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2176 string "Z", which now yields "no match" instead of "partial match". In the
2177 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2178 final character ended with (*FAIL).
2179
2180 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2181 if the pattern had a "must contain" character that was already found in the
2182 earlier partial match, unless partial matching was again requested. For
2183 example, with the pattern /dog.(body)?/, the "must contain" character is
2184 "g". If the first part-match was for the string "dog", restarting with
2185 "sbody" failed. This bug has been fixed.
2186
2187 13. The string returned by pcre_dfa_exec() after a partial match has been
2188 changed so that it starts at the first inspected character rather than the
2189 first character of the match. This makes a difference only if the pattern
2190 starts with a lookbehind assertion or \b or \B (\K is not supported by
2191 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2192 matching functions compatible, and I think it's the right thing to do.
2193
2194 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2195 so that the demonstration program is easily available in environments where
2196 PCRE has not been installed from source.
2197
2198 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2199 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2200 library.
2201
2202 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2203 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2204 is not the first non-POSIX option to be added. Clearly some people find
2205 these options useful.
2206
2207 17. If a caller to the POSIX matching function regexec() passes a non-zero
2208 value for nmatch with a NULL value for pmatch, the value of
2209 nmatch is forced to zero.
2210
2211 18. RunGrepTest did not have a test for the availability of the -u option of
2212 the diff command, as RunTest does. It now checks in the same way as
2213 RunTest, and also checks for the -b option.
2214
2215 19. If an odd number of negated classes containing just a single character
2216 interposed, within parentheses, between a forward reference to a named
2217 subpattern and the definition of the subpattern, compilation crashed with
2218 an internal error, complaining that it could not find the referenced
2219 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2220 [The bug was that it was starting one character too far in when skipping
2221 over the character class, thus treating the ] as data rather than
2222 terminating the class. This meant it could skip too much.]
2223
2224 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2225 /g option in pcretest when the pattern contains \K, which makes it possible
2226 to have an empty string match not at the start, even when the pattern is
2227 anchored. Updated pcretest and pcredemo to use this option.
2228
2229 21. If the maximum number of capturing subpatterns in a recursion was greater
2230 than the maximum at the outer level, the higher number was returned, but
2231 with unset values at the outer level. The correct (outer level) value is
2232 now given.
2233
2234 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2235 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2236 make it do so. The string so far is captured, making this feature
2237 compatible with Perl.
2238
2239 23. The tests have been re-organized, adding tests 11 and 12, to make it
2240 possible to check the Perl 5.10 features against Perl 5.10.
2241
2242 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2243 pattern matches a fixed length string. PCRE did not allow this; now it
2244 does. Neither allows recursion.
2245
2246 25. I finally figured out how to implement a request to provide the minimum
2247 length of subject string that was needed in order to match a given pattern.
2248 (It was back references and recursion that I had previously got hung up
2249 on.) This code has now been added to pcre_study(); it finds a lower bound
2250 to the length of subject needed. It is not necessarily the greatest lower
2251 bound, but using it to avoid searching strings that are too short does give
2252 some useful speed-ups. The value is available to calling programs via
2253 pcre_fullinfo().
2254
2255 26. While implementing 25, I discovered to my embarrassment that pcretest had
2256 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2257 study optimizations had never been tested with that matching function.
2258 Oops. What is worse, even when it was passed study data, there was a bug in
2259 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2260 were also very few tests of studied patterns with pcre_dfa_exec().
2261
2262 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2263 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2264 on the other side of the coin, they are no longer allowed to have different
2265 names, because these cannot be distinguished in PCRE, and this has caused
2266 confusion. (This is a difference from Perl.)
2267
2268 28. When duplicate subpattern names are present (necessarily with different
2269 numbers, as required by 27 above), and a test is made by name in a
2270 conditional pattern, either for a subpattern having been matched, or for
2271 recursion in such a pattern, all the associated numbered subpatterns are
2272 tested, and the overall condition is true if the condition is true for any
2273 one of them. This is the way Perl works, and is also more like the way
2274 testing by number works.
2275
2276
2277 Version 7.9 11-Apr-09
2278 ---------------------
2279
2280 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2281 (pcretest), all targets were linked against these libraries. This included
2282 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2283 libraries. This caused unwanted dependencies to be created. This problem
2284 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2285 pcretest is linked with readline.
2286
2287 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2288 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2289 moved outside it again, because FALSE and TRUE are already defined in AIX,
2290 but BOOL is not.
2291
2292 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2293 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2294
2295 4. The pcregrep documentation said spaces were inserted as well as colons (or
2296 hyphens) following file names and line numbers when outputting matching
2297 lines. This is not true; no spaces are inserted. I have also clarified the
2298 wording for the --colour (or --color) option.
2299
2300 5. In pcregrep, when --colour was used with -o, the list of matching strings
2301 was not coloured; this is different to GNU grep, so I have changed it to be
2302 the same.
2303
2304 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2305 each matching line was coloured. Now it goes on to look for further matches
2306 of any of the test patterns, which is the same behaviour as GNU grep.
2307
2308 7. A pattern that could match an empty string could cause pcregrep to loop; it
2309 doesn't make sense to accept an empty string match in pcregrep, so I have
2310 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2311 seems to be how GNU grep behaves. [But see later change 40 for release
2312 8.33.]
2313
2314 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2315 start or after a newline", because the conditional assertion was not being
2316 correctly handled. The rule now is that both the assertion and what follows
2317 in the first alternative must satisfy the test.
2318
2319 9. If auto-callout was enabled in a pattern with a conditional group whose
2320 condition was an assertion, PCRE could crash during matching, both with
2321 pcre_exec() and pcre_dfa_exec().
2322
2323 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2324 used for matching.
2325
2326 11. Unicode property support in character classes was not working for
2327 characters (bytes) greater than 127 when not in UTF-8 mode.
2328
2329 12. Added the -M command line option to pcretest.
2330
2331 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2332
2333 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2334
2335 16. Added comments and documentation about mis-use of no_arg in the C++
2336 wrapper.
2337
2338 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2339 from Martin Jerabek that uses macro names for all relevant character and
2340 string constants.
2341
2342 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2343 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2344 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2345 these, but not everybody uses configure.
2346
2347 19. A conditional group that had only one branch was not being correctly
2348 recognized as an item that could match an empty string. This meant that an
2349 enclosing group might also not be so recognized, causing infinite looping
2350 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2351 with the subject "ab", where knowledge that the repeated group can match
2352 nothing is needed in order to break the loop.
2353
2354 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2355 exec(), but without supplying a callout function, matching went wrong.
2356
2357 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2358 leak if the size of the offset vector was greater than 30. When the vector
2359 is smaller, the saved offsets during recursion go onto a local stack
2360 vector, but for larger vectors malloc() is used. It was failing to free
2361 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2362 error, in fact).
2363
2364 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2365 heapframe that is used only when UTF-8 support is enabled. This caused no
2366 problem, but was untidy.
2367
2368 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2369 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2370 included within another project.
2371
2372 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2373 slightly modified by me:
2374
2375 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2376 not building pcregrep.
2377
2378 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2379 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2380
2381 25. Forward references, both numeric and by name, in patterns that made use of
2382 duplicate group numbers, could behave incorrectly or give incorrect errors,
2383 because when scanning forward to find the reference group, PCRE was not
2384 taking into account the duplicate group numbers. A pattern such as
2385 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2386
2387 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2388 the feature of a custom pointer more persuasive (as requested by a user).
2389
2390 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2391 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2392 pre-defined.
2393
2394 28. Added support for (*UTF8) at the start of a pattern.
2395
2396 29. Arrange for flags added by the "release type" setting in CMake to be shown
2397 in the configuration summary.
2398
2399
2400 Version 7.8 05-Sep-08
2401 ---------------------
2402
2403 1. Replaced UCP searching code with optimized version as implemented for Ad
2404 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2405 stage table and inline lookup instead of a function, giving speed ups of 2
2406 to 5 times on some simple patterns that I tested. Permission was given to
2407 distribute the MultiStage2.py script that generates the tables (it's not in
2408 the tarball, but is in the Subversion repository).
2409
2410 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2411 scripts.
2412
2413 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2414 a group with a zero qualifier. The result of the study could be incorrect,
2415 or the function might crash, depending on the pattern.
2416
2417 4. Caseless matching was not working for non-ASCII characters in back
2418 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2419 It now works when Unicode Property Support is available.
2420
2421 5. In pcretest, an escape such as \x{de} in the data was always generating
2422 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2423 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2424 truncation.
2425
2426 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2427
2428 7. Added two (int) casts to pcregrep when printing the difference of two
2429 pointers, in case they are 64-bit values.
2430
2431 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2432 test 2 if it fails.
2433
2434 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2435 and a #define of that name to empty if it is not externally set. This is to
2436 allow users of MSVC to set it if necessary.
2437
2438 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2439 the convenience functions in the pcre_get.c source file.
2440
2441 11. An option change at the start of a pattern that had top-level alternatives
2442 could cause overwriting and/or a crash. This command provoked a crash in
2443 some environments:
2444
2445 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2446
2447 This potential security problem was recorded as CVE-2008-2371.
2448
2449 12. For a pattern where the match had to start at the beginning or immediately
2450 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2451 pcre_dfa_exec() could read past the end of the passed subject if there was
2452 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2453 pcretest so that it places the subject at the end of its malloc-ed buffer.
2454
2455 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2456 exec() might read past the end of the data buffer in UTF-8 mode.
2457
2458 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2459 the data contained the byte 0x85 as part of a UTF-8 character within its
2460 first line. This applied both to normal and DFA matching.
2461
2462 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2463 /^[^d]*?$/8 failed to match "abc".
2464
2465 16. Added a missing copyright notice to pcrecpp_internal.h.
2466
2467 17. Make it more clear in the documentation that values returned from
2468 pcre_exec() in ovector are byte offsets, not character counts.
2469
2470 18. Tidied a few places to stop certain compilers from issuing warnings.
2471
2472 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2473 supplied by Stefan Weber. I made a further small update for 7.8 because
2474 there is a change of source arrangements: the pcre_searchfuncs.c module is
2475 replaced by pcre_ucd.c.
2476
2477
2478 Version 7.7 07-May-08
2479 ---------------------
2480
2481 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2482 a string to a long long, pretend we don't even have a long long." This is
2483 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2484
2485 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2486 pre-7.6 versions, which defined a global no_arg variable instead of putting
2487 it in the RE class. (See also #8 below.)
2488
2489 3. Remove a line of dead code, identified by coverity and reported by Nuno
2490 Lopes.
2491
2492 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2493
2494 (1) The include/exclude patterns were being applied to the whole pathnames
2495 of files, instead of just to the final components.
2496
2497 (2) If there was more than one level of directory, the subdirectories were
2498 skipped unless they satisfied the include/exclude conditions. This is
2499 inconsistent with GNU grep (and could even be seen as contrary to the
2500 pcregrep specification - which I improved to make it absolutely clear).
2501 The action now is always to scan all levels of directory, and just
2502 apply the include/exclude patterns to regular files.
2503
2504 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2505 --exclude_dir in the tests to avoid scanning .svn directories.
2506
2507 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2508 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2509 doesn't support NULs in patterns.
2510
2511 7. Added some missing "const"s to declarations of static tables in
2512 pcre_compile.c and pcre_dfa_exec.c.
2513
2514 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2515 caused by fix #2 above. (Subsequently also a second patch to fix the
2516 first patch. And a third patch - this was a messy problem.)
2517
2518 9. Applied Craig's patch to remove the use of push_back().
2519
2520 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2521 matching function regexec().
2522
2523 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2524 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2525 references. PCRE supports relative numbers with this syntax (I don't think
2526 Oniguruma does).
2527
2528 12. Previously, a group with a zero repeat such as (...){0} was completely
2529 omitted from the compiled regex. However, this means that if the group
2530 was called as a subroutine from elsewhere in the pattern, things went wrong
2531 (an internal error was given). Such groups are now left in the compiled
2532 pattern, with a new opcode that causes them to be skipped at execution
2533 time.
2534
2535 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2536 to the way PCRE behaves:
2537
2538 (a) A lone ] character is dis-allowed (Perl treats it as data).
2539
2540 (b) A back reference to an unmatched subpattern matches an empty string
2541 (Perl fails the current match path).
2542
2543 (c) A data ] in a character class must be notated as \] because if the
2544 first data character in a class is ], it defines an empty class. (In
2545 Perl it is not possible to have an empty class.) The empty class []
2546 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2547 The negative empty class [^] matches any one character, independently
2548 of the DOTALL setting.
2549
2550 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2551 non-existent subpattern following a character class starting with ']' and
2552 containing () gave an internal compiling error instead of "reference to
2553 non-existent subpattern". Fortunately, when the pattern did exist, the
2554 compiled code was correct. (When scanning forwards to check for the
2555 existence of the subpattern, it was treating the data ']' as terminating
2556 the class, so got the count wrong. When actually compiling, the reference
2557 was subsequently set up correctly.)
2558
2559 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2560 it was being rejected as not supported by pcre_dfa_exec(), even though
2561 other assertions are supported. I have made pcre_dfa_exec() support
2562 (*FAIL).
2563
2564 16. The implementation of 13c above involved the invention of a new opcode,
2565 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2566 cannot be changed at match time, I realized I could make a small
2567 improvement to matching performance by compiling OP_ALLANY instead of
2568 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2569 on the OP_ANY path.
2570
2571 17. Compiling pcretest on Windows with readline support failed without the
2572 following two fixes: (1) Make the unistd.h include conditional on
2573 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2574
2575 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2576 ncurses library to be included for pcretest when ReadLine support is
2577 requested, but also to allow for it to be overridden. This patch came from
2578 Daniel Bergström.
2579
2580 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2581 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2582 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2583 spotting this.
2584
2585
2586 Version 7.6 28-Jan-08
2587 ---------------------
2588
2589 1. A character class containing a very large number of characters with
2590 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2591 overflow.
2592
2593 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2594 HAVE_LONG_LONG is not defined.
2595
2596 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2597 bring them up to date and include new features. This patch includes:
2598
2599 - Fixed PH's badly added libz and libbz2 support.
2600 - Fixed a problem with static linking.
2601 - Added pcredemo. [But later removed - see 7 below.]
2602 - Fixed dftables problem and added an option.
2603 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2604 HAVE_LONG_LONG.
2605 - Added readline support for pcretest.
2606 - Added an listing of the option settings after cmake has run.
2607
2608 4. A user submitted a patch to Makefile that makes it easy to create
2609 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2610 Makefile.am that cause it to include this special target, without
2611 affecting anything else. Note that the same mingw target plus all
2612 the other distribution libraries and programs are now supported
2613 when configuring with CMake (see 6 below) instead of with
2614 Configure/Make.
2615
2616 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2617 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2618 exported in the Windows port". It has not yet been confirmed that the patch
2619 solves the problem, but it does no harm.
2620
2621 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2622 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2623 with CMake, and also correct the comment about stack recursion.
2624
2625 7. Remove the automatic building of pcredemo from the ./configure system and
2626 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2627 of a program that users should build themselves after PCRE is installed, so
2628 building it automatically is not really right. What is more, it gave
2629 trouble in some build environments.
2630
2631 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2632
2633
2634 Version 7.5 10-Jan-08
2635 ---------------------
2636
2637 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2638 values in parens when parsing an RE using the C++ wrapper."
2639
2640 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2641 Characters greater than 255 were excluded from the class instead of being
2642 included.
2643
2644 3. The same bug as (2) above applied to negated POSIX classes such as
2645 [:^space:].
2646
2647 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2648 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2649 I have changed it.
2650
2651 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2652 first named subpattern) and a construct such as (?&a) would reference the
2653 first named subpattern whose name started with "a" (in other words, the
2654 length check was missing). Both these problems are fixed. "Subpattern name
2655 expected" is now given for (?&) (a zero-length name), and this patch also
2656 makes it give the same error for \k'' (previously it complained that that
2657 was a reference to a non-existent subpattern).
2658
2659 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2660 this is right because (?- can be followed by option settings as well as by
2661 digits. I have, however, made the messages clearer.
2662
2663 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2664 than the number used in the conditional) now cause a compile-time error.
2665 This is actually not compatible with Perl, which accepts such patterns, but
2666 treats the conditional as always being FALSE (as PCRE used to), but it
2667 seems to me that giving a diagnostic is better.
2668
2669 8. Change "alphameric" to the more common word "alphanumeric" in comments
2670 and messages.
2671
2672 9. Fix two occurrences of "backslash" in comments that should have been
2673 "backspace".
2674
2675 10. Remove two redundant lines of code that can never be obeyed (their function
2676 was moved elsewhere).
2677
2678 11. The program that makes PCRE's Unicode character property table had a bug
2679 which caused it to generate incorrect table entries for sequences of
2680 characters that have the same character type, but are in different scripts.
2681 It amalgamated them into a single range, with the script of the first of
2682 them. In other words, some characters were in the wrong script. There were
2683 thirteen such cases, affecting characters in the following ranges:
2684
2685 U+002b0 - U+002c1
2686 U+0060c - U+0060d
2687 U+0061e - U+00612
2688 U+0064b - U+0065e
2689 U+0074d - U+0076d
2690 U+01800 - U+01805
2691 U+01d00 - U+01d77
2692 U+01d9b - U+01dbf
2693 U+0200b - U+0200f
2694 U+030fc - U+030fe
2695 U+03260 - U+0327f
2696 U+0fb46 - U+0fbb1
2697 U+10450 - U+1049d
2698
2699 12. The -o option (show only the matching part of a line) for pcregrep was not
2700 compatible with GNU grep in that, if there was more than one match in a
2701 line, it showed only the first of them. It now behaves in the same way as
2702 GNU grep.
2703
2704 13. If the -o and -v options were combined for pcregrep, it printed a blank
2705 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2706 does the same. The return code can be used to tell if there were any
2707 non-matching lines.
2708
2709 14. Added --file-offsets and --line-offsets to pcregrep.
2710
2711 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2712 infinitely looping recursion. The bug was that positive lookaheads were not
2713 being skipped when checking for a possible empty match (negative lookaheads
2714 and both kinds of lookbehind were skipped).
2715
2716 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2717 inclusion of <windows.h> to before rather than after the definition of
2718 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2719
2720 17. Specifying a possessive quantifier with a specific limit for a Unicode
2721 character property caused pcre_compile() to compile bad code, which led at
2722 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2723 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2724 caused the error; without that there was no problem.
2725
2726 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2727
2728 19. Added --enable-pcretest-libreadline.
2729
2730 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2731 RE::GlobalReplace(). As a result, the number of replacements returned was
2732 double what it should be. I removed one of the increments, but Craig sent a
2733 later patch that removed the other one (the right fix) and added unit tests
2734 that check the return values (which was not done before).
2735
2736 21. Several CMake things:
2737
2738 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2739 the names libpcre and libpcreposix, not just pcre and pcreposix.
2740
2741 (2) The above change means that pcretest and pcregrep are now correctly
2742 linked with the newly-built libraries, not previously installed ones.
2743
2744 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2745
2746 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2747 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2748 UTF-8 newline character). The key issue is that the pattern starts .*;
2749 this means that the match must be either at the beginning, or after a
2750 newline. The bug was in the code for advancing after a failed match and
2751 checking that the new position followed a newline. It was not taking
2752 account of UTF-8 characters correctly.
2753
2754 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2755 character classes. PCRE was not treating the sequence [:...:] as a
2756 character class unless the ... were all letters. Perl, however, seems to
2757 allow any characters between [: and :], though of course it rejects as
2758 unknown any "names" that contain non-letters, because all the known class
2759 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2760 for example, whereas PCRE did not - it did not recognize a POSIX character
2761 class. This seemed a bit dangerous, so the code has been changed to be
2762 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2763 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2764 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2765 Perl does, and where it didn't before.
2766
2767 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2768 Windows environments %n is disabled by default.
2769
2770
2771 Version 7.4 21-Sep-07
2772 ---------------------
2773
2774 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2775 means that a class such as [\s] counted as "explicit reference to CR or
2776 LF". That isn't really right - the whole point of the change was to try to
2777 help when there was an actual mention of one of the two characters. So now
2778 the change happens only if \r or \n (or a literal CR or LF) character is
2779 encountered.
2780
2781 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2782 of both had grown to the point where there were only 3 bits left.
2783 Fortunately, there was spare space in the data structure, and so I have
2784 moved the internal flags into a new 16-bit field to free up more option
2785 bits.
2786
2787 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2788 but did not set the internal JCHANGED flag - either of these is enough to
2789 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2790 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2791 start sets both bits.
2792
2793 4. Added options (at build time, compile time, exec time) to change \R from
2794 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2795
2796 5. doc/pcresyntax.html was missing from the distribution.
2797
2798 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2799 compatibility, even though it is no longer used.
2800
2801 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2802 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2803 windows.h file is present (where different names are used). [This was
2804 reversed later after testing - see 16 below.]
2805
2806 8. Changed all #include <config.h> to #include "config.h". There were also
2807 some further <pcre.h> cases that I changed to "pcre.h".
2808
2809 9. When pcregrep was used with the --colour option, it missed the line ending
2810 sequence off the lines that it output.
2811
2812 10. It was pointed out to me that arrays of string pointers cause lots of
2813 relocations when a shared library is dynamically loaded. A technique of
2814 using a single long string with a table of offsets can drastically reduce
2815 these. I have refactored PCRE in four places to do this. The result is
2816 dramatic:
2817
2818 Originally: 290
2819 After changing UCP table: 187
2820 After changing error message table: 43
2821 After changing table of "verbs" 36
2822 After changing table of Posix names 22
2823
2824 Thanks to the folks working on Gregex for glib for this insight.
2825
2826 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2827 unicode-properties was also set.
2828
2829 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2830
2831 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2832 checked only for CRLF.
2833
2834 14. Added casts to pcretest.c to avoid compiler warnings.
2835
2836 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2837
2838 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2839 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2840 entirely. This removes changes made in 7 above.
2841
2842 17. The CMake files have been updated, and there is now more information about
2843 building with CMake in the NON-UNIX-USE document.
2844
2845
2846 Version 7.3 28-Aug-07
2847 ---------------------
2848
2849 1. In the rejigging of the build system that eventually resulted in 7.1, the
2850 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2851 brackets there is not right, since it causes compilers to look for an
2852 installed pcre.h, not the version that is in the source that is being
2853 compiled (which of course may be different). I have changed it back to:
2854
2855 #include "pcre.h"
2856
2857 I have a vague recollection that the change was concerned with compiling in
2858 different directories, but in the new build system, that is taken care of
2859 by the VPATH setting the Makefile.
2860
2861 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2862 when the subject happened to end in the byte 0x85 (e.g. if the last
2863 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2864 characters but of course it shouldn't be taken as a newline when it is part
2865 of another character. The bug was that, for an unlimited repeat of . in
2866 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2867 characters when looking for a newline.
2868
2869 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2870
2871 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2872 in debug output.
2873
2874 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2875 long printing in the pcrecpp unittest when running under MinGW.
2876
2877 6. ESC_K was left out of the EBCDIC table.
2878
2879 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2880 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2881 limit also applies to "virtual nesting" when a pattern is recursive, and in
2882 this case 1000 isn't so big. I have been able to remove this limit at the
2883 expense of backing off one optimization in certain circumstances. Normally,
2884 when pcre_exec() would call its internal match() function recursively and
2885 immediately return the result unconditionally, it uses a "tail recursion"
2886 feature to save stack. However, when a subpattern that can match an empty
2887 string has an unlimited repetition quantifier, it no longer makes this
2888 optimization. That gives it a stack frame in which to save the data for
2889 checking that an empty string has been matched. Previously this was taken
2890 from the 1000-entry workspace that had been reserved. So now there is no
2891 explicit limit, but more stack is used.
2892
2893 8. Applied Daniel's patches to solve problems with the import/export magic
2894 syntax that is required for Windows, and which was going wrong for the
2895 pcreposix and pcrecpp parts of the library. These were overlooked when this
2896 problem was solved for the main library.
2897
2898 9. There were some crude static tests to avoid integer overflow when computing
2899 the size of patterns that contain repeated groups with explicit upper
2900 limits. As the maximum quantifier is 65535, the maximum group length was
2901 set at 30,000 so that the product of these two numbers did not overflow a
2902 32-bit integer. However, it turns out that people want to use groups that
2903 are longer than 30,000 bytes (though not repeat them that many times).
2904 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2905 made it possible to implement the integer overflow checks in a much more
2906 dynamic way, which I have now done. The artificial limitation on group
2907 length has been removed - we now have only the limit on the total length of
2908 the compiled pattern, which depends on the LINK_SIZE setting.
2909
2910 10. Fixed a bug in the documentation for get/copy named substring when
2911 duplicate names are permitted. If none of the named substrings are set, the
2912 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2913 empty string.
2914
2915 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2916 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2917 because the ] is interpreted as the first data character and the
2918 terminating ] is not found. PCRE has been made compatible with Perl in this
2919 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2920 cause memory overwriting.
2921
2922 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2923 string has been matched (to stop an infinite loop). It was not recognizing
2924 a conditional subpattern that could match an empty string if that
2925 subpattern was within another subpattern. For example, it looped when
2926 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2927 condition was not nested. This bug has been fixed.
2928
2929 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2930 past the start of the subject in the presence of bytes with the top bit
2931 set, for example "\x8aBCD".
2932
2933 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2934 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2935
2936 14. Optimized (?!) to (*FAIL).
2937
2938 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2939 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2940 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2941 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2942 does: it's just the validity check that is more restrictive.
2943
2944 16. Inserted checks for integer overflows during escape sequence (backslash)
2945 processing, and also fixed erroneous offset values for syntax errors during
2946 backslash processing.
2947
2948 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2949 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2950
2951 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2952 caused an overrun.
2953
2954 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2955 something other than just ASCII characters) inside a group that had an
2956 unlimited repeat caused a loop at compile time (while checking to see
2957 whether the group could match an empty string).
2958
2959 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2960 [\P{Any}] did so. (Error in the code for printing property names.)
2961
2962 21. An orphan \E inside a character class could cause a crash.
2963
2964 22. A repeated capturing bracket such as (A)? could cause a wild memory
2965 reference during compilation.
2966
2967 23. There are several functions in pcre_compile() that scan along a compiled
2968 expression for various reasons (e.g. to see if it's fixed length for look
2969 behind). There were bugs in these functions when a repeated \p or \P was
2970 present in the pattern. These operators have additional parameters compared
2971 with \d, etc, and these were not being taken into account when moving along
2972 the compiled data. Specifically:
2973
2974 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2975 length.
2976
2977 (b) An item such as \pL+ within a repeated group could cause crashes or
2978 loops.
2979
2980 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2981 "reference to non-existent subpattern" error.
2982
2983 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2984
2985 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2986 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2987
2988 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2989
2990 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2991 character were causing crashes (broken optimization).
2992
2993 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2994 \p or \P) caused a compile-time loop.
2995
2996 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2997 break. For example, the unstudied pattern [\r\n]A does not match the string
2998 "\r\nA" because change 7.0/46 below moves the current point on by two
2999 characters after failing to match at the start. However, the pattern \nA
3000 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3001 the same is true. There doesn't seem any very clean way out of this, but
3002 what I have chosen to do makes the common cases work: PCRE now takes note
3003 of whether there can be an explicit match for \r or \n anywhere in the
3004 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3005 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3006 pattern has explicit CR or LF references.
3007
3008 29. Added (*CR) etc for changing newline setting at start of pattern.
3009
3010
3011 Version 7.2 19-Jun-07
3012 ---------------------
3013
3014 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3015 which is apparently normally available under Windows.
3016
3017 2. Re-jig the pcregrep tests with different newline settings in an attempt
3018 to make them independent of the local environment's newline setting.
3019
3020 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3021
3022 4. Some of the "internals" tests were previously cut out when the link size
3023 was not 2, because the output contained actual offsets. The recent new
3024 "Z" feature of pcretest means that these can be cut out, making the tests
3025 usable with all link sizes.
3026
3027 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3028 stack recursion. This gives a massive performance boost under BSD, but just
3029 a small improvement under Linux. However, it saves one field in the frame
3030 in all cases.
3031
3032 6. Added more features from the forthcoming Perl 5.10:
3033
3034 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3035 recursion call. It refers to the nth most recently opened parentheses.
3036
3037 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3038 to be opened parentheses.
3039
3040 (c) Conditions that refer to capturing parentheses can be specified
3041 relatively, for example, (?(-2)... or (?(+3)...
3042
3043 (d) \K resets the start of the current match so that everything before
3044 is not part of it.
3045
3046 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3047
3048 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3049 reference syntax.
3050
3051 (g) (?| introduces a group in which the numbering of parentheses in each
3052 alternative starts with the same number.
3053
3054 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3055
3056 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3057 PCRE_INFO_JCHANGED.
3058
3059 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3060 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3061 for detecting groups that can match an empty string.
3062
3063 9. A pattern with a very large number of alternatives (more than several
3064 hundred) was running out of internal workspace during the pre-compile
3065 phase, where pcre_compile() figures out how much memory will be needed. A
3066 bit of new cunning has reduced the workspace needed for groups with
3067 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3068 workspace instead of running out of the 4096 that are available.
3069
3070 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3071
3072 11. Applied patch from Google to remove an optimization that didn't quite work.
3073 The report of the bug said:
3074
3075 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3076 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3077 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3078
3079 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3080 it matched the wrong number of bytes.
3081
3082
3083 Version 7.1 24-Apr-07
3084 ---------------------
3085
3086 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3087 that is more "standard", making use of automake and other Autotools. There
3088 is some re-arrangement of the files and adjustment of comments consequent
3089 on this.
3090
3091 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3092 for recursive directory scanning broke on some systems because the files
3093 are not scanned in any specific order and on different systems the order
3094 was different. A call to "sort" has been inserted into RunGrepTest for the
3095 approprate test as a short-term fix. In the longer term there may be an
3096 alternative.
3097
3098 3. I had an email from Eric Raymond about problems translating some of PCRE's
3099 man pages to HTML (despite the fact that I distribute HTML pages, some
3100 people do their own conversions for various reasons). The problems
3101 concerned the use of low-level troff macros .br and .in. I have therefore
3102 removed all such uses from the man pages (some were redundant, some could
3103 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3104 HTML has been updated to handle .nf/.fi and to complain if it encounters
3105 .br or .in.
3106
3107 4. Updated comments in configure.ac that get placed in config.h.in and also
3108 arranged for config.h to be included in the distribution, with the name
3109 config.h.generic, for the benefit of those who have to compile without
3110 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3111
3112 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3113 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3114 makevp.bat for the current PCRE, using the additional files
3115 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3116
3117 6. A Windows user reported a minor discrepancy with test 2, which turned out
3118 to be caused by a trailing space on an input line that had got lost in his
3119 copy. The trailing space was an accident, so I've just removed it.
3120
3121 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3122 that is needed.
3123
3124 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3125 as "const" (a) because they are and (b) because it helps the PHP
3126 maintainers who have recently made a script to detect big data structures
3127 in the php code that should be moved to the .rodata section. I remembered
3128 to update Builducptable as well, so it won't revert if ucptable.h is ever
3129 re-created.
3130
3131 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3132 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3133 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3134 support is not required. This saves 1.5-2K of code, which is important in
3135 some applications.
3136
3137 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3138 so as not to refer to the tables, even though these functions will never be
3139 called when UTF-8 support is disabled. Otherwise there are problems with a
3140 shared library.
3141
3142 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3143
3144 (a) It was defining its arguments as char * instead of void *.
3145
3146 (b) It was assuming that all moves were upwards in memory; this was true
3147 a long time ago when I wrote it, but is no longer the case.
3148
3149 The emulated memove() is provided for those environments that have neither
3150 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3151 is clearly not the case, as these two bugs were recently reported.
3152
3153 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3154 and Detrail to create the HTML documentation, the .txt form of the man
3155 pages, and it removes trailing spaces from listed files. It also creates
3156 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3157 case, it wraps all the #defines with #ifndefs. This script should be run
3158 before "make dist".
3159
3160 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3161 with Unicode property support.
3162
3163 (a) For a maximizing quantifier, if the two different cases of the
3164 character were of different lengths in their UTF-8 codings (there are
3165 some cases like this - I found 11), and the matching function had to
3166 back up over a mixture of the two cases, it incorrectly assumed they
3167 were both the same length.
3168
3169 (b) When PCRE was configured to use the heap rather than the stack for
3170 recursion during matching, it was not correctly preserving the data for
3171 the other case of a UTF-8 character when checking ahead for a match
3172 while processing a minimizing repeat. If the check also involved
3173 matching a wide character, but failed, corruption could cause an
3174 erroneous result when trying to check for a repeat of the original
3175 character.
3176
3177 13. Some tidying changes to the testing mechanism:
3178
3179 (a) The RunTest script now detects the internal link size and whether there
3180 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3181 values substituted by "configure". (The RunGrepTest script already did
3182 this for UTF-8.) The configure.ac script no longer substitutes the
3183 relevant variables.
3184
3185 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3186 with length and offset values. This means that the output is different
3187 for different internal link sizes. Test 2 is skipped for link sizes
3188 other than 2 because of this, bypassing the problem. Unfortunately,
3189 there was also a test in test 3 (the locale tests) that used /B and
3190 failed for link sizes other than 2. Rather than cut the whole test out,
3191 I have added a new /Z option to pcretest that replaces the length and
3192 offset values with spaces. This is now used to make test 3 independent
3193 of link size. (Test 2 will be tidied up later.)
3194
3195 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3196 segmentation fault instead of returning the appropriate error message.
3197
3198 15. In multiline mode when the newline sequence was set to "any", the pattern
3199 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3200 This doesn't seem right; it now treats the CRLF combination as the line
3201 ending, and so does not match in that case. It's only a pattern such as ^$
3202 that would hit this one: something like ^ABC$ would have failed after \r
3203 and then tried again after \r\n.
3204
3205 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3206 in an attempt to make files that differ only in their line terminators
3207 compare equal. This works on Linux.
3208
3209 17. Under certain error circumstances pcregrep might try to free random memory
3210 as it exited. This is now fixed, thanks to valgrind.
3211
3212 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3213 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3214 was because its rules for how to advance for /g after matching an empty
3215 string at the end of a line did not allow for this case. They now check for
3216 it specially.
3217
3218 20. pcretest is supposed to handle patterns and data of any length, by
3219 extending its buffers when necessary. It was getting this wrong when the
3220 buffer for a data line had to be extended.
3221
3222 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3223 CRLF as a newline sequence.
3224
3225 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3226 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3227 I have nevertheless tidied it up.
3228
3229 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3230
3231 24. Added a man page for pcre-config.
3232
3233
3234 Version 7.0 19-Dec-06
3235 ---------------------
3236
3237 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3238 moving to gcc 4.1.1.
3239
3240 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3241 sys/time.h, which is documented as needed for this function. It doesn't
3242 seem to matter on Linux, but it showed up on some releases of OS X.
3243
3244 3. It seems that there are systems where bytes whose values are greater than
3245 127 match isprint() in the "C" locale. The "C" locale should be the
3246 default when a C program starts up. In most systems, only ASCII printing
3247 characters match isprint(). This difference caused the output from pcretest
3248 to vary, making some of the tests fail. I have changed pcretest so that:
3249
3250 (a) When it is outputting text in the compiled version of a pattern, bytes
3251 other than 32-126 are always shown as hex escapes.
3252
3253 (b) When it is outputting text that is a matched part of a subject string,
3254 it does the same, unless a different locale has been set for the match
3255 (using the /L modifier). In this case, it uses isprint() to decide.
3256
3257 4. Fixed a major bug that caused incorrect computation of the amount of memory
3258 required for a compiled pattern when options that changed within the
3259 pattern affected the logic of the preliminary scan that determines the
3260 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3261 that the computed length was too small. The symptoms of this bug were
3262 either the PCRE error "internal error: code overflow" from pcre_compile(),
3263 or a glibc crash with a message such as "pcretest: free(): invalid next
3264 size (fast)". Examples of patterns that provoked this bug (shown in
3265 pcretest format) are:
3266
3267 /(?-x: )/x
3268 /(?x)(?-x: \s*#\s*)/
3269 /((?i)[\x{c0}])/8
3270 /(?i:[\x{c0}])/8
3271
3272 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3273 is now done differently.
3274
3275 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3276 wrapper classes; (b) implement a new function in the C++ scanner that is
3277 more efficient than the old way of doing things because it avoids levels of
3278 recursion in the regex matching; (c) add a paragraph to the documentation
3279 for the FullMatch() function.
3280
3281 6. The escape sequence \n was being treated as whatever was defined as
3282 "newline". Not only was this contrary to the documentation, which states
3283 that \n is character 10 (hex 0A), but it also went horribly wrong when
3284 "newline" was defined as CRLF. This has been fixed.
3285
3286 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3287 was being set to -1 for the "end of line" case (supposedly a value that no
3288 character can have). Though this value is never used (the check for end of
3289 line is "zero bytes in current character"), it caused compiler complaints.
3290 I've changed it to 0xffffffff.
3291
3292 8. In pcre_version.c, the version string was being built by a sequence of
3293 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3294 string (as it is for production releases) called a macro with an empty
3295 argument. The C standard says the result of this is undefined. The gcc
3296 compiler treats it as an empty string (which was what was wanted) but it is
3297 reported that Visual C gives an error. The source has been hacked around to
3298 avoid this problem.
3299
3300 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3301 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3302 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3303 of them did).
3304
3305 10. Originally, pcretest opened its input and output without "b"; then I was
3306 told that "b" was needed in some environments, so it was added for release
3307 5.0 to both the input and output. (It makes no difference on Unix-like
3308 systems.) Later I was told that it is wrong for the input on Windows. I've
3309 now abstracted the modes into two macros, to make it easier to fiddle with
3310 them, and removed "b" from the input mode under Windows.
3311
3312 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3313
3314 12. Added -help and --help to pcretest as an official way of being reminded
3315 of the options.
3316
3317 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3318 and pcrecpp.cc because they annoy compilers at high warning levels.
3319
3320 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3321
3322 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3323 scripts are not C programs :-) and which was not noticed because it works
3324 on Linux.
3325
3326 16. pcretest is supposed to handle any length of pattern and data line (as one
3327 line or as a continued sequence of lines) by extending its input buffer if
3328 necessary. This feature was broken for very long pattern lines, leading to
3329 a string of junk being passed to pcre_compile() if the pattern was longer
3330 than about 50K.
3331
3332 17. I have done a major re-factoring of the way pcre_compile() computes the
3333 amount of memory needed for a compiled pattern. Previously, there was code
3334 that made a preliminary scan of the pattern in order to do this. That was
3335 OK when PCRE was new, but as the facilities have expanded, it has become
3336 harder and harder to keep it in step with the real compile phase, and there
3337 have been a number of bugs (see for example, 4 above). I have now found a
3338 cunning way of running the real compile function in a "fake" mode that
3339 enables it to compute how much memory it would need, while actually only
3340 ever using a few hundred bytes of working memory and without too many
3341 tests of the mode. This should make future maintenance and development
3342 easier. A side effect of this work is that the limit of 200 on the nesting
3343 depth of parentheses has been removed (though this was never a serious
3344 limitation, I suspect). However, there is a downside: pcre_compile() now
3345 runs more slowly than before (30% or more, depending on the pattern). I
3346 hope this isn't a big issue. There is no effect on runtime performance.
3347
3348 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3349 newline (only possible for the last line of a file) and it was a
3350 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3351
3352 19. Added additional timing features to pcretest. (1) The -tm option now times
3353 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3354 separate command line item, by a number that specifies the number of
3355 repeats to use when timing. The default is 50000; this gives better
3356 precision, but takes uncomfortably long for very large patterns.
3357
3358 20. Extended pcre_study() to be more clever in cases where a branch of a
3359 subpattern has no definite first character. For example, (a*|b*)[cd] would
3360 previously give no result from pcre_study(). Now it recognizes that the
3361 first character must be a, b, c, or d.
3362
3363 21. There was an incorrect error "recursive call could loop indefinitely" if
3364 a subpattern (or the entire pattern) that was being tested for matching an
3365 empty string contained only one non-empty item after a nested subpattern.
3366 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3367 incorrectly, because the \d was being skipped in the check.
3368
3369 22. The pcretest program now has a new pattern option /B and a command line
3370 option -b, which is equivalent to adding /B to every pattern. This causes
3371 it to show the compiled bytecode, without the additional information that
3372 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3373 is the same as /B/I).
3374
3375 23. A new optimization is now able automatically to treat some sequences such
3376 as a*b as a*+b. More specifically, if something simple (such as a character
3377 or a simple class like \d) has an unlimited quantifier, and is followed by
3378 something that cannot possibly match the quantified thing, the quantifier
3379 is automatically "possessified".
3380
3381 24. A recursive reference to a subpattern whose number was greater than 39
3382 went wrong under certain circumstances in UTF-8 mode. This bug could also
3383 have affected the operation of pcre_study().
3384
3385 25. Realized that a little bit of performance could be had by replacing
3386 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3387
3388 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3389
3390 27. Possessive quantifiers such as a++ were previously implemented by turning
3391 them into atomic groups such as ($>a+). Now they have their own opcodes,
3392 which improves performance. This includes the automatically created ones
3393 from 23 above.
3394
3395 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3396 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3397 the first matched character to be a colon. This applied both to named and
3398 numbered groups.
3399
3400 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3401
3402 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3403 building PCRE on Windows easier, so I have included it in the distribution.
3404
3405 31. There is now a check in pcretest against a ridiculously large number being
3406 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3407 loop, the loop is abandoned.
3408
3409 32. Forward references to subpatterns in conditions such as (?(2)...) where
3410 subpattern 2 is defined later cause pcre_compile() to search forwards in
3411 the pattern for the relevant set of parentheses. This search went wrong
3412 when there were unescaped parentheses in a character class, parentheses
3413 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3414
3415 33. "Subroutine" calls and backreferences were previously restricted to
3416 referencing subpatterns earlier in the regex. This restriction has now
3417 been removed.
3418
3419 34. Added a number of extra features that are going to be in Perl 5.10. On the
3420 whole, these are just syntactic alternatives for features that PCRE had
3421 previously implemented using the Python syntax or my own invention. The
3422 other formats are all retained for compatibility.
3423
3424 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3425 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3426 also .NET compatible.
3427
3428 (b) A recursion or subroutine call to a named group can now be defined as
3429 (?&name) as well as (?P>name).
3430
3431 (c) A backreference to a named group can now be defined as \k<name> or
3432 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3433 5.10, are also .NET compatible.
3434
3435 (d) A conditional reference to a named group can now use the syntax
3436 (?(<name>) or (?('name') as well as (?(name).
3437
3438 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3439 groups (named and numbered) that are never evaluated inline, but can be
3440 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3441 is always false. There may be only one alternative in such a group.
3442
3443 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3444 as the simple (?(R). The condition is true only if the most recent
3445 recursion is that of the given number or name. It does not search out
3446 through the entire recursion stack.
3447
3448 (g) The escape \gN or \g{N} has been added, where N is a positive or
3449 negative number, specifying an absolute or relative reference.
3450
3451 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3452 some "unreachable code" warnings.
3453
3454 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3455 things, this adds five new scripts.
3456
3457 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3458 There were also incompatibilities regarding the handling of \Q..\E inside
3459 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3460 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3461
3462 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3463 matches an empty string, and forcibly breaks the loop. There were bugs in
3464 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3465 against aaaa the result was just "a" rather than "aaaa", for example. Two
3466 separate and independent bugs (that affected different cases) have been
3467 fixed.
3468
3469 39. Refactored the code to abolish the use of different opcodes for small
3470 capturing bracket numbers. This is a tidy that I avoided doing when I
3471 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3472 The new approach is not only tidier, it makes it possible to reduce the
3473 memory needed to fix the previous bug (38).
3474
3475 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3476 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3477 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3478 mode.
3479
3480 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3481 report.
3482
3483 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3484 copy construction and assignment in the C++ wrapper.
3485
3486 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3487 couple of bugs that could have given wrong results in the "--newline=crlf"
3488 case.
3489
3490 44. Added a number of casts and did some reorganization of signed/unsigned int
3491 variables following suggestions from Dair Grant. Also renamed the variable
3492 "this" as "item" because it is a C++ keyword.
3493
3494 45. Arranged for dftables to add
3495
3496 #include "pcre_internal.h"
3497
3498 to pcre_chartables.c because without it, gcc 4.x may remove the array
3499 definition from the final binary if PCRE is built into a static library and
3500 dead code stripping is activated.
3501
3502 46. For an unanchored pattern, if a match attempt fails at the start of a
3503 newline sequence, and the newline setting is CRLF or ANY, and the next two
3504 characters are CRLF, advance by two characters instead of one.
3505
3506
3507 Version 6.7 04-Jul-06
3508 ---------------------
3509
3510 1. In order to handle tests when input lines are enormously long, pcretest has
3511 been re-factored so that it automatically extends its buffers when
3512 necessary. The code is crude, but this _is_ just a test program. The
3513 default size has been increased from 32K to 50K.
3514
3515 2. The code in pcre_study() was using the value of the re argument before
3516 testing it for NULL. (Of course, in any sensible call of the function, it
3517 won't be NULL.)
3518
3519 3. The memmove() emulation function in pcre_internal.h, which is used on
3520 systems that lack both memmove() and bcopy() - that is, hardly ever -
3521 was missing a "static" storage class specifier.
3522
3523 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3524 containing an extended class (one that cannot be represented by a bitmap
3525 because it contains high-valued characters or Unicode property items, e.g.
3526 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3527 pattern, but PCRE should not loop if you do not (it no longer does).
3528 [Detail: two cases were found: (a) a repeated subpattern containing an
3529 extended class; (b) a recursive reference to a subpattern that followed a
3530 previous extended class. It wasn't skipping over the extended class
3531 correctly when UTF-8 mode was not set.]
3532
3533 5. A negated single-character class was not being recognized as fixed-length
3534 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3535 compile error "lookbehind assertion is not fixed length".
3536
3537 6. The RunPerlTest auxiliary script was showing an unexpected difference
3538 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3539 write a Perl script that can interpret lines of an input file either as
3540 byte characters or as UTF-8, which is what "perltest" was being required to
3541 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3542 can't do is switch easily at run time between having the "use utf8;" pragma
3543 or not. In the end, I fudged it by using the RunPerlTest script to insert
3544 "use utf8;" explicitly for the UTF-8 tests.
3545
3546 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3547 the end of the subject string, contrary to the documentation and to what
3548 Perl does. This was true of both matching functions. Now it matches only at
3549 the start of the subject and immediately after *internal* newlines.
3550
3551 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3552 a pointer to an int instead of a pointer to an unsigned long int. This
3553 caused problems on 64-bit systems.
3554
3555 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3556 instance of the 'standard' template library not being so standard".
3557
3558 10. There was no check on the number of named subpatterns nor the maximum
3559 length of a subpattern name. The product of these values is used to compute
3560 the size of the memory block for a compiled pattern. By supplying a very
3561 long subpattern name and a large number of named subpatterns, the size
3562 computation could be caused to overflow. This is now prevented by limiting
3563 the length of names to 32 characters, and the number of named subpatterns
3564 to 10,000.
3565
3566 11. Subpatterns that are repeated with specific counts have to be replicated in
3567 the compiled pattern. The size of memory for this was computed from the
3568 length of the subpattern and the repeat count. The latter is limited to
3569 65535, but there was no limit on the former, meaning that integer overflow
3570 could in principle occur. The compiled length of a repeated subpattern is
3571 now limited to 30,000 bytes in order to prevent this.
3572
3573 12. Added the optional facility to have named substrings with the same name.
3574
3575 13. Added the ability to use a named substring as a condition, using the
3576 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3577 are numbers (not recommended). Forward references are permitted.
3578
3579 14. Added forward references in named backreferences (if you see what I mean).
3580
3581 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3582 pattern could run off the end of the subject. For example, the pattern
3583 "(?s)(.{1,5})"8 did this with the subject "ab".
3584
3585 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3586 PCRE_CASELESS was set when matching characters that were quantified with ?
3587 or *.
3588
3589 17. A character class other than a single negated character that had a minimum
3590 but no maximum quantifier - for example [ab]{6,} - was not handled
3591 correctly by pce_dfa_exec(). It would match only one character.
3592
3593 18. A valid (though odd) pattern that looked like a POSIX character
3594 class but used an invalid character after [ (for example [[,abc,]]) caused
3595 pcre_compile() to give the error "Failed: internal error: code overflow" or
3596 in some cases to crash with a glibc free() error. This could even happen if
3597 the pattern terminated after [[ but there just happened to be a sequence of
3598 letters, a binary zero, and a closing ] in the memory that followed.
3599
3600 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3601 over the years. Originally (before any Unicode support), just the bottom 8
3602 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3603 output from "man perlunicode" includes this:
3604
3605 The regular expression compiler produces polymorphic opcodes. That
3606 is, the pattern adapts to the data and automatically switches to
3607 the Unicode character scheme when presented with Unicode data--or
3608 instead uses a traditional byte scheme when presented with byte
3609 data.
3610
3611 Sadly, a wide octal escape does not cause a switch, and in a string with
3612 no other multibyte characters, these octal escapes are treated as before.
3613 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3614 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3615 Unicode string.
3616
3617 I have not perpetrated such confusion in PCRE. Up till now, it took just
3618 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3619 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3620 translate to the appropriate multibyte character.
3621
3622 29. Applied some refactoring to reduce the number of warnings from Microsoft
3623 and Borland compilers. This has included removing the fudge introduced
3624 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3625 a warning about an unused variable.
3626
3627 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3628 characters since release 4.0, because Perl (from release 5.004) does not.
3629 [Or at least, is documented not to: some releases seem to be in conflict
3630 with the documentation.] However, when a pattern was studied with
3631 pcre_study() and all its branches started with \s, PCRE still included VT
3632 as a possible starting character. Of course, this did no harm; it just
3633 caused an unnecessary match attempt.
3634
3635 22. Removed a now-redundant internal flag bit that recorded the fact that case
3636 dependency changed within the pattern. This was once needed for "required
3637 byte" processing, but is no longer used. This recovers a now-scarce options
3638 bit. Also moved the least significant internal flag bit to the most-
3639 significant bit of the word, which was not previously used (hangover from
3640 the days when it was an int rather than a uint) to free up another bit for
3641 the future.
3642
3643 23. Added support for CRLF line endings as well as CR and LF. As well as the
3644 default being selectable at build time, it can now be changed at runtime
3645 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3646 specify that it is scanning data with non-default line endings.
3647
3648 24. Changed the definition of CXXLINK to make it agree with the definition of
3649 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3650
3651 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3652 recursions. This makes a big different to stack usage for some patterns.
3653
3654 26. If a subpattern containing a named recursion or subroutine reference such
3655 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3656 the space required for the compiled pattern went wrong and gave too small a
3657 value. Depending on the environment, this could lead to "Failed: internal
3658 error: code overflow at offset 49" or "glibc detected double free or
3659 corruption" errors.
3660
3661 27. Applied patches from Google (a) to support the new newline modes and (b) to
3662 advance over multibyte UTF-8 characters in GlobalReplace.
3663
3664 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3665 difference for some implementation of PCRE in some Windows version.
3666
3667 29. Added some extra testing facilities to pcretest:
3668
3669 \q<number> in a data line sets the "match limit" value
3670 \Q<number> in a data line sets the "match recursion limt" value
3671 -S <number> sets the stack size, where <number> is in megabytes
3672
3673 The -S option isn't available for Windows.
3674
3675
3676 Version 6.6 06-Feb-06
3677 ---------------------
3678
3679 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3680 in pcreposix.h. I have copied the definition from pcre.h.
3681
3682 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3683 because pcre.h is no longer a built file.
3684
3685 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3686 not normally included in the compiled code.
3687
3688
3689 Version 6.5 01-Feb-06
3690 ---------------------
3691
3692 1. When using the partial match feature with pcre_dfa_exec(), it was not
3693 anchoring the second and subsequent partial matches at the new starting
3694 point. This could lead to incorrect results. For example, with the pattern
3695 /1234/, partially matching against "123" and then "a4" gave a match.
3696
3697 2. Changes to pcregrep:
3698
3699 (a) All non-match returns from pcre_exec() were being treated as failures
3700 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3701 error message is output. Some extra information is given for the
3702 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3703 probably the only errors that are likely to be caused by users (by
3704 specifying a regex that has nested indefinite repeats, for instance).
3705 If there are more than 20 of these errors, pcregrep is abandoned.
3706
3707 (b) A binary zero was treated as data while matching, but terminated the
3708 output line if it was written out. This has been fixed: binary zeroes
3709 are now no different to any other data bytes.
3710
3711 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3712 used to set a locale for matching. The --locale=xxxx long option has
3713 been added (no short equivalent) to specify a locale explicitly on the
3714 pcregrep command, overriding the environment variables.
3715
3716 (d) When -B was used with -n, some line numbers in the output were one less
3717 than they should have been.
3718
3719 (e) Added the -o (--only-matching) option.
3720
3721 (f) If -A or -C was used with -c (count only), some lines of context were
3722 accidentally printed for the final match.
3723
3724 (g) Added the -H (--with-filename) option.
3725
3726 (h) The combination of options -rh failed to suppress file names for files
3727 that were found from directory arguments.
3728
3729 (i) Added the -D (--devices) and -d (--directories) options.
3730
3731 (j) Added the -F (--fixed-strings) option.
3732
3733 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3734
3735 (l) Added the --colo(u)r option.
3736
3737 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3738 is not present by default.
3739
3740 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3741 items such as (?R) or (?1), when the recursion could match a number of
3742 alternatives. If it matched one of the alternatives, but subsequently,
3743 outside the recursion, there was a failure, the code tried to back up into
3744 the recursion. However, because of the way PCRE is implemented, this is not
3745 possible, and the result was an incorrect result from the match.
3746
3747 In order to prevent this happening, the specification of recursion has
3748 been changed so that all such subpatterns are automatically treated as
3749 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3750
3751 4. I had overlooked the fact that, in some locales, there are characters for
3752 which isalpha() is true but neither isupper() nor islower() are true. In
3753 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3754 and ordfeminine) are like this. This affected the treatment of \w and \W
3755 when they appeared in character classes, but not when they appeared outside
3756 a character class. The bit map for "word" characters is now created
3757 separately from the results of isalnum() instead of just taking it from the
3758 upper, lower, and digit maps. (Plus the underscore character, of course.)
3759
3760 5. The above bug also affected the handling of POSIX character classes such as
3761 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3762 permanent tables. Instead, the bit maps for such a class were previously
3763 created as the appropriate unions of the upper, lower, and digit bitmaps.
3764 Now they are created by subtraction from the [[:word:]] class, which has
3765 its own bitmap.
3766
3767 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3768 It is created by subtracting the vertical space characters (\x09, \x0a,
3769 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3770 subtraction was done in the overall bitmap for a character class, meaning
3771 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3772 be recognized. This bug has been fixed.
3773
3774 7. Patches from the folks at Google:
3775
3776 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3777 real life, but is still worth protecting against".
3778
3779 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3780 regular expressions".
3781
3782 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3783 have it.
3784
3785 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3786 "configure" and the latter not, in order to fix a problem somebody had
3787 with compiling the Arg class on HP-UX.
3788
3789 (e) Improve the error-handling of the C++ wrapper a little bit.
3790
3791 (f) New tests for checking recursion limiting.
3792
3793 8. The pcre_memmove() function, which is used only if the environment does not
3794 have a standard memmove() function (and is therefore rarely compiled),
3795 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3796 returning a result (though PCRE never actually uses the result).
3797
3798 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3799 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3800 returned instead of calling malloc() with an overflowing number that would
3801 most likely cause subsequent chaos.
3802
3803 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3804
3805 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3806 with this option is matched, the nmatch and pmatch options of regexec() are
3807 ignored.
3808
3809 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3810 provided in case anyone wants to the the POSIX interface with UTF-8
3811 strings.
3812
3813 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3814 C++ linking (needed for some HP-UX environments).
3815
3816 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3817 (unused parameter) and in the pcre_printint() function (omitted "default"
3818 switch label when the default is to do nothing).
3819
3820 15. Added some code to make it possible, when PCRE is compiled as a C++
3821 library, to replace subject pointers for pcre_exec() with a smart pointer
3822 class, thus making it possible to process discontinuous strings.
3823
3824 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3825 much the same function. They were added by different people who were trying
3826 to make PCRE easy to compile on non-Unix systems. It has been suggested
3827 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3828 for compiling on Windows systems. I have therefore replaced it with
3829 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3830 defaults to "extern" for C or "extern C" for C++, which works fine on
3831 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3832 SCOPE with something explicit in config.h. In addition:
3833
3834 (a) pcreposix.h still had just "extern" instead of either of these macros;
3835 I have replaced it with PCRE_DATA_SCOPE.
3836
3837 (b) Functions such as _pcre_xclass(), which are internal to the library,
3838 but external in the C sense, all had PCRE_EXPORT in their definitions.
3839 This is apparently wrong for the Windows case, so I have removed it.
3840 (It makes no difference on Unix-like systems.)
3841
3842 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3843 of recursive calls to match(). This is different to MATCH_LIMIT because
3844 that limits the total number of calls to match(), not all of which increase
3845 the depth of recursion. Limiting the recursion depth limits the amount of
3846 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3847 when PCRE is compiled, and changed at run time. A patch from Google adds
3848 this functionality to the C++ interface.
3849
3850 18. Changes to the handling of Unicode character properties:
3851
3852 (a) Updated the table to Unicode 4.1.0.
3853
3854 (b) Recognize characters that are not in the table as "Cn" (undefined).
3855
3856 (c) I revised the way the table is implemented to a much improved format
3857 which includes recognition of ranges. It now supports the ranges that
3858 are defined in UnicodeData.txt, and it also amalgamates other
3859 characters into ranges. This has reduced the number of entries in the
3860 table from around 16,000 to around 3,000, thus reducing its size
3861 considerably. I realized I did not need to use a tree structure after
3862 all - a binary chop search is just as efficient. Having reduced the
3863 number of entries, I extended their size from 6 bytes to 8 bytes to
3864 allow for more data.
3865
3866 (d) Added support for Unicode script names via properties such as \p{Han}.
3867
3868 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3869 matching that character.
3870
3871 20. When matching a repeated Unicode property with a minimum greater than zero,
3872 (for example \pL{2,}), PCRE could look past the end of the subject if it
3873 reached it while seeking the minimum number of characters. This could
3874 happen only if some of the characters were more than one byte long, because
3875 there is a check for at least the minimum number of bytes.
3876
3877 21. Refactored the implementation of \p and \P so as to be more general, to
3878 allow for more different types of property in future. This has changed the
3879 compiled form incompatibly. Anybody with saved compiled patterns that use
3880 \p or \P will have to recompile them.
3881
3882 22. Added "Any" and "L&" to the supported property types.
3883
3884 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3885 but give a compile time error if the value is greater than 0xff.
3886
3887 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3888 accidentally not being installed or uninstalled.
3889
3890 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3891 made were to insert the current release number. This seemed silly, because
3892 it made things harder for people building PCRE on systems that don't run
3893 "configure". I have turned pcre.h into a distributed file, no longer built
3894 by "configure", with the version identification directly included. There is
3895 no longer a pcre.h.in file.
3896
3897 However, this change necessitated a change to the pcre-config script as
3898 well. It is built from pcre-config.in, and one of the substitutions was the
3899 release number. I have updated configure.ac so that ./configure now finds
3900 the release number by grepping pcre.h.
3901
3902 26. Added the ability to run the tests under valgrind.
3903
3904
3905 Version 6.4 05-Sep-05
3906 ---------------------
3907
3908 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3909 "--" to be printed when multiple files were scanned, even when none of the
3910 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3911 consider it to be a bug, and have restored the previous behaviour.
3912
3913 2. A couple of code tidies to get rid of compiler warnings.
3914
3915 3. The pcretest program used to cheat by referring to symbols in the library
3916 whose names begin with _pcre_. These are internal symbols that are not
3917 really supposed to be visible externally, and in some environments it is
3918 possible to suppress them. The cheating is now confined to including
3919 certain files from the library's source, which is a bit cleaner.
3920
3921 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3922 file's purpose clearer.
3923
3924 5. Reorganized pcre_ucp_findchar().
3925
3926
3927 Version 6.3 15-Aug-05
3928 ---------------------
3929
3930 1. The file libpcre.pc.in did not have general read permission in the tarball.
3931
3932 2. There were some problems when building without C++ support:
3933
3934 (a) If C++ support was not built, "make install" and "make test" still
3935 tried to test it.
3936
3937 (b) There were problems when the value of CXX was explicitly set. Some
3938 changes have been made to try to fix these, and ...
3939
3940 (c) --disable-cpp can now be used to explicitly disable C++ support.
3941
3942 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3943 backslash in a target when C++ was disabled. This confuses some
3944 versions of "make", apparently. Using an intermediate variable solves
3945 this. (Same for CPP_LOBJ.)
3946
3947 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3948 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3949 necessary on certain architectures.
3950
3951 4. Added a setting of -export-symbols-regex to the link command to remove
3952 those symbols that are exported in the C sense, but actually are local
3953 within the library, and not documented. Their names all begin with
3954 "_pcre_". This is not a perfect job, because (a) we have to except some
3955 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3956 available (and never for static libraries). I have made a note to try to
3957 find a way round (a) in the future.
3958
3959
3960 Version 6.2 01-Aug-05
3961 ---------------------
3962
3963 1. There was no test for integer overflow of quantifier values. A construction
3964 such as {1111111111111111} would give undefined results. What is worse, if
3965 a minimum quantifier for a parenthesized subpattern overflowed and became
3966 negative, the calculation of the memory size went wrong. This could have
3967 led to memory overwriting.
3968
3969 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3970
3971 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3972 operating environments where this matters.
3973
3974 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3975 PCRE options from within the C++ wrapper.
3976
3977 5. Named capturing subpatterns were not being correctly counted when a pattern
3978 was compiled. This caused two problems: (a) If there were more than 100
3979 such subpatterns, the calculation of the memory needed for the whole
3980 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3981 back references of the form \12, where the number was greater than 9, were
3982 not recognized as back references, even though there were sufficient
3983 previous subpatterns.
3984
3985 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3986 versions of gcc, e.g. 2.95.4.
3987
3988
3989 Version 6.1 21-Jun-05
3990 ---------------------
3991
3992 1. There was one reference to the variable "posix" in pcretest.c that was not
3993 surrounded by "#if !defined NOPOSIX".
3994
3995 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3996 the cross-check on the old pcre_info() function, for the benefit of the
3997 cut-down version of PCRE that is currently imported into Exim.
3998
3999 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4000 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4001 patterns that start (?i) but I don't think that matters. The use of (?i) is
4002 just an example; this all applies to the other options as well.
4003
4004 4. Since libtool seems to echo the compile commands it is issuing, the output
4005 from "make" can be reduced a bit by putting "@" in front of each libtool
4006 compile command.
4007
4008 5. Patch from the folks at Google for configure.in to be a bit more thorough
4009 in checking for a suitable C++ installation before trying to compile the
4010 C++ stuff. This should fix a reported problem when a compiler was present,
4011 but no suitable headers.
4012
4013 6. The man pages all had just "PCRE" as their title. I have changed them to
4014 be the relevant file name. I have also arranged that these names are
4015 retained in the file doc/pcre.txt, which is a concatenation in text format
4016 of all the man pages except the little individual ones for each function.
4017
4018 7. The NON-UNIX-USE file had not been updated for the different set of source
4019 files that come with release 6. I also added a few comments about the C++
4020 wrapper.
4021
4022
4023 Version 6.0 07-Jun-05
4024 ---------------------
4025
4026 1. Some minor internal re-organization to help with my DFA experiments.
4027
4028 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4029 didn't matter for the library itself when fully configured, but did matter
4030 when compiling without UCP support, or within Exim, where the ucp files are
4031 not imported.
4032
4033 3. Refactoring of the library code to split up the various functions into
4034 different source modules. The addition of the new DFA matching code (see
4035 below) to a single monolithic source would have made it really too
4036 unwieldy, quite apart from causing all the code to be include in a
4037 statically linked application, when only some functions are used. This is
4038 relevant even without the DFA addition now that patterns can be compiled in
4039 one application and matched in another.
4040
4041 The downside of splitting up is that there have to be some external
4042 functions and data tables that are used internally in different modules of
4043 the library but which are not part of the API. These have all had their
4044 names changed to start with "_pcre_" so that they are unlikely to clash
4045 with other external names.
4046
4047 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4048 a different (DFA) algorithm. Although it is slower than the original
4049 function, it does have some advantages for certain types of matching
4050 problem.
4051
4052 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4053 including restarting after a partial match.
4054
4055 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4056 defined when compiling for Windows was sent to me. I have put it into the
4057 code, though I have no means of testing or verifying it.
4058
4059 7. Added the pcre_refcount() auxiliary function.
4060
4061 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4062 match before or at the first newline in the subject string. In pcretest,
4063 the /f option on a pattern can be used to set this.
4064
4065 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4066 would behave wrongly. This has been present in PCRE since release 4.0.
4067
4068 10. A number of changes to the pcregrep command:
4069
4070 (a) Refactored how -x works; insert ^(...)$ instead of setting
4071 PCRE_ANCHORED and checking the length, in preparation for adding
4072 something similar for -w.
4073
4074 (b) Added the -w (match as a word) option.
4075
4076 (c) Refactored the way lines are read and buffered so as to have more
4077 than one at a time available.
4078
4079 (d) Implemented a pcregrep test script.
4080
4081 (e) Added the -M (multiline match) option. This allows patterns to match
4082 over several lines of the subject. The buffering ensures that at least
4083 8K, or the rest of the document (whichever is the shorter) is available
4084 for matching (and similarly the previous 8K for lookbehind assertions).
4085
4086 (f) Changed the --help output so that it now says
4087
4088 -w, --word-regex(p)
4089
4090 instead of two lines, one with "regex" and the other with "regexp"
4091 because that confused at least one person since the short forms are the
4092 same. (This required a bit of code, as the output is generated
4093 automatically from a table. It wasn't just a text change.)
4094
4095 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4096 option but starts with a hyphen. Could be a pattern or a path name
4097 starting with a hyphen, for instance.
4098
4099 (h) "-" can be given as a file name to represent stdin.
4100
4101 (i) When file names are being printed, "(standard input)" is used for
4102 the standard input, for compatibility with GNU grep. Previously
4103 "<stdin>" was used.
4104
4105 (j) The option --label=xxx can be used to supply a name to be used for
4106 stdin when file names are being printed. There is no short form.
4107
4108 (k) Re-factored the options decoding logic because we are going to add
4109 two more options that take data. Such options can now be given in four
4110 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4111
4112 (l) Added the -A, -B, and -C options for requesting that lines of context
4113 around matches be printed.
4114
4115 (m) Added the -L option to print the names of files that do not contain
4116 any matching lines, that is, the complement of -l.
4117
4118 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4119 continue to scan other files.
4120
4121 (o) The -s option was incorrectly implemented. For compatibility with other
4122 greps, it now suppresses the error message for a non-existent or non-
4123 accessible file (but not the return code). There is a new option called
4124 -q that suppresses the output of matching lines, which was what -s was
4125 previously doing.
4126
4127 (p) Added --include and --exclude options to specify files for inclusion
4128 and exclusion when recursing.
4129
4130 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4131 Hopefully, it now does.
4132
4133 12. Missing cast in pcre_study().
4134
4135 13. Added an "uninstall" target to the makefile.
4136
4137 14. Replaced "extern" in the function prototypes in Makefile.in with
4138 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4139 world, but is set differently for Windows.
4140
4141 15. Added a second compiling function called pcre_compile2(). The only
4142 difference is that it has an extra argument, which is a pointer to an
4143 integer error code. When there is a compile-time failure, this is set
4144 non-zero, in addition to the error test pointer being set to point to an
4145 error message. The new argument may be NULL if no error number is required
4146 (but then you may as well call pcre_compile(), which is now just a
4147 wrapper). This facility is provided because some applications need a
4148 numeric error indication, but it has also enabled me to tidy up the way
4149 compile-time errors are handled in the POSIX wrapper.
4150
4151 16. Added VPATH=.libs to the makefile; this should help when building with one
4152 prefix path and installing with another. (Or so I'm told by someone who
4153 knows more about this stuff than I do.)
4154
4155 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4156 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4157 match everything, including newlines. This is not POSIX-compatible, but
4158 somebody wanted the feature. From pcretest it can be activated by using
4159 both the P and the s flags.
4160
4161 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4162
4163 19. libpcre.pc was being incorrectly installed as executable.
4164
4165 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4166 it now also looks for '\r' so that it will work unmodified on Windows.
4167
4168 21. Added Google's contributed C++ wrapper to the distribution.
4169
4170 22. Added some untidy missing memory free() calls in pcretest, to keep
4171 Electric Fence happy when testing.
4172
4173
4174
4175 Version 5.0 13-Sep-04
4176 ---------------------
4177
4178 1. Internal change: literal characters are no longer packed up into items
4179 containing multiple characters in a single byte-string. Each character
4180 is now matched using a separate opcode. However, there may be more than one
4181 byte in the character in UTF-8 mode.
4182
4183 2. The pcre_callout_block structure has two new fields: pattern_position and
4184 next_item_length. These contain the offset in the pattern to the next match
4185 item, and its length, respectively.
4186
4187 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4188 insertion of callouts before each pattern item. Added the /C option to
4189 pcretest to make use of this.
4190
4191 4. On the advice of a Windows user, the lines
4192
4193 #if defined(_WIN32) || defined(WIN32)
4194 _setmode( _fileno( stdout ), 0x8000 );
4195 #endif /* defined(_WIN32) || defined(WIN32) */
4196
4197 have been added to the source of pcretest. This apparently does useful
4198 magic in relation to line terminators.
4199
4200 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4201 for the benefit of those environments where the "b" makes a difference.
4202
4203 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4204 to know about it. I have put a hack into configure.in that adds in code
4205 to set GCC=yes if CC=icc. This seems to end up at a point in the
4206 generated configure script that is early enough to affect the setting of
4207 compiler options, which is what is needed, but I have no means of testing
4208 whether it really works. (The user who reported this had patched the
4209 generated configure script, which of course I cannot do.)
4210
4211 LATER: After change 22 below (new libtool files), the configure script
4212 seems to know about icc (and also ecc). Therefore, I have commented out
4213 this hack in configure.in.
4214
4215 7. Added support for pkg-config (2 patches were sent in).
4216
4217 8. Negated POSIX character classes that used a combination of internal tables
4218 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4219 [[:^ascii]]. Typically, they would match almost any characters. The other
4220 POSIX classes were not broken in this way.
4221
4222 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4223 to find the match, as PCRE was deluded into thinking that the match had to
4224 start at the start point or following a newline. The same bug applied to
4225 patterns with negative forward assertions or any backward assertions
4226 preceding ".*" at the start, unless the pattern required a fixed first
4227 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4228
4229 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4230 starting at the last subject character, bytes beyond the end of the subject
4231 string were read.
4232
4233 11. Renamed the variable "class" as "classbits" to make life easier for C++
4234 users. (Previously there was a macro definition, but it apparently wasn't
4235 enough.)
4236
4237 12. Added the new field "tables" to the extra data so that tables can be passed
4238 in at exec time, or the internal tables can be re-selected. This allows
4239 a compiled regex to be saved and re-used at a later time by a different
4240 program that might have everything at different addresses.
4241
4242 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4243 -R library as well as a -L library.
4244
4245 14. The debugging options of pcretest (-d on the command line or D on a
4246 pattern) showed incorrect output for anything following an extended class
4247 that contained multibyte characters and which was followed by a quantifier.
4248
4249 15. Added optional support for general category Unicode character properties
4250 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4251 support. It adds about 90K to the size of the library. The meanings of the
4252 inbuilt class escapes such as \d and \s have NOT been changed.
4253
4254 16. Updated pcredemo.c to include calls to free() to release the memory for the
4255 compiled pattern.
4256
4257 17. The generated file chartables.c was being created in the source directory
4258 instead of in the building directory. This caused the build to fail if the
4259 source directory was different from the building directory, and was
4260 read-only.
4261
4262 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4263 file. No doubt somebody will tell me if they don't make sense... Also added
4264 Dan Mooney's comments about building on OpenVMS.
4265
4266 19. Added support for partial matching via the PCRE_PARTIAL option for
4267 pcre_exec() and the \P data escape in pcretest.
4268
4269 20. Extended pcretest with 3 new pattern features:
4270
4271 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4272 write the compiled pattern to the file whose name is "rest-of-line".
4273 This is a straight binary dump of the data, with the saved pointer to
4274 the character tables forced to be NULL. The study data, if any, is
4275 written too. After writing, pcretest reads a new pattern.
4276
4277 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4278 compiled pattern from the given file. There must not be any
4279 occurrences of "<" in the file name (pretty unlikely); if there are,
4280 pcretest will instead treat the initial "<" as a pattern delimiter.
4281 After reading in the pattern, pcretest goes on to read data lines as
4282 usual.
4283
4284 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4285 and 16-bit fields in a compiled pattern, to simulate a pattern that
4286 was compiled on a host of opposite endianness.
4287
4288 21. The pcre-exec() function can now cope with patterns that were compiled on
4289 hosts of opposite endianness, with this restriction:
4290
4291 As for any compiled expression that is saved and used later, the tables
4292 pointer field cannot be preserved; the extra_data field in the arguments
4293 to pcre_exec() should be used to pass in a tables address if a value
4294 other than the default internal tables were used at compile time.
4295
4296 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4297 now diagnosed as an error. Previously, most of the time, a negative number
4298 would have been treated as zero, but if in addition "ovector" was passed as
4299 NULL, a crash could occur.
4300
4301 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4302 new versions from the libtool 1.5 distribution (the last one is a copy of
4303 a file called libtool.m4). This seems to have fixed the need to patch
4304 "configure" to support Darwin 1.3 (which I used to do). However, I still
4305 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4306 workstation).
4307
4308 24. Changed the PCRE licence to be the more standard "BSD" licence.
4309
4310
4311 Version 4.5 01-Dec-03
4312 ---------------------
4313
4314 1. There has been some re-arrangement of the code for the match() function so
4315 that it can be compiled in a version that does not call itself recursively.
4316 Instead, it keeps those local variables that need separate instances for
4317 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4318 needs to "recurse". Keeping track of where control must go is done by means
4319 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4320 hide most of the details from the main code, and operates only if
4321 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4322 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4323 operating.
4324
4325 To make it easier for callers to provide specially tailored get/free
4326 functions for this usage, two new functions, pcre_stack_malloc, and
4327 pcre_stack_free, are used. They are always called in strict stacking order,
4328 and the size of block requested is always the same.
4329
4330 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4331 PCRE has been compiled to use the stack or the heap for recursion. The
4332 -C option of pcretest uses this to show which version is compiled.
4333
4334 A new data escape \S, is added to pcretest; it causes the amounts of store
4335 obtained and freed by both kinds of malloc/free at match time to be added
4336 to the output.
4337
4338 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4339 what's available on my current Linux desktop machine.
4340
4341 3. When matching a UTF-8 string, the test for a valid string at the start has
4342 been extended. If start_offset is not zero, PCRE now checks that it points
4343 to a byte that is the start of a UTF-8 character. If not, it returns
4344 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4345 this is necessary because there may be backward assertions in the pattern.
4346 When matching the same subject several times, it may save resources to use
4347 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4348
4349 4. The code for checking the validity of UTF-8 strings has been tightened so
4350 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4351 containing "overlong sequences".
4352
4353 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4354 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4355 should have been "&", but it just so happened that all the cases this let
4356 through by mistake were picked up later in the function.
4357
4358 6. I had used a variable called "isblank" - this is a C99 function, causing
4359 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4360
4361 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4362 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4363 so the version is shown; (c) stop "make test" echoing "./RunTest".
4364
4365 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4366
4367 9. The prototype for memmove() for systems that don't have it was using
4368 size_t, but the inclusion of the header that defines size_t was later. I've
4369 moved the #includes for the C headers earlier to avoid this.
4370
4371 10. Added some adjustments to the code to make it easier to compiler on certain
4372 special systems:
4373
4374 (a) Some "const" qualifiers were missing.
4375 (b) Added the macro EXPORT before all exported functions; by default this
4376 is defined to be empty.
4377 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4378 that it reads its output file name as an argument instead of writing
4379 to the standard output and assuming this can be redirected.
4380
4381 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4382 class containing characters with values greater than 255, PCRE compilation
4383 went into a loop.
4384
4385 12. A recursive reference to a subpattern that was within another subpattern
4386 that had a minimum quantifier of zero caused PCRE to crash. For example,
4387 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4388 recursion. If the recursively-called subpattern itself had a zero repeat,
4389 that was OK.
4390
4391 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4392 buffer into which it was copied (for escape processing) was still set at
4393 1024, so long lines caused crashes.
4394
4395 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4396 "internal error: code overflow...". This applied to any character class
4397 that was followed by a possessive quantifier.
4398
4399 15. Modified the Makefile to add libpcre.la as a prerequisite for
4400 libpcreposix.la because I was told this is needed for a parallel build to
4401 work.
4402
4403 16. If a pattern that contained .* following optional items at the start was
4404 studied, the wrong optimizing data was generated, leading to matching
4405 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4406 matching string must start with a or b or c. The correct conclusion for
4407 this pattern is that a match can start with any character.
4408
4409
4410 Version 4.4 13-Aug-03
4411 ---------------------
4412
4413 1. In UTF-8 mode, a character class containing characters with values between
4414 127 and 255 was not handled correctly if the compiled pattern was studied.
4415 In fixing this, I have also improved the studying algorithm for such
4416 classes (slightly).
4417
4418 2. Three internal functions had redundant arguments passed to them. Removal
4419 might give a very teeny performance improvement.
4420
4421 3. Documentation bug: the value of the capture_top field in a callout is *one
4422 more than* the number of the hightest numbered captured substring.
4423
4424 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4425 in incorrectly linking with a previously installed version. They now link
4426 explicitly with libpcre.la.
4427
4428 5. configure.in no longer needs to recognize Cygwin specially.
4429
4430 6. A problem in pcre.in for Windows platforms is fixed.
4431
4432 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4433 pcretest, it used to include the size of the study block as part of its
4434 output. Unfortunately, the structure contains a field that has a different
4435 size on different hardware architectures. This meant that the tests that
4436 showed this size failed. As the block is currently always of a fixed size,
4437 this information isn't actually particularly useful in pcretest output, so
4438 I have just removed it.
4439
4440 8. Three pre-processor statements accidentally did not start in column 1.
4441 Sadly, there are *still* compilers around that complain, even though
4442 standard C has not required this for well over a decade. Sigh.
4443
4444 9. In pcretest, the code for checking callouts passed small integers in the
4445 callout_data field, which is a void * field. However, some picky compilers
4446 complained about the casts involved for this on 64-bit systems. Now
4447 pcretest passes the address of the small integer instead, which should get
4448 rid of the warnings.
4449
4450 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4451 both compile and run time, and gives an error if an invalid UTF-8 sequence
4452 is found. There is a option for disabling this check in cases where the
4453 string is known to be correct and/or the maximum performance is wanted.
4454
4455 11. In response to a bug report, I changed one line in Makefile.in from
4456
4457 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4458 to
4459 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4460
4461 to look similar to other lines, but I have no way of telling whether this
4462 is the right thing to do, as I do not use Windows. No doubt I'll get told
4463 if it's wrong...
4464
4465
4466 Version 4.3 21-May-03
4467 ---------------------
4468
4469 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4470 Makefile.
4471
4472 2. Some refactoring to improve the quality of the code:
4473
4474 (i) The utf8_table... variables are now declared "const".
4475
4476 (ii) The code for \cx, which used the "case flipping" table to upper case
4477 lower case letters, now just substracts 32. This is ASCII-specific,
4478 but the whole concept of \cx is ASCII-specific, so it seems
4479 reasonable.
4480
4481 (iii) PCRE was using its character types table to recognize decimal and
4482 hexadecimal digits in the pattern. This is silly, because it handles
4483 only 0-9, a-f, and A-F, but the character types table is locale-
4484 specific, which means strange things might happen. A private
4485 table is now used for this - though it costs 256 bytes, a table is
4486 much faster than multiple explicit tests. Of course, the standard
4487 character types table is still used for matching digits in subject
4488 strings against \d.
4489
4490 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4491 ending in _t are). So I've renamed it as ESC_tee.
4492
4493 3. The first argument for regexec() in the POSIX wrapper should have been
4494 defined as "const".
4495
4496 4. Changed pcretest to use malloc() for its buffers so that they can be
4497 Electric Fenced for debugging.
4498
4499 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4500 to read one or more bytes before the start of the subject string. Often this
4501 had no effect on PCRE's behaviour, but in some circumstances it could
4502 provoke a segmentation fault.
4503
4504 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4505 to try to read one or more bytes before the start of the subject string.
4506
4507 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4508 UTF-8 support could misbehave in various ways if the subject string
4509 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4510 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4511 back over UTF-8 characters.)
4512
4513
4514 Version 4.2 14-Apr-03
4515 ---------------------
4516
4517 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4518
4519 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4520 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4521 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4522 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4523 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4524 and BUILD_EXEEXT
4525 Note: automatic setting of the BUILD variables is not yet working
4526 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4527 compile-time but not at link-time
4528 [LINK]: use for linking executables only
4529 make different versions for Windows and non-Windows
4530 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4531 libraries
4532 [LINK_FOR_BUILD]: new variable
4533 [OBJEXT]: use throughout
4534 [EXEEXT]: use throughout
4535 <winshared>: new target
4536 <wininstall>: new target
4537 <dftables.o>: use native compiler
4538 <dftables>: use native linker
4539 <install>: handle Windows platform correctly
4540 <clean>: ditto
4541 <check>: ditto
4542 copy DLL to top builddir before testing
4543
4544 As part of these changes, -no-undefined was removed again. This was reported
4545 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4546 in any case.
4547
4548 3. Some tidies to get rid of compiler warnings:
4549
4550 . In the match_data structure, match_limit was an unsigned long int, whereas
4551 match_call_count was an int. I've made them both unsigned long ints.
4552
4553 . In pcretest the fact that a const uschar * doesn't automatically cast to
4554 a void * provoked a warning.
4555
4556 . Turning on some more compiler warnings threw up some "shadow" variables
4557 and a few more missing casts.
4558
4559 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4560 option, a class that contained a single character with a value between 128
4561 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4562
4563 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4564 option, a class that contained several characters, but with at least one
4565 whose value was between 128 and 255 caused PCRE to crash.
4566
4567
4568 Version 4.1 12-Mar-03
4569 ---------------------
4570
4571 1. Compiling with gcc -pedantic found a couple of places where casts were
4572 needed, and a string in dftables.c that was longer than standard compilers are
4573 required to support.
4574
4575 2. Compiling with Sun's compiler found a few more places where the code could
4576 be tidied up in order to avoid warnings.
4577
4578 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4579 first of these names is deprecated in the latest Autoconf in favour of the name
4580 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4581 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4582 analogy I have changed it to CFLAGS_FOR_BUILD.
4583
4584 4. Added -no-undefined to the linking command in the Makefile, because this is
4585 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4586 linking step for the pcreposix library.
4587
4588 5. PCRE was failing to diagnose the case of two named groups with the same
4589 name.
4590
4591 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4592 literal character that is needed in the subject for a match, and scans along to
4593 ensure that it is present before embarking on the full matching process. This
4594 saves time in cases of nested unlimited repeats that are never going to match.
4595 Problem: the scan can take a lot of time if the subject is very long (e.g.
4596 megabytes), thus penalizing straightforward matches. It is now done only if the
4597 amount of subject to be scanned is less than 1000 bytes.
4598
4599 7. A lesser problem with the same optimization is that it was recording the
4600 first character of an anchored pattern as "needed", thus provoking a search
4601 right along the subject, even when the first match of the pattern was going to
4602 fail. The "needed" character is now not set for anchored patterns, unless it
4603 follows something in the pattern that is of non-fixed length. Thus, it still
4604 fulfils its original purpose of finding quick non-matches in cases of nested
4605 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4606
4607
4608 Version 4.0 17-Feb-03
4609 ---------------------
4610
4611 1. If a comment in an extended regex that started immediately after a meta-item
4612 extended to the end of string, PCRE compiled incorrect data. This could lead to
4613 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4614
4615 2. Moved to autoconf 2.53 and libtool 1.4.2.
4616
4617 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4618 the special perltest8 script is no longer needed - all the tests can be run
4619 from a single perltest script.
4620
4621 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4622 by \s. It has now been removed in PCRE. This means it isn't recognized as
4623 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4624 class [:space:] *does* include VT, thereby creating a mess.
4625
4626 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4627 space and tab.
4628
4629 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4630 its new features into the main test script, reducing the number of scripts.
4631
4632 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4633 were backward compatible, and made the (?i) apply to the whole pattern, as if
4634 /i were given. Now it behaves more logically, and applies the option setting
4635 only to what follows. PCRE has been changed to follow suit. However, if it
4636 finds options settings right at the start of the pattern, it extracts them into
4637 the global options, as before. Thus, they show up in the info data.
4638
4639 8. Added support for the \Q...\E escape sequence. Characters in between are
4640 treated as literals. This is slightly different from Perl in that $ and @ are
4641 also handled as literals inside the quotes. In Perl, they will cause variable
4642 interpolation. Note the following examples:
4643
4644 Pattern PCRE matches Perl matches
4645
4646 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4647 \Qabc\$xyz\E abc\$xyz abc\$xyz
4648 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4649
4650 For compatibility with Perl, \Q...\E sequences are recognized inside character
4651 classes as well as outside them.
4652
4653 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4654 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4655 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4656 signed/unsigned warnings.
4657
4658 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4659 option for pcretest, so I've replaced it by a simple function that does just
4660 that job.
4661
4662 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4663 "pcregrep -".
4664
4665 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4666 Java package. This provides some syntactic sugar for simple cases of what my
4667 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4668 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4669 item, you can use this simplified notation. Note that only makes sense with
4670 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4671 greediness, whatever the setting of the PCRE_UNGREEDY option.
4672
4673 13. A change of greediness default within a pattern was not taking effect at
4674 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4675 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4676 was abstracted outside.
4677
4678 14. PCRE now supports the \G assertion. It is true when the current matching
4679 position is at the start point of the match. This differs from \A when the
4680 starting offset is non-zero. Used with the /g option of pcretest (or similar
4681 code), it works in the same way as it does for Perl's /g option. If all
4682 alternatives of a regex begin with \G, the expression is anchored to the start
4683 match position, and the "anchored" flag is set in the compiled expression.
4684
4685 15. Some bugs concerning the handling of certain option changes within patterns
4686 have been fixed. These applied to options other than (?ims). For example,
4687 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4688 been the other way round. Some of this was related to change 7 above.
4689
4690 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4691 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4692 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4693 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4694
4695 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4696 mode. Unlike ".", it always matches newline, whatever the setting of
4697 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4698 assertions. Perl allows it, but it doesn't (in general) work because it can't
4699 calculate the length of the lookbehind. At least, that's the case for Perl
4700 5.8.0 - I've been told they are going to document that it doesn't work in
4701 future.
4702
4703 18. Added an error diagnosis for escapes that PCRE does not support: these are
4704 \L, \l, \N, \P, \p, \U, \u, and \X.
4705
4706 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4707 reading past the end of the pattern in cases such as /[abcd/.
4708
4709 20. PCRE was getting more memory than necessary for patterns with classes that
4710 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4711
4712 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4713 compiling PCRE for use with Virtual Pascal.
4714
4715 22. Small fix to the Makefile to make it work properly if the build is done
4716 outside the source tree.
4717
4718 23. Added a new extension: a condition to go with recursion. If a conditional
4719 subpattern starts with (?(R) the "true" branch is used if recursion has
4720 happened, whereas the "false" branch is used only at the top level.
4721
4722 24. When there was a very long string of literal characters (over 255 bytes
4723 without UTF support, over 250 bytes with UTF support), the computation of how
4724 much memory was required could be incorrect, leading to segfaults or other
4725 strange effects.
4726
4727 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4728 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4729 there was a subsequent back reference to those brackets. This meant that, for
4730 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4731 possible to check for precisely this case. All we can do is abandon the
4732 optimization if .* occurs inside capturing brackets when there are any back
4733 references whatsoever. (See below for a better fix that came later.)
4734
4735 26. The handling of the optimization for finding the first character of a
4736 non-anchored pattern, and for finding a character that is required later in the
4737 match were failing in some cases. This didn't break the matching; it just
4738 failed to optimize when it could. The way this is done has been re-implemented.
4739
4740 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4741
4742 28. Added a new feature that provides some of the functionality that Perl
4743 provides with (?{...}). The facility is termed a "callout". The way it is done
4744 in PCRE is for the caller to provide an optional function, by setting
4745 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4746 global variable. By default it is unset, which disables all calling out. To get
4747 the function called, the regex must include (?C) at appropriate points. This
4748 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4749 This provides a means of identifying different callout points. When PCRE
4750 reaches such a point in the regex, if pcre_callout has been set, the external
4751 function is called. It is provided with data in a structure called
4752 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4753 matching continues; if it returns a non-zero value, the match at the current
4754 point fails. However, backtracking will occur if possible. [This was changed
4755 later and other features added - see item 49 below.]
4756
4757 29. pcretest is upgraded to test the callout functionality. It provides a
4758 callout function that displays information. By default, it shows the start of
4759 the match and the current position in the text. There are some new data escapes
4760 to vary what happens:
4761
4762 \C+ in addition, show current contents of captured substrings
4763 \C- do not supply a callout function
4764 \C!n return 1 when callout number n is reached
4765 \C!n!m return 1 when callout number n is reached for the mth time
4766
4767 30. If pcregrep was called with the -l option and just a single file name, it
4768 output "<stdin>" if a match was found, instead of the file name.
4769
4770 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4771 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4772 pcre_exec(). This saves a malloc/free per call. The default value of
4773 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4774 when configuring.
4775
4776 32. The default maximum size of a compiled pattern is 64K. There have been a
4777 few cases of people hitting this limit. The code now uses macros to handle the
4778 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4779 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4780 configuring. Tests 2 and 5 work only with 2-byte links because they output
4781 debugging information about compiled patterns.
4782
4783 33. Internal code re-arrangements:
4784
4785 (a) Moved the debugging function for printing out a compiled regex into
4786 its own source file (printint.c) and used #include to pull it into
4787 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4788 separate copies.
4789
4790 (b) Defined the list of op-code names for debugging as a macro in
4791 internal.h so that it is next to the definition of the opcodes.
4792
4793 (c) Defined a table of op-code lengths for simpler skipping along compiled
4794 code. This is again a macro in internal.h so that it is next to the
4795 definition of the opcodes.
4796
4797 34. Added support for recursive calls to individual subpatterns, along the
4798 lines of Robin Houston's patch (but implemented somewhat differently).
4799
4800 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4801 allow it to read and process whole directories in Win32. This code was
4802 contributed by Lionel Fourquaux; it has not been tested by me.
4803
4804 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4805 used to name a group. Names consist of alphanumerics and underscores, and must
4806 be unique. Back references use the syntax (?P=name) and recursive calls use
4807 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4808 numbers. The function pcre_fullinfo() can be used after compilation to extract
4809 a name/number map. There are three relevant calls:
4810
4811 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4812 PCRE_INFO_NAMECOUNT yields the number of entries
4813 PCRE_INFO_NAMETABLE yields a pointer to the map.
4814
4815 The map is a vector of fixed-size entries. The size of each entry depends on
4816 the length of the longest name used. The first two bytes of each entry are the
4817 group number, most significant byte first. There follows the corresponding
4818 name, zero terminated. The names are in alphabetical order.
4819
4820 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4821 case instead of 255. Making it the same both with and without UTF-8 support
4822 means that the same test output works with both.
4823
4824 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4825 calling malloc() with a zero argument.
4826
4827 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4828 optimization. I've improved things by keeping a bitmap of backreferences with
4829 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4830 fact referenced, the optimization can be applied. It is unlikely that a
4831 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4832 the match to follow \n) will appear inside brackets with a number greater than
4833 31, but if it does, any back reference > 31 suppresses the optimization.
4834
4835 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4836 of disabling numbered capturing parentheses. Any opening parenthesis that is
4837 not followed by ? behaves as if it were followed by ?: but named parentheses
4838 can still be used for capturing (and they will acquire numbers in the usual
4839 way).
4840
4841 41. Redesigned the return codes from the match() function into yes/no/error so
4842 that errors can be passed back from deep inside the nested calls. A malloc
4843 failure while inside a recursive subpattern call now causes the
4844 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4845
4846 42. It is now possible to set a limit on the number of times the match()
4847 function is called in a call to pcre_exec(). This facility makes it possible to
4848 limit the amount of recursion and backtracking, though not in a directly
4849 obvious way, because the match() function is used in a number of different
4850 circumstances. The count starts from zero for each position in the subject
4851 string (for non-anchored patterns). The default limit is, for compatibility, a
4852 large number, namely 10 000 000. You can change this in two ways:
4853
4854 (a) When configuring PCRE before making, you can use --with-match-limit=n
4855 to set a default value for the compiled library.
4856
4857 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4858 a different value is set. See 45 below.
4859
4860 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4861
4862 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4863 of things that can be changed at compile time. The first argument specifies
4864 what is wanted and the second points to where the information is to be placed.
4865 The current list of available information is:
4866
4867 PCRE_CONFIG_UTF8
4868
4869 The output is an integer that is set to one if UTF-8 support is available;
4870 otherwise it is set to zero.
4871
4872 PCRE_CONFIG_NEWLINE
4873
4874 The output is an integer that it set to the value of the code that is used for
4875 newline. It is either LF (10) or CR (13).
4876
4877 PCRE_CONFIG_LINK_SIZE
4878
4879 The output is an integer that contains the number of bytes used for internal
4880 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4881
4882 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4883
4884 The output is an integer that contains the threshold above which the POSIX
4885 interface uses malloc() for output vectors. See item 31 above.
4886
4887 PCRE_CONFIG_MATCH_LIMIT
4888
4889 The output is an unsigned integer that contains the default limit of the number
4890 of match() calls in a pcre_exec() execution. See 42 above.
4891
4892 44. pcretest has been upgraded by the addition of the -C option. This causes it
4893 to extract all the available output from the new pcre_config() function, and to
4894 output it. The program then exits immediately.
4895
4896 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4897 order to support additional features. One way would have been to define
4898 pcre_exec2() (for example) with extra arguments, but this would not have been
4899 extensible, and would also have required all calls to the original function to
4900 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4901 is used for passing in "extra" data from pcre_study().
4902
4903 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4904 contains the following fields:
4905
4906 flags a bitmap indicating which of the following fields are set
4907 study_data opaque data from pcre_study()
4908 match_limit a way of specifying a limit on match() calls for a specific
4909 call to pcre_exec()
4910 callout_data data for callouts (see 49 below)
4911
4912 The flag bits are also defined in pcre.h, and are
4913
4914 PCRE_EXTRA_STUDY_DATA
4915 PCRE_EXTRA_MATCH_LIMIT
4916 PCRE_EXTRA_CALLOUT_DATA
4917
4918 The pcre_study() function now returns one of these new pcre_extra blocks, with
4919 the actual study data pointed to by the study_data field, and the
4920 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4921 before. That is, this change is entirely upwards-compatible and requires no
4922 change to existing code.
4923
4924 If you want to pass in additional data to pcre_exec(), you can either place it
4925 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4926 block.
4927
4928 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4929 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4930 times with different match limits, until it finds the minimum value needed for
4931 pcre_exec() to complete. The value is then output. This can be instructive; for
4932 most simple matches the number is quite small, but for pathological cases it
4933 gets very large very quickly.
4934
4935 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4936 returns the size of the data block pointed to by the study_data field in a
4937 pcre_extra block, that is, the value that was passed as the argument to
4938 pcre_malloc() when PCRE was getting memory in which to place the information
4939 created by pcre_study(). The fourth argument should point to a size_t variable.
4940 pcretest has been extended so that this information is shown after a successful
4941 pcre_study() call when information about the compiled regex is being displayed.
4942
4943 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4944 because what follows is always an absolute path. (Later: it turns out that this
4945 is more than cosmetic for MinGW, because it doesn't like empty path
4946 components.)
4947
4948 49. Some changes have been made to the callout feature (see 28 above):
4949
4950 (i) A callout function now has three choices for what it returns:
4951
4952 0 => success, carry on matching
4953 > 0 => failure at this point, but backtrack if possible
4954 < 0 => serious error, return this value from pcre_exec()
4955
4956 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4957 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4958 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4959 use by callout functions. It will never be used by PCRE itself.
4960
4961 (ii) The pcre_extra structure (see 45 above) has a void * field called
4962 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4963 pcre_callout_block structure has a field of the same name. The contents of
4964 the field passed in the pcre_extra structure are passed to the callout
4965 function in the corresponding field in the callout block. This makes it
4966 easier to use the same callout-containing regex from multiple threads. For
4967 testing, the pcretest program has a new data escape
4968
4969 \C*n pass the number n (may be negative) as callout_data
4970
4971 If the callout function in pcretest receives a non-zero value as
4972 callout_data, it returns that value.
4973
4974 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4975 there were some redundant $(CFLAGS) in commands that are now specified as
4976 $(LINK), which already includes $(CFLAGS).
4977
4978 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4979 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4980 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4981 one-byte characters throughout. Note that case-insensitive matching applies
4982 only to characters whose values are less than 256. PCRE doesn't support the
4983 notion of cases for higher-valued characters.
4984
4985 (i) A character class whose characters are all within 0-255 is handled as
4986 a bit map, and the map is inverted for negative classes. Previously, a
4987 character > 255 always failed to match such a class; however it should
4988 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4989
4990 (ii) A negated character class with a single character < 255 is coded as
4991 "not this character" (OP_NOT). This wasn't working properly when the test
4992 character was multibyte, either singly or repeated.
4993
4994 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4995 mode, for example: \x{100}{2,3}.
4996
4997 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4998 singly or repeated) now correctly test multibyte characters. However,
4999 PCRE doesn't recognize any characters with values greater than 255 as
5000 digits, spaces, or word characters. Such characters always match \D, \S,
5001 and \W, and never match \d, \s, or \w.
5002
5003 (v) Classes may now contain characters and character ranges with values
5004 greater than 255. For example: [ab\x{100}-\x{400}].
5005
5006 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5007 PCRE in UTF-8 mode.
5008
5009 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5010 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5011 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5012 value.)
5013
5014 53. The single man page has become too large. I have therefore split it up into
5015 a number of separate man pages. These also give rise to individual HTML pages;
5016 these are now put in a separate directory, and there is an index.html page that
5017 lists them all. Some hyperlinking between the pages has been installed.
5018
5019 54. Added convenience functions for handling named capturing parentheses.
5020
5021 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5022 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5023 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5024 are faulted.
5025
5026 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5027 calling configure. These values are used when compiling the dftables.c program
5028 which is run to generate the source of the default character tables. They
5029 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5030 you will need to set these values.
5031
5032 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5033
5034
5035 Version 3.9 02-Jan-02
5036 ---------------------
5037
5038 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5039
5040 2. If --disable-static was given, the building process failed when trying to
5041 build pcretest and pcregrep. (For some reason it was using libtool to compile
5042 them, which is not right, as they aren't part of the library.)
5043
5044
5045 Version 3.8 18-Dec-01
5046 ---------------------
5047
5048 1. The experimental UTF-8 code was completely screwed up. It was packing the
5049 bytes in the wrong order. How dumb can you get?
5050
5051
5052 Version 3.7 29-Oct-01
5053 ---------------------
5054
5055 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5056 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5057 this didn't happen under Solaris 8, where I normally test things.
5058
5059 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5060 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5061 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5062 this fix an hour or so after the initial 3.7 release.)
5063
5064
5065 Version 3.6 23-Oct-01
5066 ---------------------
5067
5068 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5069 offsets passed as NULL with zero offset count.
5070
5071 2. The config.guess and config.sub files had not been updated when I moved to
5072 the latest autoconf.
5073
5074
5075 Version 3.5 15-Aug-01
5076 ---------------------
5077
5078 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5079 had been forgotten.
5080
5081 2. By using declared but undefined structures, we can avoid using "void"
5082 definitions in pcre.h while keeping the internal definitions of the structures
5083 private.
5084
5085 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5086 user point of view, this means that both static and shared libraries are built
5087 by default, but this can be individually controlled. More of the work of
5088 handling this static/shared cases is now inside libtool instead of PCRE's make
5089 file.
5090
5091 4. The pcretest utility is now installed along with pcregrep because it is
5092 useful for users (to test regexs) and by doing this, it automatically gets
5093 relinked by libtool. The documentation has been turned into a man page, so
5094 there are now .1, .txt, and .html versions in /doc.
5095
5096 5. Upgrades to pcregrep:
5097 (i) Added long-form option names like gnu grep.
5098 (ii) Added --help to list all options with an explanatory phrase.
5099 (iii) Added -r, --recursive to recurse into sub-directories.
5100 (iv) Added -f, --file to read patterns from a file.
5101
5102 6. pcre_exec() was referring to its "code" argument before testing that
5103 argument for NULL (and giving an error if it was NULL).
5104
5105 7. Upgraded Makefile.in to allow for compiling in a different directory from
5106 the source directory.
5107
5108 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5109 options bits, the pointer it was passed was to an int instead of to an unsigned
5110 long int. This mattered only on 64-bit systems.
5111
5112 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5113 generated) instead of pcre.in, which it its source. Also made the same change
5114 in several of the .c files.
5115
5116 10. A new release of gcc defines printf() as a macro, which broke pcretest
5117 because it had an ifdef in the middle of a string argument for printf(). Fixed
5118 by using separate calls to printf().
5119
5120 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5121 script, to force use of CR or LF instead of \n in the source. On non-Unix
5122 systems, the value can be set in config.h.
5123
5124 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5125 absolute limit. Changed the text of the error message to make this clear, and
5126 likewise updated the man page.
5127
5128 13. The limit of 99 on the number of capturing subpatterns has been removed.
5129 The new limit is 65535, which I hope will not be a "real" limit.
5130
5131
5132 Version 3.4 22-Aug-00
5133 ---------------------
5134
5135 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5136
5137 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5138
5139
5140 Version 3.3 01-Aug-00
5141 ---------------------
5142
5143 1. If an octal character was given, but the value was greater than \377, it
5144 was not getting masked to the least significant bits, as documented. This could
5145 lead to crashes in some systems.
5146
5147 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5148 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5149
5150 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5151 These just pass their arguments on to (pcre_free)(), but they are provided
5152 because some uses of PCRE bind it to non-C systems that can call its functions,
5153 but cannot call free() or pcre_free() directly.
5154
5155 4. Add "make test" as a synonym for "make check". Corrected some comments in
5156 the Makefile.
5157
5158 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5159 Makefile.
5160
5161 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5162 command called pgrep for grepping around the active processes.
5163
5164 7. Added the beginnings of support for UTF-8 character strings.
5165
5166 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5167 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5168 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5169 out for the ar command.)
5170
5171
5172 Version 3.2 12-May-00
5173 ---------------------
5174
5175 This is purely a bug fixing release.
5176
5177 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5178 of ZA. This was just one example of several cases that could provoke this bug,
5179 which was introduced by change 9 of version 2.00. The code for breaking
5180 infinite loops after an iteration that matches an empty string was't working
5181 correctly.
5182
5183 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5184 when matched against abbab (for example). After matching an empty string, it
5185 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5186 caused it to match further down the string than it should.
5187
5188 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5189 was there because it doesn't seem to be needed, and it causes trouble on some
5190 systems, as it is not a Standard C header. It has been removed.
5191
5192 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5193 were reported on the Macintosh. The changes were from
5194
5195 while ((c = *(++ptr)) != 0 && c != '\n');
5196 to
5197 while ((c = *(++ptr)) != 0 && c != '\n') ;
5198
5199 Totally extraordinary, but if that's what it takes...
5200
5201 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5202 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5203 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5204 assumes the way PCRE uses memmove() (always moving upwards).
5205
5206 6. PCRE is being used in one environment where strchr() is not available. There
5207 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5208 faster code anyway.
5209
5210
5211 Version 3.1 09-Feb-00
5212 ---------------------
5213
5214 The only change in this release is the fixing of some bugs in Makefile.in for
5215 the "install" target:
5216
5217 (1) It was failing to install pcreposix.h.
5218
5219 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5220
5221
5222 Version 3.0 01-Feb-00
5223 ---------------------
5224
5225 1. Add support for the /+ modifier to perltest (to output $` like it does in
5226 pcretest).
5227
5228 2. Add support for the /g modifier to perltest.
5229
5230 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5231 matches null strings.
5232
5233 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5234 pattern. Perl treats empty patterns specially - it reuses the most recent
5235 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5236 effect.
5237
5238 5. The POSIX interface was broken in that it was just handing over the POSIX
5239 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5240 required a bigger vector, with some working space on the end. This means that
5241 the POSIX wrapper now has to get and free some memory, and copy the results.
5242
5243 6. Added some simple autoconf support, placing the test data and the
5244 documentation in separate directories, re-organizing some of the
5245 information files, and making it build pcre-config (a GNU standard). Also added
5246 libtool support for building PCRE as a shared library, which is now the
5247 default.
5248
5249 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5250 09 are not valid octal constants. Single digits will be used for minor values
5251 less than 10.
5252
5253 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5254 existing programs that set these in the POSIX interface can use PCRE without
5255 modification.
5256
5257 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5258 return all that pcre_info() returns, plus additional data. The pcre_info()
5259 function is retained for compatibility, but is considered to be obsolete.
5260
5261 10. Added experimental recursion feature (?R) to handle one common case that
5262 Perl 5.6 will be able to do with (?p{...}).
5263
5264 11. Added support for POSIX character classes like [:alpha:], which Perl is
5265 adopting.
5266
5267
5268 Version 2.08 31-Aug-99
5269 ----------------------
5270
5271 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5272 trying to match at the startoffset position, but instead was moving forward to
5273 the next newline as if a previous match had failed.
5274
5275 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5276 and could get into a loop if a null string was matched other than at the start
5277 of the subject.
5278
5279 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5280 be distinguished at compile time, and for completeness also added PCRE_DATE.
5281
5282 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5283 in GnuWin32 environments.
5284
5285
5286 Version 2.07 29-Jul-99
5287 ----------------------
5288
5289 1. The documentation is now supplied in plain text form and HTML as well as in
5290 the form of man page sources.
5291
5292 2. C++ compilers don't like assigning (void *) values to other pointer types.
5293 In particular this affects malloc(). Although there is no problem in Standard
5294 C, I've put in casts to keep C++ compilers happy.
5295
5296 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5297 should be (const char *).
5298
5299 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5300 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5301 However, I haven't made this a standard facility. The documentation doesn't
5302 mention it, and the Makefile doesn't support it.
5303
5304 5. The Makefile now contains an "install" target, with editable destinations at
5305 the top of the file. The pcretest program is not installed.
5306
5307 6. pgrep -V now gives the PCRE version number and date.
5308
5309 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5310 causing the entire string to be ignored, instead of just the last character.
5311
5312 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5313 non-matching string, it can take a very, very long time, even for strings of
5314 quite modest length, because of the nested recursion. PCRE now does better in
5315 some of these cases. It does this by remembering the last required literal
5316 character in the pattern, and pre-searching the subject to ensure it is present
5317 before running the real match. In other words, it applies a heuristic to detect
5318 some types of certain failure quickly, and in the above example, if presented
5319 with a string that has no trailing " it gives "no match" very quickly.
5320
5321 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5322 other alternatives are tried instead.
5323
5324
5325 Version 2.06 09-Jun-99
5326 ----------------------
5327
5328 1. Change pcretest's output for amount of store used to show just the code
5329 space, because the remainder (the data block) varies in size between 32-bit and
5330 64-bit systems.
5331
5332 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5333 start matching at. This allows lookbehinds to work when searching for multiple
5334 occurrences in a string.
5335
5336 3. Added additional options to pcretest for testing multiple occurrences:
5337
5338 /+ outputs the rest of the string that follows a match
5339 /g loops for multiple occurrences, using the new startoffset argument
5340 /G loops for multiple occurrences by passing an incremented pointer
5341
5342 4. PCRE wasn't doing the "first character" optimization for patterns starting
5343 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5344 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5345 the letter 'x'. On long subject strings, this gives a significant speed-up.
5346
5347
5348 Version 2.05 21-Apr-99
5349 ----------------------
5350
5351 1. Changed the type of magic_number from int to long int so that it works
5352 properly on 16-bit systems.
5353
5354 2. Fixed a bug which caused patterns starting with .* not to work correctly
5355 when the subject string contained newline characters. PCRE was assuming
5356 anchoring for such patterns in all cases, which is not correct because .* will
5357 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5358 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5359 must be retried after every newline in the subject.
5360
5361
5362 Version 2.04 18-Feb-99
5363 ----------------------
5364
5365 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5366 computation of the store needed to hold the pattern was incorrect (too large).
5367 If such patterns were nested a few deep, this could multiply and become a real
5368 problem.
5369
5370 2. Added /M option to pcretest to show the memory requirement of a specific
5371 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5372
5373 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5374 compiled in such a way that the backtracking after subsequent failure was
5375 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5376 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5377
5378
5379 Version 2.03 02-Feb-99
5380 ----------------------
5381
5382 1. Fixed typo and small mistake in man page.
5383
5384 2. Added 4th condition (GPL supersedes if conflict) and created separate
5385 LICENCE file containing the conditions.
5386
5387 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5388 Perl, that is the internal \ allows the delimiter to be included in the
5389 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5390 the final delimiter, add \ to the end of the pattern (to test the error).
5391
5392 4. Added the convenience functions for extracting substrings after a successful
5393 match. Updated pcretest to make it able to test these functions.
5394
5395
5396 Version 2.02 14-Jan-99
5397 ----------------------
5398
5399 1. Initialized the working variables associated with each extraction so that
5400 their saving and restoring doesn't refer to uninitialized store.
5401
5402 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5403 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5404 fix the problem.
5405
5406 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5407 calls, and wasn't printing the correct value for compiling calls. Increased the
5408 default value of LOOPREPEAT, and the number of significant figures in the
5409 times.
5410
5411 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5412
5413 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5414 a building problem on Windows NT with a FAT file system.
5415
5416
5417 Version 2.01 21-Oct-98
5418 ----------------------
5419
5420 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5421 to character tables built by pcre_maketables() in the current locale. If NULL
5422 is passed, the default tables are used.
5423
5424
5425 Version 2.00 24-Sep-98
5426 ----------------------
5427
5428 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5429 it any more.
5430
5431 2. Allow quantification of (?>) groups, and make it work correctly.
5432
5433 3. The first character computation wasn't working for (?>) groups.
5434
5435 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5436 end of the subject) and add 5.005's \z, which really does match only at the
5437 very end of the subject.
5438
5439 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5440
5441 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5442 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5443 localized options. All options to pcre_study() were also removed.
5444
5445 7. Add other new features from 5.005:
5446
5447 $(?<= positive lookbehind
5448 $(?<! negative lookbehind
5449 (?imsx-imsx) added the unsetting capability
5450 such a setting is global if at outer level; local otherwise
5451 (?imsx-imsx:) non-capturing groups with option setting
5452 (?(cond)re|re) conditional pattern matching
5453
5454 A backreference to itself in a repeated group matches the previous
5455 captured string.
5456
5457 8. General tidying up of studying (both automatic and via "study")
5458 consequential on the addition of new assertions.
5459
5460 9. As in 5.005, unlimited repeated groups that could match an empty substring
5461 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5462 runtime if any iteration does actually match an empty substring.
5463
5464 10. Include the RunTest script in the distribution.
5465
5466 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5467 discrepancies, some of which were old and were also with respect to 5.004. They
5468 have now been fixed.
5469
5470
5471 Version 1.09 28-Apr-98
5472 ----------------------
5473
5474 1. A negated single character class followed by a quantifier with a minimum
5475 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5476 program crashes, or just wrong answers. This did not apply to negated classes
5477 containing more than one character, or to minima other than one.
5478
5479
5480 Version 1.08 27-Mar-98
5481 ----------------------
5482
5483 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5484
5485 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5486 latter must appear before anything that relies on it in the pattern.
5487
5488
5489 Version 1.07 16-Feb-98
5490 ----------------------
5491
5492 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5493 repeat of a potentially empty string).
5494
5495
5496 Version 1.06 23-Jan-98
5497 ----------------------
5498
5499 1. Added Markus Oberhumer's little patches for C++.
5500
5501 2. Literal strings longer than 255 characters were broken.
5502
5503
5504 Version 1.05 23-Dec-97
5505 ----------------------
5506
5507 1. Negated character classes containing more than one character were failing if
5508 PCRE_CASELESS was set at run time.
5509
5510
5511 Version 1.04 19-Dec-97
5512 ----------------------
5513
5514 1. Corrected the man page, where some "const" qualifiers had been omitted.
5515
5516 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5517 input syntax.
5518
5519 3. Fixed memory leak which occurred when a regex with back references was
5520 matched with an offsets vector that wasn't big enough. The temporary memory
5521 that is used in this case wasn't being freed if the match failed.
5522
5523 4. Tidied pcretest to ensure it frees memory that it gets.
5524
5525 5. Temporary memory was being obtained in the case where the passed offsets
5526 vector was exactly big enough.
5527
5528 6. Corrected definition of offsetof() from change 5 below.
5529
5530 7. I had screwed up change 6 below and broken the rules for the use of
5531 setjmp(). Now fixed.
5532
5533
5534 Version 1.03 18-Dec-97
5535 ----------------------
5536
5537 1. A erroneous regex with a missing opening parenthesis was correctly
5538 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5539 on some systems.
5540
5541 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5542 it was reported that one broken compiler failed on the former because "code" is
5543 also an independent variable.
5544
5545 3. The erroneous regex a[]b caused an array overrun reference.
5546
5547 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5548 fail on data ending with that character. (It was going on too far, and checking
5549 the next character, typically a binary zero.) This was specific to the
5550 optimized code for single-character negative classes.
5551
5552 5. Added a contributed patch from the TIN world which does the following:
5553
5554 + Add an undef for memmove, in case the the system defines a macro for it.
5555
5556 + Add a definition of offsetof(), in case there isn't one. (I don't know
5557 the reason behind this - offsetof() is part of the ANSI standard - but
5558 it does no harm).
5559
5560 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5561 most of the places where whitespace preceded '#'. I have given up and
5562 allowed the remaining 2 cases to be at the margin.
5563
5564 + Rename some variables in pcre to eliminate shadowing. This seems very
5565 pedantic, but does no harm, of course.
5566
5567 6. Moved the call to setjmp() into its own function, to get rid of warnings
5568 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5569
5570 7. Constructs such as \d{8,} were compiling into the equivalent of
5571 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5572 outcome, but in this particular case used more store than had been allocated,
5573 which caused the bug to be discovered because it threw up an internal error.
5574
5575 8. The debugging code in both pcre and pcretest for outputting the compiled
5576 form of a regex was going wrong in the case of back references followed by
5577 curly-bracketed repeats.
5578
5579
5580 Version 1.02 12-Dec-97
5581 ----------------------
5582
5583 1. Typos in pcre.3 and comments in the source fixed.
5584
5585 2. Applied a contributed patch to get rid of places where it used to remove
5586 'const' from variables, and fixed some signed/unsigned and uninitialized
5587 variable warnings.
5588
5589 3. Added the "runtest" target to Makefile.
5590
5591 4. Set default compiler flag to -O2 rather than just -O.
5592
5593
5594 Version 1.01 19-Nov-97
5595 ----------------------
5596
5597 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5598 like /([ab]*)*/, that is, for classes with more than one character in them.
5599
5600 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5601 as /((?>a*))*/ (a PCRE_EXTRA facility).
5602
5603
5604 Version 1.00 18-Nov-97
5605 ----------------------
5606
5607 1. Added compile-time macros to support systems such as SunOS4 which don't have
5608 memmove() or strerror() but have other things that can be used instead.
5609
5610 2. Arranged that "make clean" removes the executables.
5611
5612
5613 Version 0.99 27-Oct-97
5614 ----------------------
5615
5616 1. Fixed bug in code for optimizing classes with only one character. It was
5617 initializing a 32-byte map regardless, which could cause it to run off the end
5618 of the memory it had got.
5619
5620 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5621
5622
5623 Version 0.98 22-Oct-97
5624 ----------------------
5625
5626 1. Fixed bug in code for handling temporary memory usage when there are more
5627 back references than supplied space in the ovector. This could cause segfaults.
5628
5629
5630 Version 0.97 21-Oct-97
5631 ----------------------
5632
5633 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5634
5635 2. Optimized negated single characters not to use a bit map.
5636
5637 3. Brought error texts together as macro definitions; clarified some of them;
5638 fixed one that was wrong - it said "range out of order" when it meant "invalid
5639 escape sequence".
5640
5641 4. Changed some char * arguments to const char *.
5642
5643 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5644
5645 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5646 pcretest.
5647
5648
5649 Version 0.96 16-Oct-97
5650 ----------------------
5651
5652 1. Added a simple "pgrep" utility to the distribution.
5653
5654 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5655 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5656 where "ddd" means "one or more decimal digits".
5657
5658 3. Fixed serious bug. If a pattern had a back reference, but the call to
5659 pcre_exec() didn't supply a large enough ovector to record the related
5660 identifying subpattern, the match always failed. PCRE now remembers the number
5661 of the largest back reference, and gets some temporary memory in which to save
5662 the offsets during matching if necessary, in order to ensure that
5663 backreferences always work.
5664
5665 4. Increased the compatibility with Perl in a number of ways:
5666
5667 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5668 to request this handling. The option can be set at compile or exec time.
5669
5670 (b) $ matches before a terminating newline by default; an option
5671 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5672 mode). The option can be set at compile or exec time.
5673
5674 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5675 the same as Perl's. If the decimal number it represents is less than 10
5676 or there aren't that many previous left capturing parentheses, an octal
5677 escape is read. Inside a character class, it's always an octal escape,
5678 even if it is a single digit.
5679
5680 (d) An escaped but undefined alphabetic character is taken as a literal,
5681 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5682 escapes.
5683
5684 (e) {0} is now permitted. (The previous item is removed from the compiled
5685 pattern).
5686
5687 5. Changed all the names of code files so that the basic parts are no longer
5688 than 10 characters, and abolished the teeny "globals.c" file.
5689
5690 6. Changed the handling of character classes; they are now done with a 32-byte
5691 bit map always.
5692
5693 7. Added the -d and /D options to pcretest to make it possible to look at the
5694 internals of compilation without having to recompile pcre.
5695
5696
5697 Version 0.95 23-Sep-97
5698 ----------------------
5699
5700 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5701 \x20 at the start of a run of normal characters. These were being treated as
5702 real characters, instead of the source characters being re-checked.
5703
5704
5705 Version 0.94 18-Sep-97
5706 ----------------------
5707
5708 1. The functions are now thread-safe, with the caveat that the global variables
5709 containing pointers to malloc() and free() or alternative functions are the
5710 same for all threads.
5711
5712 2. Get pcre_study() to generate a bitmap of initial characters for non-
5713 anchored patterns when this is possible, and use it if passed to pcre_exec().
5714
5715
5716 Version 0.93 15-Sep-97
5717 ----------------------
5718
5719 1. /(b)|(:+)/ was computing an incorrect first character.
5720
5721 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5722 but not actually doing anything yet.
5723
5724 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5725 as Perl does (e.g. [-az] or [az-]).
5726
5727 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5728 all possible positions.
5729
5730 5. Split up into different modules to avoid including unneeded functions in a
5731 compiled binary. However, compile and exec are still in one module. The "study"
5732 function is split off.
5733
5734 6. The character tables are now in a separate module whose source is generated
5735 by an auxiliary program - but can then be edited by hand if required. There are
5736 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5737 toupper() in the code.
5738
5739 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5740 make them global. Abolish the function for setting them, as the caller can now
5741 set them directly.
5742
5743
5744 Version 0.92 11-Sep-97
5745 ----------------------
5746
5747 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5748 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5749
5750 2. Caseless matching was not working in character classes if the characters in
5751 the pattern were in upper case.
5752
5753 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5754
5755 4. Make PCRE_ANCHORED public and accept as a compile option.
5756
5757 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5758 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5759 pass them.
5760
5761 6. Give an error if bad option bits passed at compile or run time.
5762
5763 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5764 pcretest to cause it to pass that flag.
5765
5766 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5767 options, and the first character, if set.
5768
5769 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5770
5771
5772 Version 0.91 10-Sep-97
5773 ----------------------
5774
5775 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5776 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5777
5778 2. PCRE was looping on encountering an indefinitely repeated back reference to
5779 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5780 Perl does - treats the match as successful.
5781
5782 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5