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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1564 - (show annotations)
Tue Jun 9 16:33:27 2015 UTC (4 years, 4 months ago) by ph10
File size: 279490 byte(s)
Error occurred while calculating annotation data.
Fix table entry for \l in EBCDIC environments.
1 ChangeLog for PCRE
2 ------------------
3
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
6
7 Version 8.38 xx-xxx-xxxx
8 ------------------------
9
10 1. If a group that contained a recursive back reference also contained a
11 forward reference subroutine call followed by a non-forward-reference
12 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13 compile correct code, leading to undefined behaviour or an internally
14 detected error. This bug was discovered by the LLVM fuzzer.
15
16 2. Quantification of certain items (e.g. atomic back references) could cause
17 incorrect code to be compiled when recursive forward references were
18 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19 This bug was discovered by the LLVM fuzzer.
20
21 3. A repeated conditional group whose condition was a reference by name caused
22 a buffer overflow if there was more than one group with the given name.
23 This bug was discovered by the LLVM fuzzer.
24
25 4. A recursive back reference by name within a group that had the same name as
26 another group caused a buffer overflow. For example:
27 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28
29 5. A forward reference by name to a group whose number is the same as the
30 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31 a buffer overflow at compile time. This bug was discovered by the LLVM
32 fuzzer.
33
34 6. A lookbehind assertion within a set of mutually recursive subpatterns could
35 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36
37 7. Another buffer overflow bug involved duplicate named groups with a
38 reference between their definition, with a group that reset capture
39 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40 fixed by always allowing for more memory, even if not needed. (A proper fix
41 is implemented in PCRE2, but it involves more refactoring.)
42
43 8. There was no check for integer overflow in subroutine calls such as (?123).
44
45 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
46 being treated as a literal 'l' instead of causing an error.
47
48
49 Version 8.37 28-April-2015
50 --------------------------
51
52 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
53 for those parentheses to be closed with whatever has been captured so far.
54 However, it was failing to mark any other groups between the hightest
55 capture so far and the currrent group as "unset". Thus, the ovector for
56 those groups contained whatever was previously there. An example is the
57 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
58
59 2. If an assertion condition was quantified with a minimum of zero (an odd
60 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
61
62 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
63 unrecognized modifier, a crash could occur.
64
65 4. An attempt to do global matching in pcretest with a zero-length ovector
66 caused a crash.
67
68 5. Fixed a memory leak during matching that could occur for a subpattern
69 subroutine call (recursive or otherwise) if the number of captured groups
70 that had to be saved was greater than ten.
71
72 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
73 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
74 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
75
76 7. A UTF pattern containing a "not" match of a non-ASCII character and a
77 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
78
79 8. When a pattern is compiled, it remembers the highest back reference so that
80 when matching, if the ovector is too small, extra memory can be obtained to
81 use instead. A conditional subpattern whose condition is a check on a
82 capture having happened, such as, for example in the pattern
83 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
84 setting the highest backreference number. This mattered only if pcre_exec()
85 was called with an ovector that was too small to hold the capture, and there
86 was no other kind of back reference (a situation which is probably quite
87 rare). The effect of the bug was that the condition was always treated as
88 FALSE when the capture could not be consulted, leading to a incorrect
89 behaviour by pcre_exec(). This bug has been fixed.
90
91 9. A reference to a duplicated named group (either a back reference or a test
92 for being set in a conditional) that occurred in a part of the pattern where
93 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
94 to be incorrectly calculated, leading to overwriting.
95
96 10. A mutually recursive set of back references such as (\2)(\1) caused a
97 segfault at study time (while trying to find the minimum matching length).
98 The infinite loop is now broken (with the minimum length unset, that is,
99 zero).
100
101 11. If an assertion that was used as a condition was quantified with a minimum
102 of zero, matching went wrong. In particular, if the whole group had
103 unlimited repetition and could match an empty string, a segfault was
104 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
105 assertions to be quantified, but not if they are being used as conditions,
106 so the above pattern is faulted by Perl. PCRE has now been changed so that
107 it also rejects such patterns.
108
109 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
110 failed to allow the zero-repeat case if pcre2_exec() was called with an
111 ovector too small to capture the group.
112
113 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
114 Red Hat Product Security:
115
116 (a) A crash if /K and /F were both set with the option to save the compiled
117 pattern.
118
119 (b) Another crash if the option to print captured substrings in a callout
120 was combined with setting a null ovector, for example \O\C+ as a subject
121 string.
122
123 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
124 forward reference repeated a large (but limited) number of times within a
125 repeated outer group that has a zero minimum quantifier, caused incorrect
126 code to be compiled, leading to the error "internal error:
127 previously-checked referenced subpattern not found" when an incorrect
128 memory address was read. This bug was reported as "heap overflow",
129 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
130 CVE-2015-2325.
131
132 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
133 call within a group that also contained a recursive back reference caused
134 incorrect code to be compiled. This bug was reported as "heap overflow",
135 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
136 number CVE-2015-2326.
137
138 24. Computing the size of the JIT read-only data in advance has been a source
139 of various issues, and new ones are still appear unfortunately. To fix
140 existing and future issues, size computation is eliminated from the code,
141 and replaced by on-demand memory allocation.
142
143 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
144 adjacent to the end of the range, and the range contained characters with
145 more than one other case, caused incorrect behaviour when compiled in UTF
146 mode. In that example, the range a-j was left out of the class.
147
148 26. Fix JIT compilation of conditional blocks, which assertion
149 is converted to (*FAIL). E.g: /(?(?!))/.
150
151 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
152 discovered by the LLVM fuzzer.
153
154 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
155 when this assertion was used as a condition, for example (?(?!)a|b). In
156 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
157 error about an unsupported item.
158
159 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
160 possessification code could take exponential time to complete. A recursion
161 depth limit of 1000 has been imposed to limit the resources used by this
162 optimization.
163
164 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
165 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
166 because \S ensures they are all in the class. The code for doing this was
167 interacting badly with the code for computing the amount of space needed to
168 compile the pattern, leading to a buffer overflow. This bug was discovered
169 by the LLVM fuzzer.
170
171 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
172 other kinds of group caused stack overflow at compile time. This bug was
173 discovered by the LLVM fuzzer.
174
175 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
176 between a subroutine call and its quantifier was incorrectly compiled,
177 leading to buffer overflow or other errors. This bug was discovered by the
178 LLVM fuzzer.
179
180 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
181 assertion after (?(. The code was failing to check the character after
182 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
183 was discovered by the LLVM fuzzer.
184
185 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
186 a fixed maximum following a group that contains a subroutine reference was
187 incorrectly compiled and could trigger buffer overflow. This bug was
188 discovered by the LLVM fuzzer.
189
190 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
191 caused a stack overflow instead of the diagnosis of a non-fixed length
192 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
193
194 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
195 (e.g. /(?<=\Ka)/) could make pcregrep loop.
196
197 37. There was a similar problem to 36 in pcretest for global matches.
198
199 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
200 and a subsequent item in the pattern caused a non-match, backtracking over
201 the repeated \X did not stop, but carried on past the start of the subject,
202 causing reference to random memory and/or a segfault. There were also some
203 other cases where backtracking after \C could crash. This set of bugs was
204 discovered by the LLVM fuzzer.
205
206 39. The function for finding the minimum length of a matching string could take
207 a very long time if mutual recursion was present many times in a pattern,
208 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
209 method has been implemented. This infelicity was discovered by the LLVM
210 fuzzer.
211
212 40. Static linking against the PCRE library using the pkg-config module was
213 failing on missing pthread symbols.
214
215
216 Version 8.36 26-September-2014
217 ------------------------------
218
219 1. Got rid of some compiler warnings in the C++ modules that were shown up by
220 -Wmissing-field-initializers and -Wunused-parameter.
221
222 2. The tests for quantifiers being too big (greater than 65535) were being
223 applied after reading the number, and stupidly assuming that integer
224 overflow would give a negative number. The tests are now applied as the
225 numbers are read.
226
227 3. Tidy code in pcre_exec.c where two branches that used to be different are
228 now the same.
229
230 4. The JIT compiler did not generate match limit checks for certain
231 bracketed expressions with quantifiers. This may lead to exponential
232 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
233 issue should be resolved now.
234
235 5. Fixed an issue, which occures when nested alternatives are optimized
236 with table jumps.
237
238 6. Inserted two casts and changed some ints to size_t in the light of some
239 reported 64-bit compiler warnings (Bugzilla 1477).
240
241 7. Fixed a bug concerned with zero-minimum possessive groups that could match
242 an empty string, which sometimes were behaving incorrectly in the
243 interpreter (though correctly in the JIT matcher). This pcretest input is
244 an example:
245
246 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
247 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
248
249 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
250 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
251 for an empty string was breaking the inner loop and carrying on at a lower
252 level, when possessive repeated groups should always return to a higher
253 level as they have no backtrack points in them. The empty string test now
254 occurs at the outer level.
255
256 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
257 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
258
259 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
260 doesn't).
261
262 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
263 to hold a pattern to be incorrectly computed (too small) when there were
264 named back references to duplicated names. This could cause "internal
265 error: code overflow" or "double free or corruption" or other memory
266 handling errors.
267
268 11. When named subpatterns had the same prefixes, back references could be
269 confused. For example, in this pattern:
270
271 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
272
273 the reference to 'Name' was incorrectly treated as a reference to a
274 duplicate name.
275
276 12. A pattern such as /^s?c/mi8 where the optional character has more than
277 one "other case" was incorrectly compiled such that it would only try to
278 match starting at "c".
279
280 13. When a pattern starting with \s was studied, VT was not included in the
281 list of possible starting characters; this should have been part of the
282 8.34/18 patch.
283
284 14. If a character class started [\Qx]... where x is any character, the class
285 was incorrectly terminated at the ].
286
287 15. If a pattern that started with a caseless match for a character with more
288 than one "other case" was studied, PCRE did not set up the starting code
289 unit bit map for the list of possible characters. Now it does. This is an
290 optimization improvement, not a bug fix.
291
292 16. The Unicode data tables have been updated to Unicode 7.0.0.
293
294 17. Fixed a number of memory leaks in pcregrep.
295
296 18. Avoid a compiler warning (from some compilers) for a function call with
297 a cast that removes "const" from an lvalue by using an intermediate
298 variable (to which the compiler does not object).
299
300 19. Incorrect code was compiled if a group that contained an internal recursive
301 back reference was optional (had quantifier with a minimum of zero). This
302 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
303 caused segmentation faults because of stack overflows at compile time.
304
305 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
306 group that is quantified with an indefinite repeat, caused a compile-time
307 loop which used up all the system stack and provoked a segmentation fault.
308 This was not the same bug as 19 above.
309
310 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
311 Patch by Mike Frysinger.
312
313
314 Version 8.35 04-April-2014
315 --------------------------
316
317 1. A new flag is set, when property checks are present in an XCLASS.
318 When this flag is not set, PCRE can perform certain optimizations
319 such as studying these XCLASS-es.
320
321 2. The auto-possessification of character sets were improved: a normal
322 and an extended character set can be compared now. Furthermore
323 the JIT compiler optimizes more character set checks.
324
325 3. Got rid of some compiler warnings for potentially uninitialized variables
326 that show up only when compiled with -O2.
327
328 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
329 of a match later then the end of the match. The pcretest program was not
330 handling the case sensibly - it was outputting from the start to the next
331 binary zero. It now reports this situation in a message, and outputs the
332 text from the end to the start.
333
334 5. Fast forward search is improved in JIT. Instead of the first three
335 characters, any three characters with fixed position can be searched.
336 Search order: first, last, middle.
337
338 6. Improve character range checks in JIT. Characters are read by an inprecise
339 function now, which returns with an unknown value if the character code is
340 above a certain threshold (e.g: 256). The only limitation is that the value
341 must be bigger than the threshold as well. This function is useful when
342 the characters above the threshold are handled in the same way.
343
344 7. The macros whose names start with RAWUCHAR are placeholders for a future
345 mode in which only the bottom 21 bits of 32-bit data items are used. To
346 make this more memorable for those maintaining the code, the names have
347 been changed to start with UCHAR21, and an extensive comment has been added
348 to their definition.
349
350 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
351 to the export list in Makefile.am (they were accidentally omitted from the
352 8.34 tarball).
353
354 9. The informational output from pcretest used the phrase "starting byte set"
355 which is inappropriate for the 16-bit and 32-bit libraries. As the output
356 for "first char" and "need char" really means "non-UTF-char", I've changed
357 "byte" to "char", and slightly reworded the output. The documentation about
358 these values has also been (I hope) clarified.
359
360 10. Another JIT related optimization: use table jumps for selecting the correct
361 backtracking path, when more than four alternatives are present inside a
362 bracket.
363
364 11. Empty match is not possible, when the minimum length is greater than zero,
365 and there is no \K in the pattern. JIT should avoid empty match checks in
366 such cases.
367
368 12. In a caseless character class with UCP support, when a character with more
369 than one alternative case was not the first character of a range, not all
370 the alternative cases were added to the class. For example, s and \x{17f}
371 are both alternative cases for S: the class [RST] was handled correctly,
372 but [R-T] was not.
373
374 13. The configure.ac file always checked for pthread support when JIT was
375 enabled. This is not used in Windows, so I have put this test inside a
376 check for the presence of windows.h (which was already tested for).
377
378 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
379 The algorithm provides a way to skip certain starting offsets, and usually
380 faster than linear prefix searches.
381
382 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
383 as for 'fr_FR' and 'french'. For some reason, however, it then used the
384 Windows-specific input and output files, which have 'french' screwed in.
385 So this could never have worked. One of the problems with locales is that
386 they aren't always the same. I have now updated RunTest so that it checks
387 the output of the locale test (test 3) against three different output
388 files, and it allows the test to pass if any one of them matches. With luck
389 this should make the test pass on some versions of Solaris where it was
390 failing. Because of the uncertainty, the script did not used to stop if
391 test 3 failed; it now does. If further versions of a French locale ever
392 come to light, they can now easily be added.
393
394 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
395 there was a message during ./configure, but it did not stop. This now
396 provokes an error. The invalid example in README has been corrected.
397 If a value less than the minimum is given, the minimum value has always
398 been used, but now a warning is given.
399
400 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
401 was a bug in the test system, which is now fixed. Also, the list of various
402 configurations that are tested for each release did not have one with both
403 16/32 bits and --enable-bar-anycrlf. It now does.
404
405 18. pcretest was missing "-C bsr" for displaying the \R default setting.
406
407 19. Little endian PowerPC systems are supported now by the JIT compiler.
408
409 20. The fast forward newline mechanism could enter to an infinite loop on
410 certain invalid UTF-8 input. Although we don't support these cases
411 this issue can be fixed by a performance optimization.
412
413 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
414 not take account if existing stack usage. There is now a new global
415 variable called pcre_stack_guard that can be set to point to an external
416 function to check stack availability. It is called at the start of
417 processing every parenthesized group.
418
419 22. A typo in the code meant that in ungreedy mode the max/min qualifier
420 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
421 not match "ab".
422
423 23. When UTF was disabled, the JIT program reported some incorrect compile
424 errors. These messages are silenced now.
425
426 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
427 compiler.
428
429 25. Change all the temporary files used in RunGrepTest to be different to those
430 used by RunTest so that the tests can be run simultaneously, for example by
431 "make -j check".
432
433
434 Version 8.34 15-December-2013
435 -----------------------------
436
437 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
438 executable memory. Patch inspired by Carsten Klein.
439
440 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
441 this macro is never tested and has no effect, because the work to support
442 coverage involves only compiling and linking options and special targets in
443 the Makefile. The comment in config.h implied that defining the macro would
444 enable coverage support, which is totally false. There was also support for
445 setting this macro in the CMake files (my fault, I just copied it from
446 configure). SUPPORT_GCOV has now been removed.
447
448 3. Make a small performance improvement in strlen16() and strlen32() in
449 pcretest.
450
451 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
452 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
453 these cases). There was also one in pcretest.c.
454
455 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
456
457 6. In UTF mode, the code for checking whether a group could match an empty
458 string (which is used for indefinitely repeated groups to allow for
459 breaking an infinite loop) was broken when the group contained a repeated
460 negated single-character class with a character that occupied more than one
461 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
462 UTF-8 mode). The effect was undefined: the group might or might not be
463 deemed as matching an empty string, or the program might have crashed.
464
465 7. The code for checking whether a group could match an empty string was not
466 recognizing that \h, \H, \v, \V, and \R must match a character.
467
468 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
469 an empty string. If it can, pcretest shows this in its information output.
470
471 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
472 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
473 matched by pcre_exec() without using JIT:
474
475 (a) If the rest of the pattern did not match after a maximal run of
476 grapheme clusters, the code for backing up to try with fewer of them
477 did not always back up over a full grapheme when characters that do not
478 have the modifier quality were involved, e.g. Hangul syllables.
479
480 (b) If the match point in a subject started with modifier character, and
481 there was no match, the code could incorrectly back up beyond the match
482 point, and potentially beyond the first character in the subject,
483 leading to a segfault or an incorrect match result.
484
485 10. A conditional group with an assertion condition could lead to PCRE
486 recording an incorrect first data item for a match if no other first data
487 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
488 first data item, and therefore matched "ca" after "c" instead of at the
489 start.
490
491 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
492 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
493
494 12. The source of pcregrep now includes z/OS-specific code so that it can be
495 compiled for z/OS as part of the special z/OS distribution.
496
497 13. Added the -T and -TM options to pcretest.
498
499 14. The code in pcre_compile.c for creating the table of named capturing groups
500 has been refactored. Instead of creating the table dynamically during the
501 actual compiling pass, the information is remembered during the pre-compile
502 pass (on the stack unless there are more than 20 named groups, in which
503 case malloc() is used) and the whole table is created before the actual
504 compile happens. This has simplified the code (it is now nearly 150 lines
505 shorter) and prepared the way for better handling of references to groups
506 with duplicate names.
507
508 15. A back reference to a named subpattern when there is more than one of the
509 same name now checks them in the order in which they appear in the pattern.
510 The first one that is set is used for the reference. Previously only the
511 first one was inspected. This change makes PCRE more compatible with Perl.
512
513 16. Unicode character properties were updated from Unicode 6.3.0.
514
515 17. The compile-time code for auto-possessification has been refactored, based
516 on a patch by Zoltan Herczeg. It now happens after instead of during
517 compilation. The code is cleaner, and more cases are handled. The option
518 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
519 options in pcretest are provided to set it. It can also be set by
520 (*NO_AUTO_POSSESS) at the start of a pattern.
521
522 18. The character VT has been added to the default ("C" locale) set of
523 characters that match \s and are generally treated as white space,
524 following this same change in Perl 5.18. There is now no difference between
525 "Perl space" and "POSIX space". Whether VT is treated as white space in
526 other locales depends on the locale.
527
528 19. The code for checking named groups as conditions, either for being set or
529 for being recursed, has been refactored (this is related to 14 and 15
530 above). Processing unduplicated named groups should now be as fast at
531 numerical groups, and processing duplicated groups should be faster than
532 before.
533
534 20. Two patches to the CMake build system, by Alexander Barkov:
535
536 (1) Replace the "source" command by "." in CMakeLists.txt because
537 "source" is a bash-ism.
538
539 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
540 without these the CMake build does not work on Solaris.
541
542 21. Perl has changed its handling of \8 and \9. If there is no previously
543 encountered capturing group of those numbers, they are treated as the
544 literal characters 8 and 9 instead of a binary zero followed by the
545 literals. PCRE now does the same.
546
547 22. Following Perl, added \o{} to specify codepoints in octal, making it
548 possible to specify values greater than 0777 and also making them
549 unambiguous.
550
551 23. Perl now gives an error for missing closing braces after \x{... instead of
552 treating the string as literal. PCRE now does the same.
553
554 24. RunTest used to grumble if an inappropriate test was selected explicitly,
555 but just skip it when running all tests. This make it awkward to run ranges
556 of tests when one of them was inappropriate. Now it just skips any
557 inappropriate tests, as it always did when running all tests.
558
559 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
560 character types such as \d or \w, too many callouts were inserted, and the
561 data that they returned was rubbish.
562
563 26. In UCP mode, \s was not matching two of the characters that Perl matches,
564 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
565 were matched by \h. The code has now been refactored so that the lists of
566 the horizontal and vertical whitespace characters used for \h and \v (which
567 are defined only in one place) are now also used for \s.
568
569 27. Add JIT support for the 64 bit TileGX architecture.
570 Patch by Jiong Wang (Tilera Corporation).
571
572 28. Possessive quantifiers for classes (both explicit and automatically
573 generated) now use special opcodes instead of wrapping in ONCE brackets.
574
575 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
576 (because it's meaningless), this was not happening when PCRE_CASELESS was
577 set. Not wrong, but inefficient.
578
579 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
580 properties for \w, \d, etc) is present in a test regex. Otherwise if the
581 test contains no characters greater than 255, Perl doesn't realise it
582 should be using Unicode semantics.
583
584 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
585 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
586 does in Unicode mode.
587
588 32. Added the "forbid" facility to pcretest so that putting tests into the
589 wrong test files can sometimes be quickly detected.
590
591 33. There is now a limit (default 250) on the depth of nesting of parentheses.
592 This limit is imposed to control the amount of system stack used at compile
593 time. It can be changed at build time by --with-parens-nest-limit=xxx or
594 the equivalent in CMake.
595
596 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
597 errors. Perl warns for these when in warning mode, but PCRE has no facility
598 for giving warnings.
599
600 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
601 However, this was not working for (?!) because it is optimized to (*FAIL),
602 for which PCRE does not allow quantifiers. The optimization is now disabled
603 when a quantifier follows (?!). I can't see any use for this, but it makes
604 things uniform.
605
606 36. Perl no longer allows group names to start with digits, so I have made this
607 change also in PCRE. It simplifies the code a bit.
608
609 37. In extended mode, Perl ignores spaces before a + that indicates a
610 possessive quantifier. PCRE allowed a space before the quantifier, but not
611 before the possessive +. It now does.
612
613 38. The use of \K (reset reported match start) within a repeated possessive
614 group such as (a\Kb)*+ was not working.
615
616 40. Document that the same character tables must be used at compile time and
617 run time, and that the facility to pass tables to pcre_exec() and
618 pcre_dfa_exec() is for use only with saved/restored patterns.
619
620 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
621 features for Builds with MSVC:
622
623 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
624 builds). The .rc files can be used to set FileDescription and many other
625 attributes.
626
627 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
628 This allows higher-level build scripts which want .pdb files to avoid
629 hard-coding the exact files needed."
630
631 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
632 mean "start of word" and "end of word", respectively, as a transition aid.
633
634 43. A minimizing repeat of a class containing codepoints greater than 255 in
635 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
636 compiled to use the heap for recursion.
637
638 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
639 is configured.
640
641
642 Version 8.33 28-May-2013
643 ------------------------
644
645 1. Added 'U' to some constants that are compared to unsigned integers, to
646 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
647 variables that are added to signed variables, to ensure the result is
648 signed and can be negated.
649
650 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
651 CMake config files.
652
653 3. Revise the creation of config.h.generic so that all boolean macros are
654 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
655 overriding via -D on the command line possible.
656
657 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
658 to unsigned int is reported to make a quite noticeable speed difference in
659 a specific Windows environment. Testing on Linux did also appear to show
660 some benefit (and it is clearly not harmful). Also fixed the definition of
661 Xop which should be unsigned.
662
663 5. Related to (4), changing the definition of the intermediate variable cc
664 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
665 improvements.
666
667 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
668 unnecessary spaces.
669
670 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
671 and later.
672
673 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
674
675 9. Optimizing fast_forward_start_bits in JIT.
676
677 10. Adding support for callouts in JIT, and fixing some issues revealed
678 during this work. Namely:
679
680 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
681
682 (b) Minimum length was not checked before the matching is started.
683
684 11. The value of capture_last that is passed to callouts was incorrect in some
685 cases when there was a capture on one path that was subsequently abandoned
686 after a backtrack. Also, the capture_last value is now reset after a
687 recursion, since all captures are also reset in this case.
688
689 12. The interpreter no longer returns the "too many substrings" error in the
690 case when an overflowing capture is in a branch that is subsequently
691 abandoned after a backtrack.
692
693 13. In the pathological case when an offset vector of size 2 is used, pcretest
694 now prints out the matched string after a yield of 0 or 1.
695
696 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
697 Only supported by the JIT compiler at the moment.
698
699 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
700
701 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
702 offset of the starting point of the matching process, provided the offsets
703 vector is large enough.
704
705 17. The \A escape now records a lookbehind value of 1, though its execution
706 does not actually inspect the previous character. This is to ensure that,
707 in partial multi-segment matching, at least one character from the old
708 segment is retained when a new segment is processed. Otherwise, if there
709 are no lookbehinds in the pattern, \A might match incorrectly at the start
710 of a new segment.
711
712 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
713
714 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
715 gives some modest performance improvement in 8-bit mode.
716
717 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
718 be expressed in certain programming languages using Universal Character
719 Names.
720
721 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
722 which points out that "non characters" are not "characters that may not
723 appear in Unicode strings" but rather "characters that are reserved for
724 internal use and have only local meaning".
725
726 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
727 there was a conditional group that depended on an assertion, if the
728 assertion was false, the callout that immediately followed the alternation
729 in the condition was skipped when pcre_exec() was used for matching.
730
731 23. Allow an explicit callout to be inserted before an assertion that is the
732 condition for a conditional group, for compatibility with automatic
733 callouts, which always insert a callout at this point.
734
735 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
736 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
737
738 25. (*PRUNE) is now supported by the JIT compiler.
739
740 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
741
742 27. Fix the case where there are two or more SKIPs with arguments that may be
743 ignored.
744
745 28. (*SKIP) is now supported by the JIT compiler.
746
747 29. (*THEN) is now supported by the JIT compiler.
748
749 30. Update RunTest with additional test selector options.
750
751 31. The way PCRE handles backtracking verbs has been changed in two ways.
752
753 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
754 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
755 backtracking. In some cases this makes it more Perl-compatible, but Perl's
756 rather obscure rules do not always do the same thing.
757
758 (2) Previously, backtracking verbs were confined within assertions. This is
759 no longer the case for positive assertions, except for (*ACCEPT). Again,
760 this sometimes improves Perl compatibility, and sometimes does not.
761
762 32. A number of tests that were in test 2 because Perl did things differently
763 have been moved to test 1, because either Perl or PCRE has changed, and
764 these tests are now compatible.
765
766 32. Backtracking control verbs are now handled in the same way in JIT and
767 interpreter.
768
769 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
770 contained a forward subroutine reference caused a compile error.
771
772 34. Auto-detect and optimize limited repetitions in JIT.
773
774 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
775 blocking (*UTF) etc.
776
777 36. In the interpreter, maximizing pattern repetitions for characters and
778 character types now use tail recursion, which reduces stack usage.
779
780 37. The value of the max lookbehind was not correctly preserved if a compiled
781 and saved regex was reloaded on a host of different endianness.
782
783 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
784 of the compiled pattern block, expand the flags field from 16 to 32 bits
785 because it was almost full.
786
787 39. Try madvise first before posix_madvise.
788
789 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
790 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
791 original change locked out all matches of empty strings. This has been
792 changed so that one match of an empty string per line is recognized.
793 Subsequent searches on the same line (for colouring or for --only-matching,
794 for example) do not recognize empty strings.
795
796 41. Applied a user patch to fix a number of spelling mistakes in comments.
797
798 42. Data lines longer than 65536 caused pcretest to crash.
799
800 43. Clarified the data type for length and startoffset arguments for pcre_exec
801 and pcre_dfa_exec in the function-specific man pages, where they were
802 explicitly stated to be in bytes, never having been updated. I also added
803 some clarification to the pcreapi man page.
804
805 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
806 a segmentation fault.
807
808
809 Version 8.32 30-November-2012
810 -----------------------------
811
812 1. Improved JIT compiler optimizations for first character search and single
813 character iterators.
814
815 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
816 Patch by Daniel Richard G.
817
818 3. Single character iterator optimizations in the JIT compiler.
819
820 4. Improved JIT compiler optimizations for character ranges.
821
822 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
823 Reported by Giuseppe D'Angelo.
824
825 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
826 of a line, was being set incorrectly in cases where .* appeared inside
827 atomic brackets at the start of a pattern, or where there was a subsequent
828 *PRUNE or *SKIP.
829
830 7. Improved instruction cache flush for POWER/PowerPC.
831 Patch by Daniel Richard G.
832
833 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
834 grep:
835
836 (a) There is now no limit to the number of patterns to be matched.
837
838 (b) An error is given if a pattern is too long.
839
840 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
841 are now supported.
842
843 (d) --exclude-from and --include-from (multiple use) have been added.
844
845 (e) Exclusions and inclusions now apply to all files and directories, not
846 just to those obtained from scanning a directory recursively.
847
848 (f) Multiple uses of -f and --file-list are now supported.
849
850 (g) In a Windows environment, the default for -d has been changed from
851 "read" (the GNU grep default) to "skip", because otherwise the presence
852 of a directory in the file list provokes an error.
853
854 (h) The documentation has been revised and clarified in places.
855
856 9. Improve the matching speed of capturing brackets.
857
858 10. Changed the meaning of \X so that it now matches a Unicode extended
859 grapheme cluster.
860
861 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
862 out POSIX threads when JIT support is configured.
863
864 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
865
866 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
867 structure could go wrong in environments where size_t is not the same size
868 as int.
869
870 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
871 set.
872
873 15. The EBCDIC support had decayed; later updates to the code had included
874 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
875 general tidy up of EBCDIC-related issues, and the documentation was also
876 not quite right. There is now a test that can be run on ASCII systems to
877 check some of the EBCDIC-related things (but is it not a full test).
878
879 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
880 in a small tidy to the code.
881
882 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
883
884 18. If the --only-matching (-o) option in pcregrep is specified multiple
885 times, each one causes appropriate output. For example, -o1 -o2 outputs the
886 substrings matched by the 1st and 2nd capturing parentheses. A separating
887 string can be specified by --om-separator (default empty).
888
889 19. Improving the first n character searches.
890
891 20. Turn case lists for horizontal and vertical white space into macros so that
892 they are defined only once.
893
894 21. This set of changes together give more compatible Unicode case-folding
895 behaviour for characters that have more than one other case when UCP
896 support is available.
897
898 (a) The Unicode property table now has offsets into a new table of sets of
899 three or more characters that are case-equivalent. The MultiStage2.py
900 script that generates these tables (the pcre_ucd.c file) now scans
901 CaseFolding.txt instead of UnicodeData.txt for character case
902 information.
903
904 (b) The code for adding characters or ranges of characters to a character
905 class has been abstracted into a generalized function that also handles
906 case-independence. In UTF-mode with UCP support, this uses the new data
907 to handle characters with more than one other case.
908
909 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
910 whose other case is greater than 256 are now correctly matched
911 caselessly. Previously, the high codepoint matched the low one, but not
912 vice versa.
913
914 (d) The processing of \h, \H, \v, and \ in character classes now makes use
915 of the new class addition function, using character lists defined as
916 macros alongside the case definitions of 20 above.
917
918 (e) Caseless back references now work with characters that have more than
919 one other case.
920
921 (f) General caseless matching of characters with more than one other case
922 is supported.
923
924 22. Unicode character properties were updated from Unicode 6.2.0
925
926 23. Improved CMake support under Windows. Patch by Daniel Richard G.
927
928 24. Add support for 32-bit character strings, and UTF-32
929
930 25. Major JIT compiler update (code refactoring and bugfixing).
931 Experimental Sparc 32 support is added.
932
933 26. Applied a modified version of Daniel Richard G's patch to create
934 pcre.h.generic and config.h.generic by "make" instead of in the
935 PrepareRelease script.
936
937 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
938 pcre_compile.c when checking for a zero character.
939
940 28. Introducing a native interface for JIT. Through this interface, the compiled
941 machine code can be directly executed. The purpose of this interface is to
942 provide fast pattern matching, so several sanity checks are not performed.
943 However, feature tests are still performed. The new interface provides
944 1.4x speedup compared to the old one.
945
946 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
947 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
948 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
949
950 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
951 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
952 the "old" RFC 2279). Instead, it ended up passing a negative length to
953 pcre_exec().
954
955 31. Add support for GCC's visibility feature to hide internal functions.
956
957 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
958 "unknown -C option" after outputting 0 or 1.
959
960 33. There is now support for generating a code coverage report for the test
961 suite in environments where gcc is the compiler and lcov is installed. This
962 is mainly for the benefit of the developers.
963
964 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
965 unaddressable using valgrind annotations, allowing valgrind to detect
966 invalid memory accesses. This is mainly for the benefit of the developers.
967
968 25. (*UTF) can now be used to start a pattern in any of the three libraries.
969
970 26. Give configure error if --enable-cpp but no C++ compiler found.
971
972
973 Version 8.31 06-July-2012
974 -------------------------
975
976 1. Fixing a wrong JIT test case and some compiler warnings.
977
978 2. Removed a bashism from the RunTest script.
979
980 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
981 to unsigned type, result still unsigned" that was given by an MS compiler
982 on encountering the code "-sizeof(xxx)".
983
984 4. Partial matching support is added to the JIT compiler.
985
986 5. Fixed several bugs concerned with partial matching of items that consist
987 of more than one character:
988
989 (a) /^(..)\1/ did not partially match "aba" because checking references was
990 done on an "all or nothing" basis. This also applied to repeated
991 references.
992
993 (b) \R did not give a hard partial match if \r was found at the end of the
994 subject.
995
996 (c) \X did not give a hard partial match after matching one or more
997 characters at the end of the subject.
998
999 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1000 a partial match for the string "\r".
1001
1002 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1003 a partial match for a CR character at the end of the subject string.
1004
1005 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1006 running pcretest, the text "(JIT)" added to the output whenever JIT is
1007 actually used to run the match.
1008
1009 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1010 or /S+[+] with a digit between 1 and 7.
1011
1012 8. OP_NOT now supports any UTF character not just single-byte ones.
1013
1014 9. (*MARK) control verb is now supported by the JIT compiler.
1015
1016 10. The command "./RunTest list" lists the available tests without actually
1017 running any of them. (Because I keep forgetting what they all are.)
1018
1019 11. Add PCRE_INFO_MAXLOOKBEHIND.
1020
1021 12. Applied a (slightly modified) user-supplied patch that improves performance
1022 when the heap is used for recursion (compiled with --disable-stack-for-
1023 recursion). Instead of malloc and free for each heap frame each time a
1024 logical recursion happens, frames are retained on a chain and re-used where
1025 possible. This sometimes gives as much as 30% improvement.
1026
1027 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1028 call.
1029
1030 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1031
1032 15. It is now possible to link pcretest with libedit as an alternative to
1033 libreadline.
1034
1035 16. (*COMMIT) control verb is now supported by the JIT compiler.
1036
1037 17. The Unicode data tables have been updated to Unicode 6.1.0.
1038
1039 18. Added --file-list option to pcregrep.
1040
1041 19. Added binary file support to pcregrep, including the -a, --binary-files,
1042 -I, and --text options.
1043
1044 20. The madvise function is renamed for posix_madvise for QNX compatibility
1045 reasons. Fixed by Giuseppe D'Angelo.
1046
1047 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1048
1049 22. Changed the option for creating symbolic links for 16-bit man pages from
1050 -s to -sf so that re-installing does not cause issues.
1051
1052 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1053
1054 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1055 matches in certain environments (the workspace was not being correctly
1056 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1057 some of the workspace data at the beginning of a restart.
1058
1059 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1060 was not doing so when it should - probably a typo introduced by SVN 528
1061 (change 8.10/14).
1062
1063 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1064 \w+ when the character tables indicated that \x{c4} was a word character.
1065 There were several related cases, all because the tests for doing a table
1066 lookup were testing for characters less than 127 instead of 255.
1067
1068 27. If a pattern contains capturing parentheses that are not used in a match,
1069 their slots in the ovector are set to -1. For those that are higher than
1070 any matched groups, this happens at the end of processing. In the case when
1071 there were back references that the ovector was too small to contain
1072 (causing temporary malloc'd memory to be used during matching), and the
1073 highest capturing number was not used, memory off the end of the ovector
1074 was incorrectly being set to -1. (It was using the size of the temporary
1075 memory instead of the true size.)
1076
1077 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1078 ovector size, it uses memory at the end of the block that it has got.
1079
1080 29. Check for an overlong MARK name and give an error at compile time. The
1081 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1082
1083 30. JIT compiler update.
1084
1085 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1086 Rill for the patch.
1087
1088 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1089
1090 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1091
1092 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1093 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1094 was enabled.
1095
1096 35. Improve JIT code generation for greedy plus quantifier.
1097
1098 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1099 1 to "aa" instead of to an empty string. The bug affected repeated groups
1100 that could potentially match an empty string.
1101
1102 37. Optimizing single character iterators in JIT.
1103
1104 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1105 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1106 codepoints that are too big for the mode are faulted, and in a UTF mode,
1107 disallowed codepoints are also faulted.
1108
1109 39. If PCRE was compiled with UTF support, in three places in the DFA
1110 matcher there was code that should only have been obeyed in UTF mode, but
1111 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1112 processing when bytes with values greater than 127 were present. In 16-bit
1113 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1114 both cases the values are those that cannot be the first data item in a UTF
1115 character. The three items that might have provoked this were recursions,
1116 possessively repeated groups, and atomic groups.
1117
1118 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1119 and pcregrep, because some OS require shared objects to be explicitly
1120 passed to ld, causing the link step to fail if they are not.
1121
1122 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1123 mode, patterns that started with \h* or \R* might be incorrectly matched.
1124
1125
1126 Version 8.30 04-February-2012
1127 -----------------------------
1128
1129 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1130 name is defined in ctype.h.
1131
1132 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1133 only in quite long subpatterns.
1134
1135 3. Removed the function pcre_info(), which has been obsolete and deprecated
1136 since it was replaced by pcre_fullinfo() in February 2000.
1137
1138 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1139 match a (*MARK), and the match failed at the start of the subject, a
1140 reference to memory before the start of the subject could occur. This bug
1141 was introduced by fix 17 of release 8.21.
1142
1143 5. A reference to an unset group with zero minimum repetition was giving
1144 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1145 /(another)?(\1?)test/ matched against "hello world test". This bug was
1146 introduced in release 8.13.
1147
1148 6. Add support for 16-bit character strings (a large amount of work involving
1149 many changes and refactorings).
1150
1151 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1152 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1153 from a file.
1154
1155 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1156 rounding is not applied in this particular case).
1157
1158 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1159 if they appear, or are escaped, in patterns.
1160
1161 10. Get rid of a number of -Wunused-but-set-variable warnings.
1162
1163 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1164 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1165 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1166 also returns the mark "x". This bug applied to capturing parentheses,
1167 non-capturing parentheses, and atomic parentheses. It also applied to some
1168 assertions.
1169
1170 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1171 information out of configure.ac instead of relying on pcre.h.generic, which
1172 is not stored in the repository.
1173
1174 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1175 -lreadline.
1176
1177 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1178
1179 15. Applied Graycode's patch to put the top-level frame on the stack rather
1180 than the heap when not using the stack for recursion. This gives a
1181 performance improvement in many cases when recursion is not deep.
1182
1183 16. Experimental code added to "pcretest -C" to output the stack frame size.
1184
1185
1186 Version 8.21 12-Dec-2011
1187 ------------------------
1188
1189 1. Updating the JIT compiler.
1190
1191 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1192 are added as well.
1193
1194 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1195 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1196 calling _pcre_jit_exec. Some extra comments are added.
1197
1198 4. (*MARK) settings inside atomic groups that do not contain any capturing
1199 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1200 was introduced by change 18 for 8.20.
1201
1202 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1203 ECMA-262 standard.
1204
1205 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1206 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1207 This bug was probably introduced by change 9 of 8.13.
1208
1209 7. While fixing 6 above, I noticed that a number of other items were being
1210 incorrectly rejected as "not fixed length". This arose partly because newer
1211 opcodes had not been added to the fixed-length checking code. I have (a)
1212 corrected the bug and added tests for these items, and (b) arranged for an
1213 error to occur if an unknown opcode is encountered while checking for fixed
1214 length instead of just assuming "not fixed length". The items that were
1215 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1216 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1217 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1218
1219 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1220 being incorrectly compiled and would have given unpredicatble results.
1221
1222 9. A possessively repeated subpattern with minimum repeat count greater than
1223 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1224 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1225 the first (A) could occur when it should not.
1226
1227 10. Add a cast and remove a redundant test from the code.
1228
1229 11. JIT should use pcre_malloc/pcre_free for allocation.
1230
1231 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1232 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1233 is anything other than /usr, -L is still shown).
1234
1235 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1236
1237 14. Perl does not support \N without a following name in a [] class; PCRE now
1238 also gives an error.
1239
1240 15. If a forward reference was repeated with an upper limit of around 2000,
1241 it caused the error "internal error: overran compiling workspace". The
1242 maximum number of forward references (including repeats) was limited by the
1243 internal workspace, and dependent on the LINK_SIZE. The code has been
1244 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1245 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1246 of around 200,000 forward references. While doing this, I also speeded up
1247 the filling in of repeated forward references.
1248
1249 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1250 incorrectly expecting the subject to contain another "a" after the start.
1251
1252 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1253 in the match, the SKIP should be ignored. This was not happening; instead
1254 the SKIP was being treated as NOMATCH. For patterns such as
1255 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1256 tested.
1257
1258 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1259 now much more compatible with Perl, in particular in cases where the result
1260 is a non-match for a non-anchored pattern. For example, if
1261 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1262 "m", where previously it did not return a name. A side effect of this
1263 change is that for partial matches, the last encountered mark name is
1264 returned, as for non matches. A number of tests that were previously not
1265 Perl-compatible have been moved into the Perl-compatible test files. The
1266 refactoring has had the pleasing side effect of removing one argument from
1267 the match() function, thus reducing its stack requirements.
1268
1269 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1270 subsequent uses of /S (without +) incorrectly behaved like /S+.
1271
1272 21. Retrieve executable code size support for the JIT compiler and fixing
1273 some warnings.
1274
1275 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1276 not work when the shorter character appeared right at the end of the
1277 subject string.
1278
1279 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1280 systems.
1281
1282 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1283 output it when the /M option is used in pcretest.
1284
1285 25. The CheckMan script was not being included in the distribution. Also, added
1286 an explicit "perl" to run Perl scripts from the PrepareRelease script
1287 because this is reportedly needed in Windows.
1288
1289 26. If study data was being save in a file and studying had not found a set of
1290 "starts with" bytes for the pattern, the data written to the file (though
1291 never used) was taken from uninitialized memory and so caused valgrind to
1292 complain.
1293
1294 27. Updated RunTest.bat as provided by Sheri Pierce.
1295
1296 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1297
1298 29. Computation of memory usage for the table of capturing group names was
1299 giving an unnecessarily large value.
1300
1301
1302 Version 8.20 21-Oct-2011
1303 ------------------------
1304
1305 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1306 a POSIX class. After further experiments with Perl, which convinced me that
1307 Perl has bugs and confusions, a closing square bracket is no longer allowed
1308 in a POSIX name. This bug also affected patterns with classes that started
1309 with full stops.
1310
1311 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1312 captured substring, but while checking the failing first alternative,
1313 substring 1 is temporarily captured. If the output vector supplied to
1314 pcre_exec() was not big enough for this capture, the yield of the function
1315 was still zero ("insufficient space for captured substrings"). This cannot
1316 be totally fixed without adding another stack variable, which seems a lot
1317 of expense for a edge case. However, I have improved the situation in cases
1318 such as /(a)(b)x|abc/ matched against "abc", where the return code
1319 indicates that fewer than the maximum number of slots in the ovector have
1320 been set.
1321
1322 3. Related to (2) above: when there are more back references in a pattern than
1323 slots in the output vector, pcre_exec() uses temporary memory during
1324 matching, and copies in the captures as far as possible afterwards. It was
1325 using the entire output vector, but this conflicts with the specification
1326 that only 2/3 is used for passing back captured substrings. Now it uses
1327 only the first 2/3, for compatibility. This is, of course, another edge
1328 case.
1329
1330 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1331 main code base, and can be used by building with --enable-jit. When this is
1332 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1333 runtime --no-jit option is given.
1334
1335 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1336 ovector, the return from the function was zero, implying that there were
1337 other matches that did not fit. The correct "exactly full" value is now
1338 returned.
1339
1340 6. If a subpattern that was called recursively or as a subroutine contained
1341 (*PRUNE) or any other control that caused it to give a non-standard return,
1342 invalid errors such as "Error -26 (nested recursion at the same subject
1343 position)" or even infinite loops could occur.
1344
1345 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1346 computing the minimum length on reaching *ACCEPT, and so ended up with the
1347 wrong value of 1 rather than 0. Further investigation indicates that
1348 computing a minimum subject length in the presence of *ACCEPT is difficult
1349 (think back references, subroutine calls), and so I have changed the code
1350 so that no minimum is registered for a pattern that contains *ACCEPT.
1351
1352 8. If (*THEN) was present in the first (true) branch of a conditional group,
1353 it was not handled as intended. [But see 16 below.]
1354
1355 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1356 Sheri Pierce.
1357
1358 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1359 the first byte in a match must be "a".
1360
1361 11. Change 17 for 8.13 increased the recursion depth for patterns like
1362 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1363 pattern contains any instances of (*THEN). If it does not, the old
1364 optimizations are restored. It would be nice to do this on a per-group
1365 basis, but at the moment that is not feasible.
1366
1367 12. In some environments, the output of pcretest -C is CRLF terminated. This
1368 broke RunTest's code that checks for the link size. A single white space
1369 character after the value is now allowed for.
1370
1371 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1372 For "fr", it uses the Windows-specific input and output files.
1373
1374 14. If (*THEN) appeared in a group that was called recursively or as a
1375 subroutine, it did not work as intended. [But see next item.]
1376
1377 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1378 pattern fragments (but not containing any | characters). If A and B are
1379 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1380 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1381 D. In other words, Perl considers parentheses that do not contain any |
1382 characters to be part of a surrounding alternative, whereas PCRE was
1383 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1384 differently. PCRE now behaves in the same way as Perl, except in the case
1385 of subroutine/recursion calls such as (?1) which have in any case always
1386 been different (but PCRE had them first :-).
1387
1388 16. Related to 15 above: Perl does not treat the | in a conditional group as
1389 creating alternatives. Such a group is treated in the same way as an
1390 ordinary group without any | characters when processing (*THEN). PCRE has
1391 been changed to match Perl's behaviour.
1392
1393 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1394 RunGrepTest script failed.
1395
1396 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1397 inevitable for groups that contain captures, but it can lead to a lot of
1398 stack use in large patterns. The old behaviour has been restored for atomic
1399 groups that do not contain any capturing parentheses.
1400
1401 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1402 suppress the check for a minimum subject length at run time. (If it was
1403 given to pcre_exec() or pcre_dfa_exec() it did work.)
1404
1405 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1406 fail to work when decoding hex characters in data strings in EBCDIC
1407 environments.
1408
1409 21. It appears that in at least one Mac OS environment, the isxdigit() function
1410 is implemented as a macro that evaluates to its argument more than once,
1411 contravening the C 90 Standard (I haven't checked a later standard). There
1412 was an instance in pcretest which caused it to go wrong when processing
1413 \x{...} escapes in subject strings. The has been rewritten to avoid using
1414 things like p++ in the argument of isxdigit().
1415
1416
1417 Version 8.13 16-Aug-2011
1418 ------------------------
1419
1420 1. The Unicode data tables have been updated to Unicode 6.0.0.
1421
1422 2. Two minor typos in pcre_internal.h have been fixed.
1423
1424 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1425 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1426 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1427
1428 4. There were a number of related bugs in the code for matching backrefences
1429 caselessly in UTF-8 mode when codes for the characters concerned were
1430 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1431 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1432 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1433 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1434 2-byte code at the end of the subject (it thought there wasn't enough data
1435 left).
1436
1437 5. Comprehensive information about what went wrong is now returned by
1438 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1439 as the output vector has at least 2 elements. The offset of the start of
1440 the failing character and a reason code are placed in the vector.
1441
1442 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1443 now returned is for the first byte of the failing character, instead of the
1444 last byte inspected. This is an incompatible change, but I hope it is small
1445 enough not to be a problem. It makes the returned offset consistent with
1446 pcre_exec() and pcre_dfa_exec().
1447
1448 7. pcretest now gives a text phrase as well as the error number when
1449 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1450 failure, the offset and reason code are output.
1451
1452 8. When \R was used with a maximizing quantifier it failed to skip backwards
1453 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1454 back over a single character (\n). This seems wrong (because it treated the
1455 two characters as a single entity when going forwards), conflicts with the
1456 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1457 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1458 has been changed.
1459
1460 9. Some internal refactoring has changed the processing so that the handling
1461 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1462 time (the PCRE_DOTALL option was changed this way some time ago: version
1463 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1464 which was always a bit of a fudge. It also means that there is one less
1465 argument for the match() function, which reduces its stack requirements
1466 slightly. This change also fixes an incompatibility with Perl: the pattern
1467 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1468
1469 10. More internal refactoring has drastically reduced the number of recursive
1470 calls to match() for possessively repeated groups such as (abc)++ when
1471 using pcre_exec().
1472
1473 11. While implementing 10, a number of bugs in the handling of groups were
1474 discovered and fixed:
1475
1476 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1477 (a|)*(?1) gave a compile-time internal error.
1478 ((a|)+)+ did not notice that the outer group could match an empty string.
1479 (^a|^)+ was not marked as anchored.
1480 (.*a|.*)+ was not marked as matching at start or after a newline.
1481
1482 12. Yet more internal refactoring has removed another argument from the match()
1483 function. Special calls to this function are now indicated by setting a
1484 value in a variable in the "match data" data block.
1485
1486 13. Be more explicit in pcre_study() instead of relying on "default" for
1487 opcodes that mean there is no starting character; this means that when new
1488 ones are added and accidentally left out of pcre_study(), testing should
1489 pick them up.
1490
1491 14. The -s option of pcretest has been documented for ages as being an old
1492 synonym of -m (show memory usage). I have changed it to mean "force study
1493 for every regex", that is, assume /S for every regex. This is similar to -i
1494 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1495 using it. It makes it easier to run collections of tests with and without
1496 study enabled, and thereby test pcre_study() more easily. All the standard
1497 tests are now run with and without -s (but some patterns can be marked as
1498 "never study" - see 20 below).
1499
1500 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1501 restoration of the capturing data to the outer values was not happening
1502 correctly.
1503
1504 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1505 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1506 pattern had matched an empty string, and so incorrectly returned a no
1507 match.
1508
1509 17. There was optimizing code for the last branch of non-capturing parentheses,
1510 and also for the obeyed branch of a conditional subexpression, which used
1511 tail recursion to cut down on stack usage. Unfortunately, now that there is
1512 the possibility of (*THEN) occurring in these branches, tail recursion is
1513 no longer possible because the return has to be checked for (*THEN). These
1514 two optimizations have therefore been removed. [But see 8.20/11 above.]
1515
1516 18. If a pattern containing \R was studied, it was assumed that \R always
1517 matched two bytes, thus causing the minimum subject length to be
1518 incorrectly computed because \R can also match just one byte.
1519
1520 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1521 was incorrectly computed.
1522
1523 20. If /S is present twice on a test pattern in pcretest input, it now
1524 *disables* studying, thereby overriding the use of -s on the command line
1525 (see 14 above). This is necessary for one or two tests to keep the output
1526 identical in both cases.
1527
1528 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1529 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1530
1531 22. When an atomic group that contained a capturing parenthesis was
1532 successfully matched, but the branch in which it appeared failed, the
1533 capturing was not being forgotten if a higher numbered group was later
1534 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1535 group 1 to "a", when in fact it should be unset. This applied to multi-
1536 branched capturing and non-capturing groups, repeated or not, and also to
1537 positive assertions (capturing in negative assertions does not happen
1538 in PCRE) and also to nested atomic groups.
1539
1540 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1541 subject after a captured substring, to make it easier to tell which of a
1542 number of identical substrings has been captured.
1543
1544 24. The way atomic groups are processed by pcre_exec() has been changed so that
1545 if they are repeated, backtracking one repetition now resets captured
1546 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1547 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1548 "aaa". Previously, it would have been "a". As part of this code
1549 refactoring, the way recursive calls are handled has also been changed.
1550
1551 25. If an assertion condition captured any substrings, they were not passed
1552 back unless some other capturing happened later. For example, if
1553 (?(?=(a))a) was matched against "a", no capturing was returned.
1554
1555 26. When studying a pattern that contained subroutine calls or assertions,
1556 the code for finding the minimum length of a possible match was handling
1557 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1558 group 1 called group 2 while simultaneously a separate group 2 called group
1559 1). A stack overflow occurred in this case. I have fixed this by limiting
1560 the recursion depth to 10.
1561
1562 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1563 Fortmann. This supports explicit test numbers on the command line, and has
1564 argument validation and error reporting.
1565
1566 28. An instance of \X with an unlimited repeat could fail if at any point the
1567 first character it looked at was a mark character.
1568
1569 29. Some minor code refactoring concerning Unicode properties and scripts
1570 should reduce the stack requirement of match() slightly.
1571
1572 30. Added the '=' option to pcretest to check the setting of unused capturing
1573 slots at the end of the pattern, which are documented as being -1, but are
1574 not included in the return count.
1575
1576 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1577 compiled something random. Now it gives a compile-time error (as does
1578 Perl).
1579
1580 32. A *MARK encountered during the processing of a positive assertion is now
1581 recorded and passed back (compatible with Perl).
1582
1583 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1584 had alternative anchored branches, the search for a second match in a line
1585 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1586 matched the line "0102" twice. The same bug affected patterns that started
1587 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1588 twice.
1589
1590 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1591 does, and because of capturing effects, quantifying parenthesized
1592 assertions may at times be useful. Quantifiers are now allowed for
1593 parenthesized assertions.
1594
1595 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1596
1597 36. \g was being checked for fancy things in a character class, when it should
1598 just be a literal "g".
1599
1600 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1601 appearance of a nested POSIX class supersedes an apparent external class.
1602 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1603 unescaped square brackets may also appear as part of class names. For
1604 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1605 more like Perl. (But see 8.20/1 above.)
1606
1607 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1608 was because it thought it was \N{name}, which is not supported).
1609
1610 39. Add minix to OS list not supporting the -S option in pcretest.
1611
1612 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1613 cannot analyze patterns in sufficient detail to catch mutual recursions
1614 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1615 subgroup is called recursively as a subpattern for a second time at the
1616 same position in the subject string. In previous releases this might have
1617 been caught by the recursion limit, or it might have run out of stack.
1618
1619 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1620 happen only once. PCRE was, however incorrectly giving a compile time error
1621 "recursive call could loop indefinitely" because it cannot analyze the
1622 pattern in sufficient detail. The compile time test no longer happens when
1623 PCRE is compiling a conditional subpattern, but actual runaway loops are
1624 now caught at runtime (see 40 above).
1625
1626 42. It seems that Perl allows any characters other than a closing parenthesis
1627 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1628 has been changed to be the same.
1629
1630 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1631 as not to get warnings when autogen.sh is called. Also changed
1632 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1633
1634 44. To help people who use pcregrep to scan files containing exceedingly long
1635 lines, the following changes have been made:
1636
1637 (a) The default value of the buffer size parameter has been increased from
1638 8K to 20K. (The actual buffer used is three times this size.)
1639
1640 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1641 PCRE is built.
1642
1643 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1644 to be set at run time.
1645
1646 (d) Numerical values in pcregrep options can be followed by K or M, for
1647 example --buffer-size=50K.
1648
1649 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1650 given and the return code is set to 2.
1651
1652 45. Add a pointer to the latest mark to the callout data block.
1653
1654 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1655 partial match of an empty string instead of no match. This was specific to
1656 the use of ".".
1657
1658 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1659 complete match instead of a partial match. This bug was dependent on both
1660 the PCRE_UTF8 and PCRE_DOTALL options being set.
1661
1662 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1663 starting byte set, because \b was not being ignored.
1664
1665
1666 Version 8.12 15-Jan-2011
1667 ------------------------
1668
1669 1. Fixed some typos in the markup of the man pages, and wrote a script that
1670 checks for such things as part of the documentation building process.
1671
1672 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1673 --match-limit and --recursion-limit options (added for 8.11). In
1674 particular, this made one of the standard tests fail. (The integer value
1675 went into the wrong half of a long int.)
1676
1677 3. If the --colour option was given to pcregrep with -v (invert match), it
1678 did strange things, either producing crazy output, or crashing. It should,
1679 of course, ignore a request for colour when reporting lines that do not
1680 match.
1681
1682 4. Another pcregrep bug caused similar problems if --colour was specified with
1683 -M (multiline) and the pattern match finished with a line ending.
1684
1685 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1686 matched in multiline mode, the following line was shown as part of the
1687 match. This seems wrong, so I have changed it.
1688
1689 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1690 the check for further matches in the same line (so they could be coloured)
1691 to overrun the end of the current line. If another match was found, it was
1692 incorrectly shown (and then shown again when found in the next line).
1693
1694 7. If pcregrep was compiled under Windows, there was a reference to the
1695 function pcregrep_exit() before it was defined. I am assuming this was
1696 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1697 reported by a user. I've moved the definition above the reference.
1698
1699
1700 Version 8.11 10-Dec-2010
1701 ------------------------
1702
1703 1. (*THEN) was not working properly if there were untried alternatives prior
1704 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1705 backtracked to try for "b" instead of moving to the next alternative branch
1706 at the same level (in this case, to look for "c"). The Perl documentation
1707 is clear that when (*THEN) is backtracked onto, it goes to the "next
1708 alternative in the innermost enclosing group".
1709
1710 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1711 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1712 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1713 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1714 (*THEN).
1715
1716 3. If \s appeared in a character class, it removed the VT character from
1717 the class, even if it had been included by some previous item, for example
1718 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1719 of \s, but is part of the POSIX "space" class.)
1720
1721 4. A partial match never returns an empty string (because you can always
1722 match an empty string at the end of the subject); however the checking for
1723 an empty string was starting at the "start of match" point. This has been
1724 changed to the "earliest inspected character" point, because the returned
1725 data for a partial match starts at this character. This means that, for
1726 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1727 (previously it gave "no match").
1728
1729 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1730 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1731 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1732 has an implication that the given string is incomplete (because a partial
1733 match is preferred over a full match). For this reason, these items now
1734 give a partial match in this situation. [Aside: previously, the one case
1735 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1736 match rather than a full match, which was wrong by the old rules, but is
1737 now correct.]
1738
1739 6. There was a bug in the handling of #-introduced comments, recognized when
1740 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1741 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1742 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1743 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1744 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1745 places in pcre_compile().
1746
1747 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1748 comments when looking ahead for named forward references to subpatterns,
1749 the only newline sequence it recognized was NL. It now handles newlines
1750 according to the set newline convention.
1751
1752 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1753 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1754 cater for a lack of strerror(). These oversights have been fixed.
1755
1756 9. Added --match-limit and --recursion-limit to pcregrep.
1757
1758 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1759
1760 11. When the -o option was used, pcregrep was setting a return code of 1, even
1761 when matches were found, and --line-buffered was not being honoured.
1762
1763 12. Added an optional parentheses number to the -o and --only-matching options
1764 of pcregrep.
1765
1766 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1767 can match an empty string. The code to do it in pcretest and pcredemo
1768 needed fixing:
1769
1770 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1771 only one byte after an empty string match just before CRLF (this case
1772 just got forgotten; "any" and "anycrlf" were OK).
1773
1774 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1775 mode when an empty string match preceded an ASCII character followed by
1776 a non-ASCII character. (The code for advancing by one character rather
1777 than one byte was nonsense.)
1778
1779 (c) The pcredemo.c sample program did not have any code at all to handle
1780 the cases when CRLF is a valid newline sequence.
1781
1782 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1783 as a starting offset was within the subject string. There is now a new
1784 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1785 negative or greater than the length of the string. In order to test this,
1786 pcretest is extended to allow the setting of negative starting offsets.
1787
1788 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1789 starting offset points to the beginning of a UTF-8 character was
1790 unnecessarily clumsy. I tidied it up.
1791
1792 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1793 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1794
1795 17. Nobody had reported that the --include_dir option, which was added in
1796 release 7.7 should have been called --include-dir (hyphen, not underscore)
1797 for compatibility with GNU grep. I have changed it to --include-dir, but
1798 left --include_dir as an undocumented synonym, and the same for
1799 --exclude-dir, though that is not available in GNU grep, at least as of
1800 release 2.5.4.
1801
1802 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1803 characters from a string of bytes) have been redefined so as not to use
1804 loops, in order to improve performance in some environments. At the same
1805 time, I abstracted some of the common code into auxiliary macros to save
1806 repetition (this should not affect the compiled code).
1807
1808 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1809 compile-time error is now given if \c is not followed by an ASCII
1810 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1811 different, and any byte value is allowed.)
1812
1813 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1814 START_OPTIMIZE option, which is now allowed at compile time - but just
1815 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1816 to pcregrep and other applications that have no direct access to PCRE
1817 options. The new /Y option in pcretest sets this option when calling
1818 pcre_compile().
1819
1820 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1821 back references. Groups containing recursive back references were forced to
1822 be atomic by that change, but in the case of named groups, the amount of
1823 memory required was incorrectly computed, leading to "Failed: internal
1824 error: code overflow". This has been fixed.
1825
1826 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1827 pcretest.c, to avoid build problems in some Borland environments.
1828
1829
1830 Version 8.10 25-Jun-2010
1831 ------------------------
1832
1833 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1834 THEN.
1835
1836 2. (*ACCEPT) was not working when inside an atomic group.
1837
1838 3. Inside a character class, \B is treated as a literal by default, but
1839 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1840 causes the error). The code is unchanged, but I tidied the documentation.
1841
1842 4. Inside a character class, PCRE always treated \R and \X as literals,
1843 whereas Perl faults them if its -w option is set. I have changed PCRE so
1844 that it faults them when PCRE_EXTRA is set.
1845
1846 5. Added support for \N, which always matches any character other than
1847 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1848
1849 6. When compiling pcregrep with newer versions of gcc which may have
1850 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1851 declared with attribute warn_unused_result" were given. Just casting the
1852 result to (void) does not stop the warnings; a more elaborate fudge is
1853 needed. I've used a macro to implement this.
1854
1855 7. Minor change to pcretest.c to avoid a compiler warning.
1856
1857 8. Added four artifical Unicode properties to help with an option to make
1858 \s etc use properties (see next item). The new properties are: Xan
1859 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1860
1861 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1862 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1863 this option. Modified pcretest to add /W to test this facility. Added
1864 REG_UCP to make it available via the POSIX interface.
1865
1866 10. Added --line-buffered to pcregrep.
1867
1868 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1869 studied, and the match started with a letter with a code point greater than
1870 127 whose first byte was different to the first byte of the other case of
1871 the letter, the other case of this starting letter was not recognized
1872 (#976).
1873
1874 12. If a pattern that was studied started with a repeated Unicode property
1875 test, for example, \p{Nd}+, there was the theoretical possibility of
1876 setting up an incorrect bitmap of starting bytes, but fortunately it could
1877 not have actually happened in practice until change 8 above was made (it
1878 added property types that matched character-matching opcodes).
1879
1880 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1881 possible starting bytes for non-anchored patterns.
1882
1883 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1884 \R, and also a number of cases that involve Unicode properties, both
1885 explicit and implicit when PCRE_UCP is set.
1886
1887 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1888 input, it could crash or give wrong results if characters with values
1889 greater than 0xc0 were present in the subject string. (Detail: it assumed
1890 UTF-8 input when processing these items.)
1891
1892 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1893 size_t is 64-bit (#991).
1894
1895 17. Added a check for running out of memory when PCRE is compiled with
1896 --disable-stack-for-recursion (#990).
1897
1898 18. If the last data line in a file for pcretest does not have a newline on
1899 the end, a newline was missing in the output.
1900
1901 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1902 less than 128) in its various bitmaps. However, there is a facility for
1903 generating tables according to the current locale when PCRE is compiled. It
1904 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1905 characters, are recognized by isspace() and therefore were getting set in
1906 these tables, and indeed these tables seem to approximate to ISO 8859. This
1907 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1908 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1909 which of course cannot start UTF-8 characters. I have changed the code so
1910 that only real ASCII characters (less than 128) and the correct starting
1911 bytes for UTF-8 encodings are set for characters greater than 127 when in
1912 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1913 altogether.)
1914
1915 20. Added the /T option to pcretest so as to be able to run tests with non-
1916 standard character tables, thus making it possible to include the tests
1917 used for 19 above in the standard set of tests.
1918
1919 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1920 reference to a subpattern the other side of a comment that contains an
1921 opening parenthesis caused either an internal compiling error, or a
1922 reference to the wrong subpattern.
1923
1924
1925 Version 8.02 19-Mar-2010
1926 ------------------------
1927
1928 1. The Unicode data tables have been updated to Unicode 5.2.0.
1929
1930 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1931 configured.
1932
1933 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1934 original author of that file, following a query about its status.
1935
1936 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1937 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1938
1939 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1940 quantifier applied to a forward-referencing subroutine call, could compile
1941 incorrect code or give the error "internal error: previously-checked
1942 referenced subpattern not found".
1943
1944 6. Both MS Visual Studio and Symbian OS have problems with initializing
1945 variables to point to external functions. For these systems, therefore,
1946 pcre_malloc etc. are now initialized to local functions that call the
1947 relevant global functions.
1948
1949 7. There were two entries missing in the vectors called coptable and poptable
1950 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1951 I've fixed the data, and added a kludgy way of testing at compile time that
1952 the lengths are correct (equal to the number of opcodes).
1953
1954 8. Following on from 7, I added a similar kludge to check the length of the
1955 eint vector in pcreposix.c.
1956
1957 9. Error texts for pcre_compile() are held as one long string to avoid too
1958 much relocation at load time. To find a text, the string is searched,
1959 counting zeros. There was no check for running off the end of the string,
1960 which could happen if a new error number was added without updating the
1961 string.
1962
1963 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1964
1965 11. \K was not working if it appeared in an atomic group or in a group that
1966 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1967 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1968 the assertion is positive, but not if it is negative.
1969
1970 12. Change 11 fortuitously reduced the size of the stack frame used in the
1971 "match()" function of pcre_exec.c by one pointer. Forthcoming
1972 implementation of support for (*MARK) will need an extra pointer on the
1973 stack; I have reserved it now, so that the stack frame size does not
1974 decrease.
1975
1976 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1977 item in branch that calls a recursion is a subroutine call - as in the
1978 second branch in the above example - was incorrectly given the compile-
1979 time error "recursive call could loop indefinitely" because pcre_compile()
1980 was not correctly checking the subroutine for matching a non-empty string.
1981
1982 14. The checks for overrunning compiling workspace could trigger after an
1983 overrun had occurred. This is a "should never occur" error, but it can be
1984 triggered by pathological patterns such as hundreds of nested parentheses.
1985 The checks now trigger 100 bytes before the end of the workspace.
1986
1987 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1988
1989
1990 Version 8.01 19-Jan-2010
1991 ------------------------
1992
1993 1. If a pattern contained a conditional subpattern with only one branch (in
1994 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1995 computed the wrong minimum data length (which is of course zero for such
1996 subpatterns). This could cause incorrect "no match" results.
1997
1998 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1999 the pattern is reset in the first branch, pcre_compile() failed with
2000 "internal error: code overflow at offset...". This happened only when
2001 the reset was to the original external option setting. (An optimization
2002 abstracts leading options settings into an external setting, which was the
2003 cause of this.)
2004
2005 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2006 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2007 assertion pattern did not match (meaning that the assertion was true), it
2008 was incorrectly treated as false if the SKIP had been reached during the
2009 matching. This also applied to assertions used as conditions.
2010
2011 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2012 assertion subpattern, including such a pattern used as a condition,
2013 unpredictable results occurred, instead of the error return
2014 PCRE_ERROR_DFA_UITEM.
2015
2016 5. The C++ GlobalReplace function was not working like Perl for the special
2017 situation when an empty string is matched. It now does the fancy magic
2018 stuff that is necessary.
2019
2020 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2021 removed. (These were left over from very, very early versions of PCRE.)
2022
2023 7. Some cosmetic changes to the code to make life easier when compiling it
2024 as part of something else:
2025
2026 (a) Change DEBUG to PCRE_DEBUG.
2027
2028 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2029 called "current" as "current_branch", to prevent a collision with the
2030 Linux macro when compiled as a kernel module.
2031
2032 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2033 prevent a collision with the Linux macro when compiled as a kernel
2034 module.
2035
2036 8. In pcre_compile() there are some checks for integer overflows that used to
2037 cast potentially large values to (double). This has been changed to that
2038 when building, a check for int64_t is made, and if it is found, it is used
2039 instead, thus avoiding the use of floating point arithmetic. (There is no
2040 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2041 double.
2042
2043 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2044 2005 (difference between two addresses compared to an unsigned value).
2045
2046 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2047 custom one, because of the following reported problem in Windows:
2048
2049 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2050 under Win32.
2051 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2052 therefore missing the function definition.
2053 - The compiler thus generates a "C" signature for the test function.
2054 - The linker fails to find the "C" function.
2055 - PCRE fails to configure if asked to do so against libbz2.
2056
2057 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2058 messages were output:
2059
2060 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2061 rerunning libtoolize, to keep the correct libtool macros in-tree.
2062 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2063
2064 I have done both of these things.
2065
2066 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2067 most of the time, it *can* run out if it is given a pattern that contains a
2068 runaway infinite recursion. I updated the discussion in the pcrestack man
2069 page.
2070
2071 13. Now that we have gone to the x.xx style of version numbers, the minor
2072 version may start with zero. Using 08 or 09 is a bad idea because users
2073 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2074 interpreted as invalid octal numbers. I've updated the previous comment in
2075 configure.ac, and also added a check that gives an error if 08 or 09 are
2076 used.
2077
2078 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2079 causing partial matching to fail when the end of the subject matched \W
2080 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2081
2082 15. There were some discrepancies between the declarations in pcre_internal.h
2083 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2084 their definitions. The declarations used "const uschar *" and the
2085 definitions used USPTR. Even though USPTR is normally defined as "const
2086 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2087 reported that: "This difference in casting confuses some C++ compilers, for
2088 example, SunCC recognizes above declarations as different functions and
2089 generates broken code for hbpcre." I have changed the declarations to use
2090 USPTR.
2091
2092 16. GNU libtool is named differently on some systems. The autogen.sh script now
2093 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2094 (FreeBSD).
2095
2096 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2097 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2098 comment: "Figure out how to create a longlong from a string: strtoll and
2099 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2100 instance, but it only takes 2 args instead of 3!"
2101
2102 18. A subtle bug concerned with back references has been fixed by a change of
2103 specification, with a corresponding code fix. A pattern such as
2104 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2105 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2106 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2107 same bug. Such groups have to be quantified to be useful, or contained
2108 inside another quantified group. (If there's no repetition, the reference
2109 can never match.) The problem arises because, having left the group and
2110 moved on to the rest of the pattern, a later failure that backtracks into
2111 the group uses the captured value from the final iteration of the group
2112 rather than the correct earlier one. I have fixed this in PCRE by forcing
2113 any group that contains a reference to itself to be an atomic group; that
2114 is, there cannot be any backtracking into it once it has completed. This is
2115 similar to recursive and subroutine calls.
2116
2117
2118 Version 8.00 19-Oct-09
2119 ----------------------
2120
2121 1. The table for translating pcre_compile() error codes into POSIX error codes
2122 was out-of-date, and there was no check on the pcre_compile() error code
2123 being within the table. This could lead to an OK return being given in
2124 error.
2125
2126 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2127 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2128 in a Windows environment.
2129
2130 3. The pcregrep --count option prints the count for each file even when it is
2131 zero, as does GNU grep. However, pcregrep was also printing all files when
2132 --files-with-matches was added. Now, when both options are given, it prints
2133 counts only for those files that have at least one match. (GNU grep just
2134 prints the file name in this circumstance, but including the count seems
2135 more useful - otherwise, why use --count?) Also ensured that the
2136 combination -clh just lists non-zero counts, with no names.
2137
2138 4. The long form of the pcregrep -F option was incorrectly implemented as
2139 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2140 but it seems right to fix it, and I didn't think it was worth preserving
2141 the old behaviour.
2142
2143 5. The command line items --regex=pattern and --regexp=pattern were not
2144 recognized by pcregrep, which required --regex pattern or --regexp pattern
2145 (with a space rather than an '='). The man page documented the '=' forms,
2146 which are compatible with GNU grep; these now work.
2147
2148 6. No libpcreposix.pc file was created for pkg-config; there was just
2149 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2150
2151 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2152 when UCP support is not needed, by modifying the Python script that
2153 generates it from Unicode data files. This should not matter if the module
2154 is correctly used as a library, but I received one complaint about 50K of
2155 unwanted data. My guess is that the person linked everything into his
2156 program rather than using a library. Anyway, it does no harm.
2157
2158 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2159 was a minimum greater than 1 for a wide character in a possessive
2160 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2161 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2162 character. Chaos in the form of incorrect output or a compiling loop could
2163 result.
2164
2165 9. The restrictions on what a pattern can contain when partial matching is
2166 requested for pcre_exec() have been removed. All patterns can now be
2167 partially matched by this function. In addition, if there are at least two
2168 slots in the offset vector, the offset of the earliest inspected character
2169 for the match and the offset of the end of the subject are set in them when
2170 PCRE_ERROR_PARTIAL is returned.
2171
2172 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2173 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2174 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2175 and may be more useful for multi-segment matching.
2176
2177 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2178 used to be given if ever the end of the subject was reached; now it is
2179 given only if matching could not proceed because another character was
2180 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2181 string "Z", which now yields "no match" instead of "partial match". In the
2182 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2183 final character ended with (*FAIL).
2184
2185 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2186 if the pattern had a "must contain" character that was already found in the
2187 earlier partial match, unless partial matching was again requested. For
2188 example, with the pattern /dog.(body)?/, the "must contain" character is
2189 "g". If the first part-match was for the string "dog", restarting with
2190 "sbody" failed. This bug has been fixed.
2191
2192 13. The string returned by pcre_dfa_exec() after a partial match has been
2193 changed so that it starts at the first inspected character rather than the
2194 first character of the match. This makes a difference only if the pattern
2195 starts with a lookbehind assertion or \b or \B (\K is not supported by
2196 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2197 matching functions compatible, and I think it's the right thing to do.
2198
2199 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2200 so that the demonstration program is easily available in environments where
2201 PCRE has not been installed from source.
2202
2203 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2204 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2205 library.
2206
2207 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2208 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2209 is not the first non-POSIX option to be added. Clearly some people find
2210 these options useful.
2211
2212 17. If a caller to the POSIX matching function regexec() passes a non-zero
2213 value for nmatch with a NULL value for pmatch, the value of
2214 nmatch is forced to zero.
2215
2216 18. RunGrepTest did not have a test for the availability of the -u option of
2217 the diff command, as RunTest does. It now checks in the same way as
2218 RunTest, and also checks for the -b option.
2219
2220 19. If an odd number of negated classes containing just a single character
2221 interposed, within parentheses, between a forward reference to a named
2222 subpattern and the definition of the subpattern, compilation crashed with
2223 an internal error, complaining that it could not find the referenced
2224 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2225 [The bug was that it was starting one character too far in when skipping
2226 over the character class, thus treating the ] as data rather than
2227 terminating the class. This meant it could skip too much.]
2228
2229 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2230 /g option in pcretest when the pattern contains \K, which makes it possible
2231 to have an empty string match not at the start, even when the pattern is
2232 anchored. Updated pcretest and pcredemo to use this option.
2233
2234 21. If the maximum number of capturing subpatterns in a recursion was greater
2235 than the maximum at the outer level, the higher number was returned, but
2236 with unset values at the outer level. The correct (outer level) value is
2237 now given.
2238
2239 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2240 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2241 make it do so. The string so far is captured, making this feature
2242 compatible with Perl.
2243
2244 23. The tests have been re-organized, adding tests 11 and 12, to make it
2245 possible to check the Perl 5.10 features against Perl 5.10.
2246
2247 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2248 pattern matches a fixed length string. PCRE did not allow this; now it
2249 does. Neither allows recursion.
2250
2251 25. I finally figured out how to implement a request to provide the minimum
2252 length of subject string that was needed in order to match a given pattern.
2253 (It was back references and recursion that I had previously got hung up
2254 on.) This code has now been added to pcre_study(); it finds a lower bound
2255 to the length of subject needed. It is not necessarily the greatest lower
2256 bound, but using it to avoid searching strings that are too short does give
2257 some useful speed-ups. The value is available to calling programs via
2258 pcre_fullinfo().
2259
2260 26. While implementing 25, I discovered to my embarrassment that pcretest had
2261 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2262 study optimizations had never been tested with that matching function.
2263 Oops. What is worse, even when it was passed study data, there was a bug in
2264 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2265 were also very few tests of studied patterns with pcre_dfa_exec().
2266
2267 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2268 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2269 on the other side of the coin, they are no longer allowed to have different
2270 names, because these cannot be distinguished in PCRE, and this has caused
2271 confusion. (This is a difference from Perl.)
2272
2273 28. When duplicate subpattern names are present (necessarily with different
2274 numbers, as required by 27 above), and a test is made by name in a
2275 conditional pattern, either for a subpattern having been matched, or for
2276 recursion in such a pattern, all the associated numbered subpatterns are
2277 tested, and the overall condition is true if the condition is true for any
2278 one of them. This is the way Perl works, and is also more like the way
2279 testing by number works.
2280
2281
2282 Version 7.9 11-Apr-09
2283 ---------------------
2284
2285 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2286 (pcretest), all targets were linked against these libraries. This included
2287 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2288 libraries. This caused unwanted dependencies to be created. This problem
2289 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2290 pcretest is linked with readline.
2291
2292 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2293 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2294 moved outside it again, because FALSE and TRUE are already defined in AIX,
2295 but BOOL is not.
2296
2297 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2298 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2299
2300 4. The pcregrep documentation said spaces were inserted as well as colons (or
2301 hyphens) following file names and line numbers when outputting matching
2302 lines. This is not true; no spaces are inserted. I have also clarified the
2303 wording for the --colour (or --color) option.
2304
2305 5. In pcregrep, when --colour was used with -o, the list of matching strings
2306 was not coloured; this is different to GNU grep, so I have changed it to be
2307 the same.
2308
2309 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2310 each matching line was coloured. Now it goes on to look for further matches
2311 of any of the test patterns, which is the same behaviour as GNU grep.
2312
2313 7. A pattern that could match an empty string could cause pcregrep to loop; it
2314 doesn't make sense to accept an empty string match in pcregrep, so I have
2315 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2316 seems to be how GNU grep behaves. [But see later change 40 for release
2317 8.33.]
2318
2319 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2320 start or after a newline", because the conditional assertion was not being
2321 correctly handled. The rule now is that both the assertion and what follows
2322 in the first alternative must satisfy the test.
2323
2324 9. If auto-callout was enabled in a pattern with a conditional group whose
2325 condition was an assertion, PCRE could crash during matching, both with
2326 pcre_exec() and pcre_dfa_exec().
2327
2328 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2329 used for matching.
2330
2331 11. Unicode property support in character classes was not working for
2332 characters (bytes) greater than 127 when not in UTF-8 mode.
2333
2334 12. Added the -M command line option to pcretest.
2335
2336 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2337
2338 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2339
2340 16. Added comments and documentation about mis-use of no_arg in the C++
2341 wrapper.
2342
2343 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2344 from Martin Jerabek that uses macro names for all relevant character and
2345 string constants.
2346
2347 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2348 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2349 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2350 these, but not everybody uses configure.
2351
2352 19. A conditional group that had only one branch was not being correctly
2353 recognized as an item that could match an empty string. This meant that an
2354 enclosing group might also not be so recognized, causing infinite looping
2355 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2356 with the subject "ab", where knowledge that the repeated group can match
2357 nothing is needed in order to break the loop.
2358
2359 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2360 exec(), but without supplying a callout function, matching went wrong.
2361
2362 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2363 leak if the size of the offset vector was greater than 30. When the vector
2364 is smaller, the saved offsets during recursion go onto a local stack
2365 vector, but for larger vectors malloc() is used. It was failing to free
2366 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2367 error, in fact).
2368
2369 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2370 heapframe that is used only when UTF-8 support is enabled. This caused no
2371 problem, but was untidy.
2372
2373 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2374 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2375 included within another project.
2376
2377 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2378 slightly modified by me:
2379
2380 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2381 not building pcregrep.
2382
2383 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2384 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2385
2386 25. Forward references, both numeric and by name, in patterns that made use of
2387 duplicate group numbers, could behave incorrectly or give incorrect errors,
2388 because when scanning forward to find the reference group, PCRE was not
2389 taking into account the duplicate group numbers. A pattern such as
2390 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2391
2392 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2393 the feature of a custom pointer more persuasive (as requested by a user).
2394
2395 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2396 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2397 pre-defined.
2398
2399 28. Added support for (*UTF8) at the start of a pattern.
2400
2401 29. Arrange for flags added by the "release type" setting in CMake to be shown
2402 in the configuration summary.
2403
2404
2405 Version 7.8 05-Sep-08
2406 ---------------------
2407
2408 1. Replaced UCP searching code with optimized version as implemented for Ad
2409 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2410 stage table and inline lookup instead of a function, giving speed ups of 2
2411 to 5 times on some simple patterns that I tested. Permission was given to
2412 distribute the MultiStage2.py script that generates the tables (it's not in
2413 the tarball, but is in the Subversion repository).
2414
2415 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2416 scripts.
2417
2418 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2419 a group with a zero qualifier. The result of the study could be incorrect,
2420 or the function might crash, depending on the pattern.
2421
2422 4. Caseless matching was not working for non-ASCII characters in back
2423 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2424 It now works when Unicode Property Support is available.
2425
2426 5. In pcretest, an escape such as \x{de} in the data was always generating
2427 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2428 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2429 truncation.
2430
2431 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2432
2433 7. Added two (int) casts to pcregrep when printing the difference of two
2434 pointers, in case they are 64-bit values.
2435
2436 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2437 test 2 if it fails.
2438
2439 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2440 and a #define of that name to empty if it is not externally set. This is to
2441 allow users of MSVC to set it if necessary.
2442
2443 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2444 the convenience functions in the pcre_get.c source file.
2445
2446 11. An option change at the start of a pattern that had top-level alternatives
2447 could cause overwriting and/or a crash. This command provoked a crash in
2448 some environments:
2449
2450 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2451
2452 This potential security problem was recorded as CVE-2008-2371.
2453
2454 12. For a pattern where the match had to start at the beginning or immediately
2455 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2456 pcre_dfa_exec() could read past the end of the passed subject if there was
2457 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2458 pcretest so that it places the subject at the end of its malloc-ed buffer.
2459
2460 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2461 exec() might read past the end of the data buffer in UTF-8 mode.
2462
2463 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2464 the data contained the byte 0x85 as part of a UTF-8 character within its
2465 first line. This applied both to normal and DFA matching.
2466
2467 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2468 /^[^d]*?$/8 failed to match "abc".
2469
2470 16. Added a missing copyright notice to pcrecpp_internal.h.
2471
2472 17. Make it more clear in the documentation that values returned from
2473 pcre_exec() in ovector are byte offsets, not character counts.
2474
2475 18. Tidied a few places to stop certain compilers from issuing warnings.
2476
2477 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2478 supplied by Stefan Weber. I made a further small update for 7.8 because
2479 there is a change of source arrangements: the pcre_searchfuncs.c module is
2480 replaced by pcre_ucd.c.
2481
2482
2483 Version 7.7 07-May-08
2484 ---------------------
2485
2486 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2487 a string to a long long, pretend we don't even have a long long." This is
2488 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2489
2490 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2491 pre-7.6 versions, which defined a global no_arg variable instead of putting
2492 it in the RE class. (See also #8 below.)
2493
2494 3. Remove a line of dead code, identified by coverity and reported by Nuno
2495 Lopes.
2496
2497 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2498
2499 (1) The include/exclude patterns were being applied to the whole pathnames
2500 of files, instead of just to the final components.
2501
2502 (2) If there was more than one level of directory, the subdirectories were
2503 skipped unless they satisfied the include/exclude conditions. This is
2504 inconsistent with GNU grep (and could even be seen as contrary to the
2505 pcregrep specification - which I improved to make it absolutely clear).
2506 The action now is always to scan all levels of directory, and just
2507 apply the include/exclude patterns to regular files.
2508
2509 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2510 --exclude_dir in the tests to avoid scanning .svn directories.
2511
2512 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2513 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2514 doesn't support NULs in patterns.
2515
2516 7. Added some missing "const"s to declarations of static tables in
2517 pcre_compile.c and pcre_dfa_exec.c.
2518
2519 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2520 caused by fix #2 above. (Subsequently also a second patch to fix the
2521 first patch. And a third patch - this was a messy problem.)
2522
2523 9. Applied Craig's patch to remove the use of push_back().
2524
2525 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2526 matching function regexec().
2527
2528 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2529 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2530 references. PCRE supports relative numbers with this syntax (I don't think
2531 Oniguruma does).
2532
2533 12. Previously, a group with a zero repeat such as (...){0} was completely
2534 omitted from the compiled regex. However, this means that if the group
2535 was called as a subroutine from elsewhere in the pattern, things went wrong
2536 (an internal error was given). Such groups are now left in the compiled
2537 pattern, with a new opcode that causes them to be skipped at execution
2538 time.
2539
2540 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2541 to the way PCRE behaves:
2542
2543 (a) A lone ] character is dis-allowed (Perl treats it as data).
2544
2545 (b) A back reference to an unmatched subpattern matches an empty string
2546 (Perl fails the current match path).
2547
2548 (c) A data ] in a character class must be notated as \] because if the
2549 first data character in a class is ], it defines an empty class. (In
2550 Perl it is not possible to have an empty class.) The empty class []
2551 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2552 The negative empty class [^] matches any one character, independently
2553 of the DOTALL setting.
2554
2555 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2556 non-existent subpattern following a character class starting with ']' and
2557 containing () gave an internal compiling error instead of "reference to
2558 non-existent subpattern". Fortunately, when the pattern did exist, the
2559 compiled code was correct. (When scanning forwards to check for the
2560 existence of the subpattern, it was treating the data ']' as terminating
2561 the class, so got the count wrong. When actually compiling, the reference
2562 was subsequently set up correctly.)
2563
2564 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2565 it was being rejected as not supported by pcre_dfa_exec(), even though
2566 other assertions are supported. I have made pcre_dfa_exec() support
2567 (*FAIL).
2568
2569 16. The implementation of 13c above involved the invention of a new opcode,
2570 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2571 cannot be changed at match time, I realized I could make a small
2572 improvement to matching performance by compiling OP_ALLANY instead of
2573 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2574 on the OP_ANY path.
2575
2576 17. Compiling pcretest on Windows with readline support failed without the
2577 following two fixes: (1) Make the unistd.h include conditional on
2578 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2579
2580 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2581 ncurses library to be included for pcretest when ReadLine support is
2582 requested, but also to allow for it to be overridden. This patch came from
2583 Daniel Bergström.
2584
2585 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2586 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2587 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2588 spotting this.
2589
2590
2591 Version 7.6 28-Jan-08
2592 ---------------------
2593
2594 1. A character class containing a very large number of characters with
2595 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2596 overflow.
2597
2598 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2599 HAVE_LONG_LONG is not defined.
2600
2601 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2602 bring them up to date and include new features. This patch includes:
2603
2604 - Fixed PH's badly added libz and libbz2 support.
2605 - Fixed a problem with static linking.
2606 - Added pcredemo. [But later removed - see 7 below.]
2607 - Fixed dftables problem and added an option.
2608 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2609 HAVE_LONG_LONG.
2610 - Added readline support for pcretest.
2611 - Added an listing of the option settings after cmake has run.
2612
2613 4. A user submitted a patch to Makefile that makes it easy to create
2614 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2615 Makefile.am that cause it to include this special target, without
2616 affecting anything else. Note that the same mingw target plus all
2617 the other distribution libraries and programs are now supported
2618 when configuring with CMake (see 6 below) instead of with
2619 Configure/Make.
2620
2621 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2622 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2623 exported in the Windows port". It has not yet been confirmed that the patch
2624 solves the problem, but it does no harm.
2625
2626 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2627 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2628 with CMake, and also correct the comment about stack recursion.
2629
2630 7. Remove the automatic building of pcredemo from the ./configure system and
2631 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2632 of a program that users should build themselves after PCRE is installed, so
2633 building it automatically is not really right. What is more, it gave
2634 trouble in some build environments.
2635
2636 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2637
2638
2639 Version 7.5 10-Jan-08
2640 ---------------------
2641
2642 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2643 values in parens when parsing an RE using the C++ wrapper."
2644
2645 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2646 Characters greater than 255 were excluded from the class instead of being
2647 included.
2648
2649 3. The same bug as (2) above applied to negated POSIX classes such as
2650 [:^space:].
2651
2652 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2653 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2654 I have changed it.
2655
2656 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2657 first named subpattern) and a construct such as (?&a) would reference the
2658 first named subpattern whose name started with "a" (in other words, the
2659 length check was missing). Both these problems are fixed. "Subpattern name
2660 expected" is now given for (?&) (a zero-length name), and this patch also
2661 makes it give the same error for \k'' (previously it complained that that
2662 was a reference to a non-existent subpattern).
2663
2664 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2665 this is right because (?- can be followed by option settings as well as by
2666 digits. I have, however, made the messages clearer.
2667
2668 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2669 than the number used in the conditional) now cause a compile-time error.
2670 This is actually not compatible with Perl, which accepts such patterns, but
2671 treats the conditional as always being FALSE (as PCRE used to), but it
2672 seems to me that giving a diagnostic is better.
2673
2674 8. Change "alphameric" to the more common word "alphanumeric" in comments
2675 and messages.
2676
2677 9. Fix two occurrences of "backslash" in comments that should have been
2678 "backspace".
2679
2680 10. Remove two redundant lines of code that can never be obeyed (their function
2681 was moved elsewhere).
2682
2683 11. The program that makes PCRE's Unicode character property table had a bug
2684 which caused it to generate incorrect table entries for sequences of
2685 characters that have the same character type, but are in different scripts.
2686 It amalgamated them into a single range, with the script of the first of
2687 them. In other words, some characters were in the wrong script. There were
2688 thirteen such cases, affecting characters in the following ranges:
2689
2690 U+002b0 - U+002c1
2691 U+0060c - U+0060d
2692 U+0061e - U+00612
2693 U+0064b - U+0065e
2694 U+0074d - U+0076d
2695 U+01800 - U+01805
2696 U+01d00 - U+01d77
2697 U+01d9b - U+01dbf
2698 U+0200b - U+0200f
2699 U+030fc - U+030fe
2700 U+03260 - U+0327f
2701 U+0fb46 - U+0fbb1
2702 U+10450 - U+1049d
2703
2704 12. The -o option (show only the matching part of a line) for pcregrep was not
2705 compatible with GNU grep in that, if there was more than one match in a
2706 line, it showed only the first of them. It now behaves in the same way as
2707 GNU grep.
2708
2709 13. If the -o and -v options were combined for pcregrep, it printed a blank
2710 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2711 does the same. The return code can be used to tell if there were any
2712 non-matching lines.
2713
2714 14. Added --file-offsets and --line-offsets to pcregrep.
2715
2716 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2717 infinitely looping recursion. The bug was that positive lookaheads were not
2718 being skipped when checking for a possible empty match (negative lookaheads
2719 and both kinds of lookbehind were skipped).
2720
2721 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2722 inclusion of <windows.h> to before rather than after the definition of
2723 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2724
2725 17. Specifying a possessive quantifier with a specific limit for a Unicode
2726 character property caused pcre_compile() to compile bad code, which led at
2727 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2728 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2729 caused the error; without that there was no problem.
2730
2731 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2732
2733 19. Added --enable-pcretest-libreadline.
2734
2735 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2736 RE::GlobalReplace(). As a result, the number of replacements returned was
2737 double what it should be. I removed one of the increments, but Craig sent a
2738 later patch that removed the other one (the right fix) and added unit tests
2739 that check the return values (which was not done before).
2740
2741 21. Several CMake things:
2742
2743 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2744 the names libpcre and libpcreposix, not just pcre and pcreposix.
2745
2746 (2) The above change means that pcretest and pcregrep are now correctly
2747 linked with the newly-built libraries, not previously installed ones.
2748
2749 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2750
2751 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2752 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2753 UTF-8 newline character). The key issue is that the pattern starts .*;
2754 this means that the match must be either at the beginning, or after a
2755 newline. The bug was in the code for advancing after a failed match and
2756 checking that the new position followed a newline. It was not taking
2757 account of UTF-8 characters correctly.
2758
2759 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2760 character classes. PCRE was not treating the sequence [:...:] as a
2761 character class unless the ... were all letters. Perl, however, seems to
2762 allow any characters between [: and :], though of course it rejects as
2763 unknown any "names" that contain non-letters, because all the known class
2764 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2765 for example, whereas PCRE did not - it did not recognize a POSIX character
2766 class. This seemed a bit dangerous, so the code has been changed to be
2767 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2768 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2769 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2770 Perl does, and where it didn't before.
2771
2772 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2773 Windows environments %n is disabled by default.
2774
2775
2776 Version 7.4 21-Sep-07
2777 ---------------------
2778
2779 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2780 means that a class such as [\s] counted as "explicit reference to CR or
2781 LF". That isn't really right - the whole point of the change was to try to
2782 help when there was an actual mention of one of the two characters. So now
2783 the change happens only if \r or \n (or a literal CR or LF) character is
2784 encountered.
2785
2786 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2787 of both had grown to the point where there were only 3 bits left.
2788 Fortunately, there was spare space in the data structure, and so I have
2789 moved the internal flags into a new 16-bit field to free up more option
2790 bits.
2791
2792 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2793 but did not set the internal JCHANGED flag - either of these is enough to
2794 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2795 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2796 start sets both bits.
2797
2798 4. Added options (at build time, compile time, exec time) to change \R from
2799 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2800
2801 5. doc/pcresyntax.html was missing from the distribution.
2802
2803 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2804 compatibility, even though it is no longer used.
2805
2806 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2807 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2808 windows.h file is present (where different names are used). [This was
2809 reversed later after testing - see 16 below.]
2810
2811 8. Changed all #include <config.h> to #include "config.h". There were also
2812 some further <pcre.h> cases that I changed to "pcre.h".
2813
2814 9. When pcregrep was used with the --colour option, it missed the line ending
2815 sequence off the lines that it output.
2816
2817 10. It was pointed out to me that arrays of string pointers cause lots of
2818 relocations when a shared library is dynamically loaded. A technique of
2819 using a single long string with a table of offsets can drastically reduce
2820 these. I have refactored PCRE in four places to do this. The result is
2821 dramatic:
2822
2823 Originally: 290
2824 After changing UCP table: 187
2825 After changing error message table: 43
2826 After changing table of "verbs" 36
2827 After changing table of Posix names 22
2828
2829 Thanks to the folks working on Gregex for glib for this insight.
2830
2831 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2832 unicode-properties was also set.
2833
2834 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2835
2836 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2837 checked only for CRLF.
2838
2839 14. Added casts to pcretest.c to avoid compiler warnings.
2840
2841 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2842
2843 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2844 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2845 entirely. This removes changes made in 7 above.
2846
2847 17. The CMake files have been updated, and there is now more information about
2848 building with CMake in the NON-UNIX-USE document.
2849
2850
2851 Version 7.3 28-Aug-07
2852 ---------------------
2853
2854 1. In the rejigging of the build system that eventually resulted in 7.1, the
2855 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2856 brackets there is not right, since it causes compilers to look for an
2857 installed pcre.h, not the version that is in the source that is being
2858 compiled (which of course may be different). I have changed it back to:
2859
2860 #include "pcre.h"
2861
2862 I have a vague recollection that the change was concerned with compiling in
2863 different directories, but in the new build system, that is taken care of
2864 by the VPATH setting the Makefile.
2865
2866 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2867 when the subject happened to end in the byte 0x85 (e.g. if the last
2868 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2869 characters but of course it shouldn't be taken as a newline when it is part
2870 of another character. The bug was that, for an unlimited repeat of . in
2871 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2872 characters when looking for a newline.
2873
2874 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2875
2876 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2877 in debug output.
2878
2879 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2880 long printing in the pcrecpp unittest when running under MinGW.
2881
2882 6. ESC_K was left out of the EBCDIC table.
2883
2884 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2885 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2886 limit also applies to "virtual nesting" when a pattern is recursive, and in
2887 this case 1000 isn't so big. I have been able to remove this limit at the
2888 expense of backing off one optimization in certain circumstances. Normally,
2889 when pcre_exec() would call its internal match() function recursively and
2890 immediately return the result unconditionally, it uses a "tail recursion"
2891 feature to save stack. However, when a subpattern that can match an empty
2892 string has an unlimited repetition quantifier, it no longer makes this
2893 optimization. That gives it a stack frame in which to save the data for
2894 checking that an empty string has been matched. Previously this was taken
2895 from the 1000-entry workspace that had been reserved. So now there is no
2896 explicit limit, but more stack is used.
2897
2898 8. Applied Daniel's patches to solve problems with the import/export magic
2899 syntax that is required for Windows, and which was going wrong for the
2900 pcreposix and pcrecpp parts of the library. These were overlooked when this
2901 problem was solved for the main library.
2902
2903 9. There were some crude static tests to avoid integer overflow when computing
2904 the size of patterns that contain repeated groups with explicit upper
2905 limits. As the maximum quantifier is 65535, the maximum group length was
2906 set at 30,000 so that the product of these two numbers did not overflow a
2907 32-bit integer. However, it turns out that people want to use groups that
2908 are longer than 30,000 bytes (though not repeat them that many times).
2909 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2910 made it possible to implement the integer overflow checks in a much more
2911 dynamic way, which I have now done. The artificial limitation on group
2912 length has been removed - we now have only the limit on the total length of
2913 the compiled pattern, which depends on the LINK_SIZE setting.
2914
2915 10. Fixed a bug in the documentation for get/copy named substring when
2916 duplicate names are permitted. If none of the named substrings are set, the
2917 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2918 empty string.
2919
2920 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2921 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2922 because the ] is interpreted as the first data character and the
2923 terminating ] is not found. PCRE has been made compatible with Perl in this
2924 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2925 cause memory overwriting.
2926
2927 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2928 string has been matched (to stop an infinite loop). It was not recognizing
2929 a conditional subpattern that could match an empty string if that
2930 subpattern was within another subpattern. For example, it looped when
2931 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2932 condition was not nested. This bug has been fixed.
2933
2934 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2935 past the start of the subject in the presence of bytes with the top bit
2936 set, for example "\x8aBCD".
2937
2938 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2939 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2940
2941 14. Optimized (?!) to (*FAIL).
2942
2943 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2944 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2945 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2946 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2947 does: it's just the validity check that is more restrictive.
2948
2949 16. Inserted checks for integer overflows during escape sequence (backslash)
2950 processing, and also fixed erroneous offset values for syntax errors during
2951 backslash processing.
2952
2953 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2954 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2955
2956 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2957 caused an overrun.
2958
2959 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2960 something other than just ASCII characters) inside a group that had an
2961 unlimited repeat caused a loop at compile time (while checking to see
2962 whether the group could match an empty string).
2963
2964 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2965 [\P{Any}] did so. (Error in the code for printing property names.)
2966
2967 21. An orphan \E inside a character class could cause a crash.
2968
2969 22. A repeated capturing bracket such as (A)? could cause a wild memory
2970 reference during compilation.
2971
2972 23. There are several functions in pcre_compile() that scan along a compiled
2973 expression for various reasons (e.g. to see if it's fixed length for look
2974 behind). There were bugs in these functions when a repeated \p or \P was
2975 present in the pattern. These operators have additional parameters compared
2976 with \d, etc, and these were not being taken into account when moving along
2977 the compiled data. Specifically:
2978
2979 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2980 length.
2981
2982 (b) An item such as \pL+ within a repeated group could cause crashes or
2983 loops.
2984
2985 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2986 "reference to non-existent subpattern" error.
2987
2988 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2989
2990 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2991 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2992
2993 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2994
2995 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2996 character were causing crashes (broken optimization).
2997
2998 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2999 \p or \P) caused a compile-time loop.
3000
3001 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3002 break. For example, the unstudied pattern [\r\n]A does not match the string
3003 "\r\nA" because change 7.0/46 below moves the current point on by two
3004 characters after failing to match at the start. However, the pattern \nA
3005 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3006 the same is true. There doesn't seem any very clean way out of this, but
3007 what I have chosen to do makes the common cases work: PCRE now takes note
3008 of whether there can be an explicit match for \r or \n anywhere in the
3009 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3010 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3011 pattern has explicit CR or LF references.
3012
3013 29. Added (*CR) etc for changing newline setting at start of pattern.
3014
3015
3016 Version 7.2 19-Jun-07
3017 ---------------------
3018
3019 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3020 which is apparently normally available under Windows.
3021
3022 2. Re-jig the pcregrep tests with different newline settings in an attempt
3023 to make them independent of the local environment's newline setting.
3024
3025 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3026
3027 4. Some of the "internals" tests were previously cut out when the link size
3028 was not 2, because the output contained actual offsets. The recent new
3029 "Z" feature of pcretest means that these can be cut out, making the tests
3030 usable with all link sizes.
3031
3032 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3033 stack recursion. This gives a massive performance boost under BSD, but just
3034 a small improvement under Linux. However, it saves one field in the frame
3035 in all cases.
3036
3037 6. Added more features from the forthcoming Perl 5.10:
3038
3039 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3040 recursion call. It refers to the nth most recently opened parentheses.
3041
3042 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3043 to be opened parentheses.
3044
3045 (c) Conditions that refer to capturing parentheses can be specified
3046 relatively, for example, (?(-2)... or (?(+3)...
3047
3048 (d) \K resets the start of the current match so that everything before
3049 is not part of it.
3050
3051 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3052
3053 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3054 reference syntax.
3055
3056 (g) (?| introduces a group in which the numbering of parentheses in each
3057 alternative starts with the same number.
3058
3059 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3060
3061 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3062 PCRE_INFO_JCHANGED.
3063
3064 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3065 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3066 for detecting groups that can match an empty string.
3067
3068 9. A pattern with a very large number of alternatives (more than several
3069 hundred) was running out of internal workspace during the pre-compile
3070 phase, where pcre_compile() figures out how much memory will be needed. A
3071 bit of new cunning has reduced the workspace needed for groups with
3072 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3073 workspace instead of running out of the 4096 that are available.
3074
3075 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3076
3077 11. Applied patch from Google to remove an optimization that didn't quite work.
3078 The report of the bug said:
3079
3080 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3081 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3082 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3083
3084 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3085 it matched the wrong number of bytes.
3086
3087
3088 Version 7.1 24-Apr-07
3089 ---------------------
3090
3091 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3092 that is more "standard", making use of automake and other Autotools. There
3093 is some re-arrangement of the files and adjustment of comments consequent
3094 on this.
3095
3096 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3097 for recursive directory scanning broke on some systems because the files
3098 are not scanned in any specific order and on different systems the order
3099 was different. A call to "sort" has been inserted into RunGrepTest for the
3100 approprate test as a short-term fix. In the longer term there may be an
3101 alternative.
3102
3103 3. I had an email from Eric Raymond about problems translating some of PCRE's
3104 man pages to HTML (despite the fact that I distribute HTML pages, some
3105 people do their own conversions for various reasons). The problems
3106 concerned the use of low-level troff macros .br and .in. I have therefore
3107 removed all such uses from the man pages (some were redundant, some could
3108 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3109 HTML has been updated to handle .nf/.fi and to complain if it encounters
3110 .br or .in.
3111
3112 4. Updated comments in configure.ac that get placed in config.h.in and also
3113 arranged for config.h to be included in the distribution, with the name
3114 config.h.generic, for the benefit of those who have to compile without
3115 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3116
3117 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3118 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3119 makevp.bat for the current PCRE, using the additional files
3120 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3121
3122 6. A Windows user reported a minor discrepancy with test 2, which turned out
3123 to be caused by a trailing space on an input line that had got lost in his
3124 copy. The trailing space was an accident, so I've just removed it.
3125
3126 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3127 that is needed.
3128
3129 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3130 as "const" (a) because they are and (b) because it helps the PHP
3131 maintainers who have recently made a script to detect big data structures
3132 in the php code that should be moved to the .rodata section. I remembered
3133 to update Builducptable as well, so it won't revert if ucptable.h is ever
3134 re-created.
3135
3136 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3137 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3138 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3139 support is not required. This saves 1.5-2K of code, which is important in
3140 some applications.
3141
3142 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3143 so as not to refer to the tables, even though these functions will never be
3144 called when UTF-8 support is disabled. Otherwise there are problems with a
3145 shared library.
3146
3147 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3148
3149 (a) It was defining its arguments as char * instead of void *.
3150
3151 (b) It was assuming that all moves were upwards in memory; this was true
3152 a long time ago when I wrote it, but is no longer the case.
3153
3154 The emulated memove() is provided for those environments that have neither
3155 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3156 is clearly not the case, as these two bugs were recently reported.
3157
3158 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3159 and Detrail to create the HTML documentation, the .txt form of the man
3160 pages, and it removes trailing spaces from listed files. It also creates
3161 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3162 case, it wraps all the #defines with #ifndefs. This script should be run
3163 before "make dist".
3164
3165 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3166 with Unicode property support.
3167
3168 (a) For a maximizing quantifier, if the two different cases of the
3169 character were of different lengths in their UTF-8 codings (there are
3170 some cases like this - I found 11), and the matching function had to
3171 back up over a mixture of the two cases, it incorrectly assumed they
3172 were both the same length.
3173
3174 (b) When PCRE was configured to use the heap rather than the stack for
3175 recursion during matching, it was not correctly preserving the data for
3176 the other case of a UTF-8 character when checking ahead for a match
3177 while processing a minimizing repeat. If the check also involved
3178 matching a wide character, but failed, corruption could cause an
3179 erroneous result when trying to check for a repeat of the original
3180 character.
3181
3182 13. Some tidying changes to the testing mechanism:
3183
3184 (a) The RunTest script now detects the internal link size and whether there
3185 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3186 values substituted by "configure". (The RunGrepTest script already did
3187 this for UTF-8.) The configure.ac script no longer substitutes the
3188 relevant variables.
3189
3190 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3191 with length and offset values. This means that the output is different
3192 for different internal link sizes. Test 2 is skipped for link sizes
3193 other than 2 because of this, bypassing the problem. Unfortunately,
3194 there was also a test in test 3 (the locale tests) that used /B and
3195 failed for link sizes other than 2. Rather than cut the whole test out,
3196 I have added a new /Z option to pcretest that replaces the length and
3197 offset values with spaces. This is now used to make test 3 independent
3198 of link size. (Test 2 will be tidied up later.)
3199
3200 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3201 segmentation fault instead of returning the appropriate error message.
3202
3203 15. In multiline mode when the newline sequence was set to "any", the pattern
3204 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3205 This doesn't seem right; it now treats the CRLF combination as the line
3206 ending, and so does not match in that case. It's only a pattern such as ^$
3207 that would hit this one: something like ^ABC$ would have failed after \r
3208 and then tried again after \r\n.
3209
3210 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3211 in an attempt to make files that differ only in their line terminators
3212 compare equal. This works on Linux.
3213
3214 17. Under certain error circumstances pcregrep might try to free random memory
3215 as it exited. This is now fixed, thanks to valgrind.
3216
3217 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3218 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3219 was because its rules for how to advance for /g after matching an empty
3220 string at the end of a line did not allow for this case. They now check for
3221 it specially.
3222
3223 20. pcretest is supposed to handle patterns and data of any length, by
3224 extending its buffers when necessary. It was getting this wrong when the
3225 buffer for a data line had to be extended.
3226
3227 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3228 CRLF as a newline sequence.
3229
3230 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3231 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3232 I have nevertheless tidied it up.
3233
3234 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3235
3236 24. Added a man page for pcre-config.
3237
3238
3239 Version 7.0 19-Dec-06
3240 ---------------------
3241
3242 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3243 moving to gcc 4.1.1.
3244
3245 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3246 sys/time.h, which is documented as needed for this function. It doesn't
3247 seem to matter on Linux, but it showed up on some releases of OS X.
3248
3249 3. It seems that there are systems where bytes whose values are greater than
3250 127 match isprint() in the "C" locale. The "C" locale should be the
3251 default when a C program starts up. In most systems, only ASCII printing
3252 characters match isprint(). This difference caused the output from pcretest
3253 to vary, making some of the tests fail. I have changed pcretest so that:
3254
3255 (a) When it is outputting text in the compiled version of a pattern, bytes
3256 other than 32-126 are always shown as hex escapes.
3257
3258 (b) When it is outputting text that is a matched part of a subject string,
3259 it does the same, unless a different locale has been set for the match
3260 (using the /L modifier). In this case, it uses isprint() to decide.
3261
3262 4. Fixed a major bug that caused incorrect computation of the amount of memory
3263 required for a compiled pattern when options that changed within the
3264 pattern affected the logic of the preliminary scan that determines the
3265 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3266 that the computed length was too small. The symptoms of this bug were
3267 either the PCRE error "internal error: code overflow" from pcre_compile(),
3268 or a glibc crash with a message such as "pcretest: free(): invalid next
3269 size (fast)". Examples of patterns that provoked this bug (shown in
3270 pcretest format) are:
3271
3272 /(?-x: )/x
3273 /(?x)(?-x: \s*#\s*)/
3274 /((?i)[\x{c0}])/8
3275 /(?i:[\x{c0}])/8
3276
3277 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3278 is now done differently.
3279
3280 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3281 wrapper classes; (b) implement a new function in the C++ scanner that is
3282 more efficient than the old way of doing things because it avoids levels of
3283 recursion in the regex matching; (c) add a paragraph to the documentation
3284 for the FullMatch() function.
3285
3286 6. The escape sequence \n was being treated as whatever was defined as
3287 "newline". Not only was this contrary to the documentation, which states
3288 that \n is character 10 (hex 0A), but it also went horribly wrong when
3289 "newline" was defined as CRLF. This has been fixed.
3290
3291 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3292 was being set to -1 for the "end of line" case (supposedly a value that no
3293 character can have). Though this value is never used (the check for end of
3294 line is "zero bytes in current character"), it caused compiler complaints.
3295 I've changed it to 0xffffffff.
3296
3297 8. In pcre_version.c, the version string was being built by a sequence of
3298 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3299 string (as it is for production releases) called a macro with an empty
3300 argument. The C standard says the result of this is undefined. The gcc
3301 compiler treats it as an empty string (which was what was wanted) but it is
3302 reported that Visual C gives an error. The source has been hacked around to
3303 avoid this problem.
3304
3305 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3306 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3307 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3308 of them did).
3309
3310 10. Originally, pcretest opened its input and output without "b"; then I was
3311 told that "b" was needed in some environments, so it was added for release
3312 5.0 to both the input and output. (It makes no difference on Unix-like
3313 systems.) Later I was told that it is wrong for the input on Windows. I've
3314 now abstracted the modes into two macros, to make it easier to fiddle with
3315 them, and removed "b" from the input mode under Windows.
3316
3317 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3318
3319 12. Added -help and --help to pcretest as an official way of being reminded
3320 of the options.
3321
3322 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3323 and pcrecpp.cc because they annoy compilers at high warning levels.
3324
3325 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3326
3327 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3328 scripts are not C programs :-) and which was not noticed because it works
3329 on Linux.
3330
3331 16. pcretest is supposed to handle any length of pattern and data line (as one
3332 line or as a continued sequence of lines) by extending its input buffer if
3333 necessary. This feature was broken for very long pattern lines, leading to
3334 a string of junk being passed to pcre_compile() if the pattern was longer
3335 than about 50K.
3336
3337 17. I have done a major re-factoring of the way pcre_compile() computes the
3338 amount of memory needed for a compiled pattern. Previously, there was code
3339 that made a preliminary scan of the pattern in order to do this. That was
3340 OK when PCRE was new, but as the facilities have expanded, it has become
3341 harder and harder to keep it in step with the real compile phase, and there
3342 have been a number of bugs (see for example, 4 above). I have now found a
3343 cunning way of running the real compile function in a "fake" mode that
3344 enables it to compute how much memory it would need, while actually only
3345 ever using a few hundred bytes of working memory and without too many
3346 tests of the mode. This should make future maintenance and development
3347 easier. A side effect of this work is that the limit of 200 on the nesting
3348 depth of parentheses has been removed (though this was never a serious
3349 limitation, I suspect). However, there is a downside: pcre_compile() now
3350 runs more slowly than before (30% or more, depending on the pattern). I
3351 hope this isn't a big issue. There is no effect on runtime performance.
3352
3353 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3354 newline (only possible for the last line of a file) and it was a
3355 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3356
3357 19. Added additional timing features to pcretest. (1) The -tm option now times
3358 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3359 separate command line item, by a number that specifies the number of
3360 repeats to use when timing. The default is 50000; this gives better
3361 precision, but takes uncomfortably long for very large patterns.
3362
3363 20. Extended pcre_study() to be more clever in cases where a branch of a
3364 subpattern has no definite first character. For example, (a*|b*)[cd] would
3365 previously give no result from pcre_study(). Now it recognizes that the
3366 first character must be a, b, c, or d.
3367
3368 21. There was an incorrect error "recursive call could loop indefinitely" if
3369 a subpattern (or the entire pattern) that was being tested for matching an
3370 empty string contained only one non-empty item after a nested subpattern.
3371 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3372 incorrectly, because the \d was being skipped in the check.
3373
3374 22. The pcretest program now has a new pattern option /B and a command line
3375 option -b, which is equivalent to adding /B to every pattern. This causes
3376 it to show the compiled bytecode, without the additional information that
3377 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3378 is the same as /B/I).
3379
3380 23. A new optimization is now able automatically to treat some sequences such
3381 as a*b as a*+b. More specifically, if something simple (such as a character
3382 or a simple class like \d) has an unlimited quantifier, and is followed by
3383 something that cannot possibly match the quantified thing, the quantifier
3384 is automatically "possessified".
3385
3386 24. A recursive reference to a subpattern whose number was greater than 39
3387 went wrong under certain circumstances in UTF-8 mode. This bug could also
3388 have affected the operation of pcre_study().
3389
3390 25. Realized that a little bit of performance could be had by replacing
3391 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3392
3393 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3394
3395 27. Possessive quantifiers such as a++ were previously implemented by turning
3396 them into atomic groups such as ($>a+). Now they have their own opcodes,
3397 which improves performance. This includes the automatically created ones
3398 from 23 above.
3399
3400 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3401 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3402 the first matched character to be a colon. This applied both to named and
3403 numbered groups.
3404
3405 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3406
3407 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3408 building PCRE on Windows easier, so I have included it in the distribution.
3409
3410 31. There is now a check in pcretest against a ridiculously large number being
3411 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3412 loop, the loop is abandoned.
3413
3414 32. Forward references to subpatterns in conditions such as (?(2)...) where
3415 subpattern 2 is defined later cause pcre_compile() to search forwards in
3416 the pattern for the relevant set of parentheses. This search went wrong
3417 when there were unescaped parentheses in a character class, parentheses
3418 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3419
3420 33. "Subroutine" calls and backreferences were previously restricted to
3421 referencing subpatterns earlier in the regex. This restriction has now
3422 been removed.
3423
3424 34. Added a number of extra features that are going to be in Perl 5.10. On the
3425 whole, these are just syntactic alternatives for features that PCRE had
3426 previously implemented using the Python syntax or my own invention. The
3427 other formats are all retained for compatibility.
3428
3429 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3430 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3431 also .NET compatible.
3432
3433 (b) A recursion or subroutine call to a named group can now be defined as
3434 (?&name) as well as (?P>name).
3435
3436 (c) A backreference to a named group can now be defined as \k<name> or
3437 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3438 5.10, are also .NET compatible.
3439
3440 (d) A conditional reference to a named group can now use the syntax
3441 (?(<name>) or (?('name') as well as (?(name).
3442
3443 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3444 groups (named and numbered) that are never evaluated inline, but can be
3445 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3446 is always false. There may be only one alternative in such a group.
3447
3448 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3449 as the simple (?(R). The condition is true only if the most recent
3450 recursion is that of the given number or name. It does not search out
3451 through the entire recursion stack.
3452
3453 (g) The escape \gN or \g{N} has been added, where N is a positive or
3454 negative number, specifying an absolute or relative reference.
3455
3456 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3457 some "unreachable code" warnings.
3458
3459 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3460 things, this adds five new scripts.
3461
3462 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3463 There were also incompatibilities regarding the handling of \Q..\E inside
3464 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3465 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3466
3467 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3468 matches an empty string, and forcibly breaks the loop. There were bugs in
3469 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3470 against aaaa the result was just "a" rather than "aaaa", for example. Two
3471 separate and independent bugs (that affected different cases) have been
3472 fixed.
3473
3474 39. Refactored the code to abolish the use of different opcodes for small
3475 capturing bracket numbers. This is a tidy that I avoided doing when I
3476 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3477 The new approach is not only tidier, it makes it possible to reduce the
3478 memory needed to fix the previous bug (38).
3479
3480 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3481 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3482 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3483 mode.
3484
3485 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3486 report.
3487
3488 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3489 copy construction and assignment in the C++ wrapper.
3490
3491 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3492 couple of bugs that could have given wrong results in the "--newline=crlf"
3493 case.
3494
3495 44. Added a number of casts and did some reorganization of signed/unsigned int
3496 variables following suggestions from Dair Grant. Also renamed the variable
3497 "this" as "item" because it is a C++ keyword.
3498
3499 45. Arranged for dftables to add
3500
3501 #include "pcre_internal.h"
3502
3503 to pcre_chartables.c because without it, gcc 4.x may remove the array
3504 definition from the final binary if PCRE is built into a static library and
3505 dead code stripping is activated.
3506
3507 46. For an unanchored pattern, if a match attempt fails at the start of a
3508 newline sequence, and the newline setting is CRLF or ANY, and the next two
3509 characters are CRLF, advance by two characters instead of one.
3510
3511
3512 Version 6.7 04-Jul-06
3513 ---------------------
3514
3515 1. In order to handle tests when input lines are enormously long, pcretest has
3516 been re-factored so that it automatically extends its buffers when
3517 necessary. The code is crude, but this _is_ just a test program. The
3518 default size has been increased from 32K to 50K.
3519
3520 2. The code in pcre_study() was using the value of the re argument before
3521 testing it for NULL. (Of course, in any sensible call of the function, it
3522 won't be NULL.)
3523
3524 3. The memmove() emulation function in pcre_internal.h, which is used on
3525 systems that lack both memmove() and bcopy() - that is, hardly ever -
3526 was missing a "static" storage class specifier.
3527
3528 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3529 containing an extended class (one that cannot be represented by a bitmap
3530 because it contains high-valued characters or Unicode property items, e.g.
3531 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3532 pattern, but PCRE should not loop if you do not (it no longer does).
3533 [Detail: two cases were found: (a) a repeated subpattern containing an
3534 extended class; (b) a recursive reference to a subpattern that followed a
3535 previous extended class. It wasn't skipping over the extended class
3536 correctly when UTF-8 mode was not set.]
3537
3538 5. A negated single-character class was not being recognized as fixed-length
3539 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3540 compile error "lookbehind assertion is not fixed length".
3541
3542 6. The RunPerlTest auxiliary script was showing an unexpected difference
3543 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3544 write a Perl script that can interpret lines of an input file either as
3545 byte characters or as UTF-8, which is what "perltest" was being required to
3546 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3547 can't do is switch easily at run time between having the "use utf8;" pragma
3548 or not. In the end, I fudged it by using the RunPerlTest script to insert
3549 "use utf8;" explicitly for the UTF-8 tests.
3550
3551 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3552 the end of the subject string, contrary to the documentation and to what
3553 Perl does. This was true of both matching functions. Now it matches only at
3554 the start of the subject and immediately after *internal* newlines.
3555
3556 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3557 a pointer to an int instead of a pointer to an unsigned long int. This
3558 caused problems on 64-bit systems.
3559
3560 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3561 instance of the 'standard' template library not being so standard".
3562
3563 10. There was no check on the number of named subpatterns nor the maximum
3564 length of a subpattern name. The product of these values is used to compute
3565 the size of the memory block for a compiled pattern. By supplying a very
3566 long subpattern name and a large number of named subpatterns, the size
3567 computation could be caused to overflow. This is now prevented by limiting
3568 the length of names to 32 characters, and the number of named subpatterns
3569 to 10,000.
3570
3571 11. Subpatterns that are repeated with specific counts have to be replicated in
3572 the compiled pattern. The size of memory for this was computed from the
3573 length of the subpattern and the repeat count. The latter is limited to
3574 65535, but there was no limit on the former, meaning that integer overflow
3575 could in principle occur. The compiled length of a repeated subpattern is
3576 now limited to 30,000 bytes in order to prevent this.
3577
3578 12. Added the optional facility to have named substrings with the same name.
3579
3580 13. Added the ability to use a named substring as a condition, using the
3581 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3582 are numbers (not recommended). Forward references are permitted.
3583
3584 14. Added forward references in named backreferences (if you see what I mean).
3585
3586 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3587 pattern could run off the end of the subject. For example, the pattern
3588 "(?s)(.{1,5})"8 did this with the subject "ab".
3589
3590 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3591 PCRE_CASELESS was set when matching characters that were quantified with ?
3592 or *.
3593
3594 17. A character class other than a single negated character that had a minimum
3595 but no maximum quantifier - for example [ab]{6,} - was not handled
3596 correctly by pce_dfa_exec(). It would match only one character.
3597
3598 18. A valid (though odd) pattern that looked like a POSIX character
3599 class but used an invalid character after [ (for example [[,abc,]]) caused
3600 pcre_compile() to give the error "Failed: internal error: code overflow" or
3601 in some cases to crash with a glibc free() error. This could even happen if
3602 the pattern terminated after [[ but there just happened to be a sequence of
3603 letters, a binary zero, and a closing ] in the memory that followed.
3604
3605 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3606 over the years. Originally (before any Unicode support), just the bottom 8
3607 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3608 output from "man perlunicode" includes this:
3609
3610 The regular expression compiler produces polymorphic opcodes. That
3611 is, the pattern adapts to the data and automatically switches to
3612 the Unicode character scheme when presented with Unicode data--or
3613 instead uses a traditional byte scheme when presented with byte
3614 data.
3615
3616 Sadly, a wide octal escape does not cause a switch, and in a string with
3617 no other multibyte characters, these octal escapes are treated as before.
3618 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3619 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3620 Unicode string.
3621
3622 I have not perpetrated such confusion in PCRE. Up till now, it took just
3623 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3624 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3625 translate to the appropriate multibyte character.
3626
3627 29. Applied some refactoring to reduce the number of warnings from Microsoft
3628 and Borland compilers. This has included removing the fudge introduced
3629 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3630 a warning about an unused variable.
3631
3632 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3633 characters since release 4.0, because Perl (from release 5.004) does not.
3634 [Or at least, is documented not to: some releases seem to be in conflict
3635 with the documentation.] However, when a pattern was studied with
3636 pcre_study() and all its branches started with \s, PCRE still included VT
3637 as a possible starting character. Of course, this did no harm; it just
3638 caused an unnecessary match attempt.
3639
3640 22. Removed a now-redundant internal flag bit that recorded the fact that case
3641 dependency changed within the pattern. This was once needed for "required
3642 byte" processing, but is no longer used. This recovers a now-scarce options
3643 bit. Also moved the least significant internal flag bit to the most-
3644 significant bit of the word, which was not previously used (hangover from
3645 the days when it was an int rather than a uint) to free up another bit for
3646 the future.
3647
3648 23. Added support for CRLF line endings as well as CR and LF. As well as the
3649 default being selectable at build time, it can now be changed at runtime
3650 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3651 specify that it is scanning data with non-default line endings.
3652
3653 24. Changed the definition of CXXLINK to make it agree with the definition of
3654 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3655
3656 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3657 recursions. This makes a big different to stack usage for some patterns.
3658
3659 26. If a subpattern containing a named recursion or subroutine reference such
3660 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3661 the space required for the compiled pattern went wrong and gave too small a
3662 value. Depending on the environment, this could lead to "Failed: internal
3663 error: code overflow at offset 49" or "glibc detected double free or
3664 corruption" errors.
3665
3666 27. Applied patches from Google (a) to support the new newline modes and (b) to
3667 advance over multibyte UTF-8 characters in GlobalReplace.
3668
3669 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3670 difference for some implementation of PCRE in some Windows version.
3671
3672 29. Added some extra testing facilities to pcretest:
3673
3674 \q<number> in a data line sets the "match limit" value
3675 \Q<number> in a data line sets the "match recursion limt" value
3676 -S <number> sets the stack size, where <number> is in megabytes
3677
3678 The -S option isn't available for Windows.
3679
3680
3681 Version 6.6 06-Feb-06
3682 ---------------------
3683
3684 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3685 in pcreposix.h. I have copied the definition from pcre.h.
3686
3687 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3688 because pcre.h is no longer a built file.
3689
3690 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3691 not normally included in the compiled code.
3692
3693
3694 Version 6.5 01-Feb-06
3695 ---------------------
3696
3697 1. When using the partial match feature with pcre_dfa_exec(), it was not
3698 anchoring the second and subsequent partial matches at the new starting
3699 point. This could lead to incorrect results. For example, with the pattern
3700 /1234/, partially matching against "123" and then "a4" gave a match.
3701
3702 2. Changes to pcregrep:
3703
3704 (a) All non-match returns from pcre_exec() were being treated as failures
3705 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3706 error message is output. Some extra information is given for the
3707 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3708 probably the only errors that are likely to be caused by users (by
3709 specifying a regex that has nested indefinite repeats, for instance).
3710 If there are more than 20 of these errors, pcregrep is abandoned.
3711
3712 (b) A binary zero was treated as data while matching, but terminated the
3713 output line if it was written out. This has been fixed: binary zeroes
3714 are now no different to any other data bytes.
3715
3716 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3717 used to set a locale for matching. The --locale=xxxx long option has
3718 been added (no short equivalent) to specify a locale explicitly on the
3719 pcregrep command, overriding the environment variables.
3720
3721 (d) When -B was used with -n, some line numbers in the output were one less
3722 than they should have been.
3723
3724 (e) Added the -o (--only-matching) option.
3725
3726 (f) If -A or -C was used with -c (count only), some lines of context were
3727 accidentally printed for the final match.
3728
3729 (g) Added the -H (--with-filename) option.
3730
3731 (h) The combination of options -rh failed to suppress file names for files
3732 that were found from directory arguments.
3733
3734 (i) Added the -D (--devices) and -d (--directories) options.
3735
3736 (j) Added the -F (--fixed-strings) option.
3737
3738 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3739
3740 (l) Added the --colo(u)r option.
3741
3742 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3743 is not present by default.
3744
3745 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3746 items such as (?R) or (?1), when the recursion could match a number of
3747 alternatives. If it matched one of the alternatives, but subsequently,
3748 outside the recursion, there was a failure, the code tried to back up into
3749 the recursion. However, because of the way PCRE is implemented, this is not
3750 possible, and the result was an incorrect result from the match.
3751
3752 In order to prevent this happening, the specification of recursion has
3753 been changed so that all such subpatterns are automatically treated as
3754 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3755
3756 4. I had overlooked the fact that, in some locales, there are characters for
3757 which isalpha() is true but neither isupper() nor islower() are true. In
3758 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3759 and ordfeminine) are like this. This affected the treatment of \w and \W
3760 when they appeared in character classes, but not when they appeared outside
3761 a character class. The bit map for "word" characters is now created
3762 separately from the results of isalnum() instead of just taking it from the
3763 upper, lower, and digit maps. (Plus the underscore character, of course.)
3764
3765 5. The above bug also affected the handling of POSIX character classes such as
3766 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3767 permanent tables. Instead, the bit maps for such a class were previously
3768 created as the appropriate unions of the upper, lower, and digit bitmaps.
3769 Now they are created by subtraction from the [[:word:]] class, which has
3770 its own bitmap.
3771
3772 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3773 It is created by subtracting the vertical space characters (\x09, \x0a,
3774 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3775 subtraction was done in the overall bitmap for a character class, meaning
3776 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3777 be recognized. This bug has been fixed.
3778
3779 7. Patches from the folks at Google:
3780
3781 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3782 real life, but is still worth protecting against".
3783
3784 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3785 regular expressions".
3786
3787 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3788 have it.
3789
3790 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3791 "configure" and the latter not, in order to fix a problem somebody had
3792 with compiling the Arg class on HP-UX.
3793
3794 (e) Improve the error-handling of the C++ wrapper a little bit.
3795
3796 (f) New tests for checking recursion limiting.
3797
3798 8. The pcre_memmove() function, which is used only if the environment does not
3799 have a standard memmove() function (and is therefore rarely compiled),
3800 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3801 returning a result (though PCRE never actually uses the result).
3802
3803 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3804 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3805 returned instead of calling malloc() with an overflowing number that would
3806 most likely cause subsequent chaos.
3807
3808 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3809
3810 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3811 with this option is matched, the nmatch and pmatch options of regexec() are
3812 ignored.
3813
3814 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3815 provided in case anyone wants to the the POSIX interface with UTF-8
3816 strings.
3817
3818 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3819 C++ linking (needed for some HP-UX environments).
3820
3821 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3822 (unused parameter) and in the pcre_printint() function (omitted "default"
3823 switch label when the default is to do nothing).
3824
3825 15. Added some code to make it possible, when PCRE is compiled as a C++
3826 library, to replace subject pointers for pcre_exec() with a smart pointer
3827 class, thus making it possible to process discontinuous strings.
3828
3829 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3830 much the same function. They were added by different people who were trying
3831 to make PCRE easy to compile on non-Unix systems. It has been suggested
3832 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3833 for compiling on Windows systems. I have therefore replaced it with
3834 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3835 defaults to "extern" for C or "extern C" for C++, which works fine on
3836 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3837 SCOPE with something explicit in config.h. In addition:
3838
3839 (a) pcreposix.h still had just "extern" instead of either of these macros;
3840 I have replaced it with PCRE_DATA_SCOPE.
3841
3842 (b) Functions such as _pcre_xclass(), which are internal to the library,
3843 but external in the C sense, all had PCRE_EXPORT in their definitions.
3844 This is apparently wrong for the Windows case, so I have removed it.
3845 (It makes no difference on Unix-like systems.)
3846
3847 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3848 of recursive calls to match(). This is different to MATCH_LIMIT because
3849 that limits the total number of calls to match(), not all of which increase
3850 the depth of recursion. Limiting the recursion depth limits the amount of
3851 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3852 when PCRE is compiled, and changed at run time. A patch from Google adds
3853 this functionality to the C++ interface.
3854
3855 18. Changes to the handling of Unicode character properties:
3856
3857 (a) Updated the table to Unicode 4.1.0.
3858
3859 (b) Recognize characters that are not in the table as "Cn" (undefined).
3860
3861 (c) I revised the way the table is implemented to a much improved format
3862 which includes recognition of ranges. It now supports the ranges that
3863 are defined in UnicodeData.txt, and it also amalgamates other
3864 characters into ranges. This has reduced the number of entries in the
3865 table from around 16,000 to around 3,000, thus reducing its size
3866 considerably. I realized I did not need to use a tree structure after
3867 all - a binary chop search is just as efficient. Having reduced the
3868 number of entries, I extended their size from 6 bytes to 8 bytes to
3869 allow for more data.
3870
3871 (d) Added support for Unicode script names via properties such as \p{Han}.
3872
3873 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3874 matching that character.
3875
3876 20. When matching a repeated Unicode property with a minimum greater than zero,
3877 (for example \pL{2,}), PCRE could look past the end of the subject if it
3878 reached it while seeking the minimum number of characters. This could
3879 happen only if some of the characters were more than one byte long, because
3880 there is a check for at least the minimum number of bytes.
3881
3882 21. Refactored the implementation of \p and \P so as to be more general, to
3883 allow for more different types of property in future. This has changed the
3884 compiled form incompatibly. Anybody with saved compiled patterns that use
3885 \p or \P will have to recompile them.
3886
3887 22. Added "Any" and "L&" to the supported property types.
3888
3889 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3890 but give a compile time error if the value is greater than 0xff.
3891
3892 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3893 accidentally not being installed or uninstalled.
3894
3895 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3896 made were to insert the current release number. This seemed silly, because
3897 it made things harder for people building PCRE on systems that don't run
3898 "configure". I have turned pcre.h into a distributed file, no longer built
3899 by "configure", with the version identification directly included. There is
3900 no longer a pcre.h.in file.
3901
3902 However, this change necessitated a change to the pcre-config script as
3903 well. It is built from pcre-config.in, and one of the substitutions was the
3904 release number. I have updated configure.ac so that ./configure now finds
3905 the release number by grepping pcre.h.
3906
3907 26. Added the ability to run the tests under valgrind.
3908
3909
3910 Version 6.4 05-Sep-05
3911 ---------------------
3912
3913 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3914 "--" to be printed when multiple files were scanned, even when none of the
3915 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3916 consider it to be a bug, and have restored the previous behaviour.
3917
3918 2. A couple of code tidies to get rid of compiler warnings.
3919
3920 3. The pcretest program used to cheat by referring to symbols in the library
3921 whose names begin with _pcre_. These are internal symbols that are not
3922 really supposed to be visible externally, and in some environments it is
3923 possible to suppress them. The cheating is now confined to including
3924 certain files from the library's source, which is a bit cleaner.
3925
3926 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3927 file's purpose clearer.
3928
3929 5. Reorganized pcre_ucp_findchar().
3930
3931
3932 Version 6.3 15-Aug-05
3933 ---------------------
3934
3935 1. The file libpcre.pc.in did not have general read permission in the tarball.
3936
3937 2. There were some problems when building without C++ support:
3938
3939 (a) If C++ support was not built, "make install" and "make test" still
3940 tried to test it.
3941
3942 (b) There were problems when the value of CXX was explicitly set. Some
3943 changes have been made to try to fix these, and ...
3944
3945 (c) --disable-cpp can now be used to explicitly disable C++ support.
3946
3947 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3948 backslash in a target when C++ was disabled. This confuses some
3949 versions of "make", apparently. Using an intermediate variable solves
3950 this. (Same for CPP_LOBJ.)
3951
3952 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3953 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3954 necessary on certain architectures.
3955
3956 4. Added a setting of -export-symbols-regex to the link command to remove
3957 those symbols that are exported in the C sense, but actually are local
3958 within the library, and not documented. Their names all begin with
3959 "_pcre_". This is not a perfect job, because (a) we have to except some
3960 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3961 available (and never for static libraries). I have made a note to try to
3962 find a way round (a) in the future.
3963
3964
3965 Version 6.2 01-Aug-05
3966 ---------------------
3967
3968 1. There was no test for integer overflow of quantifier values. A construction
3969 such as {1111111111111111} would give undefined results. What is worse, if
3970 a minimum quantifier for a parenthesized subpattern overflowed and became
3971 negative, the calculation of the memory size went wrong. This could have
3972 led to memory overwriting.
3973
3974 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3975
3976 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3977 operating environments where this matters.
3978
3979 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3980 PCRE options from within the C++ wrapper.
3981
3982 5. Named capturing subpatterns were not being correctly counted when a pattern
3983 was compiled. This caused two problems: (a) If there were more than 100
3984 such subpatterns, the calculation of the memory needed for the whole
3985 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3986 back references of the form \12, where the number was greater than 9, were
3987 not recognized as back references, even though there were sufficient
3988 previous subpatterns.
3989
3990 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3991 versions of gcc, e.g. 2.95.4.
3992
3993
3994 Version 6.1 21-Jun-05
3995 ---------------------
3996
3997 1. There was one reference to the variable "posix" in pcretest.c that was not
3998 surrounded by "#if !defined NOPOSIX".
3999
4000 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4001 the cross-check on the old pcre_info() function, for the benefit of the
4002 cut-down version of PCRE that is currently imported into Exim.
4003
4004 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4005 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4006 patterns that start (?i) but I don't think that matters. The use of (?i) is
4007 just an example; this all applies to the other options as well.
4008
4009 4. Since libtool seems to echo the compile commands it is issuing, the output
4010 from "make" can be reduced a bit by putting "@" in front of each libtool
4011 compile command.
4012
4013 5. Patch from the folks at Google for configure.in to be a bit more thorough
4014 in checking for a suitable C++ installation before trying to compile the
4015 C++ stuff. This should fix a reported problem when a compiler was present,
4016 but no suitable headers.
4017
4018 6. The man pages all had just "PCRE" as their title. I have changed them to
4019 be the relevant file name. I have also arranged that these names are
4020 retained in the file doc/pcre.txt, which is a concatenation in text format
4021 of all the man pages except the little individual ones for each function.
4022
4023 7. The NON-UNIX-USE file had not been updated for the different set of source
4024 files that come with release 6. I also added a few comments about the C++
4025 wrapper.
4026
4027
4028 Version 6.0 07-Jun-05
4029 ---------------------
4030
4031 1. Some minor internal re-organization to help with my DFA experiments.
4032
4033 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4034 didn't matter for the library itself when fully configured, but did matter
4035 when compiling without UCP support, or within Exim, where the ucp files are
4036 not imported.
4037
4038 3. Refactoring of the library code to split up the various functions into
4039 different source modules. The addition of the new DFA matching code (see
4040 below) to a single monolithic source would have made it really too
4041 unwieldy, quite apart from causing all the code to be include in a
4042 statically linked application, when only some functions are used. This is
4043 relevant even without the DFA addition now that patterns can be compiled in
4044 one application and matched in another.
4045
4046 The downside of splitting up is that there have to be some external
4047 functions and data tables that are used internally in different modules of
4048 the library but which are not part of the API. These have all had their
4049 names changed to start with "_pcre_" so that they are unlikely to clash
4050 with other external names.
4051
4052 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4053 a different (DFA) algorithm. Although it is slower than the original
4054 function, it does have some advantages for certain types of matching
4055 problem.
4056
4057 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4058 including restarting after a partial match.
4059
4060 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4061 defined when compiling for Windows was sent to me. I have put it into the
4062 code, though I have no means of testing or verifying it.
4063
4064 7. Added the pcre_refcount() auxiliary function.
4065
4066 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4067 match before or at the first newline in the subject string. In pcretest,
4068 the /f option on a pattern can be used to set this.
4069
4070 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4071 would behave wrongly. This has been present in PCRE since release 4.0.
4072
4073 10. A number of changes to the pcregrep command:
4074
4075 (a) Refactored how -x works; insert ^(...)$ instead of setting
4076 PCRE_ANCHORED and checking the length, in preparation for adding
4077 something similar for -w.
4078
4079 (b) Added the -w (match as a word) option.
4080
4081 (c) Refactored the way lines are read and buffered so as to have more
4082 than one at a time available.
4083
4084 (d) Implemented a pcregrep test script.
4085
4086 (e) Added the -M (multiline match) option. This allows patterns to match
4087 over several lines of the subject. The buffering ensures that at least
4088 8K, or the rest of the document (whichever is the shorter) is available
4089 for matching (and similarly the previous 8K for lookbehind assertions).
4090
4091 (f) Changed the --help output so that it now says
4092
4093 -w, --word-regex(p)
4094
4095 instead of two lines, one with "regex" and the other with "regexp"
4096 because that confused at least one person since the short forms are the
4097 same. (This required a bit of code, as the output is generated
4098 automatically from a table. It wasn't just a text change.)
4099
4100 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4101 option but starts with a hyphen. Could be a pattern or a path name
4102 starting with a hyphen, for instance.
4103
4104 (h) "-" can be given as a file name to represent stdin.
4105
4106 (i) When file names are being printed, "(standard input)" is used for
4107 the standard input, for compatibility with GNU grep. Previously
4108 "<stdin>" was used.
4109
4110 (j) The option --label=xxx can be used to supply a name to be used for
4111 stdin when file names are being printed. There is no short form.
4112
4113 (k) Re-factored the options decoding logic because we are going to add
4114 two more options that take data. Such options can now be given in four
4115 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4116
4117 (l) Added the -A, -B, and -C options for requesting that lines of context
4118 around matches be printed.
4119
4120 (m) Added the -L option to print the names of files that do not contain
4121 any matching lines, that is, the complement of -l.
4122
4123 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4124 continue to scan other files.
4125
4126 (o) The -s option was incorrectly implemented. For compatibility with other
4127 greps, it now suppresses the error message for a non-existent or non-
4128 accessible file (but not the return code). There is a new option called
4129 -q that suppresses the output of matching lines, which was what -s was
4130 previously doing.
4131
4132 (p) Added --include and --exclude options to specify files for inclusion
4133 and exclusion when recursing.
4134
4135 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4136 Hopefully, it now does.
4137
4138 12. Missing cast in pcre_study().
4139
4140 13. Added an "uninstall" target to the makefile.
4141
4142 14. Replaced "extern" in the function prototypes in Makefile.in with
4143 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4144 world, but is set differently for Windows.
4145
4146 15. Added a second compiling function called pcre_compile2(). The only
4147 difference is that it has an extra argument, which is a pointer to an
4148 integer error code. When there is a compile-time failure, this is set
4149 non-zero, in addition to the error test pointer being set to point to an
4150 error message. The new argument may be NULL if no error number is required
4151 (but then you may as well call pcre_compile(), which is now just a
4152 wrapper). This facility is provided because some applications need a
4153 numeric error indication, but it has also enabled me to tidy up the way
4154 compile-time errors are handled in the POSIX wrapper.
4155
4156 16. Added VPATH=.libs to the makefile; this should help when building with one
4157 prefix path and installing with another. (Or so I'm told by someone who
4158 knows more about this stuff than I do.)
4159
4160 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4161 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4162 match everything, including newlines. This is not POSIX-compatible, but
4163 somebody wanted the feature. From pcretest it can be activated by using
4164 both the P and the s flags.
4165
4166 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4167
4168 19. libpcre.pc was being incorrectly installed as executable.
4169
4170 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4171 it now also looks for '\r' so that it will work unmodified on Windows.
4172
4173 21. Added Google's contributed C++ wrapper to the distribution.
4174
4175 22. Added some untidy missing memory free() calls in pcretest, to keep
4176 Electric Fence happy when testing.
4177
4178
4179
4180 Version 5.0 13-Sep-04
4181 ---------------------
4182
4183 1. Internal change: literal characters are no longer packed up into items
4184 containing multiple characters in a single byte-string. Each character
4185 is now matched using a separate opcode. However, there may be more than one
4186 byte in the character in UTF-8 mode.
4187
4188 2. The pcre_callout_block structure has two new fields: pattern_position and
4189 next_item_length. These contain the offset in the pattern to the next match
4190 item, and its length, respectively.
4191
4192 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4193 insertion of callouts before each pattern item. Added the /C option to
4194 pcretest to make use of this.
4195
4196 4. On the advice of a Windows user, the lines
4197
4198 #if defined(_WIN32) || defined(WIN32)
4199 _setmode( _fileno( stdout ), 0x8000 );
4200 #endif /* defined(_WIN32) || defined(WIN32) */
4201
4202 have been added to the source of pcretest. This apparently does useful
4203 magic in relation to line terminators.
4204
4205 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4206 for the benefit of those environments where the "b" makes a difference.
4207
4208 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4209 to know about it. I have put a hack into configure.in that adds in code
4210 to set GCC=yes if CC=icc. This seems to end up at a point in the
4211 generated configure script that is early enough to affect the setting of
4212 compiler options, which is what is needed, but I have no means of testing
4213 whether it really works. (The user who reported this had patched the
4214 generated configure script, which of course I cannot do.)
4215
4216 LATER: After change 22 below (new libtool files), the configure script
4217 seems to know about icc (and also ecc). Therefore, I have commented out
4218 this hack in configure.in.
4219
4220 7. Added support for pkg-config (2 patches were sent in).
4221
4222 8. Negated POSIX character classes that used a combination of internal tables
4223 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4224 [[:^ascii]]. Typically, they would match almost any characters. The other
4225 POSIX classes were not broken in this way.
4226
4227 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4228 to find the match, as PCRE was deluded into thinking that the match had to
4229 start at the start point or following a newline. The same bug applied to
4230 patterns with negative forward assertions or any backward assertions
4231 preceding ".*" at the start, unless the pattern required a fixed first
4232 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4233
4234 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4235 starting at the last subject character, bytes beyond the end of the subject
4236 string were read.
4237
4238 11. Renamed the variable "class" as "classbits" to make life easier for C++
4239 users. (Previously there was a macro definition, but it apparently wasn't
4240 enough.)
4241
4242 12. Added the new field "tables" to the extra data so that tables can be passed
4243 in at exec time, or the internal tables can be re-selected. This allows
4244 a compiled regex to be saved and re-used at a later time by a different
4245 program that might have everything at different addresses.
4246
4247 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4248 -R library as well as a -L library.
4249
4250 14. The debugging options of pcretest (-d on the command line or D on a
4251 pattern) showed incorrect output for anything following an extended class
4252 that contained multibyte characters and which was followed by a quantifier.
4253
4254 15. Added optional support for general category Unicode character properties
4255 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4256 support. It adds about 90K to the size of the library. The meanings of the
4257 inbuilt class escapes such as \d and \s have NOT been changed.
4258
4259 16. Updated pcredemo.c to include calls to free() to release the memory for the
4260 compiled pattern.
4261
4262 17. The generated file chartables.c was being created in the source directory
4263 instead of in the building directory. This caused the build to fail if the
4264 source directory was different from the building directory, and was
4265 read-only.
4266
4267 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4268 file. No doubt somebody will tell me if they don't make sense... Also added
4269 Dan Mooney's comments about building on OpenVMS.
4270
4271 19. Added support for partial matching via the PCRE_PARTIAL option for
4272 pcre_exec() and the \P data escape in pcretest.
4273
4274 20. Extended pcretest with 3 new pattern features:
4275
4276 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4277 write the compiled pattern to the file whose name is "rest-of-line".
4278 This is a straight binary dump of the data, with the saved pointer to
4279 the character tables forced to be NULL. The study data, if any, is
4280 written too. After writing, pcretest reads a new pattern.
4281
4282 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4283 compiled pattern from the given file. There must not be any
4284 occurrences of "<" in the file name (pretty unlikely); if there are,
4285 pcretest will instead treat the initial "<" as a pattern delimiter.
4286 After reading in the pattern, pcretest goes on to read data lines as
4287 usual.
4288
4289 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4290 and 16-bit fields in a compiled pattern, to simulate a pattern that
4291 was compiled on a host of opposite endianness.
4292
4293 21. The pcre-exec() function can now cope with patterns that were compiled on
4294 hosts of opposite endianness, with this restriction:
4295
4296 As for any compiled expression that is saved and used later, the tables
4297 pointer field cannot be preserved; the extra_data field in the arguments
4298 to pcre_exec() should be used to pass in a tables address if a value
4299 other than the default internal tables were used at compile time.
4300
4301 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4302 now diagnosed as an error. Previously, most of the time, a negative number
4303 would have been treated as zero, but if in addition "ovector" was passed as
4304 NULL, a crash could occur.
4305
4306 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4307 new versions from the libtool 1.5 distribution (the last one is a copy of
4308 a file called libtool.m4). This seems to have fixed the need to patch
4309 "configure" to support Darwin 1.3 (which I used to do). However, I still
4310 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4311 workstation).
4312
4313 24. Changed the PCRE licence to be the more standard "BSD" licence.
4314
4315
4316 Version 4.5 01-Dec-03
4317 ---------------------
4318
4319 1. There has been some re-arrangement of the code for the match() function so
4320 that it can be compiled in a version that does not call itself recursively.
4321 Instead, it keeps those local variables that need separate instances for
4322 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4323 needs to "recurse". Keeping track of where control must go is done by means
4324 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4325 hide most of the details from the main code, and operates only if
4326 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4327 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4328 operating.
4329
4330 To make it easier for callers to provide specially tailored get/free
4331 functions for this usage, two new functions, pcre_stack_malloc, and
4332 pcre_stack_free, are used. They are always called in strict stacking order,
4333 and the size of block requested is always the same.
4334
4335 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4336 PCRE has been compiled to use the stack or the heap for recursion. The
4337 -C option of pcretest uses this to show which version is compiled.
4338
4339 A new data escape \S, is added to pcretest; it causes the amounts of store
4340 obtained and freed by both kinds of malloc/free at match time to be added
4341 to the output.
4342
4343 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4344 what's available on my current Linux desktop machine.
4345
4346 3. When matching a UTF-8 string, the test for a valid string at the start has
4347 been extended. If start_offset is not zero, PCRE now checks that it points
4348 to a byte that is the start of a UTF-8 character. If not, it returns
4349 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4350 this is necessary because there may be backward assertions in the pattern.
4351 When matching the same subject several times, it may save resources to use
4352 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4353
4354 4. The code for checking the validity of UTF-8 strings has been tightened so
4355 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4356 containing "overlong sequences".
4357
4358 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4359 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4360 should have been "&", but it just so happened that all the cases this let
4361 through by mistake were picked up later in the function.
4362
4363 6. I had used a variable called "isblank" - this is a C99 function, causing
4364 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4365
4366 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4367 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4368 so the version is shown; (c) stop "make test" echoing "./RunTest".
4369
4370 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4371
4372 9. The prototype for memmove() for systems that don't have it was using
4373 size_t, but the inclusion of the header that defines size_t was later. I've
4374 moved the #includes for the C headers earlier to avoid this.
4375
4376 10. Added some adjustments to the code to make it easier to compiler on certain
4377 special systems:
4378
4379 (a) Some "const" qualifiers were missing.
4380 (b) Added the macro EXPORT before all exported functions; by default this
4381 is defined to be empty.
4382 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4383 that it reads its output file name as an argument instead of writing
4384 to the standard output and assuming this can be redirected.
4385
4386 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4387 class containing characters with values greater than 255, PCRE compilation
4388 went into a loop.
4389
4390 12. A recursive reference to a subpattern that was within another subpattern
4391 that had a minimum quantifier of zero caused PCRE to crash. For example,
4392 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4393 recursion. If the recursively-called subpattern itself had a zero repeat,
4394 that was OK.
4395
4396 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4397 buffer into which it was copied (for escape processing) was still set at
4398 1024, so long lines caused crashes.
4399
4400 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4401 "internal error: code overflow...". This applied to any character class
4402 that was followed by a possessive quantifier.
4403
4404 15. Modified the Makefile to add libpcre.la as a prerequisite for
4405 libpcreposix.la because I was told this is needed for a parallel build to
4406 work.
4407
4408 16. If a pattern that contained .* following optional items at the start was
4409 studied, the wrong optimizing data was generated, leading to matching
4410 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4411 matching string must start with a or b or c. The correct conclusion for
4412 this pattern is that a match can start with any character.
4413
4414
4415 Version 4.4 13-Aug-03
4416 ---------------------
4417
4418 1. In UTF-8 mode, a character class containing characters with values between
4419 127 and 255 was not handled correctly if the compiled pattern was studied.
4420 In fixing this, I have also improved the studying algorithm for such
4421 classes (slightly).
4422
4423 2. Three internal functions had redundant arguments passed to them. Removal
4424 might give a very teeny performance improvement.
4425
4426 3. Documentation bug: the value of the capture_top field in a callout is *one
4427 more than* the number of the hightest numbered captured substring.
4428
4429 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4430 in incorrectly linking with a previously installed version. They now link
4431 explicitly with libpcre.la.
4432
4433 5. configure.in no longer needs to recognize Cygwin specially.
4434
4435 6. A problem in pcre.in for Windows platforms is fixed.
4436
4437 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4438 pcretest, it used to include the size of the study block as part of its
4439 output. Unfortunately, the structure contains a field that has a different
4440 size on different hardware architectures. This meant that the tests that
4441 showed this size failed. As the block is currently always of a fixed size,
4442 this information isn't actually particularly useful in pcretest output, so
4443 I have just removed it.
4444
4445 8. Three pre-processor statements accidentally did not start in column 1.
4446 Sadly, there are *still* compilers around that complain, even though
4447 standard C has not required this for well over a decade. Sigh.
4448
4449 9. In pcretest, the code for checking callouts passed small integers in the
4450 callout_data field, which is a void * field. However, some picky compilers
4451 complained about the casts involved for this on 64-bit systems. Now
4452 pcretest passes the address of the small integer instead, which should get
4453 rid of the warnings.
4454
4455 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4456 both compile and run time, and gives an error if an invalid UTF-8 sequence
4457 is found. There is a option for disabling this check in cases where the
4458 string is known to be correct and/or the maximum performance is wanted.
4459
4460 11. In response to a bug report, I changed one line in Makefile.in from
4461
4462 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4463 to
4464 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4465
4466 to look similar to other lines, but I have no way of telling whether this
4467 is the right thing to do, as I do not use Windows. No doubt I'll get told
4468 if it's wrong...
4469
4470
4471 Version 4.3 21-May-03
4472 ---------------------
4473
4474 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4475 Makefile.
4476
4477 2. Some refactoring to improve the quality of the code:
4478
4479 (i) The utf8_table... variables are now declared "const".
4480
4481 (ii) The code for \cx, which used the "case flipping" table to upper case
4482 lower case letters, now just substracts 32. This is ASCII-specific,
4483 but the whole concept of \cx is ASCII-specific, so it seems
4484 reasonable.
4485
4486 (iii) PCRE was using its character types table to recognize decimal and
4487 hexadecimal digits in the pattern. This is silly, because it handles
4488 only 0-9, a-f, and A-F, but the character types table is locale-
4489 specific, which means strange things might happen. A private
4490 table is now used for this - though it costs 256 bytes, a table is
4491 much faster than multiple explicit tests. Of course, the standard
4492 character types table is still used for matching digits in subject
4493 strings against \d.
4494
4495 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4496 ending in _t are). So I've renamed it as ESC_tee.
4497
4498 3. The first argument for regexec() in the POSIX wrapper should have been
4499 defined as "const".
4500
4501 4. Changed pcretest to use malloc() for its buffers so that they can be
4502 Electric Fenced for debugging.
4503
4504 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4505 to read one or more bytes before the start of the subject string. Often this
4506 had no effect on PCRE's behaviour, but in some circumstances it could
4507 provoke a segmentation fault.
4508
4509 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4510 to try to read one or more bytes before the start of the subject string.
4511
4512 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4513 UTF-8 support could misbehave in various ways if the subject string
4514 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4515 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4516 back over UTF-8 characters.)
4517
4518
4519 Version 4.2 14-Apr-03
4520 ---------------------
4521
4522 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4523
4524 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4525 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4526 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4527 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4528 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4529 and BUILD_EXEEXT
4530 Note: automatic setting of the BUILD variables is not yet working
4531 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4532 compile-time but not at link-time
4533 [LINK]: use for linking executables only
4534 make different versions for Windows and non-Windows
4535 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4536 libraries
4537 [LINK_FOR_BUILD]: new variable
4538 [OBJEXT]: use throughout
4539 [EXEEXT]: use throughout
4540 <winshared>: new target
4541 <wininstall>: new target
4542 <dftables.o>: use native compiler
4543 <dftables>: use native linker
4544 <install>: handle Windows platform correctly
4545 <clean>: ditto
4546 <check>: ditto
4547 copy DLL to top builddir before testing
4548
4549 As part of these changes, -no-undefined was removed again. This was reported
4550 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4551 in any case.
4552
4553 3. Some tidies to get rid of compiler warnings:
4554
4555 . In the match_data structure, match_limit was an unsigned long int, whereas
4556 match_call_count was an int. I've made them both unsigned long ints.
4557
4558 . In pcretest the fact that a const uschar * doesn't automatically cast to
4559 a void * provoked a warning.
4560
4561 . Turning on some more compiler warnings threw up some "shadow" variables
4562 and a few more missing casts.
4563
4564 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4565 option, a class that contained a single character with a value between 128
4566 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4567
4568 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4569 option, a class that contained several characters, but with at least one
4570 whose value was between 128 and 255 caused PCRE to crash.
4571
4572
4573 Version 4.1 12-Mar-03
4574 ---------------------
4575
4576 1. Compiling with gcc -pedantic found a couple of places where casts were
4577 needed, and a string in dftables.c that was longer than standard compilers are
4578 required to support.
4579
4580 2. Compiling with Sun's compiler found a few more places where the code could
4581 be tidied up in order to avoid warnings.
4582
4583 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4584 first of these names is deprecated in the latest Autoconf in favour of the name
4585 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4586 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4587 analogy I have changed it to CFLAGS_FOR_BUILD.
4588
4589 4. Added -no-undefined to the linking command in the Makefile, because this is
4590 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4591 linking step for the pcreposix library.
4592
4593 5. PCRE was failing to diagnose the case of two named groups with the same
4594 name.
4595
4596 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4597 literal character that is needed in the subject for a match, and scans along to
4598 ensure that it is present before embarking on the full matching process. This
4599 saves time in cases of nested unlimited repeats that are never going to match.
4600 Problem: the scan can take a lot of time if the subject is very long (e.g.
4601 megabytes), thus penalizing straightforward matches. It is now done only if the
4602 amount of subject to be scanned is less than 1000 bytes.
4603
4604 7. A lesser problem with the same optimization is that it was recording the
4605 first character of an anchored pattern as "needed", thus provoking a search
4606 right along the subject, even when the first match of the pattern was going to
4607 fail. The "needed" character is now not set for anchored patterns, unless it
4608 follows something in the pattern that is of non-fixed length. Thus, it still
4609 fulfils its original purpose of finding quick non-matches in cases of nested
4610 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4611
4612
4613 Version 4.0 17-Feb-03
4614 ---------------------
4615
4616 1. If a comment in an extended regex that started immediately after a meta-item
4617 extended to the end of string, PCRE compiled incorrect data. This could lead to
4618 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4619
4620 2. Moved to autoconf 2.53 and libtool 1.4.2.
4621
4622 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4623 the special perltest8 script is no longer needed - all the tests can be run
4624 from a single perltest script.
4625
4626 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4627 by \s. It has now been removed in PCRE. This means it isn't recognized as
4628 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4629 class [:space:] *does* include VT, thereby creating a mess.
4630
4631 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4632 space and tab.
4633
4634 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4635 its new features into the main test script, reducing the number of scripts.
4636
4637 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4638 were backward compatible, and made the (?i) apply to the whole pattern, as if
4639 /i were given. Now it behaves more logically, and applies the option setting
4640 only to what follows. PCRE has been changed to follow suit. However, if it
4641 finds options settings right at the start of the pattern, it extracts them into
4642 the global options, as before. Thus, they show up in the info data.
4643
4644 8. Added support for the \Q...\E escape sequence. Characters in between are
4645 treated as literals. This is slightly different from Perl in that $ and @ are
4646 also handled as literals inside the quotes. In Perl, they will cause variable
4647 interpolation. Note the following examples:
4648
4649 Pattern PCRE matches Perl matches
4650
4651 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4652 \Qabc\$xyz\E abc\$xyz abc\$xyz
4653 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4654
4655 For compatibility with Perl, \Q...\E sequences are recognized inside character
4656 classes as well as outside them.
4657
4658 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4659 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4660 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4661 signed/unsigned warnings.
4662
4663 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4664 option for pcretest, so I've replaced it by a simple function that does just
4665 that job.
4666
4667 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4668 "pcregrep -".
4669
4670 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4671 Java package. This provides some syntactic sugar for simple cases of what my
4672 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4673 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4674 item, you can use this simplified notation. Note that only makes sense with
4675 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4676 greediness, whatever the setting of the PCRE_UNGREEDY option.
4677
4678 13. A change of greediness default within a pattern was not taking effect at
4679 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4680 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4681 was abstracted outside.
4682
4683 14. PCRE now supports the \G assertion. It is true when the current matching
4684 position is at the start point of the match. This differs from \A when the
4685 starting offset is non-zero. Used with the /g option of pcretest (or similar
4686 code), it works in the same way as it does for Perl's /g option. If all
4687 alternatives of a regex begin with \G, the expression is anchored to the start
4688 match position, and the "anchored" flag is set in the compiled expression.
4689
4690 15. Some bugs concerning the handling of certain option changes within patterns
4691 have been fixed. These applied to options other than (?ims). For example,
4692 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4693 been the other way round. Some of this was related to change 7 above.
4694
4695 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4696 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4697 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4698 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4699
4700 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4701 mode. Unlike ".", it always matches newline, whatever the setting of
4702 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4703 assertions. Perl allows it, but it doesn't (in general) work because it can't
4704 calculate the length of the lookbehind. At least, that's the case for Perl
4705 5.8.0 - I've been told they are going to document that it doesn't work in
4706 future.
4707
4708 18. Added an error diagnosis for escapes that PCRE does not support: these are
4709 \L, \l, \N, \P, \p, \U, \u, and \X.
4710
4711 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4712 reading past the end of the pattern in cases such as /[abcd/.
4713
4714 20. PCRE was getting more memory than necessary for patterns with classes that
4715 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4716
4717 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4718 compiling PCRE for use with Virtual Pascal.
4719
4720 22. Small fix to the Makefile to make it work properly if the build is done
4721 outside the source tree.
4722
4723 23. Added a new extension: a condition to go with recursion. If a conditional
4724 subpattern starts with (?(R) the "true" branch is used if recursion has
4725 happened, whereas the "false" branch is used only at the top level.
4726
4727 24. When there was a very long string of literal characters (over 255 bytes
4728 without UTF support, over 250 bytes with UTF support), the computation of how
4729 much memory was required could be incorrect, leading to segfaults or other
4730 strange effects.
4731
4732 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4733 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4734 there was a subsequent back reference to those brackets. This meant that, for
4735 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4736 possible to check for precisely this case. All we can do is abandon the
4737 optimization if .* occurs inside capturing brackets when there are any back
4738 references whatsoever. (See below for a better fix that came later.)
4739
4740 26. The handling of the optimization for finding the first character of a
4741 non-anchored pattern, and for finding a character that is required later in the
4742 match were failing in some cases. This didn't break the matching; it just
4743 failed to optimize when it could. The way this is done has been re-implemented.
4744
4745 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4746
4747 28. Added a new feature that provides some of the functionality that Perl
4748 provides with (?{...}). The facility is termed a "callout". The way it is done
4749 in PCRE is for the caller to provide an optional function, by setting
4750 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4751 global variable. By default it is unset, which disables all calling out. To get
4752 the function called, the regex must include (?C) at appropriate points. This
4753 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4754 This provides a means of identifying different callout points. When PCRE
4755 reaches such a point in the regex, if pcre_callout has been set, the external
4756 function is called. It is provided with data in a structure called
4757 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4758 matching continues; if it returns a non-zero value, the match at the current
4759 point fails. However, backtracking will occur if possible. [This was changed
4760 later and other features added - see item 49 below.]
4761
4762 29. pcretest is upgraded to test the callout functionality. It provides a
4763 callout function that displays information. By default, it shows the start of
4764 the match and the current position in the text. There are some new data escapes
4765 to vary what happens:
4766
4767 \C+ in addition, show current contents of captured substrings
4768 \C- do not supply a callout function
4769 \C!n return 1 when callout number n is reached
4770 \C!n!m return 1 when callout number n is reached for the mth time
4771
4772 30. If pcregrep was called with the -l option and just a single file name, it
4773 output "<stdin>" if a match was found, instead of the file name.
4774
4775 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4776 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4777 pcre_exec(). This saves a malloc/free per call. The default value of
4778 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4779 when configuring.
4780
4781 32. The default maximum size of a compiled pattern is 64K. There have been a
4782 few cases of people hitting this limit. The code now uses macros to handle the
4783 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4784 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4785 configuring. Tests 2 and 5 work only with 2-byte links because they output
4786 debugging information about compiled patterns.
4787
4788 33. Internal code re-arrangements:
4789
4790 (a) Moved the debugging function for printing out a compiled regex into
4791 its own source file (printint.c) and used #include to pull it into
4792 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4793 separate copies.
4794
4795 (b) Defined the list of op-code names for debugging as a macro in
4796 internal.h so that it is next to the definition of the opcodes.
4797
4798 (c) Defined a table of op-code lengths for simpler skipping along compiled
4799 code. This is again a macro in internal.h so that it is next to the
4800 definition of the opcodes.
4801
4802 34. Added support for recursive calls to individual subpatterns, along the
4803 lines of Robin Houston's patch (but implemented somewhat differently).
4804
4805 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4806 allow it to read and process whole directories in Win32. This code was
4807 contributed by Lionel Fourquaux; it has not been tested by me.
4808
4809 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4810 used to name a group. Names consist of alphanumerics and underscores, and must
4811 be unique. Back references use the syntax (?P=name) and recursive calls use
4812 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4813 numbers. The function pcre_fullinfo() can be used after compilation to extract
4814 a name/number map. There are three relevant calls:
4815
4816 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4817 PCRE_INFO_NAMECOUNT yields the number of entries
4818 PCRE_INFO_NAMETABLE yields a pointer to the map.
4819
4820 The map is a vector of fixed-size entries. The size of each entry depends on
4821 the length of the longest name used. The first two bytes of each entry are the
4822 group number, most significant byte first. There follows the corresponding
4823 name, zero terminated. The names are in alphabetical order.
4824
4825 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4826 case instead of 255. Making it the same both with and without UTF-8 support
4827 means that the same test output works with both.
4828
4829 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4830 calling malloc() with a zero argument.
4831
4832 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4833 optimization. I've improved things by keeping a bitmap of backreferences with
4834 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4835 fact referenced, the optimization can be applied. It is unlikely that a
4836 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4837 the match to follow \n) will appear inside brackets with a number greater than
4838 31, but if it does, any back reference > 31 suppresses the optimization.
4839
4840 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4841 of disabling numbered capturing parentheses. Any opening parenthesis that is
4842 not followed by ? behaves as if it were followed by ?: but named parentheses
4843 can still be used for capturing (and they will acquire numbers in the usual
4844 way).
4845
4846 41. Redesigned the return codes from the match() function into yes/no/error so
4847 that errors can be passed back from deep inside the nested calls. A malloc
4848 failure while inside a recursive subpattern call now causes the
4849 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4850
4851 42. It is now possible to set a limit on the number of times the match()
4852 function is called in a call to pcre_exec(). This facility makes it possible to
4853 limit the amount of recursion and backtracking, though not in a directly
4854 obvious way, because the match() function is used in a number of different
4855 circumstances. The count starts from zero for each position in the subject
4856 string (for non-anchored patterns). The default limit is, for compatibility, a
4857 large number, namely 10 000 000. You can change this in two ways:
4858
4859 (a) When configuring PCRE before making, you can use --with-match-limit=n
4860 to set a default value for the compiled library.
4861
4862 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4863 a different value is set. See 45 below.
4864
4865 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4866
4867 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4868 of things that can be changed at compile time. The first argument specifies
4869 what is wanted and the second points to where the information is to be placed.
4870 The current list of available information is:
4871
4872 PCRE_CONFIG_UTF8
4873
4874 The output is an integer that is set to one if UTF-8 support is available;
4875 otherwise it is set to zero.
4876
4877 PCRE_CONFIG_NEWLINE
4878
4879 The output is an integer that it set to the value of the code that is used for
4880 newline. It is either LF (10) or CR (13).
4881
4882 PCRE_CONFIG_LINK_SIZE
4883
4884 The output is an integer that contains the number of bytes used for internal
4885 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4886
4887 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4888
4889 The output is an integer that contains the threshold above which the POSIX
4890 interface uses malloc() for output vectors. See item 31 above.
4891
4892 PCRE_CONFIG_MATCH_LIMIT
4893
4894 The output is an unsigned integer that contains the default limit of the number
4895 of match() calls in a pcre_exec() execution. See 42 above.
4896
4897 44. pcretest has been upgraded by the addition of the -C option. This causes it
4898 to extract all the available output from the new pcre_config() function, and to
4899 output it. The program then exits immediately.
4900
4901 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4902 order to support additional features. One way would have been to define
4903 pcre_exec2() (for example) with extra arguments, but this would not have been
4904 extensible, and would also have required all calls to the original function to
4905 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4906 is used for passing in "extra" data from pcre_study().
4907
4908 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4909 contains the following fields:
4910
4911 flags a bitmap indicating which of the following fields are set
4912 study_data opaque data from pcre_study()
4913 match_limit a way of specifying a limit on match() calls for a specific
4914 call to pcre_exec()
4915 callout_data data for callouts (see 49 below)
4916
4917 The flag bits are also defined in pcre.h, and are
4918
4919 PCRE_EXTRA_STUDY_DATA
4920 PCRE_EXTRA_MATCH_LIMIT
4921 PCRE_EXTRA_CALLOUT_DATA
4922
4923 The pcre_study() function now returns one of these new pcre_extra blocks, with
4924 the actual study data pointed to by the study_data field, and the
4925 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4926 before. That is, this change is entirely upwards-compatible and requires no
4927 change to existing code.
4928
4929 If you want to pass in additional data to pcre_exec(), you can either place it
4930 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4931 block.
4932
4933 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4934 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4935 times with different match limits, until it finds the minimum value needed for
4936 pcre_exec() to complete. The value is then output. This can be instructive; for
4937 most simple matches the number is quite small, but for pathological cases it
4938 gets very large very quickly.
4939
4940 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4941 returns the size of the data block pointed to by the study_data field in a
4942 pcre_extra block, that is, the value that was passed as the argument to
4943 pcre_malloc() when PCRE was getting memory in which to place the information
4944 created by pcre_study(). The fourth argument should point to a size_t variable.
4945 pcretest has been extended so that this information is shown after a successful
4946 pcre_study() call when information about the compiled regex is being displayed.
4947
4948 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4949 because what follows is always an absolute path. (Later: it turns out that this
4950 is more than cosmetic for MinGW, because it doesn't like empty path
4951 components.)
4952
4953 49. Some changes have been made to the callout feature (see 28 above):
4954
4955 (i) A callout function now has three choices for what it returns:
4956
4957 0 => success, carry on matching
4958 > 0 => failure at this point, but backtrack if possible
4959 < 0 => serious error, return this value from pcre_exec()
4960
4961 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4962 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4963 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4964 use by callout functions. It will never be used by PCRE itself.
4965
4966 (ii) The pcre_extra structure (see 45 above) has a void * field called
4967 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4968 pcre_callout_block structure has a field of the same name. The contents of
4969 the field passed in the pcre_extra structure are passed to the callout
4970 function in the corresponding field in the callout block. This makes it
4971 easier to use the same callout-containing regex from multiple threads. For
4972 testing, the pcretest program has a new data escape
4973
4974 \C*n pass the number n (may be negative) as callout_data
4975
4976 If the callout function in pcretest receives a non-zero value as
4977 callout_data, it returns that value.
4978
4979 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4980 there were some redundant $(CFLAGS) in commands that are now specified as
4981 $(LINK), which already includes $(CFLAGS).
4982
4983 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4984 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4985 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4986 one-byte characters throughout. Note that case-insensitive matching applies
4987 only to characters whose values are less than 256. PCRE doesn't support the
4988 notion of cases for higher-valued characters.
4989
4990 (i) A character class whose characters are all within 0-255 is handled as
4991 a bit map, and the map is inverted for negative classes. Previously, a
4992 character > 255 always failed to match such a class; however it should
4993 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4994
4995 (ii) A negated character class with a single character < 255 is coded as
4996 "not this character" (OP_NOT). This wasn't working properly when the test
4997 character was multibyte, either singly or repeated.
4998
4999 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5000 mode, for example: \x{100}{2,3}.
5001
5002 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5003 singly or repeated) now correctly test multibyte characters. However,
5004 PCRE doesn't recognize any characters with values greater than 255 as
5005 digits, spaces, or word characters. Such characters always match \D, \S,
5006 and \W, and never match \d, \s, or \w.
5007
5008 (v) Classes may now contain characters and character ranges with values
5009 greater than 255. For example: [ab\x{100}-\x{400}].
5010
5011 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5012 PCRE in UTF-8 mode.
5013
5014 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5015 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5016 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5017 value.)
5018
5019 53. The single man page has become too large. I have therefore split it up into
5020 a number of separate man pages. These also give rise to individual HTML pages;
5021 these are now put in a separate directory, and there is an index.html page that
5022 lists them all. Some hyperlinking between the pages has been installed.
5023
5024 54. Added convenience functions for handling named capturing parentheses.
5025
5026 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5027 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5028 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5029 are faulted.
5030
5031 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5032 calling configure. These values are used when compiling the dftables.c program
5033 which is run to generate the source of the default character tables. They
5034 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5035 you will need to set these values.
5036
5037 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5038
5039
5040 Version 3.9 02-Jan-02
5041 ---------------------
5042
5043 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5044
5045 2. If --disable-static was given, the building process failed when trying to
5046 build pcretest and pcregrep. (For some reason it was using libtool to compile
5047 them, which is not right, as they aren't part of the library.)
5048
5049
5050 Version 3.8 18-Dec-01
5051 ---------------------
5052
5053 1. The experimental UTF-8 code was completely screwed up. It was packing the
5054 bytes in the wrong order. How dumb can you get?
5055
5056
5057 Version 3.7 29-Oct-01
5058 ---------------------
5059
5060 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5061 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5062 this didn't happen under Solaris 8, where I normally test things.
5063
5064 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5065 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5066 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5067 this fix an hour or so after the initial 3.7 release.)
5068
5069
5070 Version 3.6 23-Oct-01
5071 ---------------------
5072
5073 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5074 offsets passed as NULL with zero offset count.
5075
5076 2. The config.guess and config.sub files had not been updated when I moved to
5077 the latest autoconf.
5078
5079
5080 Version 3.5 15-Aug-01
5081 ---------------------
5082
5083 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5084 had been forgotten.
5085
5086 2. By using declared but undefined structures, we can avoid using "void"
5087 definitions in pcre.h while keeping the internal definitions of the structures
5088 private.
5089
5090 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5091 user point of view, this means that both static and shared libraries are built
5092 by default, but this can be individually controlled. More of the work of
5093 handling this static/shared cases is now inside libtool instead of PCRE's make
5094 file.
5095
5096 4. The pcretest utility is now installed along with pcregrep because it is
5097 useful for users (to test regexs) and by doing this, it automatically gets
5098 relinked by libtool. The documentation has been turned into a man page, so
5099 there are now .1, .txt, and .html versions in /doc.
5100
5101 5. Upgrades to pcregrep:
5102 (i) Added long-form option names like gnu grep.
5103 (ii) Added --help to list all options with an explanatory phrase.
5104 (iii) Added -r, --recursive to recurse into sub-directories.
5105 (iv) Added -f, --file to read patterns from a file.
5106
5107 6. pcre_exec() was referring to its "code" argument before testing that
5108 argument for NULL (and giving an error if it was NULL).
5109
5110 7. Upgraded Makefile.in to allow for compiling in a different directory from
5111 the source directory.
5112
5113 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5114 options bits, the pointer it was passed was to an int instead of to an unsigned
5115 long int. This mattered only on 64-bit systems.
5116
5117 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5118 generated) instead of pcre.in, which it its source. Also made the same change
5119 in several of the .c files.
5120
5121 10. A new release of gcc defines printf() as a macro, which broke pcretest
5122 because it had an ifdef in the middle of a string argument for printf(). Fixed
5123 by using separate calls to printf().
5124
5125 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5126 script, to force use of CR or LF instead of \n in the source. On non-Unix
5127 systems, the value can be set in config.h.
5128
5129 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5130 absolute limit. Changed the text of the error message to make this clear, and
5131 likewise updated the man page.
5132
5133 13. The limit of 99 on the number of capturing subpatterns has been removed.
5134 The new limit is 65535, which I hope will not be a "real" limit.
5135
5136
5137 Version 3.4 22-Aug-00
5138 ---------------------
5139
5140 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5141
5142 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5143
5144
5145 Version 3.3 01-Aug-00
5146 ---------------------
5147
5148 1. If an octal character was given, but the value was greater than \377, it
5149 was not getting masked to the least significant bits, as documented. This could
5150 lead to crashes in some systems.
5151
5152 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5153 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5154
5155 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5156 These just pass their arguments on to (pcre_free)(), but they are provided
5157 because some uses of PCRE bind it to non-C systems that can call its functions,
5158 but cannot call free() or pcre_free() directly.
5159
5160 4. Add "make test" as a synonym for "make check". Corrected some comments in
5161 the Makefile.
5162
5163 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5164 Makefile.
5165
5166 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5167 command called pgrep for grepping around the active processes.
5168
5169 7. Added the beginnings of support for UTF-8 character strings.
5170
5171 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5172 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5173 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5174 out for the ar command.)
5175
5176
5177 Version 3.2 12-May-00
5178 ---------------------
5179
5180 This is purely a bug fixing release.
5181
5182 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5183 of ZA. This was just one example of several cases that could provoke this bug,
5184 which was introduced by change 9 of version 2.00. The code for breaking
5185 infinite loops after an iteration that matches an empty string was't working
5186 correctly.
5187
5188 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5189 when matched against abbab (for example). After matching an empty string, it
5190 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5191 caused it to match further down the string than it should.
5192
5193 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5194 was there because it doesn't seem to be needed, and it causes trouble on some
5195 systems, as it is not a Standard C header. It has been removed.
5196
5197 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5198 were reported on the Macintosh. The changes were from
5199
5200 while ((c = *(++ptr)) != 0 && c != '\n');
5201 to
5202 while ((c = *(++ptr)) != 0 && c != '\n') ;
5203
5204 Totally extraordinary, but if that's what it takes...
5205
5206 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5207 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5208 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5209 assumes the way PCRE uses memmove() (always moving upwards).
5210
5211 6. PCRE is being used in one environment where strchr() is not available. There
5212 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5213 faster code anyway.
5214
5215
5216 Version 3.1 09-Feb-00
5217 ---------------------
5218
5219 The only change in this release is the fixing of some bugs in Makefile.in for
5220 the "install" target:
5221
5222 (1) It was failing to install pcreposix.h.
5223
5224 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5225
5226
5227 Version 3.0 01-Feb-00
5228 ---------------------
5229
5230 1. Add support for the /+ modifier to perltest (to output $` like it does in
5231 pcretest).
5232
5233 2. Add support for the /g modifier to perltest.
5234
5235 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5236 matches null strings.
5237
5238 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5239 pattern. Perl treats empty patterns specially - it reuses the most recent
5240 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5241 effect.
5242
5243 5. The POSIX interface was broken in that it was just handing over the POSIX
5244 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5245 required a bigger vector, with some working space on the end. This means that
5246 the POSIX wrapper now has to get and free some memory, and copy the results.
5247
5248 6. Added some simple autoconf support, placing the test data and the
5249 documentation in separate directories, re-organizing some of the
5250 information files, and making it build pcre-config (a GNU standard). Also added
5251 libtool support for building PCRE as a shared library, which is now the
5252 default.
5253
5254 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5255 09 are not valid octal constants. Single digits will be used for minor values
5256 less than 10.
5257
5258 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5259 existing programs that set these in the POSIX interface can use PCRE without
5260 modification.
5261
5262 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5263 return all that pcre_info() returns, plus additional data. The pcre_info()
5264 function is retained for compatibility, but is considered to be obsolete.
5265
5266 10. Added experimental recursion feature (?R) to handle one common case that
5267 Perl 5.6 will be able to do with (?p{...}).
5268
5269 11. Added support for POSIX character classes like [:alpha:], which Perl is
5270 adopting.
5271
5272
5273 Version 2.08 31-Aug-99
5274 ----------------------
5275
5276 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5277 trying to match at the startoffset position, but instead was moving forward to
5278 the next newline as if a previous match had failed.
5279
5280 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5281 and could get into a loop if a null string was matched other than at the start
5282 of the subject.
5283
5284 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5285 be distinguished at compile time, and for completeness also added PCRE_DATE.
5286
5287 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5288 in GnuWin32 environments.
5289
5290
5291 Version 2.07 29-Jul-99
5292 ----------------------
5293
5294 1. The documentation is now supplied in plain text form and HTML as well as in
5295 the form of man page sources.
5296
5297 2. C++ compilers don't like assigning (void *) values to other pointer types.
5298 In particular this affects malloc(). Although there is no problem in Standard
5299 C, I've put in casts to keep C++ compilers happy.
5300
5301 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5302 should be (const char *).
5303
5304 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5305 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5306 However, I haven't made this a standard facility. The documentation doesn't
5307 mention it, and the Makefile doesn't support it.
5308
5309 5. The Makefile now contains an "install" target, with editable destinations at
5310 the top of the file. The pcretest program is not installed.
5311
5312 6. pgrep -V now gives the PCRE version number and date.
5313
5314 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5315 causing the entire string to be ignored, instead of just the last character.
5316
5317 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5318 non-matching string, it can take a very, very long time, even for strings of
5319 quite modest length, because of the nested recursion. PCRE now does better in
5320 some of these cases. It does this by remembering the last required literal
5321 character in the pattern, and pre-searching the subject to ensure it is present
5322 before running the real match. In other words, it applies a heuristic to detect
5323 some types of certain failure quickly, and in the above example, if presented
5324 with a string that has no trailing " it gives "no match" very quickly.
5325
5326 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5327 other alternatives are tried instead.
5328
5329
5330 Version 2.06 09-Jun-99
5331 ----------------------
5332
5333 1. Change pcretest's output for amount of store used to show just the code
5334 space, because the remainder (the data block) varies in size between 32-bit and
5335 64-bit systems.
5336
5337 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5338 start matching at. This allows lookbehinds to work when searching for multiple
5339 occurrences in a string.
5340
5341 3. Added additional options to pcretest for testing multiple occurrences:
5342
5343 /+ outputs the rest of the string that follows a match
5344 /g loops for multiple occurrences, using the new startoffset argument
5345 /G loops for multiple occurrences by passing an incremented pointer
5346
5347 4. PCRE wasn't doing the "first character" optimization for patterns starting
5348 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5349 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5350 the letter 'x'. On long subject strings, this gives a significant speed-up.
5351
5352
5353 Version 2.05 21-Apr-99
5354 ----------------------
5355
5356 1. Changed the type of magic_number from int to long int so that it works
5357 properly on 16-bit systems.
5358
5359 2. Fixed a bug which caused patterns starting with .* not to work correctly
5360 when the subject string contained newline characters. PCRE was assuming
5361 anchoring for such patterns in all cases, which is not correct because .* will
5362 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5363 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5364 must be retried after every newline in the subject.
5365
5366
5367 Version 2.04 18-Feb-99
5368 ----------------------
5369
5370 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5371 computation of the store needed to hold the pattern was incorrect (too large).
5372 If such patterns were nested a few deep, this could multiply and become a real
5373 problem.
5374
5375 2. Added /M option to pcretest to show the memory requirement of a specific
5376 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5377
5378 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5379 compiled in such a way that the backtracking after subsequent failure was
5380 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5381 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5382
5383
5384 Version 2.03 02-Feb-99
5385 ----------------------
5386
5387 1. Fixed typo and small mistake in man page.
5388
5389 2. Added 4th condition (GPL supersedes if conflict) and created separate
5390 LICENCE file containing the conditions.
5391
5392 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5393 Perl, that is the internal \ allows the delimiter to be included in the
5394 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5395 the final delimiter, add \ to the end of the pattern (to test the error).
5396
5397 4. Added the convenience functions for extracting substrings after a successful
5398 match. Updated pcretest to make it able to test these functions.
5399
5400
5401 Version 2.02 14-Jan-99
5402 ----------------------
5403
5404 1. Initialized the working variables associated with each extraction so that
5405 their saving and restoring doesn't refer to uninitialized store.
5406
5407 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5408 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5409 fix the problem.
5410
5411 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5412 calls, and wasn't printing the correct value for compiling calls. Increased the
5413 default value of LOOPREPEAT, and the number of significant figures in the
5414 times.
5415
5416 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5417
5418 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5419 a building problem on Windows NT with a FAT file system.
5420
5421
5422 Version 2.01 21-Oct-98
5423 ----------------------
5424
5425 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5426 to character tables built by pcre_maketables() in the current locale. If NULL
5427 is passed, the default tables are used.
5428
5429
5430 Version 2.00 24-Sep-98
5431 ----------------------
5432
5433 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5434 it any more.
5435
5436 2. Allow quantification of (?>) groups, and make it work correctly.
5437
5438 3. The first character computation wasn't working for (?>) groups.
5439
5440 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5441 end of the subject) and add 5.005's \z, which really does match only at the
5442 very end of the subject.
5443
5444 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5445
5446 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5447 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5448 localized options. All options to pcre_study() were also removed.
5449
5450 7. Add other new features from 5.005:
5451
5452 $(?<= positive lookbehind
5453 $(?<! negative lookbehind
5454 (?imsx-imsx) added the unsetting capability
5455 such a setting is global if at outer level; local otherwise
5456 (?imsx-imsx:) non-capturing groups with option setting
5457 (?(cond)re|re) conditional pattern matching
5458
5459 A backreference to itself in a repeated group matches the previous
5460 captured string.
5461
5462 8. General tidying up of studying (both automatic and via "study")
5463 consequential on the addition of new assertions.
5464
5465 9. As in 5.005, unlimited repeated groups that could match an empty substring
5466 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5467 runtime if any iteration does actually match an empty substring.
5468
5469 10. Include the RunTest script in the distribution.
5470
5471 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5472 discrepancies, some of which were old and were also with respect to 5.004. They
5473 have now been fixed.
5474
5475
5476 Version 1.09 28-Apr-98
5477 ----------------------
5478
5479 1. A negated single character class followed by a quantifier with a minimum
5480 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5481 program crashes, or just wrong answers. This did not apply to negated classes
5482 containing more than one character, or to minima other than one.
5483
5484
5485 Version 1.08 27-Mar-98
5486 ----------------------
5487
5488 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5489
5490 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5491 latter must appear before anything that relies on it in the pattern.
5492
5493
5494 Version 1.07 16-Feb-98
5495 ----------------------
5496
5497 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5498 repeat of a potentially empty string).
5499
5500
5501 Version 1.06 23-Jan-98
5502 ----------------------
5503
5504 1. Added Markus Oberhumer's little patches for C++.
5505
5506 2. Literal strings longer than 255 characters were broken.
5507
5508
5509 Version 1.05 23-Dec-97
5510 ----------------------
5511
5512 1. Negated character classes containing more than one character were failing if
5513 PCRE_CASELESS was set at run time.
5514
5515
5516 Version 1.04 19-Dec-97
5517 ----------------------
5518
5519 1. Corrected the man page, where some "const" qualifiers had been omitted.
5520
5521 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5522 input syntax.
5523
5524 3. Fixed memory leak which occurred when a regex with back references was
5525 matched with an offsets vector that wasn't big enough. The temporary memory
5526 that is used in this case wasn't being freed if the match failed.
5527
5528 4. Tidied pcretest to ensure it frees memory that it gets.
5529
5530 5. Temporary memory was being obtained in the case where the passed offsets
5531 vector was exactly big enough.
5532
5533 6. Corrected definition of offsetof() from change 5 below.
5534
5535 7. I had screwed up change 6 below and broken the rules for the use of
5536 setjmp(). Now fixed.
5537
5538
5539 Version 1.03 18-Dec-97
5540 ----------------------
5541
5542 1. A erroneous regex with a missing opening parenthesis was correctly
5543 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5544 on some systems.
5545
5546 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5547 it was reported that one broken compiler failed on the former because "code" is
5548 also an independent variable.
5549
5550 3. The erroneous regex a[]b caused an array overrun reference.
5551
5552 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5553 fail on data ending with that character. (It was going on too far, and checking
5554 the next character, typically a binary zero.) This was specific to the
5555 optimized code for single-character negative classes.
5556
5557 5. Added a contributed patch from the TIN world which does the following:
5558
5559 + Add an undef for memmove, in case the the system defines a macro for it.
5560
5561 + Add a definition of offsetof(), in case there isn't one. (I don't know
5562 the reason behind this - offsetof() is part of the ANSI standard - but
5563 it does no harm).
5564
5565 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5566 most of the places where whitespace preceded '#'. I have given up and
5567 allowed the remaining 2 cases to be at the margin.
5568
5569 + Rename some variables in pcre to eliminate shadowing. This seems very
5570 pedantic, but does no harm, of course.
5571
5572 6. Moved the call to setjmp() into its own function, to get rid of warnings
5573 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5574
5575 7. Constructs such as \d{8,} were compiling into the equivalent of
5576 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5577 outcome, but in this particular case used more store than had been allocated,
5578 which caused the bug to be discovered because it threw up an internal error.
5579
5580 8. The debugging code in both pcre and pcretest for outputting the compiled
5581 form of a regex was going wrong in the case of back references followed by
5582 curly-bracketed repeats.
5583
5584
5585 Version 1.02 12-Dec-97
5586 ----------------------
5587
5588 1. Typos in pcre.3 and comments in the source fixed.
5589
5590 2. Applied a contributed patch to get rid of places where it used to remove
5591 'const' from variables, and fixed some signed/unsigned and uninitialized
5592 variable warnings.
5593
5594 3. Added the "runtest" target to Makefile.
5595
5596 4. Set default compiler flag to -O2 rather than just -O.
5597
5598
5599 Version 1.01 19-Nov-97
5600 ----------------------
5601
5602 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5603 like /([ab]*)*/, that is, for classes with more than one character in them.
5604
5605 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5606 as /((?>a*))*/ (a PCRE_EXTRA facility).
5607
5608
5609 Version 1.00 18-Nov-97
5610 ----------------------
5611
5612 1. Added compile-time macros to support systems such as SunOS4 which don't have
5613 memmove() or strerror() but have other things that can be used instead.
5614
5615 2. Arranged that "make clean" removes the executables.
5616
5617
5618 Version 0.99 27-Oct-97
5619 ----------------------
5620
5621 1. Fixed bug in code for optimizing classes with only one character. It was
5622 initializing a 32-byte map regardless, which could cause it to run off the end
5623 of the memory it had got.
5624
5625 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5626
5627
5628 Version 0.98 22-Oct-97
5629 ----------------------
5630
5631 1. Fixed bug in code for handling temporary memory usage when there are more
5632 back references than supplied space in the ovector. This could cause segfaults.
5633
5634
5635 Version 0.97 21-Oct-97
5636 ----------------------
5637
5638 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5639
5640 2. Optimized negated single characters not to use a bit map.
5641
5642 3. Brought error texts together as macro definitions; clarified some of them;
5643 fixed one that was wrong - it said "range out of order" when it meant "invalid
5644 escape sequence".
5645
5646 4. Changed some char * arguments to const char *.
5647
5648 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5649
5650 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5651 pcretest.
5652
5653
5654 Version 0.96 16-Oct-97
5655 ----------------------
5656
5657 1. Added a simple "pgrep" utility to the distribution.
5658
5659 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5660 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5661 where "ddd" means "one or more decimal digits".
5662
5663 3. Fixed serious bug. If a pattern had a back reference, but the call to
5664 pcre_exec() didn't supply a large enough ovector to record the related
5665 identifying subpattern, the match always failed. PCRE now remembers the number
5666 of the largest back reference, and gets some temporary memory in which to save
5667 the offsets during matching if necessary, in order to ensure that
5668 backreferences always work.
5669
5670 4. Increased the compatibility with Perl in a number of ways:
5671
5672 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5673 to request this handling. The option can be set at compile or exec time.
5674
5675 (b) $ matches before a terminating newline by default; an option
5676 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5677 mode). The option can be set at compile or exec time.
5678
5679 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5680 the same as Perl's. If the decimal number it represents is less than 10
5681 or there aren't that many previous left capturing parentheses, an octal
5682 escape is read. Inside a character class, it's always an octal escape,
5683 even if it is a single digit.
5684
5685 (d) An escaped but undefined alphabetic character is taken as a literal,
5686 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5687 escapes.
5688
5689 (e) {0} is now permitted. (The previous item is removed from the compiled
5690 pattern).
5691
5692 5. Changed all the names of code files so that the basic parts are no longer
5693 than 10 characters, and abolished the teeny "globals.c" file.
5694
5695 6. Changed the handling of character classes; they are now done with a 32-byte
5696 bit map always.
5697
5698 7. Added the -d and /D options to pcretest to make it possible to look at the
5699 internals of compilation without having to recompile pcre.
5700
5701
5702 Version 0.95 23-Sep-97
5703 ----------------------
5704
5705 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5706 \x20 at the start of a run of normal characters. These were being treated as
5707 real characters, instead of the source characters being re-checked.
5708
5709
5710 Version 0.94 18-Sep-97
5711 ----------------------
5712
5713 1. The functions are now thread-safe, with the caveat that the global variables
5714 containing pointers to malloc() and free() or alternative functions are the
5715 same for all threads.
5716
5717 2. Get pcre_study() to generate a bitmap of initial characters for non-
5718 anchored patterns when this is possible, and use it if passed to pcre_exec().
5719
5720
5721 Version 0.93 15-Sep-97
5722 ----------------------
5723
5724 1. /(b)|(:+)/ was computing an incorrect first character.
5725
5726 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5727 but not actually doing anything yet.
5728
5729 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5730 as Perl does (e.g. [-az] or [az-]).
5731
5732 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5733 all possible positions.
5734
5735 5. Split up into different modules to avoid including unneeded functions in a
5736 compiled binary. However, compile and exec are still in one module. The "study"
5737 function is split off.
5738
5739 6. The character tables are now in a separate module whose source is generated
5740 by an auxiliary program - but can then be edited by hand if required. There are
5741 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5742 toupper() in the code.
5743
5744 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5745 make them global. Abolish the function for setting them, as the caller can now
5746 set them directly.
5747
5748
5749 Version 0.92 11-Sep-97
5750 ----------------------
5751
5752 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5753 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5754
5755 2. Caseless matching was not working in character classes if the characters in
5756 the pattern were in upper case.
5757
5758 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5759
5760 4. Make PCRE_ANCHORED public and accept as a compile option.
5761
5762 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5763 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5764 pass them.
5765
5766 6. Give an error if bad option bits passed at compile or run time.
5767
5768 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5769 pcretest to cause it to pass that flag.
5770
5771 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5772 options, and the first character, if set.
5773
5774 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5775
5776
5777 Version 0.91 10-Sep-97
5778 ----------------------
5779
5780 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5781 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5782
5783 2. PCRE was looping on encountering an indefinitely repeated back reference to
5784 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5785 Perl does - treats the match as successful.
5786
5787 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5