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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1568 - (show annotations)
Sun Jun 14 15:53:41 2015 UTC (4 years, 5 months ago) by ph10
File size: 280225 byte(s)
Make \c in EBCDIC environments compatible with Perl.
1 ChangeLog for PCRE
2 ------------------
3
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
6
7 Version 8.38 xx-xxx-xxxx
8 ------------------------
9
10 1. If a group that contained a recursive back reference also contained a
11 forward reference subroutine call followed by a non-forward-reference
12 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13 compile correct code, leading to undefined behaviour or an internally
14 detected error. This bug was discovered by the LLVM fuzzer.
15
16 2. Quantification of certain items (e.g. atomic back references) could cause
17 incorrect code to be compiled when recursive forward references were
18 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19 This bug was discovered by the LLVM fuzzer.
20
21 3. A repeated conditional group whose condition was a reference by name caused
22 a buffer overflow if there was more than one group with the given name.
23 This bug was discovered by the LLVM fuzzer.
24
25 4. A recursive back reference by name within a group that had the same name as
26 another group caused a buffer overflow. For example:
27 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28
29 5. A forward reference by name to a group whose number is the same as the
30 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31 a buffer overflow at compile time. This bug was discovered by the LLVM
32 fuzzer.
33
34 6. A lookbehind assertion within a set of mutually recursive subpatterns could
35 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36
37 7. Another buffer overflow bug involved duplicate named groups with a
38 reference between their definition, with a group that reset capture
39 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40 fixed by always allowing for more memory, even if not needed. (A proper fix
41 is implemented in PCRE2, but it involves more refactoring.)
42
43 8. There was no check for integer overflow in subroutine calls such as (?123).
44
45 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
46 being treated as a literal 'l' instead of causing an error.
47
48 10. There was a buffer overflow if pcre_exec() was called with an ovector of
49 size 1. This bug was found by american fuzzy lop.
50
51 11. If a non-capturing group containing a conditional group that could match
52 an empty string was repeated, it was not identified as matching an empty
53 string itself. For example: /^(?:(?(1)x|)+)+$()/.
54
55 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
56 \a and \e in test subject lines.
57
58 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
59 instead of the EBCDIC value.
60
61 14. The handling of \c in an EBCDIC environment has been revised so that it is
62 now compatible with the specification in Perl's perlebcdic page.
63
64
65 Version 8.37 28-April-2015
66 --------------------------
67
68 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
69 for those parentheses to be closed with whatever has been captured so far.
70 However, it was failing to mark any other groups between the hightest
71 capture so far and the currrent group as "unset". Thus, the ovector for
72 those groups contained whatever was previously there. An example is the
73 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
74
75 2. If an assertion condition was quantified with a minimum of zero (an odd
76 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
77
78 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
79 unrecognized modifier, a crash could occur.
80
81 4. An attempt to do global matching in pcretest with a zero-length ovector
82 caused a crash.
83
84 5. Fixed a memory leak during matching that could occur for a subpattern
85 subroutine call (recursive or otherwise) if the number of captured groups
86 that had to be saved was greater than ten.
87
88 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
89 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
90 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
91
92 7. A UTF pattern containing a "not" match of a non-ASCII character and a
93 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
94
95 8. When a pattern is compiled, it remembers the highest back reference so that
96 when matching, if the ovector is too small, extra memory can be obtained to
97 use instead. A conditional subpattern whose condition is a check on a
98 capture having happened, such as, for example in the pattern
99 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
100 setting the highest backreference number. This mattered only if pcre_exec()
101 was called with an ovector that was too small to hold the capture, and there
102 was no other kind of back reference (a situation which is probably quite
103 rare). The effect of the bug was that the condition was always treated as
104 FALSE when the capture could not be consulted, leading to a incorrect
105 behaviour by pcre_exec(). This bug has been fixed.
106
107 9. A reference to a duplicated named group (either a back reference or a test
108 for being set in a conditional) that occurred in a part of the pattern where
109 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
110 to be incorrectly calculated, leading to overwriting.
111
112 10. A mutually recursive set of back references such as (\2)(\1) caused a
113 segfault at study time (while trying to find the minimum matching length).
114 The infinite loop is now broken (with the minimum length unset, that is,
115 zero).
116
117 11. If an assertion that was used as a condition was quantified with a minimum
118 of zero, matching went wrong. In particular, if the whole group had
119 unlimited repetition and could match an empty string, a segfault was
120 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
121 assertions to be quantified, but not if they are being used as conditions,
122 so the above pattern is faulted by Perl. PCRE has now been changed so that
123 it also rejects such patterns.
124
125 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
126 failed to allow the zero-repeat case if pcre2_exec() was called with an
127 ovector too small to capture the group.
128
129 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
130 Red Hat Product Security:
131
132 (a) A crash if /K and /F were both set with the option to save the compiled
133 pattern.
134
135 (b) Another crash if the option to print captured substrings in a callout
136 was combined with setting a null ovector, for example \O\C+ as a subject
137 string.
138
139 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
140 forward reference repeated a large (but limited) number of times within a
141 repeated outer group that has a zero minimum quantifier, caused incorrect
142 code to be compiled, leading to the error "internal error:
143 previously-checked referenced subpattern not found" when an incorrect
144 memory address was read. This bug was reported as "heap overflow",
145 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
146 CVE-2015-2325.
147
148 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
149 call within a group that also contained a recursive back reference caused
150 incorrect code to be compiled. This bug was reported as "heap overflow",
151 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
152 number CVE-2015-2326.
153
154 24. Computing the size of the JIT read-only data in advance has been a source
155 of various issues, and new ones are still appear unfortunately. To fix
156 existing and future issues, size computation is eliminated from the code,
157 and replaced by on-demand memory allocation.
158
159 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
160 adjacent to the end of the range, and the range contained characters with
161 more than one other case, caused incorrect behaviour when compiled in UTF
162 mode. In that example, the range a-j was left out of the class.
163
164 26. Fix JIT compilation of conditional blocks, which assertion
165 is converted to (*FAIL). E.g: /(?(?!))/.
166
167 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
168 discovered by the LLVM fuzzer.
169
170 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
171 when this assertion was used as a condition, for example (?(?!)a|b). In
172 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
173 error about an unsupported item.
174
175 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
176 possessification code could take exponential time to complete. A recursion
177 depth limit of 1000 has been imposed to limit the resources used by this
178 optimization.
179
180 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
181 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
182 because \S ensures they are all in the class. The code for doing this was
183 interacting badly with the code for computing the amount of space needed to
184 compile the pattern, leading to a buffer overflow. This bug was discovered
185 by the LLVM fuzzer.
186
187 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
188 other kinds of group caused stack overflow at compile time. This bug was
189 discovered by the LLVM fuzzer.
190
191 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
192 between a subroutine call and its quantifier was incorrectly compiled,
193 leading to buffer overflow or other errors. This bug was discovered by the
194 LLVM fuzzer.
195
196 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
197 assertion after (?(. The code was failing to check the character after
198 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
199 was discovered by the LLVM fuzzer.
200
201 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
202 a fixed maximum following a group that contains a subroutine reference was
203 incorrectly compiled and could trigger buffer overflow. This bug was
204 discovered by the LLVM fuzzer.
205
206 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
207 caused a stack overflow instead of the diagnosis of a non-fixed length
208 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
209
210 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
211 (e.g. /(?<=\Ka)/) could make pcregrep loop.
212
213 37. There was a similar problem to 36 in pcretest for global matches.
214
215 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
216 and a subsequent item in the pattern caused a non-match, backtracking over
217 the repeated \X did not stop, but carried on past the start of the subject,
218 causing reference to random memory and/or a segfault. There were also some
219 other cases where backtracking after \C could crash. This set of bugs was
220 discovered by the LLVM fuzzer.
221
222 39. The function for finding the minimum length of a matching string could take
223 a very long time if mutual recursion was present many times in a pattern,
224 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
225 method has been implemented. This infelicity was discovered by the LLVM
226 fuzzer.
227
228 40. Static linking against the PCRE library using the pkg-config module was
229 failing on missing pthread symbols.
230
231
232 Version 8.36 26-September-2014
233 ------------------------------
234
235 1. Got rid of some compiler warnings in the C++ modules that were shown up by
236 -Wmissing-field-initializers and -Wunused-parameter.
237
238 2. The tests for quantifiers being too big (greater than 65535) were being
239 applied after reading the number, and stupidly assuming that integer
240 overflow would give a negative number. The tests are now applied as the
241 numbers are read.
242
243 3. Tidy code in pcre_exec.c where two branches that used to be different are
244 now the same.
245
246 4. The JIT compiler did not generate match limit checks for certain
247 bracketed expressions with quantifiers. This may lead to exponential
248 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
249 issue should be resolved now.
250
251 5. Fixed an issue, which occures when nested alternatives are optimized
252 with table jumps.
253
254 6. Inserted two casts and changed some ints to size_t in the light of some
255 reported 64-bit compiler warnings (Bugzilla 1477).
256
257 7. Fixed a bug concerned with zero-minimum possessive groups that could match
258 an empty string, which sometimes were behaving incorrectly in the
259 interpreter (though correctly in the JIT matcher). This pcretest input is
260 an example:
261
262 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
263 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
264
265 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
266 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
267 for an empty string was breaking the inner loop and carrying on at a lower
268 level, when possessive repeated groups should always return to a higher
269 level as they have no backtrack points in them. The empty string test now
270 occurs at the outer level.
271
272 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
273 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
274
275 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
276 doesn't).
277
278 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
279 to hold a pattern to be incorrectly computed (too small) when there were
280 named back references to duplicated names. This could cause "internal
281 error: code overflow" or "double free or corruption" or other memory
282 handling errors.
283
284 11. When named subpatterns had the same prefixes, back references could be
285 confused. For example, in this pattern:
286
287 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
288
289 the reference to 'Name' was incorrectly treated as a reference to a
290 duplicate name.
291
292 12. A pattern such as /^s?c/mi8 where the optional character has more than
293 one "other case" was incorrectly compiled such that it would only try to
294 match starting at "c".
295
296 13. When a pattern starting with \s was studied, VT was not included in the
297 list of possible starting characters; this should have been part of the
298 8.34/18 patch.
299
300 14. If a character class started [\Qx]... where x is any character, the class
301 was incorrectly terminated at the ].
302
303 15. If a pattern that started with a caseless match for a character with more
304 than one "other case" was studied, PCRE did not set up the starting code
305 unit bit map for the list of possible characters. Now it does. This is an
306 optimization improvement, not a bug fix.
307
308 16. The Unicode data tables have been updated to Unicode 7.0.0.
309
310 17. Fixed a number of memory leaks in pcregrep.
311
312 18. Avoid a compiler warning (from some compilers) for a function call with
313 a cast that removes "const" from an lvalue by using an intermediate
314 variable (to which the compiler does not object).
315
316 19. Incorrect code was compiled if a group that contained an internal recursive
317 back reference was optional (had quantifier with a minimum of zero). This
318 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
319 caused segmentation faults because of stack overflows at compile time.
320
321 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
322 group that is quantified with an indefinite repeat, caused a compile-time
323 loop which used up all the system stack and provoked a segmentation fault.
324 This was not the same bug as 19 above.
325
326 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
327 Patch by Mike Frysinger.
328
329
330 Version 8.35 04-April-2014
331 --------------------------
332
333 1. A new flag is set, when property checks are present in an XCLASS.
334 When this flag is not set, PCRE can perform certain optimizations
335 such as studying these XCLASS-es.
336
337 2. The auto-possessification of character sets were improved: a normal
338 and an extended character set can be compared now. Furthermore
339 the JIT compiler optimizes more character set checks.
340
341 3. Got rid of some compiler warnings for potentially uninitialized variables
342 that show up only when compiled with -O2.
343
344 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
345 of a match later then the end of the match. The pcretest program was not
346 handling the case sensibly - it was outputting from the start to the next
347 binary zero. It now reports this situation in a message, and outputs the
348 text from the end to the start.
349
350 5. Fast forward search is improved in JIT. Instead of the first three
351 characters, any three characters with fixed position can be searched.
352 Search order: first, last, middle.
353
354 6. Improve character range checks in JIT. Characters are read by an inprecise
355 function now, which returns with an unknown value if the character code is
356 above a certain threshold (e.g: 256). The only limitation is that the value
357 must be bigger than the threshold as well. This function is useful when
358 the characters above the threshold are handled in the same way.
359
360 7. The macros whose names start with RAWUCHAR are placeholders for a future
361 mode in which only the bottom 21 bits of 32-bit data items are used. To
362 make this more memorable for those maintaining the code, the names have
363 been changed to start with UCHAR21, and an extensive comment has been added
364 to their definition.
365
366 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
367 to the export list in Makefile.am (they were accidentally omitted from the
368 8.34 tarball).
369
370 9. The informational output from pcretest used the phrase "starting byte set"
371 which is inappropriate for the 16-bit and 32-bit libraries. As the output
372 for "first char" and "need char" really means "non-UTF-char", I've changed
373 "byte" to "char", and slightly reworded the output. The documentation about
374 these values has also been (I hope) clarified.
375
376 10. Another JIT related optimization: use table jumps for selecting the correct
377 backtracking path, when more than four alternatives are present inside a
378 bracket.
379
380 11. Empty match is not possible, when the minimum length is greater than zero,
381 and there is no \K in the pattern. JIT should avoid empty match checks in
382 such cases.
383
384 12. In a caseless character class with UCP support, when a character with more
385 than one alternative case was not the first character of a range, not all
386 the alternative cases were added to the class. For example, s and \x{17f}
387 are both alternative cases for S: the class [RST] was handled correctly,
388 but [R-T] was not.
389
390 13. The configure.ac file always checked for pthread support when JIT was
391 enabled. This is not used in Windows, so I have put this test inside a
392 check for the presence of windows.h (which was already tested for).
393
394 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
395 The algorithm provides a way to skip certain starting offsets, and usually
396 faster than linear prefix searches.
397
398 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
399 as for 'fr_FR' and 'french'. For some reason, however, it then used the
400 Windows-specific input and output files, which have 'french' screwed in.
401 So this could never have worked. One of the problems with locales is that
402 they aren't always the same. I have now updated RunTest so that it checks
403 the output of the locale test (test 3) against three different output
404 files, and it allows the test to pass if any one of them matches. With luck
405 this should make the test pass on some versions of Solaris where it was
406 failing. Because of the uncertainty, the script did not used to stop if
407 test 3 failed; it now does. If further versions of a French locale ever
408 come to light, they can now easily be added.
409
410 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
411 there was a message during ./configure, but it did not stop. This now
412 provokes an error. The invalid example in README has been corrected.
413 If a value less than the minimum is given, the minimum value has always
414 been used, but now a warning is given.
415
416 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
417 was a bug in the test system, which is now fixed. Also, the list of various
418 configurations that are tested for each release did not have one with both
419 16/32 bits and --enable-bar-anycrlf. It now does.
420
421 18. pcretest was missing "-C bsr" for displaying the \R default setting.
422
423 19. Little endian PowerPC systems are supported now by the JIT compiler.
424
425 20. The fast forward newline mechanism could enter to an infinite loop on
426 certain invalid UTF-8 input. Although we don't support these cases
427 this issue can be fixed by a performance optimization.
428
429 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
430 not take account if existing stack usage. There is now a new global
431 variable called pcre_stack_guard that can be set to point to an external
432 function to check stack availability. It is called at the start of
433 processing every parenthesized group.
434
435 22. A typo in the code meant that in ungreedy mode the max/min qualifier
436 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
437 not match "ab".
438
439 23. When UTF was disabled, the JIT program reported some incorrect compile
440 errors. These messages are silenced now.
441
442 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
443 compiler.
444
445 25. Change all the temporary files used in RunGrepTest to be different to those
446 used by RunTest so that the tests can be run simultaneously, for example by
447 "make -j check".
448
449
450 Version 8.34 15-December-2013
451 -----------------------------
452
453 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
454 executable memory. Patch inspired by Carsten Klein.
455
456 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
457 this macro is never tested and has no effect, because the work to support
458 coverage involves only compiling and linking options and special targets in
459 the Makefile. The comment in config.h implied that defining the macro would
460 enable coverage support, which is totally false. There was also support for
461 setting this macro in the CMake files (my fault, I just copied it from
462 configure). SUPPORT_GCOV has now been removed.
463
464 3. Make a small performance improvement in strlen16() and strlen32() in
465 pcretest.
466
467 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
468 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
469 these cases). There was also one in pcretest.c.
470
471 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
472
473 6. In UTF mode, the code for checking whether a group could match an empty
474 string (which is used for indefinitely repeated groups to allow for
475 breaking an infinite loop) was broken when the group contained a repeated
476 negated single-character class with a character that occupied more than one
477 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
478 UTF-8 mode). The effect was undefined: the group might or might not be
479 deemed as matching an empty string, or the program might have crashed.
480
481 7. The code for checking whether a group could match an empty string was not
482 recognizing that \h, \H, \v, \V, and \R must match a character.
483
484 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
485 an empty string. If it can, pcretest shows this in its information output.
486
487 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
488 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
489 matched by pcre_exec() without using JIT:
490
491 (a) If the rest of the pattern did not match after a maximal run of
492 grapheme clusters, the code for backing up to try with fewer of them
493 did not always back up over a full grapheme when characters that do not
494 have the modifier quality were involved, e.g. Hangul syllables.
495
496 (b) If the match point in a subject started with modifier character, and
497 there was no match, the code could incorrectly back up beyond the match
498 point, and potentially beyond the first character in the subject,
499 leading to a segfault or an incorrect match result.
500
501 10. A conditional group with an assertion condition could lead to PCRE
502 recording an incorrect first data item for a match if no other first data
503 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
504 first data item, and therefore matched "ca" after "c" instead of at the
505 start.
506
507 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
508 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
509
510 12. The source of pcregrep now includes z/OS-specific code so that it can be
511 compiled for z/OS as part of the special z/OS distribution.
512
513 13. Added the -T and -TM options to pcretest.
514
515 14. The code in pcre_compile.c for creating the table of named capturing groups
516 has been refactored. Instead of creating the table dynamically during the
517 actual compiling pass, the information is remembered during the pre-compile
518 pass (on the stack unless there are more than 20 named groups, in which
519 case malloc() is used) and the whole table is created before the actual
520 compile happens. This has simplified the code (it is now nearly 150 lines
521 shorter) and prepared the way for better handling of references to groups
522 with duplicate names.
523
524 15. A back reference to a named subpattern when there is more than one of the
525 same name now checks them in the order in which they appear in the pattern.
526 The first one that is set is used for the reference. Previously only the
527 first one was inspected. This change makes PCRE more compatible with Perl.
528
529 16. Unicode character properties were updated from Unicode 6.3.0.
530
531 17. The compile-time code for auto-possessification has been refactored, based
532 on a patch by Zoltan Herczeg. It now happens after instead of during
533 compilation. The code is cleaner, and more cases are handled. The option
534 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
535 options in pcretest are provided to set it. It can also be set by
536 (*NO_AUTO_POSSESS) at the start of a pattern.
537
538 18. The character VT has been added to the default ("C" locale) set of
539 characters that match \s and are generally treated as white space,
540 following this same change in Perl 5.18. There is now no difference between
541 "Perl space" and "POSIX space". Whether VT is treated as white space in
542 other locales depends on the locale.
543
544 19. The code for checking named groups as conditions, either for being set or
545 for being recursed, has been refactored (this is related to 14 and 15
546 above). Processing unduplicated named groups should now be as fast at
547 numerical groups, and processing duplicated groups should be faster than
548 before.
549
550 20. Two patches to the CMake build system, by Alexander Barkov:
551
552 (1) Replace the "source" command by "." in CMakeLists.txt because
553 "source" is a bash-ism.
554
555 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
556 without these the CMake build does not work on Solaris.
557
558 21. Perl has changed its handling of \8 and \9. If there is no previously
559 encountered capturing group of those numbers, they are treated as the
560 literal characters 8 and 9 instead of a binary zero followed by the
561 literals. PCRE now does the same.
562
563 22. Following Perl, added \o{} to specify codepoints in octal, making it
564 possible to specify values greater than 0777 and also making them
565 unambiguous.
566
567 23. Perl now gives an error for missing closing braces after \x{... instead of
568 treating the string as literal. PCRE now does the same.
569
570 24. RunTest used to grumble if an inappropriate test was selected explicitly,
571 but just skip it when running all tests. This make it awkward to run ranges
572 of tests when one of them was inappropriate. Now it just skips any
573 inappropriate tests, as it always did when running all tests.
574
575 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
576 character types such as \d or \w, too many callouts were inserted, and the
577 data that they returned was rubbish.
578
579 26. In UCP mode, \s was not matching two of the characters that Perl matches,
580 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
581 were matched by \h. The code has now been refactored so that the lists of
582 the horizontal and vertical whitespace characters used for \h and \v (which
583 are defined only in one place) are now also used for \s.
584
585 27. Add JIT support for the 64 bit TileGX architecture.
586 Patch by Jiong Wang (Tilera Corporation).
587
588 28. Possessive quantifiers for classes (both explicit and automatically
589 generated) now use special opcodes instead of wrapping in ONCE brackets.
590
591 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
592 (because it's meaningless), this was not happening when PCRE_CASELESS was
593 set. Not wrong, but inefficient.
594
595 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
596 properties for \w, \d, etc) is present in a test regex. Otherwise if the
597 test contains no characters greater than 255, Perl doesn't realise it
598 should be using Unicode semantics.
599
600 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
601 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
602 does in Unicode mode.
603
604 32. Added the "forbid" facility to pcretest so that putting tests into the
605 wrong test files can sometimes be quickly detected.
606
607 33. There is now a limit (default 250) on the depth of nesting of parentheses.
608 This limit is imposed to control the amount of system stack used at compile
609 time. It can be changed at build time by --with-parens-nest-limit=xxx or
610 the equivalent in CMake.
611
612 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
613 errors. Perl warns for these when in warning mode, but PCRE has no facility
614 for giving warnings.
615
616 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
617 However, this was not working for (?!) because it is optimized to (*FAIL),
618 for which PCRE does not allow quantifiers. The optimization is now disabled
619 when a quantifier follows (?!). I can't see any use for this, but it makes
620 things uniform.
621
622 36. Perl no longer allows group names to start with digits, so I have made this
623 change also in PCRE. It simplifies the code a bit.
624
625 37. In extended mode, Perl ignores spaces before a + that indicates a
626 possessive quantifier. PCRE allowed a space before the quantifier, but not
627 before the possessive +. It now does.
628
629 38. The use of \K (reset reported match start) within a repeated possessive
630 group such as (a\Kb)*+ was not working.
631
632 40. Document that the same character tables must be used at compile time and
633 run time, and that the facility to pass tables to pcre_exec() and
634 pcre_dfa_exec() is for use only with saved/restored patterns.
635
636 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
637 features for Builds with MSVC:
638
639 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
640 builds). The .rc files can be used to set FileDescription and many other
641 attributes.
642
643 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
644 This allows higher-level build scripts which want .pdb files to avoid
645 hard-coding the exact files needed."
646
647 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
648 mean "start of word" and "end of word", respectively, as a transition aid.
649
650 43. A minimizing repeat of a class containing codepoints greater than 255 in
651 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
652 compiled to use the heap for recursion.
653
654 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
655 is configured.
656
657
658 Version 8.33 28-May-2013
659 ------------------------
660
661 1. Added 'U' to some constants that are compared to unsigned integers, to
662 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
663 variables that are added to signed variables, to ensure the result is
664 signed and can be negated.
665
666 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
667 CMake config files.
668
669 3. Revise the creation of config.h.generic so that all boolean macros are
670 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
671 overriding via -D on the command line possible.
672
673 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
674 to unsigned int is reported to make a quite noticeable speed difference in
675 a specific Windows environment. Testing on Linux did also appear to show
676 some benefit (and it is clearly not harmful). Also fixed the definition of
677 Xop which should be unsigned.
678
679 5. Related to (4), changing the definition of the intermediate variable cc
680 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
681 improvements.
682
683 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
684 unnecessary spaces.
685
686 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
687 and later.
688
689 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
690
691 9. Optimizing fast_forward_start_bits in JIT.
692
693 10. Adding support for callouts in JIT, and fixing some issues revealed
694 during this work. Namely:
695
696 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
697
698 (b) Minimum length was not checked before the matching is started.
699
700 11. The value of capture_last that is passed to callouts was incorrect in some
701 cases when there was a capture on one path that was subsequently abandoned
702 after a backtrack. Also, the capture_last value is now reset after a
703 recursion, since all captures are also reset in this case.
704
705 12. The interpreter no longer returns the "too many substrings" error in the
706 case when an overflowing capture is in a branch that is subsequently
707 abandoned after a backtrack.
708
709 13. In the pathological case when an offset vector of size 2 is used, pcretest
710 now prints out the matched string after a yield of 0 or 1.
711
712 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
713 Only supported by the JIT compiler at the moment.
714
715 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
716
717 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
718 offset of the starting point of the matching process, provided the offsets
719 vector is large enough.
720
721 17. The \A escape now records a lookbehind value of 1, though its execution
722 does not actually inspect the previous character. This is to ensure that,
723 in partial multi-segment matching, at least one character from the old
724 segment is retained when a new segment is processed. Otherwise, if there
725 are no lookbehinds in the pattern, \A might match incorrectly at the start
726 of a new segment.
727
728 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
729
730 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
731 gives some modest performance improvement in 8-bit mode.
732
733 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
734 be expressed in certain programming languages using Universal Character
735 Names.
736
737 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
738 which points out that "non characters" are not "characters that may not
739 appear in Unicode strings" but rather "characters that are reserved for
740 internal use and have only local meaning".
741
742 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
743 there was a conditional group that depended on an assertion, if the
744 assertion was false, the callout that immediately followed the alternation
745 in the condition was skipped when pcre_exec() was used for matching.
746
747 23. Allow an explicit callout to be inserted before an assertion that is the
748 condition for a conditional group, for compatibility with automatic
749 callouts, which always insert a callout at this point.
750
751 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
752 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
753
754 25. (*PRUNE) is now supported by the JIT compiler.
755
756 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
757
758 27. Fix the case where there are two or more SKIPs with arguments that may be
759 ignored.
760
761 28. (*SKIP) is now supported by the JIT compiler.
762
763 29. (*THEN) is now supported by the JIT compiler.
764
765 30. Update RunTest with additional test selector options.
766
767 31. The way PCRE handles backtracking verbs has been changed in two ways.
768
769 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
770 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
771 backtracking. In some cases this makes it more Perl-compatible, but Perl's
772 rather obscure rules do not always do the same thing.
773
774 (2) Previously, backtracking verbs were confined within assertions. This is
775 no longer the case for positive assertions, except for (*ACCEPT). Again,
776 this sometimes improves Perl compatibility, and sometimes does not.
777
778 32. A number of tests that were in test 2 because Perl did things differently
779 have been moved to test 1, because either Perl or PCRE has changed, and
780 these tests are now compatible.
781
782 32. Backtracking control verbs are now handled in the same way in JIT and
783 interpreter.
784
785 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
786 contained a forward subroutine reference caused a compile error.
787
788 34. Auto-detect and optimize limited repetitions in JIT.
789
790 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
791 blocking (*UTF) etc.
792
793 36. In the interpreter, maximizing pattern repetitions for characters and
794 character types now use tail recursion, which reduces stack usage.
795
796 37. The value of the max lookbehind was not correctly preserved if a compiled
797 and saved regex was reloaded on a host of different endianness.
798
799 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
800 of the compiled pattern block, expand the flags field from 16 to 32 bits
801 because it was almost full.
802
803 39. Try madvise first before posix_madvise.
804
805 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
806 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
807 original change locked out all matches of empty strings. This has been
808 changed so that one match of an empty string per line is recognized.
809 Subsequent searches on the same line (for colouring or for --only-matching,
810 for example) do not recognize empty strings.
811
812 41. Applied a user patch to fix a number of spelling mistakes in comments.
813
814 42. Data lines longer than 65536 caused pcretest to crash.
815
816 43. Clarified the data type for length and startoffset arguments for pcre_exec
817 and pcre_dfa_exec in the function-specific man pages, where they were
818 explicitly stated to be in bytes, never having been updated. I also added
819 some clarification to the pcreapi man page.
820
821 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
822 a segmentation fault.
823
824
825 Version 8.32 30-November-2012
826 -----------------------------
827
828 1. Improved JIT compiler optimizations for first character search and single
829 character iterators.
830
831 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
832 Patch by Daniel Richard G.
833
834 3. Single character iterator optimizations in the JIT compiler.
835
836 4. Improved JIT compiler optimizations for character ranges.
837
838 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
839 Reported by Giuseppe D'Angelo.
840
841 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
842 of a line, was being set incorrectly in cases where .* appeared inside
843 atomic brackets at the start of a pattern, or where there was a subsequent
844 *PRUNE or *SKIP.
845
846 7. Improved instruction cache flush for POWER/PowerPC.
847 Patch by Daniel Richard G.
848
849 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
850 grep:
851
852 (a) There is now no limit to the number of patterns to be matched.
853
854 (b) An error is given if a pattern is too long.
855
856 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
857 are now supported.
858
859 (d) --exclude-from and --include-from (multiple use) have been added.
860
861 (e) Exclusions and inclusions now apply to all files and directories, not
862 just to those obtained from scanning a directory recursively.
863
864 (f) Multiple uses of -f and --file-list are now supported.
865
866 (g) In a Windows environment, the default for -d has been changed from
867 "read" (the GNU grep default) to "skip", because otherwise the presence
868 of a directory in the file list provokes an error.
869
870 (h) The documentation has been revised and clarified in places.
871
872 9. Improve the matching speed of capturing brackets.
873
874 10. Changed the meaning of \X so that it now matches a Unicode extended
875 grapheme cluster.
876
877 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
878 out POSIX threads when JIT support is configured.
879
880 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
881
882 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
883 structure could go wrong in environments where size_t is not the same size
884 as int.
885
886 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
887 set.
888
889 15. The EBCDIC support had decayed; later updates to the code had included
890 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
891 general tidy up of EBCDIC-related issues, and the documentation was also
892 not quite right. There is now a test that can be run on ASCII systems to
893 check some of the EBCDIC-related things (but is it not a full test).
894
895 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
896 in a small tidy to the code.
897
898 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
899
900 18. If the --only-matching (-o) option in pcregrep is specified multiple
901 times, each one causes appropriate output. For example, -o1 -o2 outputs the
902 substrings matched by the 1st and 2nd capturing parentheses. A separating
903 string can be specified by --om-separator (default empty).
904
905 19. Improving the first n character searches.
906
907 20. Turn case lists for horizontal and vertical white space into macros so that
908 they are defined only once.
909
910 21. This set of changes together give more compatible Unicode case-folding
911 behaviour for characters that have more than one other case when UCP
912 support is available.
913
914 (a) The Unicode property table now has offsets into a new table of sets of
915 three or more characters that are case-equivalent. The MultiStage2.py
916 script that generates these tables (the pcre_ucd.c file) now scans
917 CaseFolding.txt instead of UnicodeData.txt for character case
918 information.
919
920 (b) The code for adding characters or ranges of characters to a character
921 class has been abstracted into a generalized function that also handles
922 case-independence. In UTF-mode with UCP support, this uses the new data
923 to handle characters with more than one other case.
924
925 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
926 whose other case is greater than 256 are now correctly matched
927 caselessly. Previously, the high codepoint matched the low one, but not
928 vice versa.
929
930 (d) The processing of \h, \H, \v, and \ in character classes now makes use
931 of the new class addition function, using character lists defined as
932 macros alongside the case definitions of 20 above.
933
934 (e) Caseless back references now work with characters that have more than
935 one other case.
936
937 (f) General caseless matching of characters with more than one other case
938 is supported.
939
940 22. Unicode character properties were updated from Unicode 6.2.0
941
942 23. Improved CMake support under Windows. Patch by Daniel Richard G.
943
944 24. Add support for 32-bit character strings, and UTF-32
945
946 25. Major JIT compiler update (code refactoring and bugfixing).
947 Experimental Sparc 32 support is added.
948
949 26. Applied a modified version of Daniel Richard G's patch to create
950 pcre.h.generic and config.h.generic by "make" instead of in the
951 PrepareRelease script.
952
953 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
954 pcre_compile.c when checking for a zero character.
955
956 28. Introducing a native interface for JIT. Through this interface, the compiled
957 machine code can be directly executed. The purpose of this interface is to
958 provide fast pattern matching, so several sanity checks are not performed.
959 However, feature tests are still performed. The new interface provides
960 1.4x speedup compared to the old one.
961
962 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
963 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
964 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
965
966 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
967 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
968 the "old" RFC 2279). Instead, it ended up passing a negative length to
969 pcre_exec().
970
971 31. Add support for GCC's visibility feature to hide internal functions.
972
973 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
974 "unknown -C option" after outputting 0 or 1.
975
976 33. There is now support for generating a code coverage report for the test
977 suite in environments where gcc is the compiler and lcov is installed. This
978 is mainly for the benefit of the developers.
979
980 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
981 unaddressable using valgrind annotations, allowing valgrind to detect
982 invalid memory accesses. This is mainly for the benefit of the developers.
983
984 25. (*UTF) can now be used to start a pattern in any of the three libraries.
985
986 26. Give configure error if --enable-cpp but no C++ compiler found.
987
988
989 Version 8.31 06-July-2012
990 -------------------------
991
992 1. Fixing a wrong JIT test case and some compiler warnings.
993
994 2. Removed a bashism from the RunTest script.
995
996 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
997 to unsigned type, result still unsigned" that was given by an MS compiler
998 on encountering the code "-sizeof(xxx)".
999
1000 4. Partial matching support is added to the JIT compiler.
1001
1002 5. Fixed several bugs concerned with partial matching of items that consist
1003 of more than one character:
1004
1005 (a) /^(..)\1/ did not partially match "aba" because checking references was
1006 done on an "all or nothing" basis. This also applied to repeated
1007 references.
1008
1009 (b) \R did not give a hard partial match if \r was found at the end of the
1010 subject.
1011
1012 (c) \X did not give a hard partial match after matching one or more
1013 characters at the end of the subject.
1014
1015 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1016 a partial match for the string "\r".
1017
1018 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1019 a partial match for a CR character at the end of the subject string.
1020
1021 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1022 running pcretest, the text "(JIT)" added to the output whenever JIT is
1023 actually used to run the match.
1024
1025 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1026 or /S+[+] with a digit between 1 and 7.
1027
1028 8. OP_NOT now supports any UTF character not just single-byte ones.
1029
1030 9. (*MARK) control verb is now supported by the JIT compiler.
1031
1032 10. The command "./RunTest list" lists the available tests without actually
1033 running any of them. (Because I keep forgetting what they all are.)
1034
1035 11. Add PCRE_INFO_MAXLOOKBEHIND.
1036
1037 12. Applied a (slightly modified) user-supplied patch that improves performance
1038 when the heap is used for recursion (compiled with --disable-stack-for-
1039 recursion). Instead of malloc and free for each heap frame each time a
1040 logical recursion happens, frames are retained on a chain and re-used where
1041 possible. This sometimes gives as much as 30% improvement.
1042
1043 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1044 call.
1045
1046 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1047
1048 15. It is now possible to link pcretest with libedit as an alternative to
1049 libreadline.
1050
1051 16. (*COMMIT) control verb is now supported by the JIT compiler.
1052
1053 17. The Unicode data tables have been updated to Unicode 6.1.0.
1054
1055 18. Added --file-list option to pcregrep.
1056
1057 19. Added binary file support to pcregrep, including the -a, --binary-files,
1058 -I, and --text options.
1059
1060 20. The madvise function is renamed for posix_madvise for QNX compatibility
1061 reasons. Fixed by Giuseppe D'Angelo.
1062
1063 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1064
1065 22. Changed the option for creating symbolic links for 16-bit man pages from
1066 -s to -sf so that re-installing does not cause issues.
1067
1068 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1069
1070 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1071 matches in certain environments (the workspace was not being correctly
1072 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1073 some of the workspace data at the beginning of a restart.
1074
1075 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1076 was not doing so when it should - probably a typo introduced by SVN 528
1077 (change 8.10/14).
1078
1079 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1080 \w+ when the character tables indicated that \x{c4} was a word character.
1081 There were several related cases, all because the tests for doing a table
1082 lookup were testing for characters less than 127 instead of 255.
1083
1084 27. If a pattern contains capturing parentheses that are not used in a match,
1085 their slots in the ovector are set to -1. For those that are higher than
1086 any matched groups, this happens at the end of processing. In the case when
1087 there were back references that the ovector was too small to contain
1088 (causing temporary malloc'd memory to be used during matching), and the
1089 highest capturing number was not used, memory off the end of the ovector
1090 was incorrectly being set to -1. (It was using the size of the temporary
1091 memory instead of the true size.)
1092
1093 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1094 ovector size, it uses memory at the end of the block that it has got.
1095
1096 29. Check for an overlong MARK name and give an error at compile time. The
1097 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1098
1099 30. JIT compiler update.
1100
1101 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1102 Rill for the patch.
1103
1104 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1105
1106 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1107
1108 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1109 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1110 was enabled.
1111
1112 35. Improve JIT code generation for greedy plus quantifier.
1113
1114 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1115 1 to "aa" instead of to an empty string. The bug affected repeated groups
1116 that could potentially match an empty string.
1117
1118 37. Optimizing single character iterators in JIT.
1119
1120 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1121 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1122 codepoints that are too big for the mode are faulted, and in a UTF mode,
1123 disallowed codepoints are also faulted.
1124
1125 39. If PCRE was compiled with UTF support, in three places in the DFA
1126 matcher there was code that should only have been obeyed in UTF mode, but
1127 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1128 processing when bytes with values greater than 127 were present. In 16-bit
1129 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1130 both cases the values are those that cannot be the first data item in a UTF
1131 character. The three items that might have provoked this were recursions,
1132 possessively repeated groups, and atomic groups.
1133
1134 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1135 and pcregrep, because some OS require shared objects to be explicitly
1136 passed to ld, causing the link step to fail if they are not.
1137
1138 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1139 mode, patterns that started with \h* or \R* might be incorrectly matched.
1140
1141
1142 Version 8.30 04-February-2012
1143 -----------------------------
1144
1145 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1146 name is defined in ctype.h.
1147
1148 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1149 only in quite long subpatterns.
1150
1151 3. Removed the function pcre_info(), which has been obsolete and deprecated
1152 since it was replaced by pcre_fullinfo() in February 2000.
1153
1154 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1155 match a (*MARK), and the match failed at the start of the subject, a
1156 reference to memory before the start of the subject could occur. This bug
1157 was introduced by fix 17 of release 8.21.
1158
1159 5. A reference to an unset group with zero minimum repetition was giving
1160 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1161 /(another)?(\1?)test/ matched against "hello world test". This bug was
1162 introduced in release 8.13.
1163
1164 6. Add support for 16-bit character strings (a large amount of work involving
1165 many changes and refactorings).
1166
1167 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1168 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1169 from a file.
1170
1171 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1172 rounding is not applied in this particular case).
1173
1174 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1175 if they appear, or are escaped, in patterns.
1176
1177 10. Get rid of a number of -Wunused-but-set-variable warnings.
1178
1179 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1180 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1181 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1182 also returns the mark "x". This bug applied to capturing parentheses,
1183 non-capturing parentheses, and atomic parentheses. It also applied to some
1184 assertions.
1185
1186 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1187 information out of configure.ac instead of relying on pcre.h.generic, which
1188 is not stored in the repository.
1189
1190 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1191 -lreadline.
1192
1193 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1194
1195 15. Applied Graycode's patch to put the top-level frame on the stack rather
1196 than the heap when not using the stack for recursion. This gives a
1197 performance improvement in many cases when recursion is not deep.
1198
1199 16. Experimental code added to "pcretest -C" to output the stack frame size.
1200
1201
1202 Version 8.21 12-Dec-2011
1203 ------------------------
1204
1205 1. Updating the JIT compiler.
1206
1207 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1208 are added as well.
1209
1210 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1211 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1212 calling _pcre_jit_exec. Some extra comments are added.
1213
1214 4. (*MARK) settings inside atomic groups that do not contain any capturing
1215 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1216 was introduced by change 18 for 8.20.
1217
1218 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1219 ECMA-262 standard.
1220
1221 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1222 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1223 This bug was probably introduced by change 9 of 8.13.
1224
1225 7. While fixing 6 above, I noticed that a number of other items were being
1226 incorrectly rejected as "not fixed length". This arose partly because newer
1227 opcodes had not been added to the fixed-length checking code. I have (a)
1228 corrected the bug and added tests for these items, and (b) arranged for an
1229 error to occur if an unknown opcode is encountered while checking for fixed
1230 length instead of just assuming "not fixed length". The items that were
1231 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1232 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1233 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1234
1235 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1236 being incorrectly compiled and would have given unpredicatble results.
1237
1238 9. A possessively repeated subpattern with minimum repeat count greater than
1239 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1240 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1241 the first (A) could occur when it should not.
1242
1243 10. Add a cast and remove a redundant test from the code.
1244
1245 11. JIT should use pcre_malloc/pcre_free for allocation.
1246
1247 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1248 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1249 is anything other than /usr, -L is still shown).
1250
1251 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1252
1253 14. Perl does not support \N without a following name in a [] class; PCRE now
1254 also gives an error.
1255
1256 15. If a forward reference was repeated with an upper limit of around 2000,
1257 it caused the error "internal error: overran compiling workspace". The
1258 maximum number of forward references (including repeats) was limited by the
1259 internal workspace, and dependent on the LINK_SIZE. The code has been
1260 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1261 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1262 of around 200,000 forward references. While doing this, I also speeded up
1263 the filling in of repeated forward references.
1264
1265 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1266 incorrectly expecting the subject to contain another "a" after the start.
1267
1268 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1269 in the match, the SKIP should be ignored. This was not happening; instead
1270 the SKIP was being treated as NOMATCH. For patterns such as
1271 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1272 tested.
1273
1274 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1275 now much more compatible with Perl, in particular in cases where the result
1276 is a non-match for a non-anchored pattern. For example, if
1277 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1278 "m", where previously it did not return a name. A side effect of this
1279 change is that for partial matches, the last encountered mark name is
1280 returned, as for non matches. A number of tests that were previously not
1281 Perl-compatible have been moved into the Perl-compatible test files. The
1282 refactoring has had the pleasing side effect of removing one argument from
1283 the match() function, thus reducing its stack requirements.
1284
1285 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1286 subsequent uses of /S (without +) incorrectly behaved like /S+.
1287
1288 21. Retrieve executable code size support for the JIT compiler and fixing
1289 some warnings.
1290
1291 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1292 not work when the shorter character appeared right at the end of the
1293 subject string.
1294
1295 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1296 systems.
1297
1298 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1299 output it when the /M option is used in pcretest.
1300
1301 25. The CheckMan script was not being included in the distribution. Also, added
1302 an explicit "perl" to run Perl scripts from the PrepareRelease script
1303 because this is reportedly needed in Windows.
1304
1305 26. If study data was being save in a file and studying had not found a set of
1306 "starts with" bytes for the pattern, the data written to the file (though
1307 never used) was taken from uninitialized memory and so caused valgrind to
1308 complain.
1309
1310 27. Updated RunTest.bat as provided by Sheri Pierce.
1311
1312 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1313
1314 29. Computation of memory usage for the table of capturing group names was
1315 giving an unnecessarily large value.
1316
1317
1318 Version 8.20 21-Oct-2011
1319 ------------------------
1320
1321 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1322 a POSIX class. After further experiments with Perl, which convinced me that
1323 Perl has bugs and confusions, a closing square bracket is no longer allowed
1324 in a POSIX name. This bug also affected patterns with classes that started
1325 with full stops.
1326
1327 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1328 captured substring, but while checking the failing first alternative,
1329 substring 1 is temporarily captured. If the output vector supplied to
1330 pcre_exec() was not big enough for this capture, the yield of the function
1331 was still zero ("insufficient space for captured substrings"). This cannot
1332 be totally fixed without adding another stack variable, which seems a lot
1333 of expense for a edge case. However, I have improved the situation in cases
1334 such as /(a)(b)x|abc/ matched against "abc", where the return code
1335 indicates that fewer than the maximum number of slots in the ovector have
1336 been set.
1337
1338 3. Related to (2) above: when there are more back references in a pattern than
1339 slots in the output vector, pcre_exec() uses temporary memory during
1340 matching, and copies in the captures as far as possible afterwards. It was
1341 using the entire output vector, but this conflicts with the specification
1342 that only 2/3 is used for passing back captured substrings. Now it uses
1343 only the first 2/3, for compatibility. This is, of course, another edge
1344 case.
1345
1346 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1347 main code base, and can be used by building with --enable-jit. When this is
1348 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1349 runtime --no-jit option is given.
1350
1351 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1352 ovector, the return from the function was zero, implying that there were
1353 other matches that did not fit. The correct "exactly full" value is now
1354 returned.
1355
1356 6. If a subpattern that was called recursively or as a subroutine contained
1357 (*PRUNE) or any other control that caused it to give a non-standard return,
1358 invalid errors such as "Error -26 (nested recursion at the same subject
1359 position)" or even infinite loops could occur.
1360
1361 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1362 computing the minimum length on reaching *ACCEPT, and so ended up with the
1363 wrong value of 1 rather than 0. Further investigation indicates that
1364 computing a minimum subject length in the presence of *ACCEPT is difficult
1365 (think back references, subroutine calls), and so I have changed the code
1366 so that no minimum is registered for a pattern that contains *ACCEPT.
1367
1368 8. If (*THEN) was present in the first (true) branch of a conditional group,
1369 it was not handled as intended. [But see 16 below.]
1370
1371 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1372 Sheri Pierce.
1373
1374 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1375 the first byte in a match must be "a".
1376
1377 11. Change 17 for 8.13 increased the recursion depth for patterns like
1378 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1379 pattern contains any instances of (*THEN). If it does not, the old
1380 optimizations are restored. It would be nice to do this on a per-group
1381 basis, but at the moment that is not feasible.
1382
1383 12. In some environments, the output of pcretest -C is CRLF terminated. This
1384 broke RunTest's code that checks for the link size. A single white space
1385 character after the value is now allowed for.
1386
1387 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1388 For "fr", it uses the Windows-specific input and output files.
1389
1390 14. If (*THEN) appeared in a group that was called recursively or as a
1391 subroutine, it did not work as intended. [But see next item.]
1392
1393 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1394 pattern fragments (but not containing any | characters). If A and B are
1395 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1396 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1397 D. In other words, Perl considers parentheses that do not contain any |
1398 characters to be part of a surrounding alternative, whereas PCRE was
1399 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1400 differently. PCRE now behaves in the same way as Perl, except in the case
1401 of subroutine/recursion calls such as (?1) which have in any case always
1402 been different (but PCRE had them first :-).
1403
1404 16. Related to 15 above: Perl does not treat the | in a conditional group as
1405 creating alternatives. Such a group is treated in the same way as an
1406 ordinary group without any | characters when processing (*THEN). PCRE has
1407 been changed to match Perl's behaviour.
1408
1409 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1410 RunGrepTest script failed.
1411
1412 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1413 inevitable for groups that contain captures, but it can lead to a lot of
1414 stack use in large patterns. The old behaviour has been restored for atomic
1415 groups that do not contain any capturing parentheses.
1416
1417 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1418 suppress the check for a minimum subject length at run time. (If it was
1419 given to pcre_exec() or pcre_dfa_exec() it did work.)
1420
1421 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1422 fail to work when decoding hex characters in data strings in EBCDIC
1423 environments.
1424
1425 21. It appears that in at least one Mac OS environment, the isxdigit() function
1426 is implemented as a macro that evaluates to its argument more than once,
1427 contravening the C 90 Standard (I haven't checked a later standard). There
1428 was an instance in pcretest which caused it to go wrong when processing
1429 \x{...} escapes in subject strings. The has been rewritten to avoid using
1430 things like p++ in the argument of isxdigit().
1431
1432
1433 Version 8.13 16-Aug-2011
1434 ------------------------
1435
1436 1. The Unicode data tables have been updated to Unicode 6.0.0.
1437
1438 2. Two minor typos in pcre_internal.h have been fixed.
1439
1440 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1441 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1442 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1443
1444 4. There were a number of related bugs in the code for matching backrefences
1445 caselessly in UTF-8 mode when codes for the characters concerned were
1446 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1447 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1448 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1449 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1450 2-byte code at the end of the subject (it thought there wasn't enough data
1451 left).
1452
1453 5. Comprehensive information about what went wrong is now returned by
1454 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1455 as the output vector has at least 2 elements. The offset of the start of
1456 the failing character and a reason code are placed in the vector.
1457
1458 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1459 now returned is for the first byte of the failing character, instead of the
1460 last byte inspected. This is an incompatible change, but I hope it is small
1461 enough not to be a problem. It makes the returned offset consistent with
1462 pcre_exec() and pcre_dfa_exec().
1463
1464 7. pcretest now gives a text phrase as well as the error number when
1465 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1466 failure, the offset and reason code are output.
1467
1468 8. When \R was used with a maximizing quantifier it failed to skip backwards
1469 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1470 back over a single character (\n). This seems wrong (because it treated the
1471 two characters as a single entity when going forwards), conflicts with the
1472 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1473 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1474 has been changed.
1475
1476 9. Some internal refactoring has changed the processing so that the handling
1477 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1478 time (the PCRE_DOTALL option was changed this way some time ago: version
1479 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1480 which was always a bit of a fudge. It also means that there is one less
1481 argument for the match() function, which reduces its stack requirements
1482 slightly. This change also fixes an incompatibility with Perl: the pattern
1483 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1484
1485 10. More internal refactoring has drastically reduced the number of recursive
1486 calls to match() for possessively repeated groups such as (abc)++ when
1487 using pcre_exec().
1488
1489 11. While implementing 10, a number of bugs in the handling of groups were
1490 discovered and fixed:
1491
1492 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1493 (a|)*(?1) gave a compile-time internal error.
1494 ((a|)+)+ did not notice that the outer group could match an empty string.
1495 (^a|^)+ was not marked as anchored.
1496 (.*a|.*)+ was not marked as matching at start or after a newline.
1497
1498 12. Yet more internal refactoring has removed another argument from the match()
1499 function. Special calls to this function are now indicated by setting a
1500 value in a variable in the "match data" data block.
1501
1502 13. Be more explicit in pcre_study() instead of relying on "default" for
1503 opcodes that mean there is no starting character; this means that when new
1504 ones are added and accidentally left out of pcre_study(), testing should
1505 pick them up.
1506
1507 14. The -s option of pcretest has been documented for ages as being an old
1508 synonym of -m (show memory usage). I have changed it to mean "force study
1509 for every regex", that is, assume /S for every regex. This is similar to -i
1510 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1511 using it. It makes it easier to run collections of tests with and without
1512 study enabled, and thereby test pcre_study() more easily. All the standard
1513 tests are now run with and without -s (but some patterns can be marked as
1514 "never study" - see 20 below).
1515
1516 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1517 restoration of the capturing data to the outer values was not happening
1518 correctly.
1519
1520 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1521 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1522 pattern had matched an empty string, and so incorrectly returned a no
1523 match.
1524
1525 17. There was optimizing code for the last branch of non-capturing parentheses,
1526 and also for the obeyed branch of a conditional subexpression, which used
1527 tail recursion to cut down on stack usage. Unfortunately, now that there is
1528 the possibility of (*THEN) occurring in these branches, tail recursion is
1529 no longer possible because the return has to be checked for (*THEN). These
1530 two optimizations have therefore been removed. [But see 8.20/11 above.]
1531
1532 18. If a pattern containing \R was studied, it was assumed that \R always
1533 matched two bytes, thus causing the minimum subject length to be
1534 incorrectly computed because \R can also match just one byte.
1535
1536 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1537 was incorrectly computed.
1538
1539 20. If /S is present twice on a test pattern in pcretest input, it now
1540 *disables* studying, thereby overriding the use of -s on the command line
1541 (see 14 above). This is necessary for one or two tests to keep the output
1542 identical in both cases.
1543
1544 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1545 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1546
1547 22. When an atomic group that contained a capturing parenthesis was
1548 successfully matched, but the branch in which it appeared failed, the
1549 capturing was not being forgotten if a higher numbered group was later
1550 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1551 group 1 to "a", when in fact it should be unset. This applied to multi-
1552 branched capturing and non-capturing groups, repeated or not, and also to
1553 positive assertions (capturing in negative assertions does not happen
1554 in PCRE) and also to nested atomic groups.
1555
1556 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1557 subject after a captured substring, to make it easier to tell which of a
1558 number of identical substrings has been captured.
1559
1560 24. The way atomic groups are processed by pcre_exec() has been changed so that
1561 if they are repeated, backtracking one repetition now resets captured
1562 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1563 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1564 "aaa". Previously, it would have been "a". As part of this code
1565 refactoring, the way recursive calls are handled has also been changed.
1566
1567 25. If an assertion condition captured any substrings, they were not passed
1568 back unless some other capturing happened later. For example, if
1569 (?(?=(a))a) was matched against "a", no capturing was returned.
1570
1571 26. When studying a pattern that contained subroutine calls or assertions,
1572 the code for finding the minimum length of a possible match was handling
1573 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1574 group 1 called group 2 while simultaneously a separate group 2 called group
1575 1). A stack overflow occurred in this case. I have fixed this by limiting
1576 the recursion depth to 10.
1577
1578 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1579 Fortmann. This supports explicit test numbers on the command line, and has
1580 argument validation and error reporting.
1581
1582 28. An instance of \X with an unlimited repeat could fail if at any point the
1583 first character it looked at was a mark character.
1584
1585 29. Some minor code refactoring concerning Unicode properties and scripts
1586 should reduce the stack requirement of match() slightly.
1587
1588 30. Added the '=' option to pcretest to check the setting of unused capturing
1589 slots at the end of the pattern, which are documented as being -1, but are
1590 not included in the return count.
1591
1592 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1593 compiled something random. Now it gives a compile-time error (as does
1594 Perl).
1595
1596 32. A *MARK encountered during the processing of a positive assertion is now
1597 recorded and passed back (compatible with Perl).
1598
1599 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1600 had alternative anchored branches, the search for a second match in a line
1601 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1602 matched the line "0102" twice. The same bug affected patterns that started
1603 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1604 twice.
1605
1606 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1607 does, and because of capturing effects, quantifying parenthesized
1608 assertions may at times be useful. Quantifiers are now allowed for
1609 parenthesized assertions.
1610
1611 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1612
1613 36. \g was being checked for fancy things in a character class, when it should
1614 just be a literal "g".
1615
1616 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1617 appearance of a nested POSIX class supersedes an apparent external class.
1618 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1619 unescaped square brackets may also appear as part of class names. For
1620 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1621 more like Perl. (But see 8.20/1 above.)
1622
1623 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1624 was because it thought it was \N{name}, which is not supported).
1625
1626 39. Add minix to OS list not supporting the -S option in pcretest.
1627
1628 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1629 cannot analyze patterns in sufficient detail to catch mutual recursions
1630 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1631 subgroup is called recursively as a subpattern for a second time at the
1632 same position in the subject string. In previous releases this might have
1633 been caught by the recursion limit, or it might have run out of stack.
1634
1635 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1636 happen only once. PCRE was, however incorrectly giving a compile time error
1637 "recursive call could loop indefinitely" because it cannot analyze the
1638 pattern in sufficient detail. The compile time test no longer happens when
1639 PCRE is compiling a conditional subpattern, but actual runaway loops are
1640 now caught at runtime (see 40 above).
1641
1642 42. It seems that Perl allows any characters other than a closing parenthesis
1643 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1644 has been changed to be the same.
1645
1646 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1647 as not to get warnings when autogen.sh is called. Also changed
1648 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1649
1650 44. To help people who use pcregrep to scan files containing exceedingly long
1651 lines, the following changes have been made:
1652
1653 (a) The default value of the buffer size parameter has been increased from
1654 8K to 20K. (The actual buffer used is three times this size.)
1655
1656 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1657 PCRE is built.
1658
1659 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1660 to be set at run time.
1661
1662 (d) Numerical values in pcregrep options can be followed by K or M, for
1663 example --buffer-size=50K.
1664
1665 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1666 given and the return code is set to 2.
1667
1668 45. Add a pointer to the latest mark to the callout data block.
1669
1670 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1671 partial match of an empty string instead of no match. This was specific to
1672 the use of ".".
1673
1674 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1675 complete match instead of a partial match. This bug was dependent on both
1676 the PCRE_UTF8 and PCRE_DOTALL options being set.
1677
1678 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1679 starting byte set, because \b was not being ignored.
1680
1681
1682 Version 8.12 15-Jan-2011
1683 ------------------------
1684
1685 1. Fixed some typos in the markup of the man pages, and wrote a script that
1686 checks for such things as part of the documentation building process.
1687
1688 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1689 --match-limit and --recursion-limit options (added for 8.11). In
1690 particular, this made one of the standard tests fail. (The integer value
1691 went into the wrong half of a long int.)
1692
1693 3. If the --colour option was given to pcregrep with -v (invert match), it
1694 did strange things, either producing crazy output, or crashing. It should,
1695 of course, ignore a request for colour when reporting lines that do not
1696 match.
1697
1698 4. Another pcregrep bug caused similar problems if --colour was specified with
1699 -M (multiline) and the pattern match finished with a line ending.
1700
1701 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1702 matched in multiline mode, the following line was shown as part of the
1703 match. This seems wrong, so I have changed it.
1704
1705 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1706 the check for further matches in the same line (so they could be coloured)
1707 to overrun the end of the current line. If another match was found, it was
1708 incorrectly shown (and then shown again when found in the next line).
1709
1710 7. If pcregrep was compiled under Windows, there was a reference to the
1711 function pcregrep_exit() before it was defined. I am assuming this was
1712 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1713 reported by a user. I've moved the definition above the reference.
1714
1715
1716 Version 8.11 10-Dec-2010
1717 ------------------------
1718
1719 1. (*THEN) was not working properly if there were untried alternatives prior
1720 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1721 backtracked to try for "b" instead of moving to the next alternative branch
1722 at the same level (in this case, to look for "c"). The Perl documentation
1723 is clear that when (*THEN) is backtracked onto, it goes to the "next
1724 alternative in the innermost enclosing group".
1725
1726 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1727 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1728 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1729 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1730 (*THEN).
1731
1732 3. If \s appeared in a character class, it removed the VT character from
1733 the class, even if it had been included by some previous item, for example
1734 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1735 of \s, but is part of the POSIX "space" class.)
1736
1737 4. A partial match never returns an empty string (because you can always
1738 match an empty string at the end of the subject); however the checking for
1739 an empty string was starting at the "start of match" point. This has been
1740 changed to the "earliest inspected character" point, because the returned
1741 data for a partial match starts at this character. This means that, for
1742 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1743 (previously it gave "no match").
1744
1745 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1746 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1747 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1748 has an implication that the given string is incomplete (because a partial
1749 match is preferred over a full match). For this reason, these items now
1750 give a partial match in this situation. [Aside: previously, the one case
1751 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1752 match rather than a full match, which was wrong by the old rules, but is
1753 now correct.]
1754
1755 6. There was a bug in the handling of #-introduced comments, recognized when
1756 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1757 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1758 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1759 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1760 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1761 places in pcre_compile().
1762
1763 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1764 comments when looking ahead for named forward references to subpatterns,
1765 the only newline sequence it recognized was NL. It now handles newlines
1766 according to the set newline convention.
1767
1768 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1769 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1770 cater for a lack of strerror(). These oversights have been fixed.
1771
1772 9. Added --match-limit and --recursion-limit to pcregrep.
1773
1774 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1775
1776 11. When the -o option was used, pcregrep was setting a return code of 1, even
1777 when matches were found, and --line-buffered was not being honoured.
1778
1779 12. Added an optional parentheses number to the -o and --only-matching options
1780 of pcregrep.
1781
1782 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1783 can match an empty string. The code to do it in pcretest and pcredemo
1784 needed fixing:
1785
1786 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1787 only one byte after an empty string match just before CRLF (this case
1788 just got forgotten; "any" and "anycrlf" were OK).
1789
1790 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1791 mode when an empty string match preceded an ASCII character followed by
1792 a non-ASCII character. (The code for advancing by one character rather
1793 than one byte was nonsense.)
1794
1795 (c) The pcredemo.c sample program did not have any code at all to handle
1796 the cases when CRLF is a valid newline sequence.
1797
1798 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1799 as a starting offset was within the subject string. There is now a new
1800 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1801 negative or greater than the length of the string. In order to test this,
1802 pcretest is extended to allow the setting of negative starting offsets.
1803
1804 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1805 starting offset points to the beginning of a UTF-8 character was
1806 unnecessarily clumsy. I tidied it up.
1807
1808 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1809 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1810
1811 17. Nobody had reported that the --include_dir option, which was added in
1812 release 7.7 should have been called --include-dir (hyphen, not underscore)
1813 for compatibility with GNU grep. I have changed it to --include-dir, but
1814 left --include_dir as an undocumented synonym, and the same for
1815 --exclude-dir, though that is not available in GNU grep, at least as of
1816 release 2.5.4.
1817
1818 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1819 characters from a string of bytes) have been redefined so as not to use
1820 loops, in order to improve performance in some environments. At the same
1821 time, I abstracted some of the common code into auxiliary macros to save
1822 repetition (this should not affect the compiled code).
1823
1824 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1825 compile-time error is now given if \c is not followed by an ASCII
1826 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1827 different, and any byte value is allowed.)
1828
1829 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1830 START_OPTIMIZE option, which is now allowed at compile time - but just
1831 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1832 to pcregrep and other applications that have no direct access to PCRE
1833 options. The new /Y option in pcretest sets this option when calling
1834 pcre_compile().
1835
1836 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1837 back references. Groups containing recursive back references were forced to
1838 be atomic by that change, but in the case of named groups, the amount of
1839 memory required was incorrectly computed, leading to "Failed: internal
1840 error: code overflow". This has been fixed.
1841
1842 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1843 pcretest.c, to avoid build problems in some Borland environments.
1844
1845
1846 Version 8.10 25-Jun-2010
1847 ------------------------
1848
1849 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1850 THEN.
1851
1852 2. (*ACCEPT) was not working when inside an atomic group.
1853
1854 3. Inside a character class, \B is treated as a literal by default, but
1855 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1856 causes the error). The code is unchanged, but I tidied the documentation.
1857
1858 4. Inside a character class, PCRE always treated \R and \X as literals,
1859 whereas Perl faults them if its -w option is set. I have changed PCRE so
1860 that it faults them when PCRE_EXTRA is set.
1861
1862 5. Added support for \N, which always matches any character other than
1863 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1864
1865 6. When compiling pcregrep with newer versions of gcc which may have
1866 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1867 declared with attribute warn_unused_result" were given. Just casting the
1868 result to (void) does not stop the warnings; a more elaborate fudge is
1869 needed. I've used a macro to implement this.
1870
1871 7. Minor change to pcretest.c to avoid a compiler warning.
1872
1873 8. Added four artifical Unicode properties to help with an option to make
1874 \s etc use properties (see next item). The new properties are: Xan
1875 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1876
1877 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1878 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1879 this option. Modified pcretest to add /W to test this facility. Added
1880 REG_UCP to make it available via the POSIX interface.
1881
1882 10. Added --line-buffered to pcregrep.
1883
1884 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1885 studied, and the match started with a letter with a code point greater than
1886 127 whose first byte was different to the first byte of the other case of
1887 the letter, the other case of this starting letter was not recognized
1888 (#976).
1889
1890 12. If a pattern that was studied started with a repeated Unicode property
1891 test, for example, \p{Nd}+, there was the theoretical possibility of
1892 setting up an incorrect bitmap of starting bytes, but fortunately it could
1893 not have actually happened in practice until change 8 above was made (it
1894 added property types that matched character-matching opcodes).
1895
1896 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1897 possible starting bytes for non-anchored patterns.
1898
1899 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1900 \R, and also a number of cases that involve Unicode properties, both
1901 explicit and implicit when PCRE_UCP is set.
1902
1903 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1904 input, it could crash or give wrong results if characters with values
1905 greater than 0xc0 were present in the subject string. (Detail: it assumed
1906 UTF-8 input when processing these items.)
1907
1908 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1909 size_t is 64-bit (#991).
1910
1911 17. Added a check for running out of memory when PCRE is compiled with
1912 --disable-stack-for-recursion (#990).
1913
1914 18. If the last data line in a file for pcretest does not have a newline on
1915 the end, a newline was missing in the output.
1916
1917 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1918 less than 128) in its various bitmaps. However, there is a facility for
1919 generating tables according to the current locale when PCRE is compiled. It
1920 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1921 characters, are recognized by isspace() and therefore were getting set in
1922 these tables, and indeed these tables seem to approximate to ISO 8859. This
1923 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1924 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1925 which of course cannot start UTF-8 characters. I have changed the code so
1926 that only real ASCII characters (less than 128) and the correct starting
1927 bytes for UTF-8 encodings are set for characters greater than 127 when in
1928 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1929 altogether.)
1930
1931 20. Added the /T option to pcretest so as to be able to run tests with non-
1932 standard character tables, thus making it possible to include the tests
1933 used for 19 above in the standard set of tests.
1934
1935 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1936 reference to a subpattern the other side of a comment that contains an
1937 opening parenthesis caused either an internal compiling error, or a
1938 reference to the wrong subpattern.
1939
1940
1941 Version 8.02 19-Mar-2010
1942 ------------------------
1943
1944 1. The Unicode data tables have been updated to Unicode 5.2.0.
1945
1946 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1947 configured.
1948
1949 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1950 original author of that file, following a query about its status.
1951
1952 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1953 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1954
1955 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1956 quantifier applied to a forward-referencing subroutine call, could compile
1957 incorrect code or give the error "internal error: previously-checked
1958 referenced subpattern not found".
1959
1960 6. Both MS Visual Studio and Symbian OS have problems with initializing
1961 variables to point to external functions. For these systems, therefore,
1962 pcre_malloc etc. are now initialized to local functions that call the
1963 relevant global functions.
1964
1965 7. There were two entries missing in the vectors called coptable and poptable
1966 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1967 I've fixed the data, and added a kludgy way of testing at compile time that
1968 the lengths are correct (equal to the number of opcodes).
1969
1970 8. Following on from 7, I added a similar kludge to check the length of the
1971 eint vector in pcreposix.c.
1972
1973 9. Error texts for pcre_compile() are held as one long string to avoid too
1974 much relocation at load time. To find a text, the string is searched,
1975 counting zeros. There was no check for running off the end of the string,
1976 which could happen if a new error number was added without updating the
1977 string.
1978
1979 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1980
1981 11. \K was not working if it appeared in an atomic group or in a group that
1982 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1983 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1984 the assertion is positive, but not if it is negative.
1985
1986 12. Change 11 fortuitously reduced the size of the stack frame used in the
1987 "match()" function of pcre_exec.c by one pointer. Forthcoming
1988 implementation of support for (*MARK) will need an extra pointer on the
1989 stack; I have reserved it now, so that the stack frame size does not
1990 decrease.
1991
1992 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1993 item in branch that calls a recursion is a subroutine call - as in the
1994 second branch in the above example - was incorrectly given the compile-
1995 time error "recursive call could loop indefinitely" because pcre_compile()
1996 was not correctly checking the subroutine for matching a non-empty string.
1997
1998 14. The checks for overrunning compiling workspace could trigger after an
1999 overrun had occurred. This is a "should never occur" error, but it can be
2000 triggered by pathological patterns such as hundreds of nested parentheses.
2001 The checks now trigger 100 bytes before the end of the workspace.
2002
2003 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2004
2005
2006 Version 8.01 19-Jan-2010
2007 ------------------------
2008
2009 1. If a pattern contained a conditional subpattern with only one branch (in
2010 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2011 computed the wrong minimum data length (which is of course zero for such
2012 subpatterns). This could cause incorrect "no match" results.
2013
2014 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2015 the pattern is reset in the first branch, pcre_compile() failed with
2016 "internal error: code overflow at offset...". This happened only when
2017 the reset was to the original external option setting. (An optimization
2018 abstracts leading options settings into an external setting, which was the
2019 cause of this.)
2020
2021 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2022 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2023 assertion pattern did not match (meaning that the assertion was true), it
2024 was incorrectly treated as false if the SKIP had been reached during the
2025 matching. This also applied to assertions used as conditions.
2026
2027 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2028 assertion subpattern, including such a pattern used as a condition,
2029 unpredictable results occurred, instead of the error return
2030 PCRE_ERROR_DFA_UITEM.
2031
2032 5. The C++ GlobalReplace function was not working like Perl for the special
2033 situation when an empty string is matched. It now does the fancy magic
2034 stuff that is necessary.
2035
2036 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2037 removed. (These were left over from very, very early versions of PCRE.)
2038
2039 7. Some cosmetic changes to the code to make life easier when compiling it
2040 as part of something else:
2041
2042 (a) Change DEBUG to PCRE_DEBUG.
2043
2044 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2045 called "current" as "current_branch", to prevent a collision with the
2046 Linux macro when compiled as a kernel module.
2047
2048 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2049 prevent a collision with the Linux macro when compiled as a kernel
2050 module.
2051
2052 8. In pcre_compile() there are some checks for integer overflows that used to
2053 cast potentially large values to (double). This has been changed to that
2054 when building, a check for int64_t is made, and if it is found, it is used
2055 instead, thus avoiding the use of floating point arithmetic. (There is no
2056 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2057 double.
2058
2059 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2060 2005 (difference between two addresses compared to an unsigned value).
2061
2062 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2063 custom one, because of the following reported problem in Windows:
2064
2065 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2066 under Win32.
2067 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2068 therefore missing the function definition.
2069 - The compiler thus generates a "C" signature for the test function.
2070 - The linker fails to find the "C" function.
2071 - PCRE fails to configure if asked to do so against libbz2.
2072
2073 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2074 messages were output:
2075
2076 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2077 rerunning libtoolize, to keep the correct libtool macros in-tree.
2078 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2079
2080 I have done both of these things.
2081
2082 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2083 most of the time, it *can* run out if it is given a pattern that contains a
2084 runaway infinite recursion. I updated the discussion in the pcrestack man
2085 page.
2086
2087 13. Now that we have gone to the x.xx style of version numbers, the minor
2088 version may start with zero. Using 08 or 09 is a bad idea because users
2089 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2090 interpreted as invalid octal numbers. I've updated the previous comment in
2091 configure.ac, and also added a check that gives an error if 08 or 09 are
2092 used.
2093
2094 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2095 causing partial matching to fail when the end of the subject matched \W
2096 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2097
2098 15. There were some discrepancies between the declarations in pcre_internal.h
2099 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2100 their definitions. The declarations used "const uschar *" and the
2101 definitions used USPTR. Even though USPTR is normally defined as "const
2102 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2103 reported that: "This difference in casting confuses some C++ compilers, for
2104 example, SunCC recognizes above declarations as different functions and
2105 generates broken code for hbpcre." I have changed the declarations to use
2106 USPTR.
2107
2108 16. GNU libtool is named differently on some systems. The autogen.sh script now
2109 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2110 (FreeBSD).
2111
2112 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2113 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2114 comment: "Figure out how to create a longlong from a string: strtoll and
2115 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2116 instance, but it only takes 2 args instead of 3!"
2117
2118 18. A subtle bug concerned with back references has been fixed by a change of
2119 specification, with a corresponding code fix. A pattern such as
2120 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2121 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2122 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2123 same bug. Such groups have to be quantified to be useful, or contained
2124 inside another quantified group. (If there's no repetition, the reference
2125 can never match.) The problem arises because, having left the group and
2126 moved on to the rest of the pattern, a later failure that backtracks into
2127 the group uses the captured value from the final iteration of the group
2128 rather than the correct earlier one. I have fixed this in PCRE by forcing
2129 any group that contains a reference to itself to be an atomic group; that
2130 is, there cannot be any backtracking into it once it has completed. This is
2131 similar to recursive and subroutine calls.
2132
2133
2134 Version 8.00 19-Oct-09
2135 ----------------------
2136
2137 1. The table for translating pcre_compile() error codes into POSIX error codes
2138 was out-of-date, and there was no check on the pcre_compile() error code
2139 being within the table. This could lead to an OK return being given in
2140 error.
2141
2142 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2143 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2144 in a Windows environment.
2145
2146 3. The pcregrep --count option prints the count for each file even when it is
2147 zero, as does GNU grep. However, pcregrep was also printing all files when
2148 --files-with-matches was added. Now, when both options are given, it prints
2149 counts only for those files that have at least one match. (GNU grep just
2150 prints the file name in this circumstance, but including the count seems
2151 more useful - otherwise, why use --count?) Also ensured that the
2152 combination -clh just lists non-zero counts, with no names.
2153
2154 4. The long form of the pcregrep -F option was incorrectly implemented as
2155 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2156 but it seems right to fix it, and I didn't think it was worth preserving
2157 the old behaviour.
2158
2159 5. The command line items --regex=pattern and --regexp=pattern were not
2160 recognized by pcregrep, which required --regex pattern or --regexp pattern
2161 (with a space rather than an '='). The man page documented the '=' forms,
2162 which are compatible with GNU grep; these now work.
2163
2164 6. No libpcreposix.pc file was created for pkg-config; there was just
2165 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2166
2167 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2168 when UCP support is not needed, by modifying the Python script that
2169 generates it from Unicode data files. This should not matter if the module
2170 is correctly used as a library, but I received one complaint about 50K of
2171 unwanted data. My guess is that the person linked everything into his
2172 program rather than using a library. Anyway, it does no harm.
2173
2174 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2175 was a minimum greater than 1 for a wide character in a possessive
2176 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2177 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2178 character. Chaos in the form of incorrect output or a compiling loop could
2179 result.
2180
2181 9. The restrictions on what a pattern can contain when partial matching is
2182 requested for pcre_exec() have been removed. All patterns can now be
2183 partially matched by this function. In addition, if there are at least two
2184 slots in the offset vector, the offset of the earliest inspected character
2185 for the match and the offset of the end of the subject are set in them when
2186 PCRE_ERROR_PARTIAL is returned.
2187
2188 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2189 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2190 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2191 and may be more useful for multi-segment matching.
2192
2193 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2194 used to be given if ever the end of the subject was reached; now it is
2195 given only if matching could not proceed because another character was
2196 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2197 string "Z", which now yields "no match" instead of "partial match". In the
2198 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2199 final character ended with (*FAIL).
2200
2201 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2202 if the pattern had a "must contain" character that was already found in the
2203 earlier partial match, unless partial matching was again requested. For
2204 example, with the pattern /dog.(body)?/, the "must contain" character is
2205 "g". If the first part-match was for the string "dog", restarting with
2206 "sbody" failed. This bug has been fixed.
2207
2208 13. The string returned by pcre_dfa_exec() after a partial match has been
2209 changed so that it starts at the first inspected character rather than the
2210 first character of the match. This makes a difference only if the pattern
2211 starts with a lookbehind assertion or \b or \B (\K is not supported by
2212 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2213 matching functions compatible, and I think it's the right thing to do.
2214
2215 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2216 so that the demonstration program is easily available in environments where
2217 PCRE has not been installed from source.
2218
2219 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2220 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2221 library.
2222
2223 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2224 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2225 is not the first non-POSIX option to be added. Clearly some people find
2226 these options useful.
2227
2228 17. If a caller to the POSIX matching function regexec() passes a non-zero
2229 value for nmatch with a NULL value for pmatch, the value of
2230 nmatch is forced to zero.
2231
2232 18. RunGrepTest did not have a test for the availability of the -u option of
2233 the diff command, as RunTest does. It now checks in the same way as
2234 RunTest, and also checks for the -b option.
2235
2236 19. If an odd number of negated classes containing just a single character
2237 interposed, within parentheses, between a forward reference to a named
2238 subpattern and the definition of the subpattern, compilation crashed with
2239 an internal error, complaining that it could not find the referenced
2240 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2241 [The bug was that it was starting one character too far in when skipping
2242 over the character class, thus treating the ] as data rather than
2243 terminating the class. This meant it could skip too much.]
2244
2245 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2246 /g option in pcretest when the pattern contains \K, which makes it possible
2247 to have an empty string match not at the start, even when the pattern is
2248 anchored. Updated pcretest and pcredemo to use this option.
2249
2250 21. If the maximum number of capturing subpatterns in a recursion was greater
2251 than the maximum at the outer level, the higher number was returned, but
2252 with unset values at the outer level. The correct (outer level) value is
2253 now given.
2254
2255 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2256 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2257 make it do so. The string so far is captured, making this feature
2258 compatible with Perl.
2259
2260 23. The tests have been re-organized, adding tests 11 and 12, to make it
2261 possible to check the Perl 5.10 features against Perl 5.10.
2262
2263 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2264 pattern matches a fixed length string. PCRE did not allow this; now it
2265 does. Neither allows recursion.
2266
2267 25. I finally figured out how to implement a request to provide the minimum
2268 length of subject string that was needed in order to match a given pattern.
2269 (It was back references and recursion that I had previously got hung up
2270 on.) This code has now been added to pcre_study(); it finds a lower bound
2271 to the length of subject needed. It is not necessarily the greatest lower
2272 bound, but using it to avoid searching strings that are too short does give
2273 some useful speed-ups. The value is available to calling programs via
2274 pcre_fullinfo().
2275
2276 26. While implementing 25, I discovered to my embarrassment that pcretest had
2277 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2278 study optimizations had never been tested with that matching function.
2279 Oops. What is worse, even when it was passed study data, there was a bug in
2280 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2281 were also very few tests of studied patterns with pcre_dfa_exec().
2282
2283 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2284 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2285 on the other side of the coin, they are no longer allowed to have different
2286 names, because these cannot be distinguished in PCRE, and this has caused
2287 confusion. (This is a difference from Perl.)
2288
2289 28. When duplicate subpattern names are present (necessarily with different
2290 numbers, as required by 27 above), and a test is made by name in a
2291 conditional pattern, either for a subpattern having been matched, or for
2292 recursion in such a pattern, all the associated numbered subpatterns are
2293 tested, and the overall condition is true if the condition is true for any
2294 one of them. This is the way Perl works, and is also more like the way
2295 testing by number works.
2296
2297
2298 Version 7.9 11-Apr-09
2299 ---------------------
2300
2301 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2302 (pcretest), all targets were linked against these libraries. This included
2303 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2304 libraries. This caused unwanted dependencies to be created. This problem
2305 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2306 pcretest is linked with readline.
2307
2308 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2309 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2310 moved outside it again, because FALSE and TRUE are already defined in AIX,
2311 but BOOL is not.
2312
2313 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2314 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2315
2316 4. The pcregrep documentation said spaces were inserted as well as colons (or
2317 hyphens) following file names and line numbers when outputting matching
2318 lines. This is not true; no spaces are inserted. I have also clarified the
2319 wording for the --colour (or --color) option.
2320
2321 5. In pcregrep, when --colour was used with -o, the list of matching strings
2322 was not coloured; this is different to GNU grep, so I have changed it to be
2323 the same.
2324
2325 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2326 each matching line was coloured. Now it goes on to look for further matches
2327 of any of the test patterns, which is the same behaviour as GNU grep.
2328
2329 7. A pattern that could match an empty string could cause pcregrep to loop; it
2330 doesn't make sense to accept an empty string match in pcregrep, so I have
2331 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2332 seems to be how GNU grep behaves. [But see later change 40 for release
2333 8.33.]
2334
2335 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2336 start or after a newline", because the conditional assertion was not being
2337 correctly handled. The rule now is that both the assertion and what follows
2338 in the first alternative must satisfy the test.
2339
2340 9. If auto-callout was enabled in a pattern with a conditional group whose
2341 condition was an assertion, PCRE could crash during matching, both with
2342 pcre_exec() and pcre_dfa_exec().
2343
2344 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2345 used for matching.
2346
2347 11. Unicode property support in character classes was not working for
2348 characters (bytes) greater than 127 when not in UTF-8 mode.
2349
2350 12. Added the -M command line option to pcretest.
2351
2352 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2353
2354 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2355
2356 16. Added comments and documentation about mis-use of no_arg in the C++
2357 wrapper.
2358
2359 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2360 from Martin Jerabek that uses macro names for all relevant character and
2361 string constants.
2362
2363 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2364 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2365 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2366 these, but not everybody uses configure.
2367
2368 19. A conditional group that had only one branch was not being correctly
2369 recognized as an item that could match an empty string. This meant that an
2370 enclosing group might also not be so recognized, causing infinite looping
2371 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2372 with the subject "ab", where knowledge that the repeated group can match
2373 nothing is needed in order to break the loop.
2374
2375 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2376 exec(), but without supplying a callout function, matching went wrong.
2377
2378 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2379 leak if the size of the offset vector was greater than 30. When the vector
2380 is smaller, the saved offsets during recursion go onto a local stack
2381 vector, but for larger vectors malloc() is used. It was failing to free
2382 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2383 error, in fact).
2384
2385 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2386 heapframe that is used only when UTF-8 support is enabled. This caused no
2387 problem, but was untidy.
2388
2389 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2390 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2391 included within another project.
2392
2393 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2394 slightly modified by me:
2395
2396 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2397 not building pcregrep.
2398
2399 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2400 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2401
2402 25. Forward references, both numeric and by name, in patterns that made use of
2403 duplicate group numbers, could behave incorrectly or give incorrect errors,
2404 because when scanning forward to find the reference group, PCRE was not
2405 taking into account the duplicate group numbers. A pattern such as
2406 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2407
2408 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2409 the feature of a custom pointer more persuasive (as requested by a user).
2410
2411 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2412 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2413 pre-defined.
2414
2415 28. Added support for (*UTF8) at the start of a pattern.
2416
2417 29. Arrange for flags added by the "release type" setting in CMake to be shown
2418 in the configuration summary.
2419
2420
2421 Version 7.8 05-Sep-08
2422 ---------------------
2423
2424 1. Replaced UCP searching code with optimized version as implemented for Ad
2425 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2426 stage table and inline lookup instead of a function, giving speed ups of 2
2427 to 5 times on some simple patterns that I tested. Permission was given to
2428 distribute the MultiStage2.py script that generates the tables (it's not in
2429 the tarball, but is in the Subversion repository).
2430
2431 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2432 scripts.
2433
2434 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2435 a group with a zero qualifier. The result of the study could be incorrect,
2436 or the function might crash, depending on the pattern.
2437
2438 4. Caseless matching was not working for non-ASCII characters in back
2439 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2440 It now works when Unicode Property Support is available.
2441
2442 5. In pcretest, an escape such as \x{de} in the data was always generating
2443 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2444 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2445 truncation.
2446
2447 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2448
2449 7. Added two (int) casts to pcregrep when printing the difference of two
2450 pointers, in case they are 64-bit values.
2451
2452 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2453 test 2 if it fails.
2454
2455 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2456 and a #define of that name to empty if it is not externally set. This is to
2457 allow users of MSVC to set it if necessary.
2458
2459 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2460 the convenience functions in the pcre_get.c source file.
2461
2462 11. An option change at the start of a pattern that had top-level alternatives
2463 could cause overwriting and/or a crash. This command provoked a crash in
2464 some environments:
2465
2466 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2467
2468 This potential security problem was recorded as CVE-2008-2371.
2469
2470 12. For a pattern where the match had to start at the beginning or immediately
2471 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2472 pcre_dfa_exec() could read past the end of the passed subject if there was
2473 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2474 pcretest so that it places the subject at the end of its malloc-ed buffer.
2475
2476 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2477 exec() might read past the end of the data buffer in UTF-8 mode.
2478
2479 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2480 the data contained the byte 0x85 as part of a UTF-8 character within its
2481 first line. This applied both to normal and DFA matching.
2482
2483 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2484 /^[^d]*?$/8 failed to match "abc".
2485
2486 16. Added a missing copyright notice to pcrecpp_internal.h.
2487
2488 17. Make it more clear in the documentation that values returned from
2489 pcre_exec() in ovector are byte offsets, not character counts.
2490
2491 18. Tidied a few places to stop certain compilers from issuing warnings.
2492
2493 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2494 supplied by Stefan Weber. I made a further small update for 7.8 because
2495 there is a change of source arrangements: the pcre_searchfuncs.c module is
2496 replaced by pcre_ucd.c.
2497
2498
2499 Version 7.7 07-May-08
2500 ---------------------
2501
2502 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2503 a string to a long long, pretend we don't even have a long long." This is
2504 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2505
2506 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2507 pre-7.6 versions, which defined a global no_arg variable instead of putting
2508 it in the RE class. (See also #8 below.)
2509
2510 3. Remove a line of dead code, identified by coverity and reported by Nuno
2511 Lopes.
2512
2513 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2514
2515 (1) The include/exclude patterns were being applied to the whole pathnames
2516 of files, instead of just to the final components.
2517
2518 (2) If there was more than one level of directory, the subdirectories were
2519 skipped unless they satisfied the include/exclude conditions. This is
2520 inconsistent with GNU grep (and could even be seen as contrary to the
2521 pcregrep specification - which I improved to make it absolutely clear).
2522 The action now is always to scan all levels of directory, and just
2523 apply the include/exclude patterns to regular files.
2524
2525 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2526 --exclude_dir in the tests to avoid scanning .svn directories.
2527
2528 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2529 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2530 doesn't support NULs in patterns.
2531
2532 7. Added some missing "const"s to declarations of static tables in
2533 pcre_compile.c and pcre_dfa_exec.c.
2534
2535 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2536 caused by fix #2 above. (Subsequently also a second patch to fix the
2537 first patch. And a third patch - this was a messy problem.)
2538
2539 9. Applied Craig's patch to remove the use of push_back().
2540
2541 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2542 matching function regexec().
2543
2544 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2545 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2546 references. PCRE supports relative numbers with this syntax (I don't think
2547 Oniguruma does).
2548
2549 12. Previously, a group with a zero repeat such as (...){0} was completely
2550 omitted from the compiled regex. However, this means that if the group
2551 was called as a subroutine from elsewhere in the pattern, things went wrong
2552 (an internal error was given). Such groups are now left in the compiled
2553 pattern, with a new opcode that causes them to be skipped at execution
2554 time.
2555
2556 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2557 to the way PCRE behaves:
2558
2559 (a) A lone ] character is dis-allowed (Perl treats it as data).
2560
2561 (b) A back reference to an unmatched subpattern matches an empty string
2562 (Perl fails the current match path).
2563
2564 (c) A data ] in a character class must be notated as \] because if the
2565 first data character in a class is ], it defines an empty class. (In
2566 Perl it is not possible to have an empty class.) The empty class []
2567 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2568 The negative empty class [^] matches any one character, independently
2569 of the DOTALL setting.
2570
2571 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2572 non-existent subpattern following a character class starting with ']' and
2573 containing () gave an internal compiling error instead of "reference to
2574 non-existent subpattern". Fortunately, when the pattern did exist, the
2575 compiled code was correct. (When scanning forwards to check for the
2576 existence of the subpattern, it was treating the data ']' as terminating
2577 the class, so got the count wrong. When actually compiling, the reference
2578 was subsequently set up correctly.)
2579
2580 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2581 it was being rejected as not supported by pcre_dfa_exec(), even though
2582 other assertions are supported. I have made pcre_dfa_exec() support
2583 (*FAIL).
2584
2585 16. The implementation of 13c above involved the invention of a new opcode,
2586 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2587 cannot be changed at match time, I realized I could make a small
2588 improvement to matching performance by compiling OP_ALLANY instead of
2589 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2590 on the OP_ANY path.
2591
2592 17. Compiling pcretest on Windows with readline support failed without the
2593 following two fixes: (1) Make the unistd.h include conditional on
2594 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2595
2596 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2597 ncurses library to be included for pcretest when ReadLine support is
2598 requested, but also to allow for it to be overridden. This patch came from
2599 Daniel Bergström.
2600
2601 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2602 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2603 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2604 spotting this.
2605
2606
2607 Version 7.6 28-Jan-08
2608 ---------------------
2609
2610 1. A character class containing a very large number of characters with
2611 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2612 overflow.
2613
2614 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2615 HAVE_LONG_LONG is not defined.
2616
2617 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2618 bring them up to date and include new features. This patch includes:
2619
2620 - Fixed PH's badly added libz and libbz2 support.
2621 - Fixed a problem with static linking.
2622 - Added pcredemo. [But later removed - see 7 below.]
2623 - Fixed dftables problem and added an option.
2624 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2625 HAVE_LONG_LONG.
2626 - Added readline support for pcretest.
2627 - Added an listing of the option settings after cmake has run.
2628
2629 4. A user submitted a patch to Makefile that makes it easy to create
2630 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2631 Makefile.am that cause it to include this special target, without
2632 affecting anything else. Note that the same mingw target plus all
2633 the other distribution libraries and programs are now supported
2634 when configuring with CMake (see 6 below) instead of with
2635 Configure/Make.
2636
2637 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2638 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2639 exported in the Windows port". It has not yet been confirmed that the patch
2640 solves the problem, but it does no harm.
2641
2642 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2643 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2644 with CMake, and also correct the comment about stack recursion.
2645
2646 7. Remove the automatic building of pcredemo from the ./configure system and
2647 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2648 of a program that users should build themselves after PCRE is installed, so
2649 building it automatically is not really right. What is more, it gave
2650 trouble in some build environments.
2651
2652 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2653
2654
2655 Version 7.5 10-Jan-08
2656 ---------------------
2657
2658 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2659 values in parens when parsing an RE using the C++ wrapper."
2660
2661 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2662 Characters greater than 255 were excluded from the class instead of being
2663 included.
2664
2665 3. The same bug as (2) above applied to negated POSIX classes such as
2666 [:^space:].
2667
2668 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2669 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2670 I have changed it.
2671
2672 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2673 first named subpattern) and a construct such as (?&a) would reference the
2674 first named subpattern whose name started with "a" (in other words, the
2675 length check was missing). Both these problems are fixed. "Subpattern name
2676 expected" is now given for (?&) (a zero-length name), and this patch also
2677 makes it give the same error for \k'' (previously it complained that that
2678 was a reference to a non-existent subpattern).
2679
2680 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2681 this is right because (?- can be followed by option settings as well as by
2682 digits. I have, however, made the messages clearer.
2683
2684 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2685 than the number used in the conditional) now cause a compile-time error.
2686 This is actually not compatible with Perl, which accepts such patterns, but
2687 treats the conditional as always being FALSE (as PCRE used to), but it
2688 seems to me that giving a diagnostic is better.
2689
2690 8. Change "alphameric" to the more common word "alphanumeric" in comments
2691 and messages.
2692
2693 9. Fix two occurrences of "backslash" in comments that should have been
2694 "backspace".
2695
2696 10. Remove two redundant lines of code that can never be obeyed (their function
2697 was moved elsewhere).
2698
2699 11. The program that makes PCRE's Unicode character property table had a bug
2700 which caused it to generate incorrect table entries for sequences of
2701 characters that have the same character type, but are in different scripts.
2702 It amalgamated them into a single range, with the script of the first of
2703 them. In other words, some characters were in the wrong script. There were
2704 thirteen such cases, affecting characters in the following ranges:
2705
2706 U+002b0 - U+002c1
2707 U+0060c - U+0060d
2708 U+0061e - U+00612
2709 U+0064b - U+0065e
2710 U+0074d - U+0076d
2711 U+01800 - U+01805
2712 U+01d00 - U+01d77
2713 U+01d9b - U+01dbf
2714 U+0200b - U+0200f
2715 U+030fc - U+030fe
2716 U+03260 - U+0327f
2717 U+0fb46 - U+0fbb1
2718 U+10450 - U+1049d
2719
2720 12. The -o option (show only the matching part of a line) for pcregrep was not
2721 compatible with GNU grep in that, if there was more than one match in a
2722 line, it showed only the first of them. It now behaves in the same way as
2723 GNU grep.
2724
2725 13. If the -o and -v options were combined for pcregrep, it printed a blank
2726 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2727 does the same. The return code can be used to tell if there were any
2728 non-matching lines.
2729
2730 14. Added --file-offsets and --line-offsets to pcregrep.
2731
2732 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2733 infinitely looping recursion. The bug was that positive lookaheads were not
2734 being skipped when checking for a possible empty match (negative lookaheads
2735 and both kinds of lookbehind were skipped).
2736
2737 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2738 inclusion of <windows.h> to before rather than after the definition of
2739 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2740
2741 17. Specifying a possessive quantifier with a specific limit for a Unicode
2742 character property caused pcre_compile() to compile bad code, which led at
2743 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2744 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2745 caused the error; without that there was no problem.
2746
2747 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2748
2749 19. Added --enable-pcretest-libreadline.
2750
2751 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2752 RE::GlobalReplace(). As a result, the number of replacements returned was
2753 double what it should be. I removed one of the increments, but Craig sent a
2754 later patch that removed the other one (the right fix) and added unit tests
2755 that check the return values (which was not done before).
2756
2757 21. Several CMake things:
2758
2759 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2760 the names libpcre and libpcreposix, not just pcre and pcreposix.
2761
2762 (2) The above change means that pcretest and pcregrep are now correctly
2763 linked with the newly-built libraries, not previously installed ones.
2764
2765 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2766
2767 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2768 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2769 UTF-8 newline character). The key issue is that the pattern starts .*;
2770 this means that the match must be either at the beginning, or after a
2771 newline. The bug was in the code for advancing after a failed match and
2772 checking that the new position followed a newline. It was not taking
2773 account of UTF-8 characters correctly.
2774
2775 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2776 character classes. PCRE was not treating the sequence [:...:] as a
2777 character class unless the ... were all letters. Perl, however, seems to
2778 allow any characters between [: and :], though of course it rejects as
2779 unknown any "names" that contain non-letters, because all the known class
2780 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2781 for example, whereas PCRE did not - it did not recognize a POSIX character
2782 class. This seemed a bit dangerous, so the code has been changed to be
2783 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2784 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2785 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2786 Perl does, and where it didn't before.
2787
2788 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2789 Windows environments %n is disabled by default.
2790
2791
2792 Version 7.4 21-Sep-07
2793 ---------------------
2794
2795 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2796 means that a class such as [\s] counted as "explicit reference to CR or
2797 LF". That isn't really right - the whole point of the change was to try to
2798 help when there was an actual mention of one of the two characters. So now
2799 the change happens only if \r or \n (or a literal CR or LF) character is
2800 encountered.
2801
2802 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2803 of both had grown to the point where there were only 3 bits left.
2804 Fortunately, there was spare space in the data structure, and so I have
2805 moved the internal flags into a new 16-bit field to free up more option
2806 bits.
2807
2808 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2809 but did not set the internal JCHANGED flag - either of these is enough to
2810 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2811 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2812 start sets both bits.
2813
2814 4. Added options (at build time, compile time, exec time) to change \R from
2815 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2816
2817 5. doc/pcresyntax.html was missing from the distribution.
2818
2819 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2820 compatibility, even though it is no longer used.
2821
2822 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2823 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2824 windows.h file is present (where different names are used). [This was
2825 reversed later after testing - see 16 below.]
2826
2827 8. Changed all #include <config.h> to #include "config.h". There were also
2828 some further <pcre.h> cases that I changed to "pcre.h".
2829
2830 9. When pcregrep was used with the --colour option, it missed the line ending
2831 sequence off the lines that it output.
2832
2833 10. It was pointed out to me that arrays of string pointers cause lots of
2834 relocations when a shared library is dynamically loaded. A technique of
2835 using a single long string with a table of offsets can drastically reduce
2836 these. I have refactored PCRE in four places to do this. The result is
2837 dramatic:
2838
2839 Originally: 290
2840 After changing UCP table: 187
2841 After changing error message table: 43
2842 After changing table of "verbs" 36
2843 After changing table of Posix names 22
2844
2845 Thanks to the folks working on Gregex for glib for this insight.
2846
2847 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2848 unicode-properties was also set.
2849
2850 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2851
2852 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2853 checked only for CRLF.
2854
2855 14. Added casts to pcretest.c to avoid compiler warnings.
2856
2857 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2858
2859 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2860 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2861 entirely. This removes changes made in 7 above.
2862
2863 17. The CMake files have been updated, and there is now more information about
2864 building with CMake in the NON-UNIX-USE document.
2865
2866
2867 Version 7.3 28-Aug-07
2868 ---------------------
2869
2870 1. In the rejigging of the build system that eventually resulted in 7.1, the
2871 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2872 brackets there is not right, since it causes compilers to look for an
2873 installed pcre.h, not the version that is in the source that is being
2874 compiled (which of course may be different). I have changed it back to:
2875
2876 #include "pcre.h"
2877
2878 I have a vague recollection that the change was concerned with compiling in
2879 different directories, but in the new build system, that is taken care of
2880 by the VPATH setting the Makefile.
2881
2882 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2883 when the subject happened to end in the byte 0x85 (e.g. if the last
2884 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2885 characters but of course it shouldn't be taken as a newline when it is part
2886 of another character. The bug was that, for an unlimited repeat of . in
2887 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2888 characters when looking for a newline.
2889
2890 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2891
2892 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2893 in debug output.
2894
2895 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2896 long printing in the pcrecpp unittest when running under MinGW.
2897
2898 6. ESC_K was left out of the EBCDIC table.
2899
2900 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2901 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2902 limit also applies to "virtual nesting" when a pattern is recursive, and in
2903 this case 1000 isn't so big. I have been able to remove this limit at the
2904 expense of backing off one optimization in certain circumstances. Normally,
2905 when pcre_exec() would call its internal match() function recursively and
2906 immediately return the result unconditionally, it uses a "tail recursion"
2907 feature to save stack. However, when a subpattern that can match an empty
2908 string has an unlimited repetition quantifier, it no longer makes this
2909 optimization. That gives it a stack frame in which to save the data for
2910 checking that an empty string has been matched. Previously this was taken
2911 from the 1000-entry workspace that had been reserved. So now there is no
2912 explicit limit, but more stack is used.
2913
2914 8. Applied Daniel's patches to solve problems with the import/export magic
2915 syntax that is required for Windows, and which was going wrong for the
2916 pcreposix and pcrecpp parts of the library. These were overlooked when this
2917 problem was solved for the main library.
2918
2919 9. There were some crude static tests to avoid integer overflow when computing
2920 the size of patterns that contain repeated groups with explicit upper
2921 limits. As the maximum quantifier is 65535, the maximum group length was
2922 set at 30,000 so that the product of these two numbers did not overflow a
2923 32-bit integer. However, it turns out that people want to use groups that
2924 are longer than 30,000 bytes (though not repeat them that many times).
2925 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2926 made it possible to implement the integer overflow checks in a much more
2927 dynamic way, which I have now done. The artificial limitation on group
2928 length has been removed - we now have only the limit on the total length of
2929 the compiled pattern, which depends on the LINK_SIZE setting.
2930
2931 10. Fixed a bug in the documentation for get/copy named substring when
2932 duplicate names are permitted. If none of the named substrings are set, the
2933 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2934 empty string.
2935
2936 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2937 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2938 because the ] is interpreted as the first data character and the
2939 terminating ] is not found. PCRE has been made compatible with Perl in this
2940 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2941 cause memory overwriting.
2942
2943 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2944 string has been matched (to stop an infinite loop). It was not recognizing
2945 a conditional subpattern that could match an empty string if that
2946 subpattern was within another subpattern. For example, it looped when
2947 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2948 condition was not nested. This bug has been fixed.
2949
2950 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2951 past the start of the subject in the presence of bytes with the top bit
2952 set, for example "\x8aBCD".
2953
2954 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2955 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2956
2957 14. Optimized (?!) to (*FAIL).
2958
2959 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2960 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2961 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2962 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2963 does: it's just the validity check that is more restrictive.
2964
2965 16. Inserted checks for integer overflows during escape sequence (backslash)
2966 processing, and also fixed erroneous offset values for syntax errors during
2967 backslash processing.
2968
2969 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2970 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2971
2972 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2973 caused an overrun.
2974
2975 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2976 something other than just ASCII characters) inside a group that had an
2977 unlimited repeat caused a loop at compile time (while checking to see
2978 whether the group could match an empty string).
2979
2980 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2981 [\P{Any}] did so. (Error in the code for printing property names.)
2982
2983 21. An orphan \E inside a character class could cause a crash.
2984
2985 22. A repeated capturing bracket such as (A)? could cause a wild memory
2986 reference during compilation.
2987
2988 23. There are several functions in pcre_compile() that scan along a compiled
2989 expression for various reasons (e.g. to see if it's fixed length for look
2990 behind). There were bugs in these functions when a repeated \p or \P was
2991 present in the pattern. These operators have additional parameters compared
2992 with \d, etc, and these were not being taken into account when moving along
2993 the compiled data. Specifically:
2994
2995 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2996 length.
2997
2998 (b) An item such as \pL+ within a repeated group could cause crashes or
2999 loops.
3000
3001 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3002 "reference to non-existent subpattern" error.
3003
3004 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3005
3006 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3007 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3008
3009 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3010
3011 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3012 character were causing crashes (broken optimization).
3013
3014 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3015 \p or \P) caused a compile-time loop.
3016
3017 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3018 break. For example, the unstudied pattern [\r\n]A does not match the string
3019 "\r\nA" because change 7.0/46 below moves the current point on by two
3020 characters after failing to match at the start. However, the pattern \nA
3021 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3022 the same is true. There doesn't seem any very clean way out of this, but
3023 what I have chosen to do makes the common cases work: PCRE now takes note
3024 of whether there can be an explicit match for \r or \n anywhere in the
3025 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3026 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3027 pattern has explicit CR or LF references.
3028
3029 29. Added (*CR) etc for changing newline setting at start of pattern.
3030
3031
3032 Version 7.2 19-Jun-07
3033 ---------------------
3034
3035 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3036 which is apparently normally available under Windows.
3037
3038 2. Re-jig the pcregrep tests with different newline settings in an attempt
3039 to make them independent of the local environment's newline setting.
3040
3041 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3042
3043 4. Some of the "internals" tests were previously cut out when the link size
3044 was not 2, because the output contained actual offsets. The recent new
3045 "Z" feature of pcretest means that these can be cut out, making the tests
3046 usable with all link sizes.
3047
3048 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3049 stack recursion. This gives a massive performance boost under BSD, but just
3050 a small improvement under Linux. However, it saves one field in the frame
3051 in all cases.
3052
3053 6. Added more features from the forthcoming Perl 5.10:
3054
3055 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3056 recursion call. It refers to the nth most recently opened parentheses.
3057
3058 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3059 to be opened parentheses.
3060
3061 (c) Conditions that refer to capturing parentheses can be specified
3062 relatively, for example, (?(-2)... or (?(+3)...
3063
3064 (d) \K resets the start of the current match so that everything before
3065 is not part of it.
3066
3067 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3068
3069 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3070 reference syntax.
3071
3072 (g) (?| introduces a group in which the numbering of parentheses in each
3073 alternative starts with the same number.
3074
3075 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3076
3077 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3078 PCRE_INFO_JCHANGED.
3079
3080 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3081 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3082 for detecting groups that can match an empty string.
3083
3084 9. A pattern with a very large number of alternatives (more than several
3085 hundred) was running out of internal workspace during the pre-compile
3086 phase, where pcre_compile() figures out how much memory will be needed. A
3087 bit of new cunning has reduced the workspace needed for groups with
3088 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3089 workspace instead of running out of the 4096 that are available.
3090
3091 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3092
3093 11. Applied patch from Google to remove an optimization that didn't quite work.
3094 The report of the bug said:
3095
3096 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3097 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3098 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3099
3100 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3101 it matched the wrong number of bytes.
3102
3103
3104 Version 7.1 24-Apr-07
3105 ---------------------
3106
3107 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3108 that is more "standard", making use of automake and other Autotools. There
3109 is some re-arrangement of the files and adjustment of comments consequent
3110 on this.
3111
3112 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3113 for recursive directory scanning broke on some systems because the files
3114 are not scanned in any specific order and on different systems the order
3115 was different. A call to "sort" has been inserted into RunGrepTest for the
3116 approprate test as a short-term fix. In the longer term there may be an
3117 alternative.
3118
3119 3. I had an email from Eric Raymond about problems translating some of PCRE's
3120 man pages to HTML (despite the fact that I distribute HTML pages, some
3121 people do their own conversions for various reasons). The problems
3122 concerned the use of low-level troff macros .br and .in. I have therefore
3123 removed all such uses from the man pages (some were redundant, some could
3124 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3125 HTML has been updated to handle .nf/.fi and to complain if it encounters
3126 .br or .in.
3127
3128 4. Updated comments in configure.ac that get placed in config.h.in and also
3129 arranged for config.h to be included in the distribution, with the name
3130 config.h.generic, for the benefit of those who have to compile without
3131 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3132
3133 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3134 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3135 makevp.bat for the current PCRE, using the additional files
3136 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3137
3138 6. A Windows user reported a minor discrepancy with test 2, which turned out
3139 to be caused by a trailing space on an input line that had got lost in his
3140 copy. The trailing space was an accident, so I've just removed it.
3141
3142 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3143 that is needed.
3144
3145 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3146 as "const" (a) because they are and (b) because it helps the PHP
3147 maintainers who have recently made a script to detect big data structures
3148 in the php code that should be moved to the .rodata section. I remembered
3149 to update Builducptable as well, so it won't revert if ucptable.h is ever
3150 re-created.
3151
3152 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3153 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3154 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3155 support is not required. This saves 1.5-2K of code, which is important in
3156 some applications.
3157
3158 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3159 so as not to refer to the tables, even though these functions will never be
3160 called when UTF-8 support is disabled. Otherwise there are problems with a
3161 shared library.
3162
3163 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3164
3165 (a) It was defining its arguments as char * instead of void *.
3166
3167 (b) It was assuming that all moves were upwards in memory; this was true
3168 a long time ago when I wrote it, but is no longer the case.
3169
3170 The emulated memove() is provided for those environments that have neither
3171 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3172 is clearly not the case, as these two bugs were recently reported.
3173
3174 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3175 and Detrail to create the HTML documentation, the .txt form of the man
3176 pages, and it removes trailing spaces from listed files. It also creates
3177 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3178 case, it wraps all the #defines with #ifndefs. This script should be run
3179 before "make dist".
3180
3181 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3182 with Unicode property support.
3183
3184 (a) For a maximizing quantifier, if the two different cases of the
3185 character were of different lengths in their UTF-8 codings (there are
3186 some cases like this - I found 11), and the matching function had to
3187 back up over a mixture of the two cases, it incorrectly assumed they
3188 were both the same length.
3189
3190 (b) When PCRE was configured to use the heap rather than the stack for
3191 recursion during matching, it was not correctly preserving the data for
3192 the other case of a UTF-8 character when checking ahead for a match
3193 while processing a minimizing repeat. If the check also involved
3194 matching a wide character, but failed, corruption could cause an
3195 erroneous result when trying to check for a repeat of the original
3196 character.
3197
3198 13. Some tidying changes to the testing mechanism:
3199
3200 (a) The RunTest script now detects the internal link size and whether there
3201 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3202 values substituted by "configure". (The RunGrepTest script already did
3203 this for UTF-8.) The configure.ac script no longer substitutes the
3204 relevant variables.
3205
3206 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3207 with length and offset values. This means that the output is different
3208 for different internal link sizes. Test 2 is skipped for link sizes
3209 other than 2 because of this, bypassing the problem. Unfortunately,
3210 there was also a test in test 3 (the locale tests) that used /B and
3211 failed for link sizes other than 2. Rather than cut the whole test out,
3212 I have added a new /Z option to pcretest that replaces the length and
3213 offset values with spaces. This is now used to make test 3 independent
3214 of link size. (Test 2 will be tidied up later.)
3215
3216 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3217 segmentation fault instead of returning the appropriate error message.
3218
3219 15. In multiline mode when the newline sequence was set to "any", the pattern
3220 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3221 This doesn't seem right; it now treats the CRLF combination as the line
3222 ending, and so does not match in that case. It's only a pattern such as ^$
3223 that would hit this one: something like ^ABC$ would have failed after \r
3224 and then tried again after \r\n.
3225
3226 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3227 in an attempt to make files that differ only in their line terminators
3228 compare equal. This works on Linux.
3229
3230 17. Under certain error circumstances pcregrep might try to free random memory
3231 as it exited. This is now fixed, thanks to valgrind.
3232
3233 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3234 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3235 was because its rules for how to advance for /g after matching an empty
3236 string at the end of a line did not allow for this case. They now check for
3237 it specially.
3238
3239 20. pcretest is supposed to handle patterns and data of any length, by
3240 extending its buffers when necessary. It was getting this wrong when the
3241 buffer for a data line had to be extended.
3242
3243 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3244 CRLF as a newline sequence.
3245
3246 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3247 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3248 I have nevertheless tidied it up.
3249
3250 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3251
3252 24. Added a man page for pcre-config.
3253
3254
3255 Version 7.0 19-Dec-06
3256 ---------------------
3257
3258 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3259 moving to gcc 4.1.1.
3260
3261 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3262 sys/time.h, which is documented as needed for this function. It doesn't
3263 seem to matter on Linux, but it showed up on some releases of OS X.
3264
3265 3. It seems that there are systems where bytes whose values are greater than
3266 127 match isprint() in the "C" locale. The "C" locale should be the
3267 default when a C program starts up. In most systems, only ASCII printing
3268 characters match isprint(). This difference caused the output from pcretest
3269 to vary, making some of the tests fail. I have changed pcretest so that:
3270
3271 (a) When it is outputting text in the compiled version of a pattern, bytes
3272 other than 32-126 are always shown as hex escapes.
3273
3274 (b) When it is outputting text that is a matched part of a subject string,
3275 it does the same, unless a different locale has been set for the match
3276 (using the /L modifier). In this case, it uses isprint() to decide.
3277
3278 4. Fixed a major bug that caused incorrect computation of the amount of memory
3279 required for a compiled pattern when options that changed within the
3280 pattern affected the logic of the preliminary scan that determines the
3281 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3282 that the computed length was too small. The symptoms of this bug were
3283 either the PCRE error "internal error: code overflow" from pcre_compile(),
3284 or a glibc crash with a message such as "pcretest: free(): invalid next
3285 size (fast)". Examples of patterns that provoked this bug (shown in
3286 pcretest format) are:
3287
3288 /(?-x: )/x
3289 /(?x)(?-x: \s*#\s*)/
3290 /((?i)[\x{c0}])/8
3291 /(?i:[\x{c0}])/8
3292
3293 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3294 is now done differently.
3295
3296 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3297 wrapper classes; (b) implement a new function in the C++ scanner that is
3298 more efficient than the old way of doing things because it avoids levels of
3299 recursion in the regex matching; (c) add a paragraph to the documentation
3300 for the FullMatch() function.
3301
3302 6. The escape sequence \n was being treated as whatever was defined as
3303 "newline". Not only was this contrary to the documentation, which states
3304 that \n is character 10 (hex 0A), but it also went horribly wrong when
3305 "newline" was defined as CRLF. This has been fixed.
3306
3307 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3308 was being set to -1 for the "end of line" case (supposedly a value that no
3309 character can have). Though this value is never used (the check for end of
3310 line is "zero bytes in current character"), it caused compiler complaints.
3311 I've changed it to 0xffffffff.
3312
3313 8. In pcre_version.c, the version string was being built by a sequence of
3314 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3315 string (as it is for production releases) called a macro with an empty
3316 argument. The C standard says the result of this is undefined. The gcc
3317 compiler treats it as an empty string (which was what was wanted) but it is
3318 reported that Visual C gives an error. The source has been hacked around to
3319 avoid this problem.
3320
3321 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3322 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3323 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3324 of them did).
3325
3326 10. Originally, pcretest opened its input and output without "b"; then I was
3327 told that "b" was needed in some environments, so it was added for release
3328 5.0 to both the input and output. (It makes no difference on Unix-like
3329 systems.) Later I was told that it is wrong for the input on Windows. I've
3330 now abstracted the modes into two macros, to make it easier to fiddle with
3331 them, and removed "b" from the input mode under Windows.
3332
3333 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3334
3335 12. Added -help and --help to pcretest as an official way of being reminded
3336 of the options.
3337
3338 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3339 and pcrecpp.cc because they annoy compilers at high warning levels.
3340
3341 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3342
3343 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3344 scripts are not C programs :-) and which was not noticed because it works
3345 on Linux.
3346
3347 16. pcretest is supposed to handle any length of pattern and data line (as one
3348 line or as a continued sequence of lines) by extending its input buffer if
3349 necessary. This feature was broken for very long pattern lines, leading to
3350 a string of junk being passed to pcre_compile() if the pattern was longer
3351 than about 50K.
3352
3353 17. I have done a major re-factoring of the way pcre_compile() computes the
3354 amount of memory needed for a compiled pattern. Previously, there was code
3355 that made a preliminary scan of the pattern in order to do this. That was
3356 OK when PCRE was new, but as the facilities have expanded, it has become
3357 harder and harder to keep it in step with the real compile phase, and there
3358 have been a number of bugs (see for example, 4 above). I have now found a
3359 cunning way of running the real compile function in a "fake" mode that
3360 enables it to compute how much memory it would need, while actually only
3361 ever using a few hundred bytes of working memory and without too many
3362 tests of the mode. This should make future maintenance and development
3363 easier. A side effect of this work is that the limit of 200 on the nesting
3364 depth of parentheses has been removed (though this was never a serious
3365 limitation, I suspect). However, there is a downside: pcre_compile() now
3366 runs more slowly than before (30% or more, depending on the pattern). I
3367 hope this isn't a big issue. There is no effect on runtime performance.
3368
3369 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3370 newline (only possible for the last line of a file) and it was a
3371 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3372
3373 19. Added additional timing features to pcretest. (1) The -tm option now times
3374 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3375 separate command line item, by a number that specifies the number of
3376 repeats to use when timing. The default is 50000; this gives better
3377 precision, but takes uncomfortably long for very large patterns.
3378
3379 20. Extended pcre_study() to be more clever in cases where a branch of a
3380 subpattern has no definite first character. For example, (a*|b*)[cd] would
3381 previously give no result from pcre_study(). Now it recognizes that the
3382 first character must be a, b, c, or d.
3383
3384 21. There was an incorrect error "recursive call could loop indefinitely" if
3385 a subpattern (or the entire pattern) that was being tested for matching an
3386 empty string contained only one non-empty item after a nested subpattern.
3387 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3388 incorrectly, because the \d was being skipped in the check.
3389
3390 22. The pcretest program now has a new pattern option /B and a command line
3391 option -b, which is equivalent to adding /B to every pattern. This causes
3392 it to show the compiled bytecode, without the additional information that
3393 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3394 is the same as /B/I).
3395
3396 23. A new optimization is now able automatically to treat some sequences such
3397 as a*b as a*+b. More specifically, if something simple (such as a character
3398 or a simple class like \d) has an unlimited quantifier, and is followed by
3399 something that cannot possibly match the quantified thing, the quantifier
3400 is automatically "possessified".
3401
3402 24. A recursive reference to a subpattern whose number was greater than 39
3403 went wrong under certain circumstances in UTF-8 mode. This bug could also
3404 have affected the operation of pcre_study().
3405
3406 25. Realized that a little bit of performance could be had by replacing
3407 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3408
3409 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3410
3411 27. Possessive quantifiers such as a++ were previously implemented by turning
3412 them into atomic groups such as ($>a+). Now they have their own opcodes,
3413 which improves performance. This includes the automatically created ones
3414 from 23 above.
3415
3416 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3417 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3418 the first matched character to be a colon. This applied both to named and
3419 numbered groups.
3420
3421 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3422
3423 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3424 building PCRE on Windows easier, so I have included it in the distribution.
3425
3426 31. There is now a check in pcretest against a ridiculously large number being
3427 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3428 loop, the loop is abandoned.
3429
3430 32. Forward references to subpatterns in conditions such as (?(2)...) where
3431 subpattern 2 is defined later cause pcre_compile() to search forwards in
3432 the pattern for the relevant set of parentheses. This search went wrong
3433 when there were unescaped parentheses in a character class, parentheses
3434 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3435
3436 33. "Subroutine" calls and backreferences were previously restricted to
3437 referencing subpatterns earlier in the regex. This restriction has now
3438 been removed.
3439
3440 34. Added a number of extra features that are going to be in Perl 5.10. On the
3441 whole, these are just syntactic alternatives for features that PCRE had
3442 previously implemented using the Python syntax or my own invention. The
3443 other formats are all retained for compatibility.
3444
3445 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3446 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3447 also .NET compatible.
3448
3449 (b) A recursion or subroutine call to a named group can now be defined as
3450 (?&name) as well as (?P>name).
3451
3452 (c) A backreference to a named group can now be defined as \k<name> or
3453 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3454 5.10, are also .NET compatible.
3455
3456 (d) A conditional reference to a named group can now use the syntax
3457 (?(<name>) or (?('name') as well as (?(name).
3458
3459 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3460 groups (named and numbered) that are never evaluated inline, but can be
3461 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3462 is always false. There may be only one alternative in such a group.
3463
3464 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3465 as the simple (?(R). The condition is true only if the most recent
3466 recursion is that of the given number or name. It does not search out
3467 through the entire recursion stack.
3468
3469 (g) The escape \gN or \g{N} has been added, where N is a positive or
3470 negative number, specifying an absolute or relative reference.
3471
3472 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3473 some "unreachable code" warnings.
3474
3475 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3476 things, this adds five new scripts.
3477
3478 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3479 There were also incompatibilities regarding the handling of \Q..\E inside
3480 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3481 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3482
3483 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3484 matches an empty string, and forcibly breaks the loop. There were bugs in
3485 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3486 against aaaa the result was just "a" rather than "aaaa", for example. Two
3487 separate and independent bugs (that affected different cases) have been
3488 fixed.
3489
3490 39. Refactored the code to abolish the use of different opcodes for small
3491 capturing bracket numbers. This is a tidy that I avoided doing when I
3492 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3493 The new approach is not only tidier, it makes it possible to reduce the
3494 memory needed to fix the previous bug (38).
3495
3496 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3497 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3498 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3499 mode.
3500
3501 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3502 report.
3503
3504 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3505 copy construction and assignment in the C++ wrapper.
3506
3507 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3508 couple of bugs that could have given wrong results in the "--newline=crlf"
3509 case.
3510
3511 44. Added a number of casts and did some reorganization of signed/unsigned int
3512 variables following suggestions from Dair Grant. Also renamed the variable
3513 "this" as "item" because it is a C++ keyword.
3514
3515 45. Arranged for dftables to add
3516
3517 #include "pcre_internal.h"
3518
3519 to pcre_chartables.c because without it, gcc 4.x may remove the array
3520 definition from the final binary if PCRE is built into a static library and
3521 dead code stripping is activated.
3522
3523 46. For an unanchored pattern, if a match attempt fails at the start of a
3524 newline sequence, and the newline setting is CRLF or ANY, and the next two
3525 characters are CRLF, advance by two characters instead of one.
3526
3527
3528 Version 6.7 04-Jul-06
3529 ---------------------
3530
3531 1. In order to handle tests when input lines are enormously long, pcretest has
3532 been re-factored so that it automatically extends its buffers when
3533 necessary. The code is crude, but this _is_ just a test program. The
3534 default size has been increased from 32K to 50K.
3535
3536 2. The code in pcre_study() was using the value of the re argument before
3537 testing it for NULL. (Of course, in any sensible call of the function, it
3538 won't be NULL.)
3539
3540 3. The memmove() emulation function in pcre_internal.h, which is used on
3541 systems that lack both memmove() and bcopy() - that is, hardly ever -
3542 was missing a "static" storage class specifier.
3543
3544 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3545 containing an extended class (one that cannot be represented by a bitmap
3546 because it contains high-valued characters or Unicode property items, e.g.
3547 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3548 pattern, but PCRE should not loop if you do not (it no longer does).
3549 [Detail: two cases were found: (a) a repeated subpattern containing an
3550 extended class; (b) a recursive reference to a subpattern that followed a
3551 previous extended class. It wasn't skipping over the extended class
3552 correctly when UTF-8 mode was not set.]
3553
3554 5. A negated single-character class was not being recognized as fixed-length
3555 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3556 compile error "lookbehind assertion is not fixed length".
3557
3558 6. The RunPerlTest auxiliary script was showing an unexpected difference
3559 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3560 write a Perl script that can interpret lines of an input file either as
3561 byte characters or as UTF-8, which is what "perltest" was being required to
3562 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3563 can't do is switch easily at run time between having the "use utf8;" pragma
3564 or not. In the end, I fudged it by using the RunPerlTest script to insert
3565 "use utf8;" explicitly for the UTF-8 tests.
3566
3567 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3568 the end of the subject string, contrary to the documentation and to what
3569 Perl does. This was true of both matching functions. Now it matches only at
3570 the start of the subject and immediately after *internal* newlines.
3571
3572 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3573 a pointer to an int instead of a pointer to an unsigned long int. This
3574 caused problems on 64-bit systems.
3575
3576 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3577 instance of the 'standard' template library not being so standard".
3578
3579 10. There was no check on the number of named subpatterns nor the maximum
3580 length of a subpattern name. The product of these values is used to compute
3581 the size of the memory block for a compiled pattern. By supplying a very
3582 long subpattern name and a large number of named subpatterns, the size
3583 computation could be caused to overflow. This is now prevented by limiting
3584 the length of names to 32 characters, and the number of named subpatterns
3585 to 10,000.
3586
3587 11. Subpatterns that are repeated with specific counts have to be replicated in
3588 the compiled pattern. The size of memory for this was computed from the
3589 length of the subpattern and the repeat count. The latter is limited to
3590 65535, but there was no limit on the former, meaning that integer overflow
3591 could in principle occur. The compiled length of a repeated subpattern is
3592 now limited to 30,000 bytes in order to prevent this.
3593
3594 12. Added the optional facility to have named substrings with the same name.
3595
3596 13. Added the ability to use a named substring as a condition, using the
3597 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3598 are numbers (not recommended). Forward references are permitted.
3599
3600 14. Added forward references in named backreferences (if you see what I mean).
3601
3602 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3603 pattern could run off the end of the subject. For example, the pattern
3604 "(?s)(.{1,5})"8 did this with the subject "ab".
3605
3606 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3607 PCRE_CASELESS was set when matching characters that were quantified with ?
3608 or *.
3609
3610 17. A character class other than a single negated character that had a minimum
3611 but no maximum quantifier - for example [ab]{6,} - was not handled
3612 correctly by pce_dfa_exec(). It would match only one character.
3613
3614 18. A valid (though odd) pattern that looked like a POSIX character
3615 class but used an invalid character after [ (for example [[,abc,]]) caused
3616 pcre_compile() to give the error "Failed: internal error: code overflow" or
3617 in some cases to crash with a glibc free() error. This could even happen if
3618 the pattern terminated after [[ but there just happened to be a sequence of
3619 letters, a binary zero, and a closing ] in the memory that followed.
3620
3621 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3622 over the years. Originally (before any Unicode support), just the bottom 8
3623 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3624 output from "man perlunicode" includes this:
3625
3626 The regular expression compiler produces polymorphic opcodes. That
3627 is, the pattern adapts to the data and automatically switches to
3628 the Unicode character scheme when presented with Unicode data--or
3629 instead uses a traditional byte scheme when presented with byte
3630 data.
3631
3632 Sadly, a wide octal escape does not cause a switch, and in a string with
3633 no other multibyte characters, these octal escapes are treated as before.
3634 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3635 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3636 Unicode string.
3637
3638 I have not perpetrated such confusion in PCRE. Up till now, it took just
3639 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3640 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3641 translate to the appropriate multibyte character.
3642
3643 29. Applied some refactoring to reduce the number of warnings from Microsoft
3644 and Borland compilers. This has included removing the fudge introduced
3645 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3646 a warning about an unused variable.
3647
3648 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3649 characters since release 4.0, because Perl (from release 5.004) does not.
3650 [Or at least, is documented not to: some releases seem to be in conflict
3651 with the documentation.] However, when a pattern was studied with
3652 pcre_study() and all its branches started with \s, PCRE still included VT
3653 as a possible starting character. Of course, this did no harm; it just
3654 caused an unnecessary match attempt.
3655
3656 22. Removed a now-redundant internal flag bit that recorded the fact that case
3657 dependency changed within the pattern. This was once needed for "required
3658 byte" processing, but is no longer used. This recovers a now-scarce options
3659 bit. Also moved the least significant internal flag bit to the most-
3660 significant bit of the word, which was not previously used (hangover from
3661 the days when it was an int rather than a uint) to free up another bit for
3662 the future.
3663
3664 23. Added support for CRLF line endings as well as CR and LF. As well as the
3665 default being selectable at build time, it can now be changed at runtime
3666 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3667 specify that it is scanning data with non-default line endings.
3668
3669 24. Changed the definition of CXXLINK to make it agree with the definition of
3670 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3671
3672 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3673 recursions. This makes a big different to stack usage for some patterns.
3674
3675 26. If a subpattern containing a named recursion or subroutine reference such
3676 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3677 the space required for the compiled pattern went wrong and gave too small a
3678 value. Depending on the environment, this could lead to "Failed: internal
3679 error: code overflow at offset 49" or "glibc detected double free or
3680 corruption" errors.
3681
3682 27. Applied patches from Google (a) to support the new newline modes and (b) to
3683 advance over multibyte UTF-8 characters in GlobalReplace.
3684
3685 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3686 difference for some implementation of PCRE in some Windows version.
3687
3688 29. Added some extra testing facilities to pcretest:
3689
3690 \q<number> in a data line sets the "match limit" value
3691 \Q<number> in a data line sets the "match recursion limt" value
3692 -S <number> sets the stack size, where <number> is in megabytes
3693
3694 The -S option isn't available for Windows.
3695
3696
3697 Version 6.6 06-Feb-06
3698 ---------------------
3699
3700 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3701 in pcreposix.h. I have copied the definition from pcre.h.
3702
3703 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3704 because pcre.h is no longer a built file.
3705
3706 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3707 not normally included in the compiled code.
3708
3709
3710 Version 6.5 01-Feb-06
3711 ---------------------
3712
3713 1. When using the partial match feature with pcre_dfa_exec(), it was not
3714 anchoring the second and subsequent partial matches at the new starting
3715 point. This could lead to incorrect results. For example, with the pattern
3716 /1234/, partially matching against "123" and then "a4" gave a match.
3717
3718 2. Changes to pcregrep:
3719
3720 (a) All non-match returns from pcre_exec() were being treated as failures
3721 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3722 error message is output. Some extra information is given for the
3723 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3724 probably the only errors that are likely to be caused by users (by
3725 specifying a regex that has nested indefinite repeats, for instance).
3726 If there are more than 20 of these errors, pcregrep is abandoned.
3727
3728 (b) A binary zero was treated as data while matching, but terminated the
3729 output line if it was written out. This has been fixed: binary zeroes
3730 are now no different to any other data bytes.
3731
3732 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3733 used to set a locale for matching. The --locale=xxxx long option has
3734 been added (no short equivalent) to specify a locale explicitly on the
3735 pcregrep command, overriding the environment variables.
3736
3737 (d) When -B was used with -n, some line numbers in the output were one less
3738 than they should have been.
3739
3740 (e) Added the -o (--only-matching) option.
3741
3742 (f) If -A or -C was used with -c (count only), some lines of context were
3743 accidentally printed for the final match.
3744
3745 (g) Added the -H (--with-filename) option.
3746
3747 (h) The combination of options -rh failed to suppress file names for files
3748 that were found from directory arguments.
3749
3750 (i) Added the -D (--devices) and -d (--directories) options.
3751
3752 (j) Added the -F (--fixed-strings) option.
3753
3754 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3755
3756 (l) Added the --colo(u)r option.
3757
3758 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3759 is not present by default.
3760
3761 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3762 items such as (?R) or (?1), when the recursion could match a number of
3763 alternatives. If it matched one of the alternatives, but subsequently,
3764 outside the recursion, there was a failure, the code tried to back up into
3765 the recursion. However, because of the way PCRE is implemented, this is not
3766 possible, and the result was an incorrect result from the match.
3767
3768 In order to prevent this happening, the specification of recursion has
3769 been changed so that all such subpatterns are automatically treated as
3770 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3771
3772 4. I had overlooked the fact that, in some locales, there are characters for
3773 which isalpha() is true but neither isupper() nor islower() are true. In
3774 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3775 and ordfeminine) are like this. This affected the treatment of \w and \W
3776 when they appeared in character classes, but not when they appeared outside
3777 a character class. The bit map for "word" characters is now created
3778 separately from the results of isalnum() instead of just taking it from the
3779 upper, lower, and digit maps. (Plus the underscore character, of course.)
3780
3781 5. The above bug also affected the handling of POSIX character classes such as
3782 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3783 permanent tables. Instead, the bit maps for such a class were previously
3784 created as the appropriate unions of the upper, lower, and digit bitmaps.
3785 Now they are created by subtraction from the [[:word:]] class, which has
3786 its own bitmap.
3787
3788 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3789 It is created by subtracting the vertical space characters (\x09, \x0a,
3790 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3791 subtraction was done in the overall bitmap for a character class, meaning
3792 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3793 be recognized. This bug has been fixed.
3794
3795 7. Patches from the folks at Google:
3796
3797 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3798 real life, but is still worth protecting against".
3799
3800 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3801 regular expressions".
3802
3803 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3804 have it.
3805
3806 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3807 "configure" and the latter not, in order to fix a problem somebody had
3808 with compiling the Arg class on HP-UX.
3809
3810 (e) Improve the error-handling of the C++ wrapper a little bit.
3811
3812 (f) New tests for checking recursion limiting.
3813
3814 8. The pcre_memmove() function, which is used only if the environment does not
3815 have a standard memmove() function (and is therefore rarely compiled),
3816 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3817 returning a result (though PCRE never actually uses the result).
3818
3819 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3820 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3821 returned instead of calling malloc() with an overflowing number that would
3822 most likely cause subsequent chaos.
3823
3824 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3825
3826 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3827 with this option is matched, the nmatch and pmatch options of regexec() are
3828 ignored.
3829
3830 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3831 provided in case anyone wants to the the POSIX interface with UTF-8
3832 strings.
3833
3834 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3835 C++ linking (needed for some HP-UX environments).
3836
3837 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3838 (unused parameter) and in the pcre_printint() function (omitted "default"
3839 switch label when the default is to do nothing).
3840
3841 15. Added some code to make it possible, when PCRE is compiled as a C++
3842 library, to replace subject pointers for pcre_exec() with a smart pointer
3843 class, thus making it possible to process discontinuous strings.
3844
3845 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3846 much the same function. They were added by different people who were trying
3847 to make PCRE easy to compile on non-Unix systems. It has been suggested
3848 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3849 for compiling on Windows systems. I have therefore replaced it with
3850 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3851 defaults to "extern" for C or "extern C" for C++, which works fine on
3852 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3853 SCOPE with something explicit in config.h. In addition:
3854
3855 (a) pcreposix.h still had just "extern" instead of either of these macros;
3856 I have replaced it with PCRE_DATA_SCOPE.
3857
3858 (b) Functions such as _pcre_xclass(), which are internal to the library,
3859 but external in the C sense, all had PCRE_EXPORT in their definitions.
3860 This is apparently wrong for the Windows case, so I have removed it.
3861 (It makes no difference on Unix-like systems.)
3862
3863 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3864 of recursive calls to match(). This is different to MATCH_LIMIT because
3865 that limits the total number of calls to match(), not all of which increase
3866 the depth of recursion. Limiting the recursion depth limits the amount of
3867 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3868 when PCRE is compiled, and changed at run time. A patch from Google adds
3869 this functionality to the C++ interface.
3870
3871 18. Changes to the handling of Unicode character properties:
3872
3873 (a) Updated the table to Unicode 4.1.0.
3874
3875 (b) Recognize characters that are not in the table as "Cn" (undefined).
3876
3877 (c) I revised the way the table is implemented to a much improved format
3878 which includes recognition of ranges. It now supports the ranges that
3879 are defined in UnicodeData.txt, and it also amalgamates other
3880 characters into ranges. This has reduced the number of entries in the
3881 table from around 16,000 to around 3,000, thus reducing its size
3882 considerably. I realized I did not need to use a tree structure after
3883 all - a binary chop search is just as efficient. Having reduced the
3884 number of entries, I extended their size from 6 bytes to 8 bytes to
3885 allow for more data.
3886
3887 (d) Added support for Unicode script names via properties such as \p{Han}.
3888
3889 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3890 matching that character.
3891
3892 20. When matching a repeated Unicode property with a minimum greater than zero,
3893 (for example \pL{2,}), PCRE could look past the end of the subject if it
3894 reached it while seeking the minimum number of characters. This could
3895 happen only if some of the characters were more than one byte long, because
3896 there is a check for at least the minimum number of bytes.
3897
3898 21. Refactored the implementation of \p and \P so as to be more general, to
3899 allow for more different types of property in future. This has changed the
3900 compiled form incompatibly. Anybody with saved compiled patterns that use
3901 \p or \P will have to recompile them.
3902
3903 22. Added "Any" and "L&" to the supported property types.
3904
3905 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3906 but give a compile time error if the value is greater than 0xff.
3907
3908 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3909 accidentally not being installed or uninstalled.
3910
3911 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3912 made were to insert the current release number. This seemed silly, because
3913 it made things harder for people building PCRE on systems that don't run
3914 "configure". I have turned pcre.h into a distributed file, no longer built
3915 by "configure", with the version identification directly included. There is
3916 no longer a pcre.h.in file.
3917
3918 However, this change necessitated a change to the pcre-config script as
3919 well. It is built from pcre-config.in, and one of the substitutions was the
3920 release number. I have updated configure.ac so that ./configure now finds
3921 the release number by grepping pcre.h.
3922
3923 26. Added the ability to run the tests under valgrind.
3924
3925
3926 Version 6.4 05-Sep-05
3927 ---------------------
3928
3929 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3930 "--" to be printed when multiple files were scanned, even when none of the
3931 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3932 consider it to be a bug, and have restored the previous behaviour.
3933
3934 2. A couple of code tidies to get rid of compiler warnings.
3935
3936 3. The pcretest program used to cheat by referring to symbols in the library
3937 whose names begin with _pcre_. These are internal symbols that are not
3938 really supposed to be visible externally, and in some environments it is
3939 possible to suppress them. The cheating is now confined to including
3940 certain files from the library's source, which is a bit cleaner.
3941
3942 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3943 file's purpose clearer.
3944
3945 5. Reorganized pcre_ucp_findchar().
3946
3947
3948 Version 6.3 15-Aug-05
3949 ---------------------
3950
3951 1. The file libpcre.pc.in did not have general read permission in the tarball.
3952
3953 2. There were some problems when building without C++ support:
3954
3955 (a) If C++ support was not built, "make install" and "make test" still
3956 tried to test it.
3957
3958 (b) There were problems when the value of CXX was explicitly set. Some
3959 changes have been made to try to fix these, and ...
3960
3961 (c) --disable-cpp can now be used to explicitly disable C++ support.
3962
3963 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3964 backslash in a target when C++ was disabled. This confuses some
3965 versions of "make", apparently. Using an intermediate variable solves
3966 this. (Same for CPP_LOBJ.)
3967
3968 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3969 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3970 necessary on certain architectures.
3971
3972 4. Added a setting of -export-symbols-regex to the link command to remove
3973 those symbols that are exported in the C sense, but actually are local
3974 within the library, and not documented. Their names all begin with
3975 "_pcre_". This is not a perfect job, because (a) we have to except some
3976 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3977 available (and never for static libraries). I have made a note to try to
3978 find a way round (a) in the future.
3979
3980
3981 Version 6.2 01-Aug-05
3982 ---------------------
3983
3984 1. There was no test for integer overflow of quantifier values. A construction
3985 such as {1111111111111111} would give undefined results. What is worse, if
3986 a minimum quantifier for a parenthesized subpattern overflowed and became
3987 negative, the calculation of the memory size went wrong. This could have
3988 led to memory overwriting.
3989
3990 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3991
3992 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3993 operating environments where this matters.
3994
3995 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3996 PCRE options from within the C++ wrapper.
3997
3998 5. Named capturing subpatterns were not being correctly counted when a pattern
3999 was compiled. This caused two problems: (a) If there were more than 100
4000 such subpatterns, the calculation of the memory needed for the whole
4001 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4002 back references of the form \12, where the number was greater than 9, were
4003 not recognized as back references, even though there were sufficient
4004 previous subpatterns.
4005
4006 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4007 versions of gcc, e.g. 2.95.4.
4008
4009
4010 Version 6.1 21-Jun-05
4011 ---------------------
4012
4013 1. There was one reference to the variable "posix" in pcretest.c that was not
4014 surrounded by "#if !defined NOPOSIX".
4015
4016 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4017 the cross-check on the old pcre_info() function, for the benefit of the
4018 cut-down version of PCRE that is currently imported into Exim.
4019
4020 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4021 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4022 patterns that start (?i) but I don't think that matters. The use of (?i) is
4023 just an example; this all applies to the other options as well.
4024
4025 4. Since libtool seems to echo the compile commands it is issuing, the output
4026 from "make" can be reduced a bit by putting "@" in front of each libtool
4027 compile command.
4028
4029 5. Patch from the folks at Google for configure.in to be a bit more thorough
4030 in checking for a suitable C++ installation before trying to compile the
4031 C++ stuff. This should fix a reported problem when a compiler was present,
4032 but no suitable headers.
4033
4034 6. The man pages all had just "PCRE" as their title. I have changed them to
4035 be the relevant file name. I have also arranged that these names are
4036 retained in the file doc/pcre.txt, which is a concatenation in text format
4037 of all the man pages except the little individual ones for each function.
4038
4039 7. The NON-UNIX-USE file had not been updated for the different set of source
4040 files that come with release 6. I also added a few comments about the C++
4041 wrapper.
4042
4043
4044 Version 6.0 07-Jun-05
4045 ---------------------
4046
4047 1. Some minor internal re-organization to help with my DFA experiments.
4048
4049 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4050 didn't matter for the library itself when fully configured, but did matter
4051 when compiling without UCP support, or within Exim, where the ucp files are
4052 not imported.
4053
4054 3. Refactoring of the library code to split up the various functions into
4055 different source modules. The addition of the new DFA matching code (see
4056 below) to a single monolithic source would have made it really too
4057 unwieldy, quite apart from causing all the code to be include in a
4058 statically linked application, when only some functions are used. This is
4059 relevant even without the DFA addition now that patterns can be compiled in
4060 one application and matched in another.
4061
4062 The downside of splitting up is that there have to be some external
4063 functions and data tables that are used internally in different modules of
4064 the library but which are not part of the API. These have all had their
4065 names changed to start with "_pcre_" so that they are unlikely to clash
4066 with other external names.
4067
4068 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4069 a different (DFA) algorithm. Although it is slower than the original
4070 function, it does have some advantages for certain types of matching
4071 problem.
4072
4073 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4074 including restarting after a partial match.
4075
4076 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4077 defined when compiling for Windows was sent to me. I have put it into the
4078 code, though I have no means of testing or verifying it.
4079
4080 7. Added the pcre_refcount() auxiliary function.
4081
4082 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4083 match before or at the first newline in the subject string. In pcretest,
4084 the /f option on a pattern can be used to set this.
4085
4086 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4087 would behave wrongly. This has been present in PCRE since release 4.0.
4088
4089 10. A number of changes to the pcregrep command:
4090
4091 (a) Refactored how -x works; insert ^(...)$ instead of setting
4092 PCRE_ANCHORED and checking the length, in preparation for adding
4093 something similar for -w.
4094
4095 (b) Added the -w (match as a word) option.
4096
4097 (c) Refactored the way lines are read and buffered so as to have more
4098 than one at a time available.
4099
4100 (d) Implemented a pcregrep test script.
4101
4102 (e) Added the -M (multiline match) option. This allows patterns to match
4103 over several lines of the subject. The buffering ensures that at least
4104 8K, or the rest of the document (whichever is the shorter) is available
4105 for matching (and similarly the previous 8K for lookbehind assertions).
4106
4107 (f) Changed the --help output so that it now says
4108
4109 -w, --word-regex(p)
4110
4111 instead of two lines, one with "regex" and the other with "regexp"
4112 because that confused at least one person since the short forms are the
4113 same. (This required a bit of code, as the output is generated
4114 automatically from a table. It wasn't just a text change.)
4115
4116 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4117 option but starts with a hyphen. Could be a pattern or a path name
4118 starting with a hyphen, for instance.
4119
4120 (h) "-" can be given as a file name to represent stdin.
4121
4122 (i) When file names are being printed, "(standard input)" is used for
4123 the standard input, for compatibility with GNU grep. Previously
4124 "<stdin>" was used.
4125
4126 (j) The option --label=xxx can be used to supply a name to be used for
4127 stdin when file names are being printed. There is no short form.
4128
4129 (k) Re-factored the options decoding logic because we are going to add
4130 two more options that take data. Such options can now be given in four
4131 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4132
4133 (l) Added the -A, -B, and -C options for requesting that lines of context
4134 around matches be printed.
4135
4136 (m) Added the -L option to print the names of files that do not contain
4137 any matching lines, that is, the complement of -l.
4138
4139 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4140 continue to scan other files.
4141
4142 (o) The -s option was incorrectly implemented. For compatibility with other
4143 greps, it now suppresses the error message for a non-existent or non-
4144 accessible file (but not the return code). There is a new option called
4145 -q that suppresses the output of matching lines, which was what -s was
4146 previously doing.
4147
4148 (p) Added --include and --exclude options to specify files for inclusion
4149 and exclusion when recursing.
4150
4151 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4152 Hopefully, it now does.
4153
4154 12. Missing cast in pcre_study().
4155
4156 13. Added an "uninstall" target to the makefile.
4157
4158 14. Replaced "extern" in the function prototypes in Makefile.in with
4159 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4160 world, but is set differently for Windows.
4161
4162 15. Added a second compiling function called pcre_compile2(). The only
4163 difference is that it has an extra argument, which is a pointer to an
4164 integer error code. When there is a compile-time failure, this is set
4165 non-zero, in addition to the error test pointer being set to point to an
4166 error message. The new argument may be NULL if no error number is required
4167 (but then you may as well call pcre_compile(), which is now just a
4168 wrapper). This facility is provided because some applications need a
4169 numeric error indication, but it has also enabled me to tidy up the way
4170 compile-time errors are handled in the POSIX wrapper.
4171
4172 16. Added VPATH=.libs to the makefile; this should help when building with one
4173 prefix path and installing with another. (Or so I'm told by someone who
4174 knows more about this stuff than I do.)
4175
4176 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4177 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4178 match everything, including newlines. This is not POSIX-compatible, but
4179 somebody wanted the feature. From pcretest it can be activated by using
4180 both the P and the s flags.
4181
4182 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4183
4184 19. libpcre.pc was being incorrectly installed as executable.
4185
4186 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4187 it now also looks for '\r' so that it will work unmodified on Windows.
4188
4189 21. Added Google's contributed C++ wrapper to the distribution.
4190
4191 22. Added some untidy missing memory free() calls in pcretest, to keep
4192 Electric Fence happy when testing.
4193
4194
4195
4196 Version 5.0 13-Sep-04
4197 ---------------------
4198
4199 1. Internal change: literal characters are no longer packed up into items
4200 containing multiple characters in a single byte-string. Each character
4201 is now matched using a separate opcode. However, there may be more than one
4202 byte in the character in UTF-8 mode.
4203
4204 2. The pcre_callout_block structure has two new fields: pattern_position and
4205 next_item_length. These contain the offset in the pattern to the next match
4206 item, and its length, respectively.
4207
4208 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4209 insertion of callouts before each pattern item. Added the /C option to
4210 pcretest to make use of this.
4211
4212 4. On the advice of a Windows user, the lines
4213
4214 #if defined(_WIN32) || defined(WIN32)
4215 _setmode( _fileno( stdout ), 0x8000 );
4216 #endif /* defined(_WIN32) || defined(WIN32) */
4217
4218 have been added to the source of pcretest. This apparently does useful
4219 magic in relation to line terminators.
4220
4221 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4222 for the benefit of those environments where the "b" makes a difference.
4223
4224 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4225 to know about it. I have put a hack into configure.in that adds in code
4226 to set GCC=yes if CC=icc. This seems to end up at a point in the
4227 generated configure script that is early enough to affect the setting of
4228 compiler options, which is what is needed, but I have no means of testing
4229 whether it really works. (The user who reported this had patched the
4230 generated configure script, which of course I cannot do.)
4231
4232 LATER: After change 22 below (new libtool files), the configure script
4233 seems to know about icc (and also ecc). Therefore, I have commented out
4234 this hack in configure.in.
4235
4236 7. Added support for pkg-config (2 patches were sent in).
4237
4238 8. Negated POSIX character classes that used a combination of internal tables
4239 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4240 [[:^ascii]]. Typically, they would match almost any characters. The other
4241 POSIX classes were not broken in this way.
4242
4243 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4244 to find the match, as PCRE was deluded into thinking that the match had to
4245 start at the start point or following a newline. The same bug applied to
4246 patterns with negative forward assertions or any backward assertions
4247 preceding ".*" at the start, unless the pattern required a fixed first
4248 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4249
4250 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4251 starting at the last subject character, bytes beyond the end of the subject
4252 string were read.
4253
4254 11. Renamed the variable "class" as "classbits" to make life easier for C++
4255 users. (Previously there was a macro definition, but it apparently wasn't
4256 enough.)
4257
4258 12. Added the new field "tables" to the extra data so that tables can be passed
4259 in at exec time, or the internal tables can be re-selected. This allows
4260 a compiled regex to be saved and re-used at a later time by a different
4261 program that might have everything at different addresses.
4262
4263 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4264 -R library as well as a -L library.
4265
4266 14. The debugging options of pcretest (-d on the command line or D on a
4267 pattern) showed incorrect output for anything following an extended class
4268 that contained multibyte characters and which was followed by a quantifier.
4269
4270 15. Added optional support for general category Unicode character properties
4271 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4272 support. It adds about 90K to the size of the library. The meanings of the
4273 inbuilt class escapes such as \d and \s have NOT been changed.
4274
4275 16. Updated pcredemo.c to include calls to free() to release the memory for the
4276 compiled pattern.
4277
4278 17. The generated file chartables.c was being created in the source directory
4279 instead of in the building directory. This caused the build to fail if the
4280 source directory was different from the building directory, and was
4281 read-only.
4282
4283 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4284 file. No doubt somebody will tell me if they don't make sense... Also added
4285 Dan Mooney's comments about building on OpenVMS.
4286
4287 19. Added support for partial matching via the PCRE_PARTIAL option for
4288 pcre_exec() and the \P data escape in pcretest.
4289
4290 20. Extended pcretest with 3 new pattern features:
4291
4292 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4293 write the compiled pattern to the file whose name is "rest-of-line".
4294 This is a straight binary dump of the data, with the saved pointer to
4295 the character tables forced to be NULL. The study data, if any, is
4296 written too. After writing, pcretest reads a new pattern.
4297
4298 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4299 compiled pattern from the given file. There must not be any
4300 occurrences of "<" in the file name (pretty unlikely); if there are,
4301 pcretest will instead treat the initial "<" as a pattern delimiter.
4302 After reading in the pattern, pcretest goes on to read data lines as
4303 usual.
4304
4305 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4306 and 16-bit fields in a compiled pattern, to simulate a pattern that
4307 was compiled on a host of opposite endianness.
4308
4309 21. The pcre-exec() function can now cope with patterns that were compiled on
4310 hosts of opposite endianness, with this restriction:
4311
4312 As for any compiled expression that is saved and used later, the tables
4313 pointer field cannot be preserved; the extra_data field in the arguments
4314 to pcre_exec() should be used to pass in a tables address if a value
4315 other than the default internal tables were used at compile time.
4316
4317 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4318 now diagnosed as an error. Previously, most of the time, a negative number
4319 would have been treated as zero, but if in addition "ovector" was passed as
4320 NULL, a crash could occur.
4321
4322 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4323 new versions from the libtool 1.5 distribution (the last one is a copy of
4324 a file called libtool.m4). This seems to have fixed the need to patch
4325 "configure" to support Darwin 1.3 (which I used to do). However, I still
4326 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4327 workstation).
4328
4329 24. Changed the PCRE licence to be the more standard "BSD" licence.
4330
4331
4332 Version 4.5 01-Dec-03
4333 ---------------------
4334
4335 1. There has been some re-arrangement of the code for the match() function so
4336 that it can be compiled in a version that does not call itself recursively.
4337 Instead, it keeps those local variables that need separate instances for
4338 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4339 needs to "recurse". Keeping track of where control must go is done by means
4340 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4341 hide most of the details from the main code, and operates only if
4342 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4343 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4344 operating.
4345
4346 To make it easier for callers to provide specially tailored get/free
4347 functions for this usage, two new functions, pcre_stack_malloc, and
4348 pcre_stack_free, are used. They are always called in strict stacking order,
4349 and the size of block requested is always the same.
4350
4351 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4352 PCRE has been compiled to use the stack or the heap for recursion. The
4353 -C option of pcretest uses this to show which version is compiled.
4354
4355 A new data escape \S, is added to pcretest; it causes the amounts of store
4356 obtained and freed by both kinds of malloc/free at match time to be added
4357 to the output.
4358
4359 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4360 what's available on my current Linux desktop machine.
4361
4362 3. When matching a UTF-8 string, the test for a valid string at the start has
4363 been extended. If start_offset is not zero, PCRE now checks that it points
4364 to a byte that is the start of a UTF-8 character. If not, it returns
4365 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4366 this is necessary because there may be backward assertions in the pattern.
4367 When matching the same subject several times, it may save resources to use
4368 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4369
4370 4. The code for checking the validity of UTF-8 strings has been tightened so
4371 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4372 containing "overlong sequences".
4373
4374 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4375 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4376 should have been "&", but it just so happened that all the cases this let
4377 through by mistake were picked up later in the function.
4378
4379 6. I had used a variable called "isblank" - this is a C99 function, causing
4380 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4381
4382 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4383 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4384 so the version is shown; (c) stop "make test" echoing "./RunTest".
4385
4386 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4387
4388 9. The prototype for memmove() for systems that don't have it was using
4389 size_t, but the inclusion of the header that defines size_t was later. I've
4390 moved the #includes for the C headers earlier to avoid this.
4391
4392 10. Added some adjustments to the code to make it easier to compiler on certain
4393 special systems:
4394
4395 (a) Some "const" qualifiers were missing.
4396 (b) Added the macro EXPORT before all exported functions; by default this
4397 is defined to be empty.
4398 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4399 that it reads its output file name as an argument instead of writing
4400 to the standard output and assuming this can be redirected.
4401
4402 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4403 class containing characters with values greater than 255, PCRE compilation
4404 went into a loop.
4405
4406 12. A recursive reference to a subpattern that was within another subpattern
4407 that had a minimum quantifier of zero caused PCRE to crash. For example,
4408 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4409 recursion. If the recursively-called subpattern itself had a zero repeat,
4410 that was OK.
4411
4412 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4413 buffer into which it was copied (for escape processing) was still set at
4414 1024, so long lines caused crashes.
4415
4416 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4417 "internal error: code overflow...". This applied to any character class
4418 that was followed by a possessive quantifier.
4419
4420 15. Modified the Makefile to add libpcre.la as a prerequisite for
4421 libpcreposix.la because I was told this is needed for a parallel build to
4422 work.
4423
4424 16. If a pattern that contained .* following optional items at the start was
4425 studied, the wrong optimizing data was generated, leading to matching
4426 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4427 matching string must start with a or b or c. The correct conclusion for
4428 this pattern is that a match can start with any character.
4429
4430
4431 Version 4.4 13-Aug-03
4432 ---------------------
4433
4434 1. In UTF-8 mode, a character class containing characters with values between
4435 127 and 255 was not handled correctly if the compiled pattern was studied.
4436 In fixing this, I have also improved the studying algorithm for such
4437 classes (slightly).
4438
4439 2. Three internal functions had redundant arguments passed to them. Removal
4440 might give a very teeny performance improvement.
4441
4442 3. Documentation bug: the value of the capture_top field in a callout is *one
4443 more than* the number of the hightest numbered captured substring.
4444
4445 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4446 in incorrectly linking with a previously installed version. They now link
4447 explicitly with libpcre.la.
4448
4449 5. configure.in no longer needs to recognize Cygwin specially.
4450
4451 6. A problem in pcre.in for Windows platforms is fixed.
4452
4453 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4454 pcretest, it used to include the size of the study block as part of its
4455 output. Unfortunately, the structure contains a field that has a different
4456 size on different hardware architectures. This meant that the tests that
4457 showed this size failed. As the block is currently always of a fixed size,
4458 this information isn't actually particularly useful in pcretest output, so
4459 I have just removed it.
4460
4461 8. Three pre-processor statements accidentally did not start in column 1.
4462 Sadly, there are *still* compilers around that complain, even though
4463 standard C has not required this for well over a decade. Sigh.
4464
4465 9. In pcretest, the code for checking callouts passed small integers in the
4466 callout_data field, which is a void * field. However, some picky compilers
4467 complained about the casts involved for this on 64-bit systems. Now
4468 pcretest passes the address of the small integer instead, which should get
4469 rid of the warnings.
4470
4471 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4472 both compile and run time, and gives an error if an invalid UTF-8 sequence
4473 is found. There is a option for disabling this check in cases where the
4474 string is known to be correct and/or the maximum performance is wanted.
4475
4476 11. In response to a bug report, I changed one line in Makefile.in from
4477
4478 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4479 to
4480 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4481
4482 to look similar to other lines, but I have no way of telling whether this
4483 is the right thing to do, as I do not use Windows. No doubt I'll get told
4484 if it's wrong...
4485
4486
4487 Version 4.3 21-May-03
4488 ---------------------
4489
4490 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4491 Makefile.
4492
4493 2. Some refactoring to improve the quality of the code:
4494
4495 (i) The utf8_table... variables are now declared "const".
4496
4497 (ii) The code for \cx, which used the "case flipping" table to upper case
4498 lower case letters, now just substracts 32. This is ASCII-specific,
4499 but the whole concept of \cx is ASCII-specific, so it seems
4500 reasonable.
4501
4502 (iii) PCRE was using its character types table to recognize decimal and
4503 hexadecimal digits in the pattern. This is silly, because it handles
4504 only 0-9, a-f, and A-F, but the character types table is locale-
4505 specific, which means strange things might happen. A private
4506 table is now used for this - though it costs 256 bytes, a table is
4507 much faster than multiple explicit tests. Of course, the standard
4508 character types table is still used for matching digits in subject
4509 strings against \d.
4510
4511 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4512 ending in _t are). So I've renamed it as ESC_tee.
4513
4514 3. The first argument for regexec() in the POSIX wrapper should have been
4515 defined as "const".
4516
4517 4. Changed pcretest to use malloc() for its buffers so that they can be
4518 Electric Fenced for debugging.
4519
4520 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4521 to read one or more bytes before the start of the subject string. Often this
4522 had no effect on PCRE's behaviour, but in some circumstances it could
4523 provoke a segmentation fault.
4524
4525 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4526 to try to read one or more bytes before the start of the subject string.
4527
4528 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4529 UTF-8 support could misbehave in various ways if the subject string
4530 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4531 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4532 back over UTF-8 characters.)
4533
4534
4535 Version 4.2 14-Apr-03
4536 ---------------------
4537
4538 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4539
4540 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4541 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4542 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4543 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4544 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4545 and BUILD_EXEEXT
4546 Note: automatic setting of the BUILD variables is not yet working
4547 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4548 compile-time but not at link-time
4549 [LINK]: use for linking executables only
4550 make different versions for Windows and non-Windows
4551 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4552 libraries
4553 [LINK_FOR_BUILD]: new variable
4554 [OBJEXT]: use throughout
4555 [EXEEXT]: use throughout
4556 <winshared>: new target
4557 <wininstall>: new target
4558 <dftables.o>: use native compiler
4559 <dftables>: use native linker
4560 <install>: handle Windows platform correctly
4561 <clean>: ditto
4562 <check>: ditto
4563 copy DLL to top builddir before testing
4564
4565 As part of these changes, -no-undefined was removed again. This was reported
4566 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4567 in any case.
4568
4569 3. Some tidies to get rid of compiler warnings:
4570
4571 . In the match_data structure, match_limit was an unsigned long int, whereas
4572 match_call_count was an int. I've made them both unsigned long ints.
4573
4574 . In pcretest the fact that a const uschar * doesn't automatically cast to
4575 a void * provoked a warning.
4576
4577 . Turning on some more compiler warnings threw up some "shadow" variables
4578 and a few more missing casts.
4579
4580 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4581 option, a class that contained a single character with a value between 128
4582 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4583
4584 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4585 option, a class that contained several characters, but with at least one
4586 whose value was between 128 and 255 caused PCRE to crash.
4587
4588
4589 Version 4.1 12-Mar-03
4590 ---------------------
4591
4592 1. Compiling with gcc -pedantic found a couple of places where casts were
4593 needed, and a string in dftables.c that was longer than standard compilers are
4594 required to support.
4595
4596 2. Compiling with Sun's compiler found a few more places where the code could
4597 be tidied up in order to avoid warnings.
4598
4599 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4600 first of these names is deprecated in the latest Autoconf in favour of the name
4601 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4602 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4603 analogy I have changed it to CFLAGS_FOR_BUILD.
4604
4605 4. Added -no-undefined to the linking command in the Makefile, because this is
4606 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4607 linking step for the pcreposix library.
4608
4609 5. PCRE was failing to diagnose the case of two named groups with the same
4610 name.
4611
4612 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4613 literal character that is needed in the subject for a match, and scans along to
4614 ensure that it is present before embarking on the full matching process. This
4615 saves time in cases of nested unlimited repeats that are never going to match.
4616 Problem: the scan can take a lot of time if the subject is very long (e.g.
4617 megabytes), thus penalizing straightforward matches. It is now done only if the
4618 amount of subject to be scanned is less than 1000 bytes.
4619
4620 7. A lesser problem with the same optimization is that it was recording the
4621 first character of an anchored pattern as "needed", thus provoking a search
4622 right along the subject, even when the first match of the pattern was going to
4623 fail. The "needed" character is now not set for anchored patterns, unless it
4624 follows something in the pattern that is of non-fixed length. Thus, it still
4625 fulfils its original purpose of finding quick non-matches in cases of nested
4626 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4627
4628
4629 Version 4.0 17-Feb-03
4630 ---------------------
4631
4632 1. If a comment in an extended regex that started immediately after a meta-item
4633 extended to the end of string, PCRE compiled incorrect data. This could lead to
4634 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4635
4636 2. Moved to autoconf 2.53 and libtool 1.4.2.
4637
4638 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4639 the special perltest8 script is no longer needed - all the tests can be run
4640 from a single perltest script.
4641
4642 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4643 by \s. It has now been removed in PCRE. This means it isn't recognized as
4644 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4645 class [:space:] *does* include VT, thereby creating a mess.
4646
4647 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4648 space and tab.
4649
4650 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4651 its new features into the main test script, reducing the number of scripts.
4652
4653 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4654 were backward compatible, and made the (?i) apply to the whole pattern, as if
4655 /i were given. Now it behaves more logically, and applies the option setting
4656 only to what follows. PCRE has been changed to follow suit. However, if it
4657 finds options settings right at the start of the pattern, it extracts them into
4658 the global options, as before. Thus, they show up in the info data.
4659
4660 8. Added support for the \Q...\E escape sequence. Characters in between are
4661 treated as literals. This is slightly different from Perl in that $ and @ are
4662 also handled as literals inside the quotes. In Perl, they will cause variable
4663 interpolation. Note the following examples:
4664
4665 Pattern PCRE matches Perl matches
4666
4667 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4668 \Qabc\$xyz\E abc\$xyz abc\$xyz
4669 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4670
4671 For compatibility with Perl, \Q...\E sequences are recognized inside character
4672 classes as well as outside them.
4673
4674 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4675 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4676 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4677 signed/unsigned warnings.
4678
4679 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4680 option for pcretest, so I've replaced it by a simple function that does just
4681 that job.
4682
4683 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4684 "pcregrep -".
4685
4686 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4687 Java package. This provides some syntactic sugar for simple cases of what my
4688 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4689 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4690 item, you can use this simplified notation. Note that only makes sense with
4691 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4692 greediness, whatever the setting of the PCRE_UNGREEDY option.
4693
4694 13. A change of greediness default within a pattern was not taking effect at
4695 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4696 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4697 was abstracted outside.
4698
4699 14. PCRE now supports the \G assertion. It is true when the current matching
4700 position is at the start point of the match. This differs from \A when the
4701 starting offset is non-zero. Used with the /g option of pcretest (or similar
4702 code), it works in the same way as it does for Perl's /g option. If all
4703 alternatives of a regex begin with \G, the expression is anchored to the start
4704 match position, and the "anchored" flag is set in the compiled expression.
4705
4706 15. Some bugs concerning the handling of certain option changes within patterns
4707 have been fixed. These applied to options other than (?ims). For example,
4708 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4709 been the other way round. Some of this was related to change 7 above.
4710
4711 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4712 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4713 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4714 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4715
4716 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4717 mode. Unlike ".", it always matches newline, whatever the setting of
4718 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4719 assertions. Perl allows it, but it doesn't (in general) work because it can't
4720 calculate the length of the lookbehind. At least, that's the case for Perl
4721 5.8.0 - I've been told they are going to document that it doesn't work in
4722 future.
4723
4724 18. Added an error diagnosis for escapes that PCRE does not support: these are
4725 \L, \l, \N, \P, \p, \U, \u, and \X.
4726
4727 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4728 reading past the end of the pattern in cases such as /[abcd/.
4729
4730 20. PCRE was getting more memory than necessary for patterns with classes that
4731 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4732
4733 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4734 compiling PCRE for use with Virtual Pascal.
4735
4736 22. Small fix to the Makefile to make it work properly if the build is done
4737 outside the source tree.
4738
4739 23. Added a new extension: a condition to go with recursion. If a conditional
4740 subpattern starts with (?(R) the "true" branch is used if recursion has
4741 happened, whereas the "false" branch is used only at the top level.
4742
4743 24. When there was a very long string of literal characters (over 255 bytes
4744 without UTF support, over 250 bytes with UTF support), the computation of how
4745 much memory was required could be incorrect, leading to segfaults or other
4746 strange effects.
4747
4748 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4749 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4750 there was a subsequent back reference to those brackets. This meant that, for
4751 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4752 possible to check for precisely this case. All we can do is abandon the
4753 optimization if .* occurs inside capturing brackets when there are any back
4754 references whatsoever. (See below for a better fix that came later.)
4755
4756 26. The handling of the optimization for finding the first character of a
4757 non-anchored pattern, and for finding a character that is required later in the
4758 match were failing in some cases. This didn't break the matching; it just
4759 failed to optimize when it could. The way this is done has been re-implemented.
4760
4761 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4762
4763 28. Added a new feature that provides some of the functionality that Perl
4764 provides with (?{...}). The facility is termed a "callout". The way it is done
4765 in PCRE is for the caller to provide an optional function, by setting
4766 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4767 global variable. By default it is unset, which disables all calling out. To get
4768 the function called, the regex must include (?C) at appropriate points. This
4769 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4770 This provides a means of identifying different callout points. When PCRE
4771 reaches such a point in the regex, if pcre_callout has been set, the external
4772 function is called. It is provided with data in a structure called
4773 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4774 matching continues; if it returns a non-zero value, the match at the current
4775 point fails. However, backtracking will occur if possible. [This was changed
4776 later and other features added - see item 49 below.]
4777
4778 29. pcretest is upgraded to test the callout functionality. It provides a
4779 callout function that displays information. By default, it shows the start of
4780 the match and the current position in the text. There are some new data escapes
4781 to vary what happens:
4782
4783 \C+ in addition, show current contents of captured substrings
4784 \C- do not supply a callout function
4785 \C!n return 1 when callout number n is reached
4786 \C!n!m return 1 when callout number n is reached for the mth time
4787
4788 30. If pcregrep was called with the -l option and just a single file name, it
4789 output "<stdin>" if a match was found, instead of the file name.
4790
4791 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4792 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4793 pcre_exec(). This saves a malloc/free per call. The default value of
4794 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4795 when configuring.
4796
4797 32. The default maximum size of a compiled pattern is 64K. There have been a
4798 few cases of people hitting this limit. The code now uses macros to handle the
4799 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4800 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4801 configuring. Tests 2 and 5 work only with 2-byte links because they output
4802 debugging information about compiled patterns.
4803
4804 33. Internal code re-arrangements:
4805
4806 (a) Moved the debugging function for printing out a compiled regex into
4807 its own source file (printint.c) and used #include to pull it into
4808 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4809 separate copies.
4810
4811 (b) Defined the list of op-code names for debugging as a macro in
4812 internal.h so that it is next to the definition of the opcodes.
4813
4814 (c) Defined a table of op-code lengths for simpler skipping along compiled
4815 code. This is again a macro in internal.h so that it is next to the
4816 definition of the opcodes.
4817
4818 34. Added support for recursive calls to individual subpatterns, along the
4819 lines of Robin Houston's patch (but implemented somewhat differently).
4820
4821 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4822 allow it to read and process whole directories in Win32. This code was
4823 contributed by Lionel Fourquaux; it has not been tested by me.
4824
4825 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4826 used to name a group. Names consist of alphanumerics and underscores, and must
4827 be unique. Back references use the syntax (?P=name) and recursive calls use
4828 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4829 numbers. The function pcre_fullinfo() can be used after compilation to extract
4830 a name/number map. There are three relevant calls:
4831
4832 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4833 PCRE_INFO_NAMECOUNT yields the number of entries
4834 PCRE_INFO_NAMETABLE yields a pointer to the map.
4835
4836 The map is a vector of fixed-size entries. The size of each entry depends on
4837 the length of the longest name used. The first two bytes of each entry are the
4838 group number, most significant byte first. There follows the corresponding
4839 name, zero terminated. The names are in alphabetical order.
4840
4841 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4842 case instead of 255. Making it the same both with and without UTF-8 support
4843 means that the same test output works with both.
4844
4845 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4846 calling malloc() with a zero argument.
4847
4848 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4849 optimization. I've improved things by keeping a bitmap of backreferences with
4850 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4851 fact referenced, the optimization can be applied. It is unlikely that a
4852 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4853 the match to follow \n) will appear inside brackets with a number greater than
4854 31, but if it does, any back reference > 31 suppresses the optimization.
4855
4856 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4857 of disabling numbered capturing parentheses. Any opening parenthesis that is
4858 not followed by ? behaves as if it were followed by ?: but named parentheses
4859 can still be used for capturing (and they will acquire numbers in the usual
4860 way).
4861
4862 41. Redesigned the return codes from the match() function into yes/no/error so
4863 that errors can be passed back from deep inside the nested calls. A malloc
4864 failure while inside a recursive subpattern call now causes the
4865 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4866
4867 42. It is now possible to set a limit on the number of times the match()
4868 function is called in a call to pcre_exec(). This facility makes it possible to
4869 limit the amount of recursion and backtracking, though not in a directly
4870 obvious way, because the match() function is used in a number of different
4871 circumstances. The count starts from zero for each position in the subject
4872 string (for non-anchored patterns). The default limit is, for compatibility, a
4873 large number, namely 10 000 000. You can change this in two ways:
4874
4875 (a) When configuring PCRE before making, you can use --with-match-limit=n
4876 to set a default value for the compiled library.
4877
4878 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4879 a different value is set. See 45 below.
4880
4881 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4882
4883 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4884 of things that can be changed at compile time. The first argument specifies
4885 what is wanted and the second points to where the information is to be placed.
4886 The current list of available information is:
4887
4888 PCRE_CONFIG_UTF8
4889
4890 The output is an integer that is set to one if UTF-8 support is available;
4891 otherwise it is set to zero.
4892
4893 PCRE_CONFIG_NEWLINE
4894
4895 The output is an integer that it set to the value of the code that is used for
4896 newline. It is either LF (10) or CR (13).
4897
4898 PCRE_CONFIG_LINK_SIZE
4899
4900 The output is an integer that contains the number of bytes used for internal
4901 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4902
4903 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4904
4905 The output is an integer that contains the threshold above which the POSIX
4906 interface uses malloc() for output vectors. See item 31 above.
4907
4908 PCRE_CONFIG_MATCH_LIMIT
4909
4910 The output is an unsigned integer that contains the default limit of the number
4911 of match() calls in a pcre_exec() execution. See 42 above.
4912
4913 44. pcretest has been upgraded by the addition of the -C option. This causes it
4914 to extract all the available output from the new pcre_config() function, and to
4915 output it. The program then exits immediately.
4916
4917 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4918 order to support additional features. One way would have been to define
4919 pcre_exec2() (for example) with extra arguments, but this would not have been
4920 extensible, and would also have required all calls to the original function to
4921 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4922 is used for passing in "extra" data from pcre_study().
4923
4924 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4925 contains the following fields:
4926
4927 flags a bitmap indicating which of the following fields are set
4928 study_data opaque data from pcre_study()
4929 match_limit a way of specifying a limit on match() calls for a specific
4930 call to pcre_exec()
4931 callout_data data for callouts (see 49 below)
4932
4933 The flag bits are also defined in pcre.h, and are
4934
4935 PCRE_EXTRA_STUDY_DATA
4936 PCRE_EXTRA_MATCH_LIMIT
4937 PCRE_EXTRA_CALLOUT_DATA
4938
4939 The pcre_study() function now returns one of these new pcre_extra blocks, with
4940 the actual study data pointed to by the study_data field, and the
4941 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4942 before. That is, this change is entirely upwards-compatible and requires no
4943 change to existing code.
4944
4945 If you want to pass in additional data to pcre_exec(), you can either place it
4946 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4947 block.
4948
4949 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4950 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4951 times with different match limits, until it finds the minimum value needed for
4952 pcre_exec() to complete. The value is then output. This can be instructive; for
4953 most simple matches the number is quite small, but for pathological cases it
4954 gets very large very quickly.
4955
4956 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4957 returns the size of the data block pointed to by the study_data field in a
4958 pcre_extra block, that is, the value that was passed as the argument to
4959 pcre_malloc() when PCRE was getting memory in which to place the information
4960 created by pcre_study(). The fourth argument should point to a size_t variable.
4961 pcretest has been extended so that this information is shown after a successful
4962 pcre_study() call when information about the compiled regex is being displayed.
4963
4964 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4965 because what follows is always an absolute path. (Later: it turns out that this
4966 is more than cosmetic for MinGW, because it doesn't like empty path
4967 components.)
4968
4969 49. Some changes have been made to the callout feature (see 28 above):
4970
4971 (i) A callout function now has three choices for what it returns:
4972
4973 0 => success, carry on matching
4974 > 0 => failure at this point, but backtrack if possible
4975 < 0 => serious error, return this value from pcre_exec()
4976
4977 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4978 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4979 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4980 use by callout functions. It will never be used by PCRE itself.
4981
4982 (ii) The pcre_extra structure (see 45 above) has a void * field called
4983 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4984 pcre_callout_block structure has a field of the same name. The contents of
4985 the field passed in the pcre_extra structure are passed to the callout
4986 function in the corresponding field in the callout block. This makes it
4987 easier to use the same callout-containing regex from multiple threads. For
4988 testing, the pcretest program has a new data escape
4989
4990 \C*n pass the number n (may be negative) as callout_data
4991
4992 If the callout function in pcretest receives a non-zero value as
4993 callout_data, it returns that value.
4994
4995 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4996 there were some redundant $(CFLAGS) in commands that are now specified as
4997 $(LINK), which already includes $(CFLAGS).
4998
4999 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5000 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5001 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5002 one-byte characters throughout. Note that case-insensitive matching applies
5003 only to characters whose values are less than 256. PCRE doesn't support the
5004 notion of cases for higher-valued characters.
5005
5006 (i) A character class whose characters are all within 0-255 is handled as
5007 a bit map, and the map is inverted for negative classes. Previously, a
5008 character > 255 always failed to match such a class; however it should
5009 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5010
5011 (ii) A negated character class with a single character < 255 is coded as
5012 "not this character" (OP_NOT). This wasn't working properly when the test
5013 character was multibyte, either singly or repeated.
5014
5015 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5016 mode, for example: \x{100}{2,3}.
5017
5018 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5019 singly or repeated) now correctly test multibyte characters. However,
5020 PCRE doesn't recognize any characters with values greater than 255 as
5021 digits, spaces, or word characters. Such characters always match \D, \S,
5022 and \W, and never match \d, \s, or \w.
5023
5024 (v) Classes may now contain characters and character ranges with values
5025 greater than 255. For example: [ab\x{100}-\x{400}].
5026
5027 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5028 PCRE in UTF-8 mode.
5029
5030 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5031 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5032 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5033 value.)
5034
5035 53. The single man page has become too large. I have therefore split it up into
5036 a number of separate man pages. These also give rise to individual HTML pages;
5037 these are now put in a separate directory, and there is an index.html page that
5038 lists them all. Some hyperlinking between the pages has been installed.
5039
5040 54. Added convenience functions for handling named capturing parentheses.
5041
5042 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5043 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5044 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5045 are faulted.
5046
5047 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5048 calling configure. These values are used when compiling the dftables.c program
5049 which is run to generate the source of the default character tables. They
5050 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5051 you will need to set these values.
5052
5053 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5054
5055
5056 Version 3.9 02-Jan-02
5057 ---------------------
5058
5059 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5060
5061 2. If --disable-static was given, the building process failed when trying to
5062 build pcretest and pcregrep. (For some reason it was using libtool to compile
5063 them, which is not right, as they aren't part of the library.)
5064
5065
5066 Version 3.8 18-Dec-01
5067 ---------------------
5068
5069 1. The experimental UTF-8 code was completely screwed up. It was packing the
5070 bytes in the wrong order. How dumb can you get?
5071
5072
5073 Version 3.7 29-Oct-01
5074 ---------------------
5075
5076 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5077 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5078 this didn't happen under Solaris 8, where I normally test things.
5079
5080 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5081 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5082 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5083 this fix an hour or so after the initial 3.7 release.)
5084
5085
5086 Version 3.6 23-Oct-01
5087 ---------------------
5088
5089 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5090 offsets passed as NULL with zero offset count.
5091
5092 2. The config.guess and config.sub files had not been updated when I moved to
5093 the latest autoconf.
5094
5095
5096 Version 3.5 15-Aug-01
5097 ---------------------
5098
5099 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5100 had been forgotten.
5101
5102 2. By using declared but undefined structures, we can avoid using "void"
5103 definitions in pcre.h while keeping the internal definitions of the structures
5104 private.
5105
5106 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5107 user point of view, this means that both static and shared libraries are built
5108 by default, but this can be individually controlled. More of the work of
5109 handling this static/shared cases is now inside libtool instead of PCRE's make
5110 file.
5111
5112 4. The pcretest utility is now installed along with pcregrep because it is
5113 useful for users (to test regexs) and by doing this, it automatically gets
5114 relinked by libtool. The documentation has been turned into a man page, so
5115 there are now .1, .txt, and .html versions in /doc.
5116
5117 5. Upgrades to pcregrep:
5118 (i) Added long-form option names like gnu grep.
5119 (ii) Added --help to list all options with an explanatory phrase.
5120 (iii) Added -r, --recursive to recurse into sub-directories.
5121 (iv) Added -f, --file to read patterns from a file.
5122
5123 6. pcre_exec() was referring to its "code" argument before testing that
5124 argument for NULL (and giving an error if it was NULL).
5125
5126 7. Upgraded Makefile.in to allow for compiling in a different directory from
5127 the source directory.
5128
5129 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5130 options bits, the pointer it was passed was to an int instead of to an unsigned
5131 long int. This mattered only on 64-bit systems.
5132
5133 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5134 generated) instead of pcre.in, which it its source. Also made the same change
5135 in several of the .c files.
5136
5137 10. A new release of gcc defines printf() as a macro, which broke pcretest
5138 because it had an ifdef in the middle of a string argument for printf(). Fixed
5139 by using separate calls to printf().
5140
5141 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5142 script, to force use of CR or LF instead of \n in the source. On non-Unix
5143 systems, the value can be set in config.h.
5144
5145 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5146 absolute limit. Changed the text of the error message to make this clear, and
5147 likewise updated the man page.
5148
5149 13. The limit of 99 on the number of capturing subpatterns has been removed.
5150 The new limit is 65535, which I hope will not be a "real" limit.
5151
5152
5153 Version 3.4 22-Aug-00
5154 ---------------------
5155
5156 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5157
5158 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5159
5160
5161 Version 3.3 01-Aug-00
5162 ---------------------
5163
5164 1. If an octal character was given, but the value was greater than \377, it
5165 was not getting masked to the least significant bits, as documented. This could
5166 lead to crashes in some systems.
5167
5168 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5169 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5170
5171 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5172 These just pass their arguments on to (pcre_free)(), but they are provided
5173 because some uses of PCRE bind it to non-C systems that can call its functions,
5174 but cannot call free() or pcre_free() directly.
5175
5176 4. Add "make test" as a synonym for "make check". Corrected some comments in
5177 the Makefile.
5178
5179 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5180 Makefile.
5181
5182 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5183 command called pgrep for grepping around the active processes.
5184
5185 7. Added the beginnings of support for UTF-8 character strings.
5186
5187 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5188 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5189 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5190 out for the ar command.)
5191
5192
5193 Version 3.2 12-May-00
5194 ---------------------
5195
5196 This is purely a bug fixing release.
5197
5198 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5199 of ZA. This was just one example of several cases that could provoke this bug,
5200 which was introduced by change 9 of version 2.00. The code for breaking
5201 infinite loops after an iteration that matches an empty string was't working
5202 correctly.
5203
5204 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5205 when matched against abbab (for example). After matching an empty string, it
5206 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5207 caused it to match further down the string than it should.
5208
5209 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5210 was there because it doesn't seem to be needed, and it causes trouble on some
5211 systems, as it is not a Standard C header. It has been removed.
5212
5213 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5214 were reported on the Macintosh. The changes were from
5215
5216 while ((c = *(++ptr)) != 0 && c != '\n');
5217 to
5218 while ((c = *(++ptr)) != 0 && c != '\n') ;
5219
5220 Totally extraordinary, but if that's what it takes...
5221
5222 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5223 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5224 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5225 assumes the way PCRE uses memmove() (always moving upwards).
5226
5227 6. PCRE is being used in one environment where strchr() is not available. There
5228 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5229 faster code anyway.
5230
5231
5232 Version 3.1 09-Feb-00
5233 ---------------------
5234
5235 The only change in this release is the fixing of some bugs in Makefile.in for
5236 the "install" target:
5237
5238 (1) It was failing to install pcreposix.h.
5239
5240 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5241
5242
5243 Version 3.0 01-Feb-00
5244 ---------------------
5245
5246 1. Add support for the /+ modifier to perltest (to output $` like it does in
5247 pcretest).
5248
5249 2. Add support for the /g modifier to perltest.
5250
5251 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5252 matches null strings.
5253
5254 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5255 pattern. Perl treats empty patterns specially - it reuses the most recent
5256 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5257 effect.
5258
5259 5. The POSIX interface was broken in that it was just handing over the POSIX
5260 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5261 required a bigger vector, with some working space on the end. This means that
5262 the POSIX wrapper now has to get and free some memory, and copy the results.
5263
5264 6. Added some simple autoconf support, placing the test data and the
5265 documentation in separate directories, re-organizing some of the
5266 information files, and making it build pcre-config (a GNU standard). Also added
5267 libtool support for building PCRE as a shared library, which is now the
5268 default.
5269
5270 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5271 09 are not valid octal constants. Single digits will be used for minor values
5272 less than 10.
5273
5274 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5275 existing programs that set these in the POSIX interface can use PCRE without
5276 modification.
5277
5278 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5279 return all that pcre_info() returns, plus additional data. The pcre_info()
5280 function is retained for compatibility, but is considered to be obsolete.
5281
5282 10. Added experimental recursion feature (?R) to handle one common case that
5283 Perl 5.6 will be able to do with (?p{...}).
5284
5285 11. Added support for POSIX character classes like [:alpha:], which Perl is
5286 adopting.
5287
5288
5289 Version 2.08 31-Aug-99
5290 ----------------------
5291
5292 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5293 trying to match at the startoffset position, but instead was moving forward to
5294 the next newline as if a previous match had failed.
5295
5296 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5297 and could get into a loop if a null string was matched other than at the start
5298 of the subject.
5299
5300 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5301 be distinguished at compile time, and for completeness also added PCRE_DATE.
5302
5303 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5304 in GnuWin32 environments.
5305
5306
5307 Version 2.07 29-Jul-99
5308 ----------------------
5309
5310 1. The documentation is now supplied in plain text form and HTML as well as in
5311 the form of man page sources.
5312
5313 2. C++ compilers don't like assigning (void *) values to other pointer types.
5314 In particular this affects malloc(). Although there is no problem in Standard
5315 C, I've put in casts to keep C++ compilers happy.
5316
5317 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5318 should be (const char *).
5319
5320 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5321 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5322 However, I haven't made this a standard facility. The documentation doesn't
5323 mention it, and the Makefile doesn't support it.
5324
5325 5. The Makefile now contains an "install" target, with editable destinations at
5326 the top of the file. The pcretest program is not installed.
5327
5328 6. pgrep -V now gives the PCRE version number and date.
5329
5330 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5331 causing the entire string to be ignored, instead of just the last character.
5332
5333 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5334 non-matching string, it can take a very, very long time, even for strings of
5335 quite modest length, because of the nested recursion. PCRE now does better in
5336 some of these cases. It does this by remembering the last required literal
5337 character in the pattern, and pre-searching the subject to ensure it is present
5338 before running the real match. In other words, it applies a heuristic to detect
5339 some types of certain failure quickly, and in the above example, if presented
5340 with a string that has no trailing " it gives "no match" very quickly.
5341
5342 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5343 other alternatives are tried instead.
5344
5345
5346 Version 2.06 09-Jun-99
5347 ----------------------
5348
5349 1. Change pcretest's output for amount of store used to show just the code
5350 space, because the remainder (the data block) varies in size between 32-bit and
5351 64-bit systems.
5352
5353 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5354 start matching at. This allows lookbehinds to work when searching for multiple
5355 occurrences in a string.
5356
5357 3. Added additional options to pcretest for testing multiple occurrences:
5358
5359 /+ outputs the rest of the string that follows a match
5360 /g loops for multiple occurrences, using the new startoffset argument
5361 /G loops for multiple occurrences by passing an incremented pointer
5362
5363 4. PCRE wasn't doing the "first character" optimization for patterns starting
5364 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5365 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5366 the letter 'x'. On long subject strings, this gives a significant speed-up.
5367
5368
5369 Version 2.05 21-Apr-99
5370 ----------------------
5371
5372 1. Changed the type of magic_number from int to long int so that it works
5373 properly on 16-bit systems.
5374
5375 2. Fixed a bug which caused patterns starting with .* not to work correctly
5376 when the subject string contained newline characters. PCRE was assuming
5377 anchoring for such patterns in all cases, which is not correct because .* will
5378 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5379 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5380 must be retried after every newline in the subject.
5381
5382
5383 Version 2.04 18-Feb-99
5384 ----------------------
5385
5386 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5387 computation of the store needed to hold the pattern was incorrect (too large).
5388 If such patterns were nested a few deep, this could multiply and become a real
5389 problem.
5390
5391 2. Added /M option to pcretest to show the memory requirement of a specific
5392 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5393
5394 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5395 compiled in such a way that the backtracking after subsequent failure was
5396 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5397 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5398
5399
5400 Version 2.03 02-Feb-99
5401 ----------------------
5402
5403 1. Fixed typo and small mistake in man page.
5404
5405 2. Added 4th condition (GPL supersedes if conflict) and created separate
5406 LICENCE file containing the conditions.
5407
5408 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5409 Perl, that is the internal \ allows the delimiter to be included in the
5410 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5411 the final delimiter, add \ to the end of the pattern (to test the error).
5412
5413 4. Added the convenience functions for extracting substrings after a successful
5414 match. Updated pcretest to make it able to test these functions.
5415
5416
5417 Version 2.02 14-Jan-99
5418 ----------------------
5419
5420 1. Initialized the working variables associated with each extraction so that
5421 their saving and restoring doesn't refer to uninitialized store.
5422
5423 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5424 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5425 fix the problem.
5426
5427 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5428 calls, and wasn't printing the correct value for compiling calls. Increased the
5429 default value of LOOPREPEAT, and the number of significant figures in the
5430 times.
5431
5432 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5433
5434 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5435 a building problem on Windows NT with a FAT file system.
5436
5437
5438 Version 2.01 21-Oct-98
5439 ----------------------
5440
5441 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5442 to character tables built by pcre_maketables() in the current locale. If NULL
5443 is passed, the default tables are used.
5444
5445
5446 Version 2.00 24-Sep-98
5447 ----------------------
5448
5449 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5450 it any more.
5451
5452 2. Allow quantification of (?>) groups, and make it work correctly.
5453
5454 3. The first character computation wasn't working for (?>) groups.
5455
5456 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5457 end of the subject) and add 5.005's \z, which really does match only at the
5458 very end of the subject.
5459
5460 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5461
5462 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5463 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5464 localized options. All options to pcre_study() were also removed.
5465
5466 7. Add other new features from 5.005:
5467
5468 $(?<= positive lookbehind
5469 $(?<! negative lookbehind
5470 (?imsx-imsx) added the unsetting capability
5471 such a setting is global if at outer level; local otherwise
5472 (?imsx-imsx:) non-capturing groups with option setting
5473 (?(cond)re|re) conditional pattern matching
5474
5475 A backreference to itself in a repeated group matches the previous
5476 captured string.
5477
5478 8. General tidying up of studying (both automatic and via "study")
5479 consequential on the addition of new assertions.
5480
5481 9. As in 5.005, unlimited repeated groups that could match an empty substring
5482 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5483 runtime if any iteration does actually match an empty substring.
5484
5485 10. Include the RunTest script in the distribution.
5486
5487 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5488 discrepancies, some of which were old and were also with respect to 5.004. They
5489 have now been fixed.
5490
5491
5492 Version 1.09 28-Apr-98
5493 ----------------------
5494
5495 1. A negated single character class followed by a quantifier with a minimum
5496 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5497 program crashes, or just wrong answers. This did not apply to negated classes
5498 containing more than one character, or to minima other than one.
5499
5500
5501 Version 1.08 27-Mar-98
5502 ----------------------
5503
5504 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5505
5506 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5507 latter must appear before anything that relies on it in the pattern.
5508
5509
5510 Version 1.07 16-Feb-98
5511 ----------------------
5512
5513 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5514 repeat of a potentially empty string).
5515
5516
5517 Version 1.06 23-Jan-98
5518 ----------------------
5519
5520 1. Added Markus Oberhumer's little patches for C++.
5521
5522 2. Literal strings longer than 255 characters were broken.
5523
5524
5525 Version 1.05 23-Dec-97
5526 ----------------------
5527
5528 1. Negated character classes containing more than one character were failing if
5529 PCRE_CASELESS was set at run time.
5530
5531
5532 Version 1.04 19-Dec-97
5533 ----------------------
5534
5535 1. Corrected the man page, where some "const" qualifiers had been omitted.
5536
5537 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5538 input syntax.
5539
5540 3. Fixed memory leak which occurred when a regex with back references was
5541 matched with an offsets vector that wasn't big enough. The temporary memory
5542 that is used in this case wasn't being freed if the match failed.
5543
5544 4. Tidied pcretest to ensure it frees memory that it gets.
5545
5546 5. Temporary memory was being obtained in the case where the passed offsets
5547 vector was exactly big enough.
5548
5549 6. Corrected definition of offsetof() from change 5 below.
5550
5551 7. I had screwed up change 6 below and broken the rules for the use of
5552 setjmp(). Now fixed.
5553
5554
5555 Version 1.03 18-Dec-97
5556 ----------------------
5557
5558 1. A erroneous regex with a missing opening parenthesis was correctly
5559 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5560 on some systems.
5561
5562 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5563 it was reported that one broken compiler failed on the former because "code" is
5564 also an independent variable.
5565
5566 3. The erroneous regex a[]b caused an array overrun reference.
5567
5568 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5569 fail on data ending with that character. (It was going on too far, and checking
5570 the next character, typically a binary zero.) This was specific to the
5571 optimized code for single-character negative classes.
5572
5573 5. Added a contributed patch from the TIN world which does the following:
5574
5575 + Add an undef for memmove, in case the the system defines a macro for it.
5576
5577 + Add a definition of offsetof(), in case there isn't one. (I don't know
5578 the reason behind this - offsetof() is part of the ANSI standard - but
5579 it does no harm).
5580
5581 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5582 most of the places where whitespace preceded '#'. I have given up and
5583 allowed the remaining 2 cases to be at the margin.
5584
5585 + Rename some variables in pcre to eliminate shadowing. This seems very
5586 pedantic, but does no harm, of course.
5587
5588 6. Moved the call to setjmp() into its own function, to get rid of warnings
5589 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5590
5591 7. Constructs such as \d{8,} were compiling into the equivalent of
5592 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5593 outcome, but in this particular case used more store than had been allocated,
5594 which caused the bug to be discovered because it threw up an internal error.
5595
5596 8. The debugging code in both pcre and pcretest for outputting the compiled
5597 form of a regex was going wrong in the case of back references followed by
5598 curly-bracketed repeats.
5599
5600
5601 Version 1.02 12-Dec-97
5602 ----------------------
5603
5604 1. Typos in pcre.3 and comments in the source fixed.
5605
5606 2. Applied a contributed patch to get rid of places where it used to remove
5607 'const' from variables, and fixed some signed/unsigned and uninitialized
5608 variable warnings.
5609
5610 3. Added the "runtest" target to Makefile.
5611
5612 4. Set default compiler flag to -O2 rather than just -O.
5613
5614
5615 Version 1.01 19-Nov-97
5616 ----------------------
5617
5618 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5619 like /([ab]*)*/, that is, for classes with more than one character in them.
5620
5621 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5622 as /((?>a*))*/ (a PCRE_EXTRA facility).
5623
5624
5625 Version 1.00 18-Nov-97
5626 ----------------------
5627
5628 1. Added compile-time macros to support systems such as SunOS4 which don't have
5629 memmove() or strerror() but have other things that can be used instead.
5630
5631 2. Arranged that "make clean" removes the executables.
5632
5633
5634 Version 0.99 27-Oct-97
5635 ----------------------
5636
5637 1. Fixed bug in code for optimizing classes with only one character. It was
5638 initializing a 32-byte map regardless, which could cause it to run off the end
5639 of the memory it had got.
5640
5641 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5642
5643
5644 Version 0.98 22-Oct-97
5645 ----------------------
5646
5647 1. Fixed bug in code for handling temporary memory usage when there are more
5648 back references than supplied space in the ovector. This could cause segfaults.
5649
5650
5651 Version 0.97 21-Oct-97
5652 ----------------------
5653
5654 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5655
5656 2. Optimized negated single characters not to use a bit map.
5657
5658 3. Brought error texts together as macro definitions; clarified some of them;
5659 fixed one that was wrong - it said "range out of order" when it meant "invalid
5660 escape sequence".
5661
5662 4. Changed some char * arguments to const char *.
5663
5664 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5665
5666 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5667 pcretest.
5668
5669
5670 Version 0.96 16-Oct-97
5671 ----------------------
5672
5673 1. Added a simple "pgrep" utility to the distribution.
5674
5675 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5676 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5677 where "ddd" means "one or more decimal digits".
5678
5679 3. Fixed serious bug. If a pattern had a back reference, but the call to
5680 pcre_exec() didn't supply a large enough ovector to record the related
5681 identifying subpattern, the match always failed. PCRE now remembers the number
5682 of the largest back reference, and gets some temporary memory in which to save
5683 the offsets during matching if necessary, in order to ensure that
5684 backreferences always work.
5685
5686 4. Increased the compatibility with Perl in a number of ways:
5687
5688 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5689 to request this handling. The option can be set at compile or exec time.
5690
5691 (b) $ matches before a terminating newline by default; an option
5692 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5693 mode). The option can be set at compile or exec time.
5694
5695 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5696 the same as Perl's. If the decimal number it represents is less than 10
5697 or there aren't that many previous left capturing parentheses, an octal
5698 escape is read. Inside a character class, it's always an octal escape,
5699 even if it is a single digit.
5700
5701 (d) An escaped but undefined alphabetic character is taken as a literal,
5702 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5703 escapes.
5704
5705 (e) {0} is now permitted. (The previous item is removed from the compiled
5706 pattern).
5707
5708 5. Changed all the names of code files so that the basic parts are no longer
5709 than 10 characters, and abolished the teeny "globals.c" file.
5710
5711 6. Changed the handling of character classes; they are now done with a 32-byte
5712 bit map always.
5713
5714 7. Added the -d and /D options to pcretest to make it possible to look at the
5715 internals of compilation without having to recompile pcre.
5716
5717
5718 Version 0.95 23-Sep-97
5719 ----------------------
5720
5721 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5722 \x20 at the start of a run of normal characters. These were being treated as
5723 real characters, instead of the source characters being re-checked.
5724
5725
5726 Version 0.94 18-Sep-97
5727 ----------------------
5728
5729 1. The functions are now thread-safe, with the caveat that the global variables
5730 containing pointers to malloc() and free() or alternative functions are the
5731 same for all threads.
5732
5733 2. Get pcre_study() to generate a bitmap of initial characters for non-
5734 anchored patterns when this is possible, and use it if passed to pcre_exec().
5735
5736
5737 Version 0.93 15-Sep-97
5738 ----------------------
5739
5740 1. /(b)|(:+)/ was computing an incorrect first character.
5741
5742 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5743 but not actually doing anything yet.
5744
5745 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5746 as Perl does (e.g. [-az] or [az-]).
5747
5748 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5749 all possible positions.
5750
5751 5. Split up into different modules to avoid including unneeded functions in a
5752 compiled binary. However, compile and exec are still in one module. The "study"
5753 function is split off.
5754
5755 6. The character tables are now in a separate module whose source is generated
5756 by an auxiliary program - but can then be edited by hand if required. There are
5757 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5758 toupper() in the code.
5759
5760 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5761 make them global. Abolish the function for setting them, as the caller can now
5762 set them directly.
5763
5764
5765 Version 0.92 11-Sep-97
5766 ----------------------
5767
5768 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5769 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5770
5771 2. Caseless matching was not working in character classes if the characters in
5772 the pattern were in upper case.
5773
5774 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5775
5776 4. Make PCRE_ANCHORED public and accept as a compile option.
5777
5778 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5779 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5780 pass them.
5781
5782 6. Give an error if bad option bits passed at compile or run time.
5783
5784 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5785 pcretest to cause it to pass that flag.
5786
5787 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5788 options, and the first character, if set.
5789
5790 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5791
5792
5793 Version 0.91 10-Sep-97
5794 ----------------------
5795
5796 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5797 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5798
5799 2. PCRE was looping on encountering an indefinitely repeated back reference to
5800 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5801 Perl does - treats the match as successful.
5802
5803 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5