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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1597 - (show annotations)
Sat Aug 29 15:23:48 2015 UTC (3 days, 13 hours ago) by ph10
File size: 284334 byte(s)
Error occurred while calculating annotation data.
Fix bad error check for workspace overrun.
1 ChangeLog for PCRE
2 ------------------
3
4 Note that the PCRE 8.xx series (PCRE1) is now in a bugfix-only state. All
5 development is happening in the PCRE2 10.xx series.
6
7 Version 8.38 xx-xxx-xxxx
8 ------------------------
9
10 1. If a group that contained a recursive back reference also contained a
11 forward reference subroutine call followed by a non-forward-reference
12 subroutine call, for example /.((?2)(?R)\1)()/, pcre2_compile() failed to
13 compile correct code, leading to undefined behaviour or an internally
14 detected error. This bug was discovered by the LLVM fuzzer.
15
16 2. Quantification of certain items (e.g. atomic back references) could cause
17 incorrect code to be compiled when recursive forward references were
18 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
19 This bug was discovered by the LLVM fuzzer.
20
21 3. A repeated conditional group whose condition was a reference by name caused
22 a buffer overflow if there was more than one group with the given name.
23 This bug was discovered by the LLVM fuzzer.
24
25 4. A recursive back reference by name within a group that had the same name as
26 another group caused a buffer overflow. For example:
27 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
28
29 5. A forward reference by name to a group whose number is the same as the
30 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
31 a buffer overflow at compile time. This bug was discovered by the LLVM
32 fuzzer.
33
34 6. A lookbehind assertion within a set of mutually recursive subpatterns could
35 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
36
37 7. Another buffer overflow bug involved duplicate named groups with a
38 reference between their definition, with a group that reset capture
39 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
40 fixed by always allowing for more memory, even if not needed. (A proper fix
41 is implemented in PCRE2, but it involves more refactoring.)
42
43 8. There was no check for integer overflow in subroutine calls such as (?123).
44
45 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
46 being treated as a literal 'l' instead of causing an error.
47
48 10. There was a buffer overflow if pcre_exec() was called with an ovector of
49 size 1. This bug was found by american fuzzy lop.
50
51 11. If a non-capturing group containing a conditional group that could match
52 an empty string was repeated, it was not identified as matching an empty
53 string itself. For example: /^(?:(?(1)x|)+)+$()/.
54
55 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
56 \a and \e in test subject lines.
57
58 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
59 instead of the EBCDIC value.
60
61 14. The handling of \c in an EBCDIC environment has been revised so that it is
62 now compatible with the specification in Perl's perlebcdic page.
63
64 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
65 ASCII/Unicode. This has now been added to the list of characters that are
66 recognized as white space in EBCDIC.
67
68 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
69 error (correctly) when used outside a class, but did not give an error
70 within a class.
71
72 17. \h within a class was incorrectly compiled in EBCDIC environments.
73
74 18. A pattern with an unmatched closing parenthesis that contained a backward
75 assertion which itself contained a forward reference caused buffer
76 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
77
78 19. JIT should return with error when the compiled pattern requires more stack
79 space than the maximum.
80
81 20. A possessively repeated conditional group that could match an empty string,
82 for example, /(?(R))*+/, was incorrectly compiled.
83
84 21. Fix infinite recursion in the JIT compiler when certain patterns such as
85 /(?:|a|){100}x/ are analysed.
86
87 22. Some patterns with character classes involving [: and \\ were incorrectly
88 compiled and could cause reading from uninitialized memory or an incorrect
89 error diagnosis.
90
91 23. Pathological patterns containing many nested occurrences of [: caused
92 pcre_compile() to run for a very long time.
93
94 24. A conditional group with only one branch has an implicit empty alternative
95 branch and must therefore be treated as potentially matching an empty
96 string.
97
98 25. If (?R was followed by - or + incorrect behaviour happened instead of a
99 diagnostic.
100
101 26. Arrange to give up on finding the minimum matching length for overly
102 complex patterns.
103
104 27. Similar to (4) above: in a pattern with duplicated named groups and an
105 occurrence of (?| it is possible for an apparently non-recursive back
106 reference to become recursive if a later named group with the relevant
107 number is encountered. This could lead to a buffer overflow. Wen Guanxing
108 from Venustech ADLAB discovered this bug.
109
110 28. If pcregrep was given the -q option with -c or -l, or when handling a
111 binary file, it incorrectly wrote output to stdout.
112
113 29. The JIT compiler did not restore the control verb head in case of *THEN
114 control verbs. This issue was found by Karl Skomski with a custom LLVM
115 fuzzer.
116
117 30. Error messages for syntax errors following \g and \k were giving inaccurate
118 offsets in the pattern.
119
120 31. Added a check for integer overflow in conditions (?(<digits>) and
121 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
122 fuzzer.
123
124 32. Handling recursive references such as (?2) when the reference is to a group
125 later in the pattern uses code that is very hacked about and error-prone.
126 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
127 give an internal error if it is obvious that compiling has gone wrong.
128
129 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
130 This issue was found by Karl Skomski with a custom LLVM fuzzer.
131
132 34. The JIT compiler should restore the control chain for empty possessive
133 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
134
135 35. Match limit check added to JIT recursion. This issue was found by Karl
136 Skomski with a custom LLVM fuzzer.
137
138 36. Yet another case similar to 27 above has been circumvented by an
139 unconditional allocation of extra memory. This issue is fixed "properly" in
140 PCRE2 by refactoring the way references are handled. Wen Guanxing
141 from Venustech ADLAB discovered this bug.
142
143 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
144 with a custom LLVM fuzzer.
145
146 38. Fixed a corner case of range optimization in JIT.
147
148 39. An incorrect error "overran compiling workspace" was given if there were
149 exactly enough group forward references such that the last one extended
150 into the workspace safety margin. The next one would have expanded the
151 workspace. The test for overflow was not including the safety margin.
152
153
154 Version 8.37 28-April-2015
155 --------------------------
156
157 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
158 for those parentheses to be closed with whatever has been captured so far.
159 However, it was failing to mark any other groups between the hightest
160 capture so far and the currrent group as "unset". Thus, the ovector for
161 those groups contained whatever was previously there. An example is the
162 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
163
164 2. If an assertion condition was quantified with a minimum of zero (an odd
165 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
166
167 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
168 unrecognized modifier, a crash could occur.
169
170 4. An attempt to do global matching in pcretest with a zero-length ovector
171 caused a crash.
172
173 5. Fixed a memory leak during matching that could occur for a subpattern
174 subroutine call (recursive or otherwise) if the number of captured groups
175 that had to be saved was greater than ten.
176
177 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
178 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
179 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
180
181 7. A UTF pattern containing a "not" match of a non-ASCII character and a
182 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
183
184 8. When a pattern is compiled, it remembers the highest back reference so that
185 when matching, if the ovector is too small, extra memory can be obtained to
186 use instead. A conditional subpattern whose condition is a check on a
187 capture having happened, such as, for example in the pattern
188 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
189 setting the highest backreference number. This mattered only if pcre_exec()
190 was called with an ovector that was too small to hold the capture, and there
191 was no other kind of back reference (a situation which is probably quite
192 rare). The effect of the bug was that the condition was always treated as
193 FALSE when the capture could not be consulted, leading to a incorrect
194 behaviour by pcre_exec(). This bug has been fixed.
195
196 9. A reference to a duplicated named group (either a back reference or a test
197 for being set in a conditional) that occurred in a part of the pattern where
198 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
199 to be incorrectly calculated, leading to overwriting.
200
201 10. A mutually recursive set of back references such as (\2)(\1) caused a
202 segfault at study time (while trying to find the minimum matching length).
203 The infinite loop is now broken (with the minimum length unset, that is,
204 zero).
205
206 11. If an assertion that was used as a condition was quantified with a minimum
207 of zero, matching went wrong. In particular, if the whole group had
208 unlimited repetition and could match an empty string, a segfault was
209 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
210 assertions to be quantified, but not if they are being used as conditions,
211 so the above pattern is faulted by Perl. PCRE has now been changed so that
212 it also rejects such patterns.
213
214 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
215 failed to allow the zero-repeat case if pcre2_exec() was called with an
216 ovector too small to capture the group.
217
218 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
219 Red Hat Product Security:
220
221 (a) A crash if /K and /F were both set with the option to save the compiled
222 pattern.
223
224 (b) Another crash if the option to print captured substrings in a callout
225 was combined with setting a null ovector, for example \O\C+ as a subject
226 string.
227
228 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
229 forward reference repeated a large (but limited) number of times within a
230 repeated outer group that has a zero minimum quantifier, caused incorrect
231 code to be compiled, leading to the error "internal error:
232 previously-checked referenced subpattern not found" when an incorrect
233 memory address was read. This bug was reported as "heap overflow",
234 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
235 CVE-2015-2325.
236
237 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
238 call within a group that also contained a recursive back reference caused
239 incorrect code to be compiled. This bug was reported as "heap overflow",
240 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
241 number CVE-2015-2326.
242
243 24. Computing the size of the JIT read-only data in advance has been a source
244 of various issues, and new ones are still appear unfortunately. To fix
245 existing and future issues, size computation is eliminated from the code,
246 and replaced by on-demand memory allocation.
247
248 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
249 adjacent to the end of the range, and the range contained characters with
250 more than one other case, caused incorrect behaviour when compiled in UTF
251 mode. In that example, the range a-j was left out of the class.
252
253 26. Fix JIT compilation of conditional blocks, which assertion
254 is converted to (*FAIL). E.g: /(?(?!))/.
255
256 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
257 discovered by the LLVM fuzzer.
258
259 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
260 when this assertion was used as a condition, for example (?(?!)a|b). In
261 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
262 error about an unsupported item.
263
264 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
265 possessification code could take exponential time to complete. A recursion
266 depth limit of 1000 has been imposed to limit the resources used by this
267 optimization.
268
269 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
270 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
271 because \S ensures they are all in the class. The code for doing this was
272 interacting badly with the code for computing the amount of space needed to
273 compile the pattern, leading to a buffer overflow. This bug was discovered
274 by the LLVM fuzzer.
275
276 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
277 other kinds of group caused stack overflow at compile time. This bug was
278 discovered by the LLVM fuzzer.
279
280 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
281 between a subroutine call and its quantifier was incorrectly compiled,
282 leading to buffer overflow or other errors. This bug was discovered by the
283 LLVM fuzzer.
284
285 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
286 assertion after (?(. The code was failing to check the character after
287 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
288 was discovered by the LLVM fuzzer.
289
290 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
291 a fixed maximum following a group that contains a subroutine reference was
292 incorrectly compiled and could trigger buffer overflow. This bug was
293 discovered by the LLVM fuzzer.
294
295 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
296 caused a stack overflow instead of the diagnosis of a non-fixed length
297 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
298
299 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
300 (e.g. /(?<=\Ka)/) could make pcregrep loop.
301
302 37. There was a similar problem to 36 in pcretest for global matches.
303
304 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
305 and a subsequent item in the pattern caused a non-match, backtracking over
306 the repeated \X did not stop, but carried on past the start of the subject,
307 causing reference to random memory and/or a segfault. There were also some
308 other cases where backtracking after \C could crash. This set of bugs was
309 discovered by the LLVM fuzzer.
310
311 39. The function for finding the minimum length of a matching string could take
312 a very long time if mutual recursion was present many times in a pattern,
313 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
314 method has been implemented. This infelicity was discovered by the LLVM
315 fuzzer.
316
317 40. Static linking against the PCRE library using the pkg-config module was
318 failing on missing pthread symbols.
319
320
321 Version 8.36 26-September-2014
322 ------------------------------
323
324 1. Got rid of some compiler warnings in the C++ modules that were shown up by
325 -Wmissing-field-initializers and -Wunused-parameter.
326
327 2. The tests for quantifiers being too big (greater than 65535) were being
328 applied after reading the number, and stupidly assuming that integer
329 overflow would give a negative number. The tests are now applied as the
330 numbers are read.
331
332 3. Tidy code in pcre_exec.c where two branches that used to be different are
333 now the same.
334
335 4. The JIT compiler did not generate match limit checks for certain
336 bracketed expressions with quantifiers. This may lead to exponential
337 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
338 issue should be resolved now.
339
340 5. Fixed an issue, which occures when nested alternatives are optimized
341 with table jumps.
342
343 6. Inserted two casts and changed some ints to size_t in the light of some
344 reported 64-bit compiler warnings (Bugzilla 1477).
345
346 7. Fixed a bug concerned with zero-minimum possessive groups that could match
347 an empty string, which sometimes were behaving incorrectly in the
348 interpreter (though correctly in the JIT matcher). This pcretest input is
349 an example:
350
351 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
352 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
353
354 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
355 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
356 for an empty string was breaking the inner loop and carrying on at a lower
357 level, when possessive repeated groups should always return to a higher
358 level as they have no backtrack points in them. The empty string test now
359 occurs at the outer level.
360
361 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
362 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
363
364 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
365 doesn't).
366
367 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
368 to hold a pattern to be incorrectly computed (too small) when there were
369 named back references to duplicated names. This could cause "internal
370 error: code overflow" or "double free or corruption" or other memory
371 handling errors.
372
373 11. When named subpatterns had the same prefixes, back references could be
374 confused. For example, in this pattern:
375
376 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
377
378 the reference to 'Name' was incorrectly treated as a reference to a
379 duplicate name.
380
381 12. A pattern such as /^s?c/mi8 where the optional character has more than
382 one "other case" was incorrectly compiled such that it would only try to
383 match starting at "c".
384
385 13. When a pattern starting with \s was studied, VT was not included in the
386 list of possible starting characters; this should have been part of the
387 8.34/18 patch.
388
389 14. If a character class started [\Qx]... where x is any character, the class
390 was incorrectly terminated at the ].
391
392 15. If a pattern that started with a caseless match for a character with more
393 than one "other case" was studied, PCRE did not set up the starting code
394 unit bit map for the list of possible characters. Now it does. This is an
395 optimization improvement, not a bug fix.
396
397 16. The Unicode data tables have been updated to Unicode 7.0.0.
398
399 17. Fixed a number of memory leaks in pcregrep.
400
401 18. Avoid a compiler warning (from some compilers) for a function call with
402 a cast that removes "const" from an lvalue by using an intermediate
403 variable (to which the compiler does not object).
404
405 19. Incorrect code was compiled if a group that contained an internal recursive
406 back reference was optional (had quantifier with a minimum of zero). This
407 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
408 caused segmentation faults because of stack overflows at compile time.
409
410 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
411 group that is quantified with an indefinite repeat, caused a compile-time
412 loop which used up all the system stack and provoked a segmentation fault.
413 This was not the same bug as 19 above.
414
415 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
416 Patch by Mike Frysinger.
417
418
419 Version 8.35 04-April-2014
420 --------------------------
421
422 1. A new flag is set, when property checks are present in an XCLASS.
423 When this flag is not set, PCRE can perform certain optimizations
424 such as studying these XCLASS-es.
425
426 2. The auto-possessification of character sets were improved: a normal
427 and an extended character set can be compared now. Furthermore
428 the JIT compiler optimizes more character set checks.
429
430 3. Got rid of some compiler warnings for potentially uninitialized variables
431 that show up only when compiled with -O2.
432
433 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
434 of a match later then the end of the match. The pcretest program was not
435 handling the case sensibly - it was outputting from the start to the next
436 binary zero. It now reports this situation in a message, and outputs the
437 text from the end to the start.
438
439 5. Fast forward search is improved in JIT. Instead of the first three
440 characters, any three characters with fixed position can be searched.
441 Search order: first, last, middle.
442
443 6. Improve character range checks in JIT. Characters are read by an inprecise
444 function now, which returns with an unknown value if the character code is
445 above a certain threshold (e.g: 256). The only limitation is that the value
446 must be bigger than the threshold as well. This function is useful when
447 the characters above the threshold are handled in the same way.
448
449 7. The macros whose names start with RAWUCHAR are placeholders for a future
450 mode in which only the bottom 21 bits of 32-bit data items are used. To
451 make this more memorable for those maintaining the code, the names have
452 been changed to start with UCHAR21, and an extensive comment has been added
453 to their definition.
454
455 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
456 to the export list in Makefile.am (they were accidentally omitted from the
457 8.34 tarball).
458
459 9. The informational output from pcretest used the phrase "starting byte set"
460 which is inappropriate for the 16-bit and 32-bit libraries. As the output
461 for "first char" and "need char" really means "non-UTF-char", I've changed
462 "byte" to "char", and slightly reworded the output. The documentation about
463 these values has also been (I hope) clarified.
464
465 10. Another JIT related optimization: use table jumps for selecting the correct
466 backtracking path, when more than four alternatives are present inside a
467 bracket.
468
469 11. Empty match is not possible, when the minimum length is greater than zero,
470 and there is no \K in the pattern. JIT should avoid empty match checks in
471 such cases.
472
473 12. In a caseless character class with UCP support, when a character with more
474 than one alternative case was not the first character of a range, not all
475 the alternative cases were added to the class. For example, s and \x{17f}
476 are both alternative cases for S: the class [RST] was handled correctly,
477 but [R-T] was not.
478
479 13. The configure.ac file always checked for pthread support when JIT was
480 enabled. This is not used in Windows, so I have put this test inside a
481 check for the presence of windows.h (which was already tested for).
482
483 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
484 The algorithm provides a way to skip certain starting offsets, and usually
485 faster than linear prefix searches.
486
487 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
488 as for 'fr_FR' and 'french'. For some reason, however, it then used the
489 Windows-specific input and output files, which have 'french' screwed in.
490 So this could never have worked. One of the problems with locales is that
491 they aren't always the same. I have now updated RunTest so that it checks
492 the output of the locale test (test 3) against three different output
493 files, and it allows the test to pass if any one of them matches. With luck
494 this should make the test pass on some versions of Solaris where it was
495 failing. Because of the uncertainty, the script did not used to stop if
496 test 3 failed; it now does. If further versions of a French locale ever
497 come to light, they can now easily be added.
498
499 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
500 there was a message during ./configure, but it did not stop. This now
501 provokes an error. The invalid example in README has been corrected.
502 If a value less than the minimum is given, the minimum value has always
503 been used, but now a warning is given.
504
505 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
506 was a bug in the test system, which is now fixed. Also, the list of various
507 configurations that are tested for each release did not have one with both
508 16/32 bits and --enable-bar-anycrlf. It now does.
509
510 18. pcretest was missing "-C bsr" for displaying the \R default setting.
511
512 19. Little endian PowerPC systems are supported now by the JIT compiler.
513
514 20. The fast forward newline mechanism could enter to an infinite loop on
515 certain invalid UTF-8 input. Although we don't support these cases
516 this issue can be fixed by a performance optimization.
517
518 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
519 not take account if existing stack usage. There is now a new global
520 variable called pcre_stack_guard that can be set to point to an external
521 function to check stack availability. It is called at the start of
522 processing every parenthesized group.
523
524 22. A typo in the code meant that in ungreedy mode the max/min qualifier
525 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
526 not match "ab".
527
528 23. When UTF was disabled, the JIT program reported some incorrect compile
529 errors. These messages are silenced now.
530
531 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
532 compiler.
533
534 25. Change all the temporary files used in RunGrepTest to be different to those
535 used by RunTest so that the tests can be run simultaneously, for example by
536 "make -j check".
537
538
539 Version 8.34 15-December-2013
540 -----------------------------
541
542 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
543 executable memory. Patch inspired by Carsten Klein.
544
545 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
546 this macro is never tested and has no effect, because the work to support
547 coverage involves only compiling and linking options and special targets in
548 the Makefile. The comment in config.h implied that defining the macro would
549 enable coverage support, which is totally false. There was also support for
550 setting this macro in the CMake files (my fault, I just copied it from
551 configure). SUPPORT_GCOV has now been removed.
552
553 3. Make a small performance improvement in strlen16() and strlen32() in
554 pcretest.
555
556 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
557 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
558 these cases). There was also one in pcretest.c.
559
560 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
561
562 6. In UTF mode, the code for checking whether a group could match an empty
563 string (which is used for indefinitely repeated groups to allow for
564 breaking an infinite loop) was broken when the group contained a repeated
565 negated single-character class with a character that occupied more than one
566 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
567 UTF-8 mode). The effect was undefined: the group might or might not be
568 deemed as matching an empty string, or the program might have crashed.
569
570 7. The code for checking whether a group could match an empty string was not
571 recognizing that \h, \H, \v, \V, and \R must match a character.
572
573 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
574 an empty string. If it can, pcretest shows this in its information output.
575
576 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
577 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
578 matched by pcre_exec() without using JIT:
579
580 (a) If the rest of the pattern did not match after a maximal run of
581 grapheme clusters, the code for backing up to try with fewer of them
582 did not always back up over a full grapheme when characters that do not
583 have the modifier quality were involved, e.g. Hangul syllables.
584
585 (b) If the match point in a subject started with modifier character, and
586 there was no match, the code could incorrectly back up beyond the match
587 point, and potentially beyond the first character in the subject,
588 leading to a segfault or an incorrect match result.
589
590 10. A conditional group with an assertion condition could lead to PCRE
591 recording an incorrect first data item for a match if no other first data
592 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
593 first data item, and therefore matched "ca" after "c" instead of at the
594 start.
595
596 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
597 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
598
599 12. The source of pcregrep now includes z/OS-specific code so that it can be
600 compiled for z/OS as part of the special z/OS distribution.
601
602 13. Added the -T and -TM options to pcretest.
603
604 14. The code in pcre_compile.c for creating the table of named capturing groups
605 has been refactored. Instead of creating the table dynamically during the
606 actual compiling pass, the information is remembered during the pre-compile
607 pass (on the stack unless there are more than 20 named groups, in which
608 case malloc() is used) and the whole table is created before the actual
609 compile happens. This has simplified the code (it is now nearly 150 lines
610 shorter) and prepared the way for better handling of references to groups
611 with duplicate names.
612
613 15. A back reference to a named subpattern when there is more than one of the
614 same name now checks them in the order in which they appear in the pattern.
615 The first one that is set is used for the reference. Previously only the
616 first one was inspected. This change makes PCRE more compatible with Perl.
617
618 16. Unicode character properties were updated from Unicode 6.3.0.
619
620 17. The compile-time code for auto-possessification has been refactored, based
621 on a patch by Zoltan Herczeg. It now happens after instead of during
622 compilation. The code is cleaner, and more cases are handled. The option
623 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
624 options in pcretest are provided to set it. It can also be set by
625 (*NO_AUTO_POSSESS) at the start of a pattern.
626
627 18. The character VT has been added to the default ("C" locale) set of
628 characters that match \s and are generally treated as white space,
629 following this same change in Perl 5.18. There is now no difference between
630 "Perl space" and "POSIX space". Whether VT is treated as white space in
631 other locales depends on the locale.
632
633 19. The code for checking named groups as conditions, either for being set or
634 for being recursed, has been refactored (this is related to 14 and 15
635 above). Processing unduplicated named groups should now be as fast at
636 numerical groups, and processing duplicated groups should be faster than
637 before.
638
639 20. Two patches to the CMake build system, by Alexander Barkov:
640
641 (1) Replace the "source" command by "." in CMakeLists.txt because
642 "source" is a bash-ism.
643
644 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
645 without these the CMake build does not work on Solaris.
646
647 21. Perl has changed its handling of \8 and \9. If there is no previously
648 encountered capturing group of those numbers, they are treated as the
649 literal characters 8 and 9 instead of a binary zero followed by the
650 literals. PCRE now does the same.
651
652 22. Following Perl, added \o{} to specify codepoints in octal, making it
653 possible to specify values greater than 0777 and also making them
654 unambiguous.
655
656 23. Perl now gives an error for missing closing braces after \x{... instead of
657 treating the string as literal. PCRE now does the same.
658
659 24. RunTest used to grumble if an inappropriate test was selected explicitly,
660 but just skip it when running all tests. This make it awkward to run ranges
661 of tests when one of them was inappropriate. Now it just skips any
662 inappropriate tests, as it always did when running all tests.
663
664 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
665 character types such as \d or \w, too many callouts were inserted, and the
666 data that they returned was rubbish.
667
668 26. In UCP mode, \s was not matching two of the characters that Perl matches,
669 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
670 were matched by \h. The code has now been refactored so that the lists of
671 the horizontal and vertical whitespace characters used for \h and \v (which
672 are defined only in one place) are now also used for \s.
673
674 27. Add JIT support for the 64 bit TileGX architecture.
675 Patch by Jiong Wang (Tilera Corporation).
676
677 28. Possessive quantifiers for classes (both explicit and automatically
678 generated) now use special opcodes instead of wrapping in ONCE brackets.
679
680 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
681 (because it's meaningless), this was not happening when PCRE_CASELESS was
682 set. Not wrong, but inefficient.
683
684 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
685 properties for \w, \d, etc) is present in a test regex. Otherwise if the
686 test contains no characters greater than 255, Perl doesn't realise it
687 should be using Unicode semantics.
688
689 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
690 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
691 does in Unicode mode.
692
693 32. Added the "forbid" facility to pcretest so that putting tests into the
694 wrong test files can sometimes be quickly detected.
695
696 33. There is now a limit (default 250) on the depth of nesting of parentheses.
697 This limit is imposed to control the amount of system stack used at compile
698 time. It can be changed at build time by --with-parens-nest-limit=xxx or
699 the equivalent in CMake.
700
701 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
702 errors. Perl warns for these when in warning mode, but PCRE has no facility
703 for giving warnings.
704
705 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
706 However, this was not working for (?!) because it is optimized to (*FAIL),
707 for which PCRE does not allow quantifiers. The optimization is now disabled
708 when a quantifier follows (?!). I can't see any use for this, but it makes
709 things uniform.
710
711 36. Perl no longer allows group names to start with digits, so I have made this
712 change also in PCRE. It simplifies the code a bit.
713
714 37. In extended mode, Perl ignores spaces before a + that indicates a
715 possessive quantifier. PCRE allowed a space before the quantifier, but not
716 before the possessive +. It now does.
717
718 38. The use of \K (reset reported match start) within a repeated possessive
719 group such as (a\Kb)*+ was not working.
720
721 40. Document that the same character tables must be used at compile time and
722 run time, and that the facility to pass tables to pcre_exec() and
723 pcre_dfa_exec() is for use only with saved/restored patterns.
724
725 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
726 features for Builds with MSVC:
727
728 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
729 builds). The .rc files can be used to set FileDescription and many other
730 attributes.
731
732 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
733 This allows higher-level build scripts which want .pdb files to avoid
734 hard-coding the exact files needed."
735
736 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
737 mean "start of word" and "end of word", respectively, as a transition aid.
738
739 43. A minimizing repeat of a class containing codepoints greater than 255 in
740 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
741 compiled to use the heap for recursion.
742
743 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
744 is configured.
745
746
747 Version 8.33 28-May-2013
748 ------------------------
749
750 1. Added 'U' to some constants that are compared to unsigned integers, to
751 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
752 variables that are added to signed variables, to ensure the result is
753 signed and can be negated.
754
755 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
756 CMake config files.
757
758 3. Revise the creation of config.h.generic so that all boolean macros are
759 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
760 overriding via -D on the command line possible.
761
762 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
763 to unsigned int is reported to make a quite noticeable speed difference in
764 a specific Windows environment. Testing on Linux did also appear to show
765 some benefit (and it is clearly not harmful). Also fixed the definition of
766 Xop which should be unsigned.
767
768 5. Related to (4), changing the definition of the intermediate variable cc
769 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
770 improvements.
771
772 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
773 unnecessary spaces.
774
775 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
776 and later.
777
778 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
779
780 9. Optimizing fast_forward_start_bits in JIT.
781
782 10. Adding support for callouts in JIT, and fixing some issues revealed
783 during this work. Namely:
784
785 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
786
787 (b) Minimum length was not checked before the matching is started.
788
789 11. The value of capture_last that is passed to callouts was incorrect in some
790 cases when there was a capture on one path that was subsequently abandoned
791 after a backtrack. Also, the capture_last value is now reset after a
792 recursion, since all captures are also reset in this case.
793
794 12. The interpreter no longer returns the "too many substrings" error in the
795 case when an overflowing capture is in a branch that is subsequently
796 abandoned after a backtrack.
797
798 13. In the pathological case when an offset vector of size 2 is used, pcretest
799 now prints out the matched string after a yield of 0 or 1.
800
801 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
802 Only supported by the JIT compiler at the moment.
803
804 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
805
806 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
807 offset of the starting point of the matching process, provided the offsets
808 vector is large enough.
809
810 17. The \A escape now records a lookbehind value of 1, though its execution
811 does not actually inspect the previous character. This is to ensure that,
812 in partial multi-segment matching, at least one character from the old
813 segment is retained when a new segment is processed. Otherwise, if there
814 are no lookbehinds in the pattern, \A might match incorrectly at the start
815 of a new segment.
816
817 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
818
819 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
820 gives some modest performance improvement in 8-bit mode.
821
822 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
823 be expressed in certain programming languages using Universal Character
824 Names.
825
826 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
827 which points out that "non characters" are not "characters that may not
828 appear in Unicode strings" but rather "characters that are reserved for
829 internal use and have only local meaning".
830
831 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
832 there was a conditional group that depended on an assertion, if the
833 assertion was false, the callout that immediately followed the alternation
834 in the condition was skipped when pcre_exec() was used for matching.
835
836 23. Allow an explicit callout to be inserted before an assertion that is the
837 condition for a conditional group, for compatibility with automatic
838 callouts, which always insert a callout at this point.
839
840 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
841 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
842
843 25. (*PRUNE) is now supported by the JIT compiler.
844
845 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
846
847 27. Fix the case where there are two or more SKIPs with arguments that may be
848 ignored.
849
850 28. (*SKIP) is now supported by the JIT compiler.
851
852 29. (*THEN) is now supported by the JIT compiler.
853
854 30. Update RunTest with additional test selector options.
855
856 31. The way PCRE handles backtracking verbs has been changed in two ways.
857
858 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
859 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
860 backtracking. In some cases this makes it more Perl-compatible, but Perl's
861 rather obscure rules do not always do the same thing.
862
863 (2) Previously, backtracking verbs were confined within assertions. This is
864 no longer the case for positive assertions, except for (*ACCEPT). Again,
865 this sometimes improves Perl compatibility, and sometimes does not.
866
867 32. A number of tests that were in test 2 because Perl did things differently
868 have been moved to test 1, because either Perl or PCRE has changed, and
869 these tests are now compatible.
870
871 32. Backtracking control verbs are now handled in the same way in JIT and
872 interpreter.
873
874 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
875 contained a forward subroutine reference caused a compile error.
876
877 34. Auto-detect and optimize limited repetitions in JIT.
878
879 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
880 blocking (*UTF) etc.
881
882 36. In the interpreter, maximizing pattern repetitions for characters and
883 character types now use tail recursion, which reduces stack usage.
884
885 37. The value of the max lookbehind was not correctly preserved if a compiled
886 and saved regex was reloaded on a host of different endianness.
887
888 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
889 of the compiled pattern block, expand the flags field from 16 to 32 bits
890 because it was almost full.
891
892 39. Try madvise first before posix_madvise.
893
894 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
895 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
896 original change locked out all matches of empty strings. This has been
897 changed so that one match of an empty string per line is recognized.
898 Subsequent searches on the same line (for colouring or for --only-matching,
899 for example) do not recognize empty strings.
900
901 41. Applied a user patch to fix a number of spelling mistakes in comments.
902
903 42. Data lines longer than 65536 caused pcretest to crash.
904
905 43. Clarified the data type for length and startoffset arguments for pcre_exec
906 and pcre_dfa_exec in the function-specific man pages, where they were
907 explicitly stated to be in bytes, never having been updated. I also added
908 some clarification to the pcreapi man page.
909
910 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
911 a segmentation fault.
912
913
914 Version 8.32 30-November-2012
915 -----------------------------
916
917 1. Improved JIT compiler optimizations for first character search and single
918 character iterators.
919
920 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
921 Patch by Daniel Richard G.
922
923 3. Single character iterator optimizations in the JIT compiler.
924
925 4. Improved JIT compiler optimizations for character ranges.
926
927 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
928 Reported by Giuseppe D'Angelo.
929
930 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
931 of a line, was being set incorrectly in cases where .* appeared inside
932 atomic brackets at the start of a pattern, or where there was a subsequent
933 *PRUNE or *SKIP.
934
935 7. Improved instruction cache flush for POWER/PowerPC.
936 Patch by Daniel Richard G.
937
938 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
939 grep:
940
941 (a) There is now no limit to the number of patterns to be matched.
942
943 (b) An error is given if a pattern is too long.
944
945 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
946 are now supported.
947
948 (d) --exclude-from and --include-from (multiple use) have been added.
949
950 (e) Exclusions and inclusions now apply to all files and directories, not
951 just to those obtained from scanning a directory recursively.
952
953 (f) Multiple uses of -f and --file-list are now supported.
954
955 (g) In a Windows environment, the default for -d has been changed from
956 "read" (the GNU grep default) to "skip", because otherwise the presence
957 of a directory in the file list provokes an error.
958
959 (h) The documentation has been revised and clarified in places.
960
961 9. Improve the matching speed of capturing brackets.
962
963 10. Changed the meaning of \X so that it now matches a Unicode extended
964 grapheme cluster.
965
966 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
967 out POSIX threads when JIT support is configured.
968
969 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
970
971 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
972 structure could go wrong in environments where size_t is not the same size
973 as int.
974
975 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
976 set.
977
978 15. The EBCDIC support had decayed; later updates to the code had included
979 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
980 general tidy up of EBCDIC-related issues, and the documentation was also
981 not quite right. There is now a test that can be run on ASCII systems to
982 check some of the EBCDIC-related things (but is it not a full test).
983
984 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
985 in a small tidy to the code.
986
987 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
988
989 18. If the --only-matching (-o) option in pcregrep is specified multiple
990 times, each one causes appropriate output. For example, -o1 -o2 outputs the
991 substrings matched by the 1st and 2nd capturing parentheses. A separating
992 string can be specified by --om-separator (default empty).
993
994 19. Improving the first n character searches.
995
996 20. Turn case lists for horizontal and vertical white space into macros so that
997 they are defined only once.
998
999 21. This set of changes together give more compatible Unicode case-folding
1000 behaviour for characters that have more than one other case when UCP
1001 support is available.
1002
1003 (a) The Unicode property table now has offsets into a new table of sets of
1004 three or more characters that are case-equivalent. The MultiStage2.py
1005 script that generates these tables (the pcre_ucd.c file) now scans
1006 CaseFolding.txt instead of UnicodeData.txt for character case
1007 information.
1008
1009 (b) The code for adding characters or ranges of characters to a character
1010 class has been abstracted into a generalized function that also handles
1011 case-independence. In UTF-mode with UCP support, this uses the new data
1012 to handle characters with more than one other case.
1013
1014 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1015 whose other case is greater than 256 are now correctly matched
1016 caselessly. Previously, the high codepoint matched the low one, but not
1017 vice versa.
1018
1019 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1020 of the new class addition function, using character lists defined as
1021 macros alongside the case definitions of 20 above.
1022
1023 (e) Caseless back references now work with characters that have more than
1024 one other case.
1025
1026 (f) General caseless matching of characters with more than one other case
1027 is supported.
1028
1029 22. Unicode character properties were updated from Unicode 6.2.0
1030
1031 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1032
1033 24. Add support for 32-bit character strings, and UTF-32
1034
1035 25. Major JIT compiler update (code refactoring and bugfixing).
1036 Experimental Sparc 32 support is added.
1037
1038 26. Applied a modified version of Daniel Richard G's patch to create
1039 pcre.h.generic and config.h.generic by "make" instead of in the
1040 PrepareRelease script.
1041
1042 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1043 pcre_compile.c when checking for a zero character.
1044
1045 28. Introducing a native interface for JIT. Through this interface, the compiled
1046 machine code can be directly executed. The purpose of this interface is to
1047 provide fast pattern matching, so several sanity checks are not performed.
1048 However, feature tests are still performed. The new interface provides
1049 1.4x speedup compared to the old one.
1050
1051 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1052 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1053 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1054
1055 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1056 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1057 the "old" RFC 2279). Instead, it ended up passing a negative length to
1058 pcre_exec().
1059
1060 31. Add support for GCC's visibility feature to hide internal functions.
1061
1062 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1063 "unknown -C option" after outputting 0 or 1.
1064
1065 33. There is now support for generating a code coverage report for the test
1066 suite in environments where gcc is the compiler and lcov is installed. This
1067 is mainly for the benefit of the developers.
1068
1069 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1070 unaddressable using valgrind annotations, allowing valgrind to detect
1071 invalid memory accesses. This is mainly for the benefit of the developers.
1072
1073 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1074
1075 26. Give configure error if --enable-cpp but no C++ compiler found.
1076
1077
1078 Version 8.31 06-July-2012
1079 -------------------------
1080
1081 1. Fixing a wrong JIT test case and some compiler warnings.
1082
1083 2. Removed a bashism from the RunTest script.
1084
1085 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1086 to unsigned type, result still unsigned" that was given by an MS compiler
1087 on encountering the code "-sizeof(xxx)".
1088
1089 4. Partial matching support is added to the JIT compiler.
1090
1091 5. Fixed several bugs concerned with partial matching of items that consist
1092 of more than one character:
1093
1094 (a) /^(..)\1/ did not partially match "aba" because checking references was
1095 done on an "all or nothing" basis. This also applied to repeated
1096 references.
1097
1098 (b) \R did not give a hard partial match if \r was found at the end of the
1099 subject.
1100
1101 (c) \X did not give a hard partial match after matching one or more
1102 characters at the end of the subject.
1103
1104 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1105 a partial match for the string "\r".
1106
1107 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1108 a partial match for a CR character at the end of the subject string.
1109
1110 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1111 running pcretest, the text "(JIT)" added to the output whenever JIT is
1112 actually used to run the match.
1113
1114 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1115 or /S+[+] with a digit between 1 and 7.
1116
1117 8. OP_NOT now supports any UTF character not just single-byte ones.
1118
1119 9. (*MARK) control verb is now supported by the JIT compiler.
1120
1121 10. The command "./RunTest list" lists the available tests without actually
1122 running any of them. (Because I keep forgetting what they all are.)
1123
1124 11. Add PCRE_INFO_MAXLOOKBEHIND.
1125
1126 12. Applied a (slightly modified) user-supplied patch that improves performance
1127 when the heap is used for recursion (compiled with --disable-stack-for-
1128 recursion). Instead of malloc and free for each heap frame each time a
1129 logical recursion happens, frames are retained on a chain and re-used where
1130 possible. This sometimes gives as much as 30% improvement.
1131
1132 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1133 call.
1134
1135 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1136
1137 15. It is now possible to link pcretest with libedit as an alternative to
1138 libreadline.
1139
1140 16. (*COMMIT) control verb is now supported by the JIT compiler.
1141
1142 17. The Unicode data tables have been updated to Unicode 6.1.0.
1143
1144 18. Added --file-list option to pcregrep.
1145
1146 19. Added binary file support to pcregrep, including the -a, --binary-files,
1147 -I, and --text options.
1148
1149 20. The madvise function is renamed for posix_madvise for QNX compatibility
1150 reasons. Fixed by Giuseppe D'Angelo.
1151
1152 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1153
1154 22. Changed the option for creating symbolic links for 16-bit man pages from
1155 -s to -sf so that re-installing does not cause issues.
1156
1157 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1158
1159 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1160 matches in certain environments (the workspace was not being correctly
1161 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1162 some of the workspace data at the beginning of a restart.
1163
1164 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1165 was not doing so when it should - probably a typo introduced by SVN 528
1166 (change 8.10/14).
1167
1168 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1169 \w+ when the character tables indicated that \x{c4} was a word character.
1170 There were several related cases, all because the tests for doing a table
1171 lookup were testing for characters less than 127 instead of 255.
1172
1173 27. If a pattern contains capturing parentheses that are not used in a match,
1174 their slots in the ovector are set to -1. For those that are higher than
1175 any matched groups, this happens at the end of processing. In the case when
1176 there were back references that the ovector was too small to contain
1177 (causing temporary malloc'd memory to be used during matching), and the
1178 highest capturing number was not used, memory off the end of the ovector
1179 was incorrectly being set to -1. (It was using the size of the temporary
1180 memory instead of the true size.)
1181
1182 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1183 ovector size, it uses memory at the end of the block that it has got.
1184
1185 29. Check for an overlong MARK name and give an error at compile time. The
1186 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1187
1188 30. JIT compiler update.
1189
1190 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1191 Rill for the patch.
1192
1193 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1194
1195 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1196
1197 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1198 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1199 was enabled.
1200
1201 35. Improve JIT code generation for greedy plus quantifier.
1202
1203 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1204 1 to "aa" instead of to an empty string. The bug affected repeated groups
1205 that could potentially match an empty string.
1206
1207 37. Optimizing single character iterators in JIT.
1208
1209 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1210 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1211 codepoints that are too big for the mode are faulted, and in a UTF mode,
1212 disallowed codepoints are also faulted.
1213
1214 39. If PCRE was compiled with UTF support, in three places in the DFA
1215 matcher there was code that should only have been obeyed in UTF mode, but
1216 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1217 processing when bytes with values greater than 127 were present. In 16-bit
1218 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1219 both cases the values are those that cannot be the first data item in a UTF
1220 character. The three items that might have provoked this were recursions,
1221 possessively repeated groups, and atomic groups.
1222
1223 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1224 and pcregrep, because some OS require shared objects to be explicitly
1225 passed to ld, causing the link step to fail if they are not.
1226
1227 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1228 mode, patterns that started with \h* or \R* might be incorrectly matched.
1229
1230
1231 Version 8.30 04-February-2012
1232 -----------------------------
1233
1234 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1235 name is defined in ctype.h.
1236
1237 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1238 only in quite long subpatterns.
1239
1240 3. Removed the function pcre_info(), which has been obsolete and deprecated
1241 since it was replaced by pcre_fullinfo() in February 2000.
1242
1243 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1244 match a (*MARK), and the match failed at the start of the subject, a
1245 reference to memory before the start of the subject could occur. This bug
1246 was introduced by fix 17 of release 8.21.
1247
1248 5. A reference to an unset group with zero minimum repetition was giving
1249 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1250 /(another)?(\1?)test/ matched against "hello world test". This bug was
1251 introduced in release 8.13.
1252
1253 6. Add support for 16-bit character strings (a large amount of work involving
1254 many changes and refactorings).
1255
1256 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1257 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1258 from a file.
1259
1260 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1261 rounding is not applied in this particular case).
1262
1263 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1264 if they appear, or are escaped, in patterns.
1265
1266 10. Get rid of a number of -Wunused-but-set-variable warnings.
1267
1268 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1269 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1270 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1271 also returns the mark "x". This bug applied to capturing parentheses,
1272 non-capturing parentheses, and atomic parentheses. It also applied to some
1273 assertions.
1274
1275 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1276 information out of configure.ac instead of relying on pcre.h.generic, which
1277 is not stored in the repository.
1278
1279 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1280 -lreadline.
1281
1282 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1283
1284 15. Applied Graycode's patch to put the top-level frame on the stack rather
1285 than the heap when not using the stack for recursion. This gives a
1286 performance improvement in many cases when recursion is not deep.
1287
1288 16. Experimental code added to "pcretest -C" to output the stack frame size.
1289
1290
1291 Version 8.21 12-Dec-2011
1292 ------------------------
1293
1294 1. Updating the JIT compiler.
1295
1296 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1297 are added as well.
1298
1299 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1300 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1301 calling _pcre_jit_exec. Some extra comments are added.
1302
1303 4. (*MARK) settings inside atomic groups that do not contain any capturing
1304 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1305 was introduced by change 18 for 8.20.
1306
1307 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1308 ECMA-262 standard.
1309
1310 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1311 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1312 This bug was probably introduced by change 9 of 8.13.
1313
1314 7. While fixing 6 above, I noticed that a number of other items were being
1315 incorrectly rejected as "not fixed length". This arose partly because newer
1316 opcodes had not been added to the fixed-length checking code. I have (a)
1317 corrected the bug and added tests for these items, and (b) arranged for an
1318 error to occur if an unknown opcode is encountered while checking for fixed
1319 length instead of just assuming "not fixed length". The items that were
1320 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1321 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1322 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1323
1324 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1325 being incorrectly compiled and would have given unpredicatble results.
1326
1327 9. A possessively repeated subpattern with minimum repeat count greater than
1328 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1329 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1330 the first (A) could occur when it should not.
1331
1332 10. Add a cast and remove a redundant test from the code.
1333
1334 11. JIT should use pcre_malloc/pcre_free for allocation.
1335
1336 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1337 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1338 is anything other than /usr, -L is still shown).
1339
1340 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1341
1342 14. Perl does not support \N without a following name in a [] class; PCRE now
1343 also gives an error.
1344
1345 15. If a forward reference was repeated with an upper limit of around 2000,
1346 it caused the error "internal error: overran compiling workspace". The
1347 maximum number of forward references (including repeats) was limited by the
1348 internal workspace, and dependent on the LINK_SIZE. The code has been
1349 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1350 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1351 of around 200,000 forward references. While doing this, I also speeded up
1352 the filling in of repeated forward references.
1353
1354 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1355 incorrectly expecting the subject to contain another "a" after the start.
1356
1357 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1358 in the match, the SKIP should be ignored. This was not happening; instead
1359 the SKIP was being treated as NOMATCH. For patterns such as
1360 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1361 tested.
1362
1363 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1364 now much more compatible with Perl, in particular in cases where the result
1365 is a non-match for a non-anchored pattern. For example, if
1366 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1367 "m", where previously it did not return a name. A side effect of this
1368 change is that for partial matches, the last encountered mark name is
1369 returned, as for non matches. A number of tests that were previously not
1370 Perl-compatible have been moved into the Perl-compatible test files. The
1371 refactoring has had the pleasing side effect of removing one argument from
1372 the match() function, thus reducing its stack requirements.
1373
1374 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1375 subsequent uses of /S (without +) incorrectly behaved like /S+.
1376
1377 21. Retrieve executable code size support for the JIT compiler and fixing
1378 some warnings.
1379
1380 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1381 not work when the shorter character appeared right at the end of the
1382 subject string.
1383
1384 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1385 systems.
1386
1387 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1388 output it when the /M option is used in pcretest.
1389
1390 25. The CheckMan script was not being included in the distribution. Also, added
1391 an explicit "perl" to run Perl scripts from the PrepareRelease script
1392 because this is reportedly needed in Windows.
1393
1394 26. If study data was being save in a file and studying had not found a set of
1395 "starts with" bytes for the pattern, the data written to the file (though
1396 never used) was taken from uninitialized memory and so caused valgrind to
1397 complain.
1398
1399 27. Updated RunTest.bat as provided by Sheri Pierce.
1400
1401 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1402
1403 29. Computation of memory usage for the table of capturing group names was
1404 giving an unnecessarily large value.
1405
1406
1407 Version 8.20 21-Oct-2011
1408 ------------------------
1409
1410 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1411 a POSIX class. After further experiments with Perl, which convinced me that
1412 Perl has bugs and confusions, a closing square bracket is no longer allowed
1413 in a POSIX name. This bug also affected patterns with classes that started
1414 with full stops.
1415
1416 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1417 captured substring, but while checking the failing first alternative,
1418 substring 1 is temporarily captured. If the output vector supplied to
1419 pcre_exec() was not big enough for this capture, the yield of the function
1420 was still zero ("insufficient space for captured substrings"). This cannot
1421 be totally fixed without adding another stack variable, which seems a lot
1422 of expense for a edge case. However, I have improved the situation in cases
1423 such as /(a)(b)x|abc/ matched against "abc", where the return code
1424 indicates that fewer than the maximum number of slots in the ovector have
1425 been set.
1426
1427 3. Related to (2) above: when there are more back references in a pattern than
1428 slots in the output vector, pcre_exec() uses temporary memory during
1429 matching, and copies in the captures as far as possible afterwards. It was
1430 using the entire output vector, but this conflicts with the specification
1431 that only 2/3 is used for passing back captured substrings. Now it uses
1432 only the first 2/3, for compatibility. This is, of course, another edge
1433 case.
1434
1435 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1436 main code base, and can be used by building with --enable-jit. When this is
1437 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1438 runtime --no-jit option is given.
1439
1440 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1441 ovector, the return from the function was zero, implying that there were
1442 other matches that did not fit. The correct "exactly full" value is now
1443 returned.
1444
1445 6. If a subpattern that was called recursively or as a subroutine contained
1446 (*PRUNE) or any other control that caused it to give a non-standard return,
1447 invalid errors such as "Error -26 (nested recursion at the same subject
1448 position)" or even infinite loops could occur.
1449
1450 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1451 computing the minimum length on reaching *ACCEPT, and so ended up with the
1452 wrong value of 1 rather than 0. Further investigation indicates that
1453 computing a minimum subject length in the presence of *ACCEPT is difficult
1454 (think back references, subroutine calls), and so I have changed the code
1455 so that no minimum is registered for a pattern that contains *ACCEPT.
1456
1457 8. If (*THEN) was present in the first (true) branch of a conditional group,
1458 it was not handled as intended. [But see 16 below.]
1459
1460 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1461 Sheri Pierce.
1462
1463 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1464 the first byte in a match must be "a".
1465
1466 11. Change 17 for 8.13 increased the recursion depth for patterns like
1467 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1468 pattern contains any instances of (*THEN). If it does not, the old
1469 optimizations are restored. It would be nice to do this on a per-group
1470 basis, but at the moment that is not feasible.
1471
1472 12. In some environments, the output of pcretest -C is CRLF terminated. This
1473 broke RunTest's code that checks for the link size. A single white space
1474 character after the value is now allowed for.
1475
1476 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1477 For "fr", it uses the Windows-specific input and output files.
1478
1479 14. If (*THEN) appeared in a group that was called recursively or as a
1480 subroutine, it did not work as intended. [But see next item.]
1481
1482 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1483 pattern fragments (but not containing any | characters). If A and B are
1484 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1485 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1486 D. In other words, Perl considers parentheses that do not contain any |
1487 characters to be part of a surrounding alternative, whereas PCRE was
1488 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1489 differently. PCRE now behaves in the same way as Perl, except in the case
1490 of subroutine/recursion calls such as (?1) which have in any case always
1491 been different (but PCRE had them first :-).
1492
1493 16. Related to 15 above: Perl does not treat the | in a conditional group as
1494 creating alternatives. Such a group is treated in the same way as an
1495 ordinary group without any | characters when processing (*THEN). PCRE has
1496 been changed to match Perl's behaviour.
1497
1498 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1499 RunGrepTest script failed.
1500
1501 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1502 inevitable for groups that contain captures, but it can lead to a lot of
1503 stack use in large patterns. The old behaviour has been restored for atomic
1504 groups that do not contain any capturing parentheses.
1505
1506 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1507 suppress the check for a minimum subject length at run time. (If it was
1508 given to pcre_exec() or pcre_dfa_exec() it did work.)
1509
1510 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1511 fail to work when decoding hex characters in data strings in EBCDIC
1512 environments.
1513
1514 21. It appears that in at least one Mac OS environment, the isxdigit() function
1515 is implemented as a macro that evaluates to its argument more than once,
1516 contravening the C 90 Standard (I haven't checked a later standard). There
1517 was an instance in pcretest which caused it to go wrong when processing
1518 \x{...} escapes in subject strings. The has been rewritten to avoid using
1519 things like p++ in the argument of isxdigit().
1520
1521
1522 Version 8.13 16-Aug-2011
1523 ------------------------
1524
1525 1. The Unicode data tables have been updated to Unicode 6.0.0.
1526
1527 2. Two minor typos in pcre_internal.h have been fixed.
1528
1529 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1530 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1531 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1532
1533 4. There were a number of related bugs in the code for matching backrefences
1534 caselessly in UTF-8 mode when codes for the characters concerned were
1535 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1536 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1537 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1538 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1539 2-byte code at the end of the subject (it thought there wasn't enough data
1540 left).
1541
1542 5. Comprehensive information about what went wrong is now returned by
1543 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1544 as the output vector has at least 2 elements. The offset of the start of
1545 the failing character and a reason code are placed in the vector.
1546
1547 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1548 now returned is for the first byte of the failing character, instead of the
1549 last byte inspected. This is an incompatible change, but I hope it is small
1550 enough not to be a problem. It makes the returned offset consistent with
1551 pcre_exec() and pcre_dfa_exec().
1552
1553 7. pcretest now gives a text phrase as well as the error number when
1554 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1555 failure, the offset and reason code are output.
1556
1557 8. When \R was used with a maximizing quantifier it failed to skip backwards
1558 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1559 back over a single character (\n). This seems wrong (because it treated the
1560 two characters as a single entity when going forwards), conflicts with the
1561 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1562 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1563 has been changed.
1564
1565 9. Some internal refactoring has changed the processing so that the handling
1566 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1567 time (the PCRE_DOTALL option was changed this way some time ago: version
1568 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1569 which was always a bit of a fudge. It also means that there is one less
1570 argument for the match() function, which reduces its stack requirements
1571 slightly. This change also fixes an incompatibility with Perl: the pattern
1572 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1573
1574 10. More internal refactoring has drastically reduced the number of recursive
1575 calls to match() for possessively repeated groups such as (abc)++ when
1576 using pcre_exec().
1577
1578 11. While implementing 10, a number of bugs in the handling of groups were
1579 discovered and fixed:
1580
1581 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1582 (a|)*(?1) gave a compile-time internal error.
1583 ((a|)+)+ did not notice that the outer group could match an empty string.
1584 (^a|^)+ was not marked as anchored.
1585 (.*a|.*)+ was not marked as matching at start or after a newline.
1586
1587 12. Yet more internal refactoring has removed another argument from the match()
1588 function. Special calls to this function are now indicated by setting a
1589 value in a variable in the "match data" data block.
1590
1591 13. Be more explicit in pcre_study() instead of relying on "default" for
1592 opcodes that mean there is no starting character; this means that when new
1593 ones are added and accidentally left out of pcre_study(), testing should
1594 pick them up.
1595
1596 14. The -s option of pcretest has been documented for ages as being an old
1597 synonym of -m (show memory usage). I have changed it to mean "force study
1598 for every regex", that is, assume /S for every regex. This is similar to -i
1599 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1600 using it. It makes it easier to run collections of tests with and without
1601 study enabled, and thereby test pcre_study() more easily. All the standard
1602 tests are now run with and without -s (but some patterns can be marked as
1603 "never study" - see 20 below).
1604
1605 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1606 restoration of the capturing data to the outer values was not happening
1607 correctly.
1608
1609 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1610 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1611 pattern had matched an empty string, and so incorrectly returned a no
1612 match.
1613
1614 17. There was optimizing code for the last branch of non-capturing parentheses,
1615 and also for the obeyed branch of a conditional subexpression, which used
1616 tail recursion to cut down on stack usage. Unfortunately, now that there is
1617 the possibility of (*THEN) occurring in these branches, tail recursion is
1618 no longer possible because the return has to be checked for (*THEN). These
1619 two optimizations have therefore been removed. [But see 8.20/11 above.]
1620
1621 18. If a pattern containing \R was studied, it was assumed that \R always
1622 matched two bytes, thus causing the minimum subject length to be
1623 incorrectly computed because \R can also match just one byte.
1624
1625 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1626 was incorrectly computed.
1627
1628 20. If /S is present twice on a test pattern in pcretest input, it now
1629 *disables* studying, thereby overriding the use of -s on the command line
1630 (see 14 above). This is necessary for one or two tests to keep the output
1631 identical in both cases.
1632
1633 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1634 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1635
1636 22. When an atomic group that contained a capturing parenthesis was
1637 successfully matched, but the branch in which it appeared failed, the
1638 capturing was not being forgotten if a higher numbered group was later
1639 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1640 group 1 to "a", when in fact it should be unset. This applied to multi-
1641 branched capturing and non-capturing groups, repeated or not, and also to
1642 positive assertions (capturing in negative assertions does not happen
1643 in PCRE) and also to nested atomic groups.
1644
1645 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1646 subject after a captured substring, to make it easier to tell which of a
1647 number of identical substrings has been captured.
1648
1649 24. The way atomic groups are processed by pcre_exec() has been changed so that
1650 if they are repeated, backtracking one repetition now resets captured
1651 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1652 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1653 "aaa". Previously, it would have been "a". As part of this code
1654 refactoring, the way recursive calls are handled has also been changed.
1655
1656 25. If an assertion condition captured any substrings, they were not passed
1657 back unless some other capturing happened later. For example, if
1658 (?(?=(a))a) was matched against "a", no capturing was returned.
1659
1660 26. When studying a pattern that contained subroutine calls or assertions,
1661 the code for finding the minimum length of a possible match was handling
1662 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1663 group 1 called group 2 while simultaneously a separate group 2 called group
1664 1). A stack overflow occurred in this case. I have fixed this by limiting
1665 the recursion depth to 10.
1666
1667 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1668 Fortmann. This supports explicit test numbers on the command line, and has
1669 argument validation and error reporting.
1670
1671 28. An instance of \X with an unlimited repeat could fail if at any point the
1672 first character it looked at was a mark character.
1673
1674 29. Some minor code refactoring concerning Unicode properties and scripts
1675 should reduce the stack requirement of match() slightly.
1676
1677 30. Added the '=' option to pcretest to check the setting of unused capturing
1678 slots at the end of the pattern, which are documented as being -1, but are
1679 not included in the return count.
1680
1681 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1682 compiled something random. Now it gives a compile-time error (as does
1683 Perl).
1684
1685 32. A *MARK encountered during the processing of a positive assertion is now
1686 recorded and passed back (compatible with Perl).
1687
1688 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1689 had alternative anchored branches, the search for a second match in a line
1690 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1691 matched the line "0102" twice. The same bug affected patterns that started
1692 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1693 twice.
1694
1695 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1696 does, and because of capturing effects, quantifying parenthesized
1697 assertions may at times be useful. Quantifiers are now allowed for
1698 parenthesized assertions.
1699
1700 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1701
1702 36. \g was being checked for fancy things in a character class, when it should
1703 just be a literal "g".
1704
1705 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1706 appearance of a nested POSIX class supersedes an apparent external class.
1707 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1708 unescaped square brackets may also appear as part of class names. For
1709 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1710 more like Perl. (But see 8.20/1 above.)
1711
1712 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1713 was because it thought it was \N{name}, which is not supported).
1714
1715 39. Add minix to OS list not supporting the -S option in pcretest.
1716
1717 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1718 cannot analyze patterns in sufficient detail to catch mutual recursions
1719 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1720 subgroup is called recursively as a subpattern for a second time at the
1721 same position in the subject string. In previous releases this might have
1722 been caught by the recursion limit, or it might have run out of stack.
1723
1724 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1725 happen only once. PCRE was, however incorrectly giving a compile time error
1726 "recursive call could loop indefinitely" because it cannot analyze the
1727 pattern in sufficient detail. The compile time test no longer happens when
1728 PCRE is compiling a conditional subpattern, but actual runaway loops are
1729 now caught at runtime (see 40 above).
1730
1731 42. It seems that Perl allows any characters other than a closing parenthesis
1732 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1733 has been changed to be the same.
1734
1735 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1736 as not to get warnings when autogen.sh is called. Also changed
1737 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1738
1739 44. To help people who use pcregrep to scan files containing exceedingly long
1740 lines, the following changes have been made:
1741
1742 (a) The default value of the buffer size parameter has been increased from
1743 8K to 20K. (The actual buffer used is three times this size.)
1744
1745 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1746 PCRE is built.
1747
1748 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1749 to be set at run time.
1750
1751 (d) Numerical values in pcregrep options can be followed by K or M, for
1752 example --buffer-size=50K.
1753
1754 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1755 given and the return code is set to 2.
1756
1757 45. Add a pointer to the latest mark to the callout data block.
1758
1759 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1760 partial match of an empty string instead of no match. This was specific to
1761 the use of ".".
1762
1763 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1764 complete match instead of a partial match. This bug was dependent on both
1765 the PCRE_UTF8 and PCRE_DOTALL options being set.
1766
1767 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1768 starting byte set, because \b was not being ignored.
1769
1770
1771 Version 8.12 15-Jan-2011
1772 ------------------------
1773
1774 1. Fixed some typos in the markup of the man pages, and wrote a script that
1775 checks for such things as part of the documentation building process.
1776
1777 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1778 --match-limit and --recursion-limit options (added for 8.11). In
1779 particular, this made one of the standard tests fail. (The integer value
1780 went into the wrong half of a long int.)
1781
1782 3. If the --colour option was given to pcregrep with -v (invert match), it
1783 did strange things, either producing crazy output, or crashing. It should,
1784 of course, ignore a request for colour when reporting lines that do not
1785 match.
1786
1787 4. Another pcregrep bug caused similar problems if --colour was specified with
1788 -M (multiline) and the pattern match finished with a line ending.
1789
1790 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1791 matched in multiline mode, the following line was shown as part of the
1792 match. This seems wrong, so I have changed it.
1793
1794 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1795 the check for further matches in the same line (so they could be coloured)
1796 to overrun the end of the current line. If another match was found, it was
1797 incorrectly shown (and then shown again when found in the next line).
1798
1799 7. If pcregrep was compiled under Windows, there was a reference to the
1800 function pcregrep_exit() before it was defined. I am assuming this was
1801 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1802 reported by a user. I've moved the definition above the reference.
1803
1804
1805 Version 8.11 10-Dec-2010
1806 ------------------------
1807
1808 1. (*THEN) was not working properly if there were untried alternatives prior
1809 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1810 backtracked to try for "b" instead of moving to the next alternative branch
1811 at the same level (in this case, to look for "c"). The Perl documentation
1812 is clear that when (*THEN) is backtracked onto, it goes to the "next
1813 alternative in the innermost enclosing group".
1814
1815 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1816 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1817 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1818 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1819 (*THEN).
1820
1821 3. If \s appeared in a character class, it removed the VT character from
1822 the class, even if it had been included by some previous item, for example
1823 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1824 of \s, but is part of the POSIX "space" class.)
1825
1826 4. A partial match never returns an empty string (because you can always
1827 match an empty string at the end of the subject); however the checking for
1828 an empty string was starting at the "start of match" point. This has been
1829 changed to the "earliest inspected character" point, because the returned
1830 data for a partial match starts at this character. This means that, for
1831 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1832 (previously it gave "no match").
1833
1834 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1835 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1836 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1837 has an implication that the given string is incomplete (because a partial
1838 match is preferred over a full match). For this reason, these items now
1839 give a partial match in this situation. [Aside: previously, the one case
1840 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1841 match rather than a full match, which was wrong by the old rules, but is
1842 now correct.]
1843
1844 6. There was a bug in the handling of #-introduced comments, recognized when
1845 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1846 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1847 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1848 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1849 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1850 places in pcre_compile().
1851
1852 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1853 comments when looking ahead for named forward references to subpatterns,
1854 the only newline sequence it recognized was NL. It now handles newlines
1855 according to the set newline convention.
1856
1857 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1858 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1859 cater for a lack of strerror(). These oversights have been fixed.
1860
1861 9. Added --match-limit and --recursion-limit to pcregrep.
1862
1863 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1864
1865 11. When the -o option was used, pcregrep was setting a return code of 1, even
1866 when matches were found, and --line-buffered was not being honoured.
1867
1868 12. Added an optional parentheses number to the -o and --only-matching options
1869 of pcregrep.
1870
1871 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1872 can match an empty string. The code to do it in pcretest and pcredemo
1873 needed fixing:
1874
1875 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1876 only one byte after an empty string match just before CRLF (this case
1877 just got forgotten; "any" and "anycrlf" were OK).
1878
1879 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1880 mode when an empty string match preceded an ASCII character followed by
1881 a non-ASCII character. (The code for advancing by one character rather
1882 than one byte was nonsense.)
1883
1884 (c) The pcredemo.c sample program did not have any code at all to handle
1885 the cases when CRLF is a valid newline sequence.
1886
1887 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1888 as a starting offset was within the subject string. There is now a new
1889 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1890 negative or greater than the length of the string. In order to test this,
1891 pcretest is extended to allow the setting of negative starting offsets.
1892
1893 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1894 starting offset points to the beginning of a UTF-8 character was
1895 unnecessarily clumsy. I tidied it up.
1896
1897 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1898 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1899
1900 17. Nobody had reported that the --include_dir option, which was added in
1901 release 7.7 should have been called --include-dir (hyphen, not underscore)
1902 for compatibility with GNU grep. I have changed it to --include-dir, but
1903 left --include_dir as an undocumented synonym, and the same for
1904 --exclude-dir, though that is not available in GNU grep, at least as of
1905 release 2.5.4.
1906
1907 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1908 characters from a string of bytes) have been redefined so as not to use
1909 loops, in order to improve performance in some environments. At the same
1910 time, I abstracted some of the common code into auxiliary macros to save
1911 repetition (this should not affect the compiled code).
1912
1913 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1914 compile-time error is now given if \c is not followed by an ASCII
1915 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1916 different, and any byte value is allowed.)
1917
1918 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1919 START_OPTIMIZE option, which is now allowed at compile time - but just
1920 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1921 to pcregrep and other applications that have no direct access to PCRE
1922 options. The new /Y option in pcretest sets this option when calling
1923 pcre_compile().
1924
1925 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1926 back references. Groups containing recursive back references were forced to
1927 be atomic by that change, but in the case of named groups, the amount of
1928 memory required was incorrectly computed, leading to "Failed: internal
1929 error: code overflow". This has been fixed.
1930
1931 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1932 pcretest.c, to avoid build problems in some Borland environments.
1933
1934
1935 Version 8.10 25-Jun-2010
1936 ------------------------
1937
1938 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1939 THEN.
1940
1941 2. (*ACCEPT) was not working when inside an atomic group.
1942
1943 3. Inside a character class, \B is treated as a literal by default, but
1944 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1945 causes the error). The code is unchanged, but I tidied the documentation.
1946
1947 4. Inside a character class, PCRE always treated \R and \X as literals,
1948 whereas Perl faults them if its -w option is set. I have changed PCRE so
1949 that it faults them when PCRE_EXTRA is set.
1950
1951 5. Added support for \N, which always matches any character other than
1952 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1953
1954 6. When compiling pcregrep with newer versions of gcc which may have
1955 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1956 declared with attribute warn_unused_result" were given. Just casting the
1957 result to (void) does not stop the warnings; a more elaborate fudge is
1958 needed. I've used a macro to implement this.
1959
1960 7. Minor change to pcretest.c to avoid a compiler warning.
1961
1962 8. Added four artifical Unicode properties to help with an option to make
1963 \s etc use properties (see next item). The new properties are: Xan
1964 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1965
1966 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1967 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1968 this option. Modified pcretest to add /W to test this facility. Added
1969 REG_UCP to make it available via the POSIX interface.
1970
1971 10. Added --line-buffered to pcregrep.
1972
1973 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1974 studied, and the match started with a letter with a code point greater than
1975 127 whose first byte was different to the first byte of the other case of
1976 the letter, the other case of this starting letter was not recognized
1977 (#976).
1978
1979 12. If a pattern that was studied started with a repeated Unicode property
1980 test, for example, \p{Nd}+, there was the theoretical possibility of
1981 setting up an incorrect bitmap of starting bytes, but fortunately it could
1982 not have actually happened in practice until change 8 above was made (it
1983 added property types that matched character-matching opcodes).
1984
1985 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1986 possible starting bytes for non-anchored patterns.
1987
1988 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1989 \R, and also a number of cases that involve Unicode properties, both
1990 explicit and implicit when PCRE_UCP is set.
1991
1992 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1993 input, it could crash or give wrong results if characters with values
1994 greater than 0xc0 were present in the subject string. (Detail: it assumed
1995 UTF-8 input when processing these items.)
1996
1997 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1998 size_t is 64-bit (#991).
1999
2000 17. Added a check for running out of memory when PCRE is compiled with
2001 --disable-stack-for-recursion (#990).
2002
2003 18. If the last data line in a file for pcretest does not have a newline on
2004 the end, a newline was missing in the output.
2005
2006 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2007 less than 128) in its various bitmaps. However, there is a facility for
2008 generating tables according to the current locale when PCRE is compiled. It
2009 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2010 characters, are recognized by isspace() and therefore were getting set in
2011 these tables, and indeed these tables seem to approximate to ISO 8859. This
2012 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2013 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2014 which of course cannot start UTF-8 characters. I have changed the code so
2015 that only real ASCII characters (less than 128) and the correct starting
2016 bytes for UTF-8 encodings are set for characters greater than 127 when in
2017 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2018 altogether.)
2019
2020 20. Added the /T option to pcretest so as to be able to run tests with non-
2021 standard character tables, thus making it possible to include the tests
2022 used for 19 above in the standard set of tests.
2023
2024 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2025 reference to a subpattern the other side of a comment that contains an
2026 opening parenthesis caused either an internal compiling error, or a
2027 reference to the wrong subpattern.
2028
2029
2030 Version 8.02 19-Mar-2010
2031 ------------------------
2032
2033 1. The Unicode data tables have been updated to Unicode 5.2.0.
2034
2035 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2036 configured.
2037
2038 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2039 original author of that file, following a query about its status.
2040
2041 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2042 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2043
2044 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2045 quantifier applied to a forward-referencing subroutine call, could compile
2046 incorrect code or give the error "internal error: previously-checked
2047 referenced subpattern not found".
2048
2049 6. Both MS Visual Studio and Symbian OS have problems with initializing
2050 variables to point to external functions. For these systems, therefore,
2051 pcre_malloc etc. are now initialized to local functions that call the
2052 relevant global functions.
2053
2054 7. There were two entries missing in the vectors called coptable and poptable
2055 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2056 I've fixed the data, and added a kludgy way of testing at compile time that
2057 the lengths are correct (equal to the number of opcodes).
2058
2059 8. Following on from 7, I added a similar kludge to check the length of the
2060 eint vector in pcreposix.c.
2061
2062 9. Error texts for pcre_compile() are held as one long string to avoid too
2063 much relocation at load time. To find a text, the string is searched,
2064 counting zeros. There was no check for running off the end of the string,
2065 which could happen if a new error number was added without updating the
2066 string.
2067
2068 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2069
2070 11. \K was not working if it appeared in an atomic group or in a group that
2071 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2072 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2073 the assertion is positive, but not if it is negative.
2074
2075 12. Change 11 fortuitously reduced the size of the stack frame used in the
2076 "match()" function of pcre_exec.c by one pointer. Forthcoming
2077 implementation of support for (*MARK) will need an extra pointer on the
2078 stack; I have reserved it now, so that the stack frame size does not
2079 decrease.
2080
2081 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2082 item in branch that calls a recursion is a subroutine call - as in the
2083 second branch in the above example - was incorrectly given the compile-
2084 time error "recursive call could loop indefinitely" because pcre_compile()
2085 was not correctly checking the subroutine for matching a non-empty string.
2086
2087 14. The checks for overrunning compiling workspace could trigger after an
2088 overrun had occurred. This is a "should never occur" error, but it can be
2089 triggered by pathological patterns such as hundreds of nested parentheses.
2090 The checks now trigger 100 bytes before the end of the workspace.
2091
2092 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2093
2094
2095 Version 8.01 19-Jan-2010
2096 ------------------------
2097
2098 1. If a pattern contained a conditional subpattern with only one branch (in
2099 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2100 computed the wrong minimum data length (which is of course zero for such
2101 subpatterns). This could cause incorrect "no match" results.
2102
2103 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2104 the pattern is reset in the first branch, pcre_compile() failed with
2105 "internal error: code overflow at offset...". This happened only when
2106 the reset was to the original external option setting. (An optimization
2107 abstracts leading options settings into an external setting, which was the
2108 cause of this.)
2109
2110 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2111 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2112 assertion pattern did not match (meaning that the assertion was true), it
2113 was incorrectly treated as false if the SKIP had been reached during the
2114 matching. This also applied to assertions used as conditions.
2115
2116 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2117 assertion subpattern, including such a pattern used as a condition,
2118 unpredictable results occurred, instead of the error return
2119 PCRE_ERROR_DFA_UITEM.
2120
2121 5. The C++ GlobalReplace function was not working like Perl for the special
2122 situation when an empty string is matched. It now does the fancy magic
2123 stuff that is necessary.
2124
2125 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2126 removed. (These were left over from very, very early versions of PCRE.)
2127
2128 7. Some cosmetic changes to the code to make life easier when compiling it
2129 as part of something else:
2130
2131 (a) Change DEBUG to PCRE_DEBUG.
2132
2133 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2134 called "current" as "current_branch", to prevent a collision with the
2135 Linux macro when compiled as a kernel module.
2136
2137 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2138 prevent a collision with the Linux macro when compiled as a kernel
2139 module.
2140
2141 8. In pcre_compile() there are some checks for integer overflows that used to
2142 cast potentially large values to (double). This has been changed to that
2143 when building, a check for int64_t is made, and if it is found, it is used
2144 instead, thus avoiding the use of floating point arithmetic. (There is no
2145 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2146 double.
2147
2148 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2149 2005 (difference between two addresses compared to an unsigned value).
2150
2151 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2152 custom one, because of the following reported problem in Windows:
2153
2154 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2155 under Win32.
2156 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2157 therefore missing the function definition.
2158 - The compiler thus generates a "C" signature for the test function.
2159 - The linker fails to find the "C" function.
2160 - PCRE fails to configure if asked to do so against libbz2.
2161
2162 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2163 messages were output:
2164
2165 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2166 rerunning libtoolize, to keep the correct libtool macros in-tree.
2167 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2168
2169 I have done both of these things.
2170
2171 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2172 most of the time, it *can* run out if it is given a pattern that contains a
2173 runaway infinite recursion. I updated the discussion in the pcrestack man
2174 page.
2175
2176 13. Now that we have gone to the x.xx style of version numbers, the minor
2177 version may start with zero. Using 08 or 09 is a bad idea because users
2178 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2179 interpreted as invalid octal numbers. I've updated the previous comment in
2180 configure.ac, and also added a check that gives an error if 08 or 09 are
2181 used.
2182
2183 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2184 causing partial matching to fail when the end of the subject matched \W
2185 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2186
2187 15. There were some discrepancies between the declarations in pcre_internal.h
2188 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2189 their definitions. The declarations used "const uschar *" and the
2190 definitions used USPTR. Even though USPTR is normally defined as "const
2191 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2192 reported that: "This difference in casting confuses some C++ compilers, for
2193 example, SunCC recognizes above declarations as different functions and
2194 generates broken code for hbpcre." I have changed the declarations to use
2195 USPTR.
2196
2197 16. GNU libtool is named differently on some systems. The autogen.sh script now
2198 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2199 (FreeBSD).
2200
2201 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2202 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2203 comment: "Figure out how to create a longlong from a string: strtoll and
2204 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2205 instance, but it only takes 2 args instead of 3!"
2206
2207 18. A subtle bug concerned with back references has been fixed by a change of
2208 specification, with a corresponding code fix. A pattern such as
2209 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2210 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2211 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2212 same bug. Such groups have to be quantified to be useful, or contained
2213 inside another quantified group. (If there's no repetition, the reference
2214 can never match.) The problem arises because, having left the group and
2215 moved on to the rest of the pattern, a later failure that backtracks into
2216 the group uses the captured value from the final iteration of the group
2217 rather than the correct earlier one. I have fixed this in PCRE by forcing
2218 any group that contains a reference to itself to be an atomic group; that
2219 is, there cannot be any backtracking into it once it has completed. This is
2220 similar to recursive and subroutine calls.
2221
2222
2223 Version 8.00 19-Oct-09
2224 ----------------------
2225
2226 1. The table for translating pcre_compile() error codes into POSIX error codes
2227 was out-of-date, and there was no check on the pcre_compile() error code
2228 being within the table. This could lead to an OK return being given in
2229 error.
2230
2231 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2232 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2233 in a Windows environment.
2234
2235 3. The pcregrep --count option prints the count for each file even when it is
2236 zero, as does GNU grep. However, pcregrep was also printing all files when
2237 --files-with-matches was added. Now, when both options are given, it prints
2238 counts only for those files that have at least one match. (GNU grep just
2239 prints the file name in this circumstance, but including the count seems
2240 more useful - otherwise, why use --count?) Also ensured that the
2241 combination -clh just lists non-zero counts, with no names.
2242
2243 4. The long form of the pcregrep -F option was incorrectly implemented as
2244 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2245 but it seems right to fix it, and I didn't think it was worth preserving
2246 the old behaviour.
2247
2248 5. The command line items --regex=pattern and --regexp=pattern were not
2249 recognized by pcregrep, which required --regex pattern or --regexp pattern
2250 (with a space rather than an '='). The man page documented the '=' forms,
2251 which are compatible with GNU grep; these now work.
2252
2253 6. No libpcreposix.pc file was created for pkg-config; there was just
2254 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2255
2256 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2257 when UCP support is not needed, by modifying the Python script that
2258 generates it from Unicode data files. This should not matter if the module
2259 is correctly used as a library, but I received one complaint about 50K of
2260 unwanted data. My guess is that the person linked everything into his
2261 program rather than using a library. Anyway, it does no harm.
2262
2263 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2264 was a minimum greater than 1 for a wide character in a possessive
2265 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2266 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2267 character. Chaos in the form of incorrect output or a compiling loop could
2268 result.
2269
2270 9. The restrictions on what a pattern can contain when partial matching is
2271 requested for pcre_exec() have been removed. All patterns can now be
2272 partially matched by this function. In addition, if there are at least two
2273 slots in the offset vector, the offset of the earliest inspected character
2274 for the match and the offset of the end of the subject are set in them when
2275 PCRE_ERROR_PARTIAL is returned.
2276
2277 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2278 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2279 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2280 and may be more useful for multi-segment matching.
2281
2282 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2283 used to be given if ever the end of the subject was reached; now it is
2284 given only if matching could not proceed because another character was
2285 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2286 string "Z", which now yields "no match" instead of "partial match". In the
2287 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2288 final character ended with (*FAIL).
2289
2290 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2291 if the pattern had a "must contain" character that was already found in the
2292 earlier partial match, unless partial matching was again requested. For
2293 example, with the pattern /dog.(body)?/, the "must contain" character is
2294 "g". If the first part-match was for the string "dog", restarting with
2295 "sbody" failed. This bug has been fixed.
2296
2297 13. The string returned by pcre_dfa_exec() after a partial match has been
2298 changed so that it starts at the first inspected character rather than the
2299 first character of the match. This makes a difference only if the pattern
2300 starts with a lookbehind assertion or \b or \B (\K is not supported by
2301 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2302 matching functions compatible, and I think it's the right thing to do.
2303
2304 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2305 so that the demonstration program is easily available in environments where
2306 PCRE has not been installed from source.
2307
2308 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2309 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2310 library.
2311
2312 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2313 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2314 is not the first non-POSIX option to be added. Clearly some people find
2315 these options useful.
2316
2317 17. If a caller to the POSIX matching function regexec() passes a non-zero
2318 value for nmatch with a NULL value for pmatch, the value of
2319 nmatch is forced to zero.
2320
2321 18. RunGrepTest did not have a test for the availability of the -u option of
2322 the diff command, as RunTest does. It now checks in the same way as
2323 RunTest, and also checks for the -b option.
2324
2325 19. If an odd number of negated classes containing just a single character
2326 interposed, within parentheses, between a forward reference to a named
2327 subpattern and the definition of the subpattern, compilation crashed with
2328 an internal error, complaining that it could not find the referenced
2329 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2330 [The bug was that it was starting one character too far in when skipping
2331 over the character class, thus treating the ] as data rather than
2332 terminating the class. This meant it could skip too much.]
2333
2334 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2335 /g option in pcretest when the pattern contains \K, which makes it possible
2336 to have an empty string match not at the start, even when the pattern is
2337 anchored. Updated pcretest and pcredemo to use this option.
2338
2339 21. If the maximum number of capturing subpatterns in a recursion was greater
2340 than the maximum at the outer level, the higher number was returned, but
2341 with unset values at the outer level. The correct (outer level) value is
2342 now given.
2343
2344 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2345 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2346 make it do so. The string so far is captured, making this feature
2347 compatible with Perl.
2348
2349 23. The tests have been re-organized, adding tests 11 and 12, to make it
2350 possible to check the Perl 5.10 features against Perl 5.10.
2351
2352 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2353 pattern matches a fixed length string. PCRE did not allow this; now it
2354 does. Neither allows recursion.
2355
2356 25. I finally figured out how to implement a request to provide the minimum
2357 length of subject string that was needed in order to match a given pattern.
2358 (It was back references and recursion that I had previously got hung up
2359 on.) This code has now been added to pcre_study(); it finds a lower bound
2360 to the length of subject needed. It is not necessarily the greatest lower
2361 bound, but using it to avoid searching strings that are too short does give
2362 some useful speed-ups. The value is available to calling programs via
2363 pcre_fullinfo().
2364
2365 26. While implementing 25, I discovered to my embarrassment that pcretest had
2366 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2367 study optimizations had never been tested with that matching function.
2368 Oops. What is worse, even when it was passed study data, there was a bug in
2369 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2370 were also very few tests of studied patterns with pcre_dfa_exec().
2371
2372 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2373 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2374 on the other side of the coin, they are no longer allowed to have different
2375 names, because these cannot be distinguished in PCRE, and this has caused
2376 confusion. (This is a difference from Perl.)
2377
2378 28. When duplicate subpattern names are present (necessarily with different
2379 numbers, as required by 27 above), and a test is made by name in a
2380 conditional pattern, either for a subpattern having been matched, or for
2381 recursion in such a pattern, all the associated numbered subpatterns are
2382 tested, and the overall condition is true if the condition is true for any
2383 one of them. This is the way Perl works, and is also more like the way
2384 testing by number works.
2385
2386
2387 Version 7.9 11-Apr-09
2388 ---------------------
2389
2390 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2391 (pcretest), all targets were linked against these libraries. This included
2392 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2393 libraries. This caused unwanted dependencies to be created. This problem
2394 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2395 pcretest is linked with readline.
2396
2397 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2398 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2399 moved outside it again, because FALSE and TRUE are already defined in AIX,
2400 but BOOL is not.
2401
2402 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2403 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2404
2405 4. The pcregrep documentation said spaces were inserted as well as colons (or
2406 hyphens) following file names and line numbers when outputting matching
2407 lines. This is not true; no spaces are inserted. I have also clarified the
2408 wording for the --colour (or --color) option.
2409
2410 5. In pcregrep, when --colour was used with -o, the list of matching strings
2411 was not coloured; this is different to GNU grep, so I have changed it to be
2412 the same.
2413
2414 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2415 each matching line was coloured. Now it goes on to look for further matches
2416 of any of the test patterns, which is the same behaviour as GNU grep.
2417
2418 7. A pattern that could match an empty string could cause pcregrep to loop; it
2419 doesn't make sense to accept an empty string match in pcregrep, so I have
2420 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2421 seems to be how GNU grep behaves. [But see later change 40 for release
2422 8.33.]
2423
2424 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2425 start or after a newline", because the conditional assertion was not being
2426 correctly handled. The rule now is that both the assertion and what follows
2427 in the first alternative must satisfy the test.
2428
2429 9. If auto-callout was enabled in a pattern with a conditional group whose
2430 condition was an assertion, PCRE could crash during matching, both with
2431 pcre_exec() and pcre_dfa_exec().
2432
2433 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2434 used for matching.
2435
2436 11. Unicode property support in character classes was not working for
2437 characters (bytes) greater than 127 when not in UTF-8 mode.
2438
2439 12. Added the -M command line option to pcretest.
2440
2441 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2442
2443 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2444
2445 16. Added comments and documentation about mis-use of no_arg in the C++
2446 wrapper.
2447
2448 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2449 from Martin Jerabek that uses macro names for all relevant character and
2450 string constants.
2451
2452 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2453 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2454 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2455 these, but not everybody uses configure.
2456
2457 19. A conditional group that had only one branch was not being correctly
2458 recognized as an item that could match an empty string. This meant that an
2459 enclosing group might also not be so recognized, causing infinite looping
2460 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2461 with the subject "ab", where knowledge that the repeated group can match
2462 nothing is needed in order to break the loop.
2463
2464 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2465 exec(), but without supplying a callout function, matching went wrong.
2466
2467 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2468 leak if the size of the offset vector was greater than 30. When the vector
2469 is smaller, the saved offsets during recursion go onto a local stack
2470 vector, but for larger vectors malloc() is used. It was failing to free
2471 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2472 error, in fact).
2473
2474 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2475 heapframe that is used only when UTF-8 support is enabled. This caused no
2476 problem, but was untidy.
2477
2478 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2479 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2480 included within another project.
2481
2482 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2483 slightly modified by me:
2484
2485 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2486 not building pcregrep.
2487
2488 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2489 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2490
2491 25. Forward references, both numeric and by name, in patterns that made use of
2492 duplicate group numbers, could behave incorrectly or give incorrect errors,
2493 because when scanning forward to find the reference group, PCRE was not
2494 taking into account the duplicate group numbers. A pattern such as
2495 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2496
2497 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2498 the feature of a custom pointer more persuasive (as requested by a user).
2499
2500 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2501 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2502 pre-defined.
2503
2504 28. Added support for (*UTF8) at the start of a pattern.
2505
2506 29. Arrange for flags added by the "release type" setting in CMake to be shown
2507 in the configuration summary.
2508
2509
2510 Version 7.8 05-Sep-08
2511 ---------------------
2512
2513 1. Replaced UCP searching code with optimized version as implemented for Ad
2514 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2515 stage table and inline lookup instead of a function, giving speed ups of 2
2516 to 5 times on some simple patterns that I tested. Permission was given to
2517 distribute the MultiStage2.py script that generates the tables (it's not in
2518 the tarball, but is in the Subversion repository).
2519
2520 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2521 scripts.
2522
2523 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2524 a group with a zero qualifier. The result of the study could be incorrect,
2525 or the function might crash, depending on the pattern.
2526
2527 4. Caseless matching was not working for non-ASCII characters in back
2528 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2529 It now works when Unicode Property Support is available.
2530
2531 5. In pcretest, an escape such as \x{de} in the data was always generating
2532 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2533 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2534 truncation.
2535
2536 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2537
2538 7. Added two (int) casts to pcregrep when printing the difference of two
2539 pointers, in case they are 64-bit values.
2540
2541 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2542 test 2 if it fails.
2543
2544 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2545 and a #define of that name to empty if it is not externally set. This is to
2546 allow users of MSVC to set it if necessary.
2547
2548 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2549 the convenience functions in the pcre_get.c source file.
2550
2551 11. An option change at the start of a pattern that had top-level alternatives
2552 could cause overwriting and/or a crash. This command provoked a crash in
2553 some environments:
2554
2555 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2556
2557 This potential security problem was recorded as CVE-2008-2371.
2558
2559 12. For a pattern where the match had to start at the beginning or immediately
2560 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2561 pcre_dfa_exec() could read past the end of the passed subject if there was
2562 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2563 pcretest so that it places the subject at the end of its malloc-ed buffer.
2564
2565 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2566 exec() might read past the end of the data buffer in UTF-8 mode.
2567
2568 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2569 the data contained the byte 0x85 as part of a UTF-8 character within its
2570 first line. This applied both to normal and DFA matching.
2571
2572 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2573 /^[^d]*?$/8 failed to match "abc".
2574
2575 16. Added a missing copyright notice to pcrecpp_internal.h.
2576
2577 17. Make it more clear in the documentation that values returned from
2578 pcre_exec() in ovector are byte offsets, not character counts.
2579
2580 18. Tidied a few places to stop certain compilers from issuing warnings.
2581
2582 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2583 supplied by Stefan Weber. I made a further small update for 7.8 because
2584 there is a change of source arrangements: the pcre_searchfuncs.c module is
2585 replaced by pcre_ucd.c.
2586
2587
2588 Version 7.7 07-May-08
2589 ---------------------
2590
2591 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2592 a string to a long long, pretend we don't even have a long long." This is
2593 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2594
2595 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2596 pre-7.6 versions, which defined a global no_arg variable instead of putting
2597 it in the RE class. (See also #8 below.)
2598
2599 3. Remove a line of dead code, identified by coverity and reported by Nuno
2600 Lopes.
2601
2602 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2603
2604 (1) The include/exclude patterns were being applied to the whole pathnames
2605 of files, instead of just to the final components.
2606
2607 (2) If there was more than one level of directory, the subdirectories were
2608 skipped unless they satisfied the include/exclude conditions. This is
2609 inconsistent with GNU grep (and could even be seen as contrary to the
2610 pcregrep specification - which I improved to make it absolutely clear).
2611 The action now is always to scan all levels of directory, and just
2612 apply the include/exclude patterns to regular files.
2613
2614 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2615 --exclude_dir in the tests to avoid scanning .svn directories.
2616
2617 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2618 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2619 doesn't support NULs in patterns.
2620
2621 7. Added some missing "const"s to declarations of static tables in
2622 pcre_compile.c and pcre_dfa_exec.c.
2623
2624 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2625 caused by fix #2 above. (Subsequently also a second patch to fix the
2626 first patch. And a third patch - this was a messy problem.)
2627
2628 9. Applied Craig's patch to remove the use of push_back().
2629
2630 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2631 matching function regexec().
2632
2633 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2634 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2635 references. PCRE supports relative numbers with this syntax (I don't think
2636 Oniguruma does).
2637
2638 12. Previously, a group with a zero repeat such as (...){0} was completely
2639 omitted from the compiled regex. However, this means that if the group
2640 was called as a subroutine from elsewhere in the pattern, things went wrong
2641 (an internal error was given). Such groups are now left in the compiled
2642 pattern, with a new opcode that causes them to be skipped at execution
2643 time.
2644
2645 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2646 to the way PCRE behaves:
2647
2648 (a) A lone ] character is dis-allowed (Perl treats it as data).
2649
2650 (b) A back reference to an unmatched subpattern matches an empty string
2651 (Perl fails the current match path).
2652
2653 (c) A data ] in a character class must be notated as \] because if the
2654 first data character in a class is ], it defines an empty class. (In
2655 Perl it is not possible to have an empty class.) The empty class []
2656 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2657 The negative empty class [^] matches any one character, independently
2658 of the DOTALL setting.
2659
2660 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2661 non-existent subpattern following a character class starting with ']' and
2662 containing () gave an internal compiling error instead of "reference to
2663 non-existent subpattern". Fortunately, when the pattern did exist, the
2664 compiled code was correct. (When scanning forwards to check for the
2665 existence of the subpattern, it was treating the data ']' as terminating
2666 the class, so got the count wrong. When actually compiling, the reference
2667 was subsequently set up correctly.)
2668
2669 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2670 it was being rejected as not supported by pcre_dfa_exec(), even though
2671 other assertions are supported. I have made pcre_dfa_exec() support
2672 (*FAIL).
2673
2674 16. The implementation of 13c above involved the invention of a new opcode,
2675 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2676 cannot be changed at match time, I realized I could make a small
2677 improvement to matching performance by compiling OP_ALLANY instead of
2678 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2679 on the OP_ANY path.
2680
2681 17. Compiling pcretest on Windows with readline support failed without the
2682 following two fixes: (1) Make the unistd.h include conditional on
2683 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2684
2685 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2686 ncurses library to be included for pcretest when ReadLine support is
2687 requested, but also to allow for it to be overridden. This patch came from
2688 Daniel Bergström.
2689
2690 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2691 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2692 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2693 spotting this.
2694
2695
2696 Version 7.6 28-Jan-08
2697 ---------------------
2698
2699 1. A character class containing a very large number of characters with
2700 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2701 overflow.
2702
2703 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2704 HAVE_LONG_LONG is not defined.
2705
2706 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2707 bring them up to date and include new features. This patch includes:
2708
2709 - Fixed PH's badly added libz and libbz2 support.
2710 - Fixed a problem with static linking.
2711 - Added pcredemo. [But later removed - see 7 below.]
2712 - Fixed dftables problem and added an option.
2713 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2714 HAVE_LONG_LONG.
2715 - Added readline support for pcretest.
2716 - Added an listing of the option settings after cmake has run.
2717
2718 4. A user submitted a patch to Makefile that makes it easy to create
2719 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2720 Makefile.am that cause it to include this special target, without
2721 affecting anything else. Note that the same mingw target plus all
2722 the other distribution libraries and programs are now supported
2723 when configuring with CMake (see 6 below) instead of with
2724 Configure/Make.
2725
2726 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2727 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2728 exported in the Windows port". It has not yet been confirmed that the patch
2729 solves the problem, but it does no harm.
2730
2731 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2732 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2733 with CMake, and also correct the comment about stack recursion.
2734
2735 7. Remove the automatic building of pcredemo from the ./configure system and
2736 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2737 of a program that users should build themselves after PCRE is installed, so
2738 building it automatically is not really right. What is more, it gave
2739 trouble in some build environments.
2740
2741 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2742
2743
2744 Version 7.5 10-Jan-08
2745 ---------------------
2746
2747 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2748 values in parens when parsing an RE using the C++ wrapper."
2749
2750 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2751 Characters greater than 255 were excluded from the class instead of being
2752 included.
2753
2754 3. The same bug as (2) above applied to negated POSIX classes such as
2755 [:^space:].
2756
2757 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2758 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2759 I have changed it.
2760
2761 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2762 first named subpattern) and a construct such as (?&a) would reference the
2763 first named subpattern whose name started with "a" (in other words, the
2764 length check was missing). Both these problems are fixed. "Subpattern name
2765 expected" is now given for (?&) (a zero-length name), and this patch also
2766 makes it give the same error for \k'' (previously it complained that that
2767 was a reference to a non-existent subpattern).
2768
2769 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2770 this is right because (?- can be followed by option settings as well as by
2771 digits. I have, however, made the messages clearer.
2772
2773 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2774 than the number used in the conditional) now cause a compile-time error.
2775 This is actually not compatible with Perl, which accepts such patterns, but
2776 treats the conditional as always being FALSE (as PCRE used to), but it
2777 seems to me that giving a diagnostic is better.
2778
2779 8. Change "alphameric" to the more common word "alphanumeric" in comments
2780 and messages.
2781
2782 9. Fix two occurrences of "backslash" in comments that should have been
2783 "backspace".
2784
2785 10. Remove two redundant lines of code that can never be obeyed (their function
2786 was moved elsewhere).
2787
2788 11. The program that makes PCRE's Unicode character property table had a bug
2789 which caused it to generate incorrect table entries for sequences of
2790 characters that have the same character type, but are in different scripts.
2791 It amalgamated them into a single range, with the script of the first of
2792 them. In other words, some characters were in the wrong script. There were
2793 thirteen such cases, affecting characters in the following ranges:
2794
2795 U+002b0 - U+002c1
2796 U+0060c - U+0060d
2797 U+0061e - U+00612
2798 U+0064b - U+0065e
2799 U+0074d - U+0076d
2800 U+01800 - U+01805
2801 U+01d00 - U+01d77
2802 U+01d9b - U+01dbf
2803 U+0200b - U+0200f
2804 U+030fc - U+030fe
2805 U+03260 - U+0327f
2806 U+0fb46 - U+0fbb1
2807 U+10450 - U+1049d
2808
2809 12. The -o option (show only the matching part of a line) for pcregrep was not
2810 compatible with GNU grep in that, if there was more than one match in a
2811 line, it showed only the first of them. It now behaves in the same way as
2812 GNU grep.
2813
2814 13. If the -o and -v options were combined for pcregrep, it printed a blank
2815 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2816 does the same. The return code can be used to tell if there were any
2817 non-matching lines.
2818
2819 14. Added --file-offsets and --line-offsets to pcregrep.
2820
2821 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2822 infinitely looping recursion. The bug was that positive lookaheads were not
2823 being skipped when checking for a possible empty match (negative lookaheads
2824 and both kinds of lookbehind were skipped).
2825
2826 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2827 inclusion of <windows.h> to before rather than after the definition of
2828 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2829
2830 17. Specifying a possessive quantifier with a specific limit for a Unicode
2831 character property caused pcre_compile() to compile bad code, which led at
2832 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2833 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2834 caused the error; without that there was no problem.
2835
2836 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2837
2838 19. Added --enable-pcretest-libreadline.
2839
2840 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2841 RE::GlobalReplace(). As a result, the number of replacements returned was
2842 double what it should be. I removed one of the increments, but Craig sent a
2843 later patch that removed the other one (the right fix) and added unit tests
2844 that check the return values (which was not done before).
2845
2846 21. Several CMake things:
2847
2848 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2849 the names libpcre and libpcreposix, not just pcre and pcreposix.
2850
2851 (2) The above change means that pcretest and pcregrep are now correctly
2852 linked with the newly-built libraries, not previously installed ones.
2853
2854 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2855
2856 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2857 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2858 UTF-8 newline character). The key issue is that the pattern starts .*;
2859 this means that the match must be either at the beginning, or after a
2860 newline. The bug was in the code for advancing after a failed match and
2861 checking that the new position followed a newline. It was not taking
2862 account of UTF-8 characters correctly.
2863
2864 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2865 character classes. PCRE was not treating the sequence [:...:] as a
2866 character class unless the ... were all letters. Perl, however, seems to
2867 allow any characters between [: and :], though of course it rejects as
2868 unknown any "names" that contain non-letters, because all the known class
2869 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2870 for example, whereas PCRE did not - it did not recognize a POSIX character
2871 class. This seemed a bit dangerous, so the code has been changed to be
2872 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2873 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2874 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2875 Perl does, and where it didn't before.
2876
2877 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2878 Windows environments %n is disabled by default.
2879
2880
2881 Version 7.4 21-Sep-07
2882 ---------------------
2883
2884 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2885 means that a class such as [\s] counted as "explicit reference to CR or
2886 LF". That isn't really right - the whole point of the change was to try to
2887 help when there was an actual mention of one of the two characters. So now
2888 the change happens only if \r or \n (or a literal CR or LF) character is
2889 encountered.
2890
2891 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2892 of both had grown to the point where there were only 3 bits left.
2893 Fortunately, there was spare space in the data structure, and so I have
2894 moved the internal flags into a new 16-bit field to free up more option
2895 bits.
2896
2897 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2898 but did not set the internal JCHANGED flag - either of these is enough to
2899 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2900 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2901 start sets both bits.
2902
2903 4. Added options (at build time, compile time, exec time) to change \R from
2904 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2905
2906 5. doc/pcresyntax.html was missing from the distribution.
2907
2908 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2909 compatibility, even though it is no longer used.
2910
2911 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2912 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2913 windows.h file is present (where different names are used). [This was
2914 reversed later after testing - see 16 below.]
2915
2916 8. Changed all #include <config.h> to #include "config.h". There were also
2917 some further <pcre.h> cases that I changed to "pcre.h".
2918
2919 9. When pcregrep was used with the --colour option, it missed the line ending
2920 sequence off the lines that it output.
2921
2922 10. It was pointed out to me that arrays of string pointers cause lots of
2923 relocations when a shared library is dynamically loaded. A technique of
2924 using a single long string with a table of offsets can drastically reduce
2925 these. I have refactored PCRE in four places to do this. The result is
2926 dramatic:
2927
2928 Originally: 290
2929 After changing UCP table: 187
2930 After changing error message table: 43
2931 After changing table of "verbs" 36
2932 After changing table of Posix names 22
2933
2934 Thanks to the folks working on Gregex for glib for this insight.
2935
2936 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2937 unicode-properties was also set.
2938
2939 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2940
2941 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2942 checked only for CRLF.
2943
2944 14. Added casts to pcretest.c to avoid compiler warnings.
2945
2946 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2947
2948 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2949 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2950 entirely. This removes changes made in 7 above.
2951
2952 17. The CMake files have been updated, and there is now more information about
2953 building with CMake in the NON-UNIX-USE document.
2954
2955
2956 Version 7.3 28-Aug-07
2957 ---------------------
2958
2959 1. In the rejigging of the build system that eventually resulted in 7.1, the
2960 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2961 brackets there is not right, since it causes compilers to look for an
2962 installed pcre.h, not the version that is in the source that is being
2963 compiled (which of course may be different). I have changed it back to:
2964
2965 #include "pcre.h"
2966
2967 I have a vague recollection that the change was concerned with compiling in
2968 different directories, but in the new build system, that is taken care of
2969 by the VPATH setting the Makefile.
2970
2971 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2972 when the subject happened to end in the byte 0x85 (e.g. if the last
2973 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2974 characters but of course it shouldn't be taken as a newline when it is part
2975 of another character. The bug was that, for an unlimited repeat of . in
2976 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2977 characters when looking for a newline.
2978
2979 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2980
2981 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2982 in debug output.
2983
2984 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2985 long printing in the pcrecpp unittest when running under MinGW.
2986
2987 6. ESC_K was left out of the EBCDIC table.
2988
2989 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2990 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2991 limit also applies to "virtual nesting" when a pattern is recursive, and in
2992 this case 1000 isn't so big. I have been able to remove this limit at the
2993 expense of backing off one optimization in certain circumstances. Normally,
2994 when pcre_exec() would call its internal match() function recursively and
2995 immediately return the result unconditionally, it uses a "tail recursion"
2996 feature to save stack. However, when a subpattern that can match an empty
2997 string has an unlimited repetition quantifier, it no longer makes this
2998 optimization. That gives it a stack frame in which to save the data for
2999 checking that an empty string has been matched. Previously this was taken
3000 from the 1000-entry workspace that had been reserved. So now there is no
3001 explicit limit, but more stack is used.
3002
3003 8. Applied Daniel's patches to solve problems with the import/export magic
3004 syntax that is required for Windows, and which was going wrong for the
3005 pcreposix and pcrecpp parts of the library. These were overlooked when this
3006 problem was solved for the main library.
3007
3008 9. There were some crude static tests to avoid integer overflow when computing
3009 the size of patterns that contain repeated groups with explicit upper
3010 limits. As the maximum quantifier is 65535, the maximum group length was
3011 set at 30,000 so that the product of these two numbers did not overflow a
3012 32-bit integer. However, it turns out that people want to use groups that
3013 are longer than 30,000 bytes (though not repeat them that many times).
3014 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3015 made it possible to implement the integer overflow checks in a much more
3016 dynamic way, which I have now done. The artificial limitation on group
3017 length has been removed - we now have only the limit on the total length of
3018 the compiled pattern, which depends on the LINK_SIZE setting.
3019
3020 10. Fixed a bug in the documentation for get/copy named substring when
3021 duplicate names are permitted. If none of the named substrings are set, the
3022 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3023 empty string.
3024
3025 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3026 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3027 because the ] is interpreted as the first data character and the
3028 terminating ] is not found. PCRE has been made compatible with Perl in this
3029 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3030 cause memory overwriting.
3031
3032 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3033 string has been matched (to stop an infinite loop). It was not recognizing
3034 a conditional subpattern that could match an empty string if that
3035 subpattern was within another subpattern. For example, it looped when
3036 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3037 condition was not nested. This bug has been fixed.
3038
3039 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3040 past the start of the subject in the presence of bytes with the top bit
3041 set, for example "\x8aBCD".
3042
3043 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3044 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3045
3046 14. Optimized (?!) to (*FAIL).
3047
3048 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3049 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3050 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3051 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3052 does: it's just the validity check that is more restrictive.
3053
3054 16. Inserted checks for integer overflows during escape sequence (backslash)
3055 processing, and also fixed erroneous offset values for syntax errors during
3056 backslash processing.
3057
3058 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3059 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3060
3061 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3062 caused an overrun.
3063
3064 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3065 something other than just ASCII characters) inside a group that had an
3066 unlimited repeat caused a loop at compile time (while checking to see
3067 whether the group could match an empty string).
3068
3069 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3070 [\P{Any}] did so. (Error in the code for printing property names.)
3071
3072 21. An orphan \E inside a character class could cause a crash.
3073
3074 22. A repeated capturing bracket such as (A)? could cause a wild memory
3075 reference during compilation.
3076
3077 23. There are several functions in pcre_compile() that scan along a compiled
3078 expression for various reasons (e.g. to see if it's fixed length for look
3079 behind). There were bugs in these functions when a repeated \p or \P was
3080 present in the pattern. These operators have additional parameters compared
3081 with \d, etc, and these were not being taken into account when moving along
3082 the compiled data. Specifically:
3083
3084 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3085 length.
3086
3087 (b) An item such as \pL+ within a repeated group could cause crashes or
3088 loops.
3089
3090 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3091 "reference to non-existent subpattern" error.
3092
3093 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3094
3095 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3096 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3097
3098 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3099
3100 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3101 character were causing crashes (broken optimization).
3102
3103 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3104 \p or \P) caused a compile-time loop.
3105
3106 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3107 break. For example, the unstudied pattern [\r\n]A does not match the string
3108 "\r\nA" because change 7.0/46 below moves the current point on by two
3109 characters after failing to match at the start. However, the pattern \nA
3110 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3111 the same is true. There doesn't seem any very clean way out of this, but
3112 what I have chosen to do makes the common cases work: PCRE now takes note
3113 of whether there can be an explicit match for \r or \n anywhere in the
3114 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3115 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3116 pattern has explicit CR or LF references.
3117
3118 29. Added (*CR) etc for changing newline setting at start of pattern.
3119
3120
3121 Version 7.2 19-Jun-07
3122 ---------------------
3123
3124 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3125 which is apparently normally available under Windows.
3126
3127 2. Re-jig the pcregrep tests with different newline settings in an attempt
3128 to make them independent of the local environment's newline setting.
3129
3130 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3131
3132 4. Some of the "internals" tests were previously cut out when the link size
3133 was not 2, because the output contained actual offsets. The recent new
3134 "Z" feature of pcretest means that these can be cut out, making the tests
3135 usable with all link sizes.
3136
3137 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3138 stack recursion. This gives a massive performance boost under BSD, but just
3139 a small improvement under Linux. However, it saves one field in the frame
3140 in all cases.
3141
3142 6. Added more features from the forthcoming Perl 5.10:
3143
3144 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3145 recursion call. It refers to the nth most recently opened parentheses.
3146
3147 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3148 to be opened parentheses.
3149
3150 (c) Conditions that refer to capturing parentheses can be specified
3151 relatively, for example, (?(-2)... or (?(+3)...
3152
3153 (d) \K resets the start of the current match so that everything before
3154 is not part of it.
3155
3156 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3157
3158 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3159 reference syntax.
3160
3161 (g) (?| introduces a group in which the numbering of parentheses in each
3162 alternative starts with the same number.
3163
3164 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3165
3166 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3167 PCRE_INFO_JCHANGED.
3168
3169 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3170 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3171 for detecting groups that can match an empty string.
3172
3173 9. A pattern with a very large number of alternatives (more than several
3174 hundred) was running out of internal workspace during the pre-compile
3175 phase, where pcre_compile() figures out how much memory will be needed. A
3176 bit of new cunning has reduced the workspace needed for groups with
3177 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3178 workspace instead of running out of the 4096 that are available.
3179
3180 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3181
3182 11. Applied patch from Google to remove an optimization that didn't quite work.
3183 The report of the bug said:
3184
3185 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3186 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3187 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3188
3189 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3190 it matched the wrong number of bytes.
3191
3192
3193 Version 7.1 24-Apr-07
3194 ---------------------
3195
3196 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3197 that is more "standard", making use of automake and other Autotools. There
3198 is some re-arrangement of the files and adjustment of comments consequent
3199 on this.
3200
3201 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3202 for recursive directory scanning broke on some systems because the files
3203 are not scanned in any specific order and on different systems the order
3204 was different. A call to "sort" has been inserted into RunGrepTest for the
3205 approprate test as a short-term fix. In the longer term there may be an
3206 alternative.
3207
3208 3. I had an email from Eric Raymond about problems translating some of PCRE's
3209 man pages to HTML (despite the fact that I distribute HTML pages, some
3210 people do their own conversions for various reasons). The problems
3211 concerned the use of low-level troff macros .br and .in. I have therefore
3212 removed all such uses from the man pages (some were redundant, some could
3213 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3214 HTML has been updated to handle .nf/.fi and to complain if it encounters
3215 .br or .in.
3216
3217 4. Updated comments in configure.ac that get placed in config.h.in and also
3218 arranged for config.h to be included in the distribution, with the name
3219 config.h.generic, for the benefit of those who have to compile without
3220 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3221
3222 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3223 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3224 makevp.bat for the current PCRE, using the additional files
3225 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3226
3227 6. A Windows user reported a minor discrepancy with test 2, which turned out
3228 to be caused by a trailing space on an input line that had got lost in his
3229 copy. The trailing space was an accident, so I've just removed it.
3230
3231 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3232 that is needed.
3233
3234 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3235 as "const" (a) because they are and (b) because it helps the PHP
3236 maintainers who have recently made a script to detect big data structures
3237 in the php code that should be moved to the .rodata section. I remembered
3238 to update Builducptable as well, so it won't revert if ucptable.h is ever
3239 re-created.
3240
3241 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3242 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3243 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3244 support is not required. This saves 1.5-2K of code, which is important in
3245 some applications.
3246
3247 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3248 so as not to refer to the tables, even though these functions will never be
3249 called when UTF-8 support is disabled. Otherwise there are problems with a
3250 shared library.
3251
3252 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3253
3254 (a) It was defining its arguments as char * instead of void *.
3255
3256 (b) It was assuming that all moves were upwards in memory; this was true
3257 a long time ago when I wrote it, but is no longer the case.
3258
3259 The emulated memove() is provided for those environments that have neither
3260 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3261 is clearly not the case, as these two bugs were recently reported.
3262
3263 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3264 and Detrail to create the HTML documentation, the .txt form of the man
3265 pages, and it removes trailing spaces from listed files. It also creates
3266 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3267 case, it wraps all the #defines with #ifndefs. This script should be run
3268 before "make dist".
3269
3270 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3271 with Unicode property support.
3272
3273 (a) For a maximizing quantifier, if the two different cases of the
3274 character were of different lengths in their UTF-8 codings (there are
3275 some cases like this - I found 11), and the matching function had to
3276 back up over a mixture of the two cases, it incorrectly assumed they
3277 were both the same length.
3278
3279 (b) When PCRE was configured to use the heap rather than the stack for
3280 recursion during matching, it was not correctly preserving the data for
3281 the other case of a UTF-8 character when checking ahead for a match
3282 while processing a minimizing repeat. If the check also involved
3283 matching a wide character, but failed, corruption could cause an
3284 erroneous result when trying to check for a repeat of the original
3285 character.
3286
3287 13. Some tidying changes to the testing mechanism:
3288
3289 (a) The RunTest script now detects the internal link size and whether there
3290 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3291 values substituted by "configure". (The RunGrepTest script already did
3292 this for UTF-8.) The configure.ac script no longer substitutes the
3293 relevant variables.
3294
3295 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3296 with length and offset values. This means that the output is different
3297 for different internal link sizes. Test 2 is skipped for link sizes
3298 other than 2 because of this, bypassing the problem. Unfortunately,
3299 there was also a test in test 3 (the locale tests) that used /B and
3300 failed for link sizes other than 2. Rather than cut the whole test out,
3301 I have added a new /Z option to pcretest that replaces the length and
3302 offset values with spaces. This is now used to make test 3 independent
3303 of link size. (Test 2 will be tidied up later.)
3304
3305 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3306 segmentation fault instead of returning the appropriate error message.
3307
3308 15. In multiline mode when the newline sequence was set to "any", the pattern
3309 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3310 This doesn't seem right; it now treats the CRLF combination as the line
3311 ending, and so does not match in that case. It's only a pattern such as ^$
3312 that would hit this one: something like ^ABC$ would have failed after \r
3313 and then tried again after \r\n.
3314
3315 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3316 in an attempt to make files that differ only in their line terminators
3317 compare equal. This works on Linux.
3318
3319 17. Under certain error circumstances pcregrep might try to free random memory
3320 as it exited. This is now fixed, thanks to valgrind.
3321
3322 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3323 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3324 was because its rules for how to advance for /g after matching an empty
3325 string at the end of a line did not allow for this case. They now check for
3326 it specially.
3327
3328 20. pcretest is supposed to handle patterns and data of any length, by
3329 extending its buffers when necessary. It was getting this wrong when the
3330 buffer for a data line had to be extended.
3331
3332 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3333 CRLF as a newline sequence.
3334
3335 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3336 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3337 I have nevertheless tidied it up.
3338
3339 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3340
3341 24. Added a man page for pcre-config.
3342
3343
3344 Version 7.0 19-Dec-06
3345 ---------------------
3346
3347 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3348 moving to gcc 4.1.1.
3349
3350 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3351 sys/time.h, which is documented as needed for this function. It doesn't
3352 seem to matter on Linux, but it showed up on some releases of OS X.
3353
3354 3. It seems that there are systems where bytes whose values are greater than
3355 127 match isprint() in the "C" locale. The "C" locale should be the
3356 default when a C program starts up. In most systems, only ASCII printing
3357 characters match isprint(). This difference caused the output from pcretest
3358 to vary, making some of the tests fail. I have changed pcretest so that:
3359
3360 (a) When it is outputting text in the compiled version of a pattern, bytes
3361 other than 32-126 are always shown as hex escapes.
3362
3363 (b) When it is outputting text that is a matched part of a subject string,
3364 it does the same, unless a different locale has been set for the match
3365 (using the /L modifier). In this case, it uses isprint() to decide.
3366
3367 4. Fixed a major bug that caused incorrect computation of the amount of memory
3368 required for a compiled pattern when options that changed within the
3369 pattern affected the logic of the preliminary scan that determines the
3370 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3371 that the computed length was too small. The symptoms of this bug were
3372 either the PCRE error "internal error: code overflow" from pcre_compile(),
3373 or a glibc crash with a message such as "pcretest: free(): invalid next
3374 size (fast)". Examples of patterns that provoked this bug (shown in
3375 pcretest format) are:
3376
3377 /(?-x: )/x
3378 /(?x)(?-x: \s*#\s*)/
3379 /((?i)[\x{c0}])/8
3380 /(?i:[\x{c0}])/8
3381
3382 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3383 is now done differently.
3384
3385 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3386 wrapper classes; (b) implement a new function in the C++ scanner that is
3387 more efficient than the old way of doing things because it avoids levels of
3388 recursion in the regex matching; (c) add a paragraph to the documentation
3389 for the FullMatch() function.
3390
3391 6. The escape sequence \n was being treated as whatever was defined as
3392 "newline". Not only was this contrary to the documentation, which states
3393 that \n is character 10 (hex 0A), but it also went horribly wrong when
3394 "newline" was defined as CRLF. This has been fixed.
3395
3396 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3397 was being set to -1 for the "end of line" case (supposedly a value that no
3398 character can have). Though this value is never used (the check for end of
3399 line is "zero bytes in current character"), it caused compiler complaints.
3400 I've changed it to 0xffffffff.
3401
3402 8. In pcre_version.c, the version string was being built by a sequence of
3403 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3404 string (as it is for production releases) called a macro with an empty
3405 argument. The C standard says the result of this is undefined. The gcc
3406 compiler treats it as an empty string (which was what was wanted) but it is
3407 reported that Visual C gives an error. The source has been hacked around to
3408 avoid this problem.
3409
3410 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3411 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3412 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3413 of them did).
3414
3415 10. Originally, pcretest opened its input and output without "b"; then I was
3416 told that "b" was needed in some environments, so it was added for release
3417 5.0 to both the input and output. (It makes no difference on Unix-like
3418 systems.) Later I was told that it is wrong for the input on Windows. I've
3419 now abstracted the modes into two macros, to make it easier to fiddle with
3420 them, and removed "b" from the input mode under Windows.
3421
3422 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3423
3424 12. Added -help and --help to pcretest as an official way of being reminded
3425 of the options.
3426
3427 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3428 and pcrecpp.cc because they annoy compilers at high warning levels.
3429
3430 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3431
3432 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3433 scripts are not C programs :-) and which was not noticed because it works
3434 on Linux.
3435
3436 16. pcretest is supposed to handle any length of pattern and data line (as one
3437 line or as a continued sequence of lines) by extending its input buffer if
3438 necessary. This feature was broken for very long pattern lines, leading to
3439 a string of junk being passed to pcre_compile() if the pattern was longer
3440 than about 50K.
3441
3442 17. I have done a major re-factoring of the way pcre_compile() computes the
3443 amount of memory needed for a compiled pattern. Previously, there was code
3444 that made a preliminary scan of the pattern in order to do this. That was
3445 OK when PCRE was new, but as the facilities have expanded, it has become
3446 harder and harder to keep it in step with the real compile phase, and there
3447 have been a number of bugs (see for example, 4 above). I have now found a
3448 cunning way of running the real compile function in a "fake" mode that
3449 enables it to compute how much memory it would need, while actually only
3450 ever using a few hundred bytes of working memory and without too many
3451 tests of the mode. This should make future maintenance and development
3452 easier. A side effect of this work is that the limit of 200 on the nesting
3453 depth of parentheses has been removed (though this was never a serious
3454 limitation, I suspect). However, there is a downside: pcre_compile() now
3455 runs more slowly than before (30% or more, depending on the pattern). I
3456 hope this isn't a big issue. There is no effect on runtime performance.
3457
3458 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3459 newline (only possible for the last line of a file) and it was a
3460 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3461
3462 19. Added additional timing features to pcretest. (1) The -tm option now times
3463 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3464 separate command line item, by a number that specifies the number of
3465 repeats to use when timing. The default is 50000; this gives better
3466 precision, but takes uncomfortably long for very large patterns.
3467
3468 20. Extended pcre_study() to be more clever in cases where a branch of a
3469 subpattern has no definite first character. For example, (a*|b*)[cd] would
3470 previously give no result from pcre_study(). Now it recognizes that the
3471 first character must be a, b, c, or d.
3472
3473 21. There was an incorrect error "recursive call could loop indefinitely" if
3474 a subpattern (or the entire pattern) that was being tested for matching an
3475 empty string contained only one non-empty item after a nested subpattern.
3476 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3477 incorrectly, because the \d was being skipped in the check.
3478
3479 22. The pcretest program now has a new pattern option /B and a command line
3480 option -b, which is equivalent to adding /B to every pattern. This causes
3481 it to show the compiled bytecode, without the additional information that
3482 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3483 is the same as /B/I).
3484
3485 23. A new optimization is now able automatically to treat some sequences such
3486 as a*b as a*+b. More specifically, if something simple (such as a character
3487 or a simple class like \d) has an unlimited quantifier, and is followed by
3488 something that cannot possibly match the quantified thing, the quantifier
3489 is automatically "possessified".
3490
3491 24. A recursive reference to a subpattern whose number was greater than 39
3492 went wrong under certain circumstances in UTF-8 mode. This bug could also
3493 have affected the operation of pcre_study().
3494
3495 25. Realized that a little bit of performance could be had by replacing
3496 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3497
3498 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3499
3500 27. Possessive quantifiers such as a++ were previously implemented by turning
3501 them into atomic groups such as ($>a+). Now they have their own opcodes,
3502 which improves performance. This includes the automatically created ones
3503 from 23 above.
3504
3505 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3506 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3507 the first matched character to be a colon. This applied both to named and
3508 numbered groups.
3509
3510 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3511
3512 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3513 building PCRE on Windows easier, so I have included it in the distribution.
3514
3515 31. There is now a check in pcretest against a ridiculously large number being
3516 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3517 loop, the loop is abandoned.
3518
3519 32. Forward references to subpatterns in conditions such as (?(2)...) where
3520 subpattern 2 is defined later cause pcre_compile() to search forwards in
3521 the pattern for the relevant set of parentheses. This search went wrong
3522 when there were unescaped parentheses in a character class, parentheses
3523 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3524
3525 33. "Subroutine" calls and backreferences were previously restricted to
3526 referencing subpatterns earlier in the regex. This restriction has now
3527 been removed.
3528
3529 34. Added a number of extra features that are going to be in Perl 5.10. On the
3530 whole, these are just syntactic alternatives for features that PCRE had
3531 previously implemented using the Python syntax or my own invention. The
3532 other formats are all retained for compatibility.
3533
3534 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3535 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3536 also .NET compatible.
3537
3538 (b) A recursion or subroutine call to a named group can now be defined as
3539 (?&name) as well as (?P>name).
3540
3541 (c) A backreference to a named group can now be defined as \k<name> or
3542 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3543 5.10, are also .NET compatible.
3544
3545 (d) A conditional reference to a named group can now use the syntax
3546 (?(<name>) or (?('name') as well as (?(name).
3547
3548 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3549 groups (named and numbered) that are never evaluated inline, but can be
3550 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3551 is always false. There may be only one alternative in such a group.
3552
3553 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3554 as the simple (?(R). The condition is true only if the most recent
3555 recursion is that of the given number or name. It does not search out
3556 through the entire recursion stack.
3557
3558 (g) The escape \gN or \g{N} has been added, where N is a positive or
3559 negative number, specifying an absolute or relative reference.
3560
3561 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3562 some "unreachable code" warnings.
3563
3564 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3565 things, this adds five new scripts.
3566
3567 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3568 There were also incompatibilities regarding the handling of \Q..\E inside
3569 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3570 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3571
3572 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3573 matches an empty string, and forcibly breaks the loop. There were bugs in
3574 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3575 against aaaa the result was just "a" rather than "aaaa", for example. Two
3576 separate and independent bugs (that affected different cases) have been
3577 fixed.
3578
3579 39. Refactored the code to abolish the use of different opcodes for small
3580 capturing bracket numbers. This is a tidy that I avoided doing when I
3581 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3582 The new approach is not only tidier, it makes it possible to reduce the
3583 memory needed to fix the previous bug (38).
3584
3585 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3586 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3587 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3588 mode.
3589
3590 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3591 report.
3592
3593 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3594 copy construction and assignment in the C++ wrapper.
3595
3596 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3597 couple of bugs that could have given wrong results in the "--newline=crlf"
3598 case.
3599
3600 44. Added a number of casts and did some reorganization of signed/unsigned int
3601 variables following suggestions from Dair Grant. Also renamed the variable
3602 "this" as "item" because it is a C++ keyword.
3603
3604 45. Arranged for dftables to add
3605
3606 #include "pcre_internal.h"
3607
3608 to pcre_chartables.c because without it, gcc 4.x may remove the array
3609 definition from the final binary if PCRE is built into a static library and
3610 dead code stripping is activated.
3611
3612 46. For an unanchored pattern, if a match attempt fails at the start of a
3613 newline sequence, and the newline setting is CRLF or ANY, and the next two
3614 characters are CRLF, advance by two characters instead of one.
3615
3616
3617 Version 6.7 04-Jul-06
3618 ---------------------
3619
3620 1. In order to handle tests when input lines are enormously long, pcretest has
3621 been re-factored so that it automatically extends its buffers when
3622 necessary. The code is crude, but this _is_ just a test program. The
3623 default size has been increased from 32K to 50K.
3624
3625 2. The code in pcre_study() was using the value of the re argument before
3626 testing it for NULL. (Of course, in any sensible call of the function, it
3627 won't be NULL.)
3628
3629 3. The memmove() emulation function in pcre_internal.h, which is used on
3630 systems that lack both memmove() and bcopy() - that is, hardly ever -
3631 was missing a "static" storage class specifier.
3632
3633 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3634 containing an extended class (one that cannot be represented by a bitmap
3635 because it contains high-valued characters or Unicode property items, e.g.
3636 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3637 pattern, but PCRE should not loop if you do not (it no longer does).
3638 [Detail: two cases were found: (a) a repeated subpattern containing an
3639 extended class; (b) a recursive reference to a subpattern that followed a
3640 previous extended class. It wasn't skipping over the extended class
3641 correctly when UTF-8 mode was not set.]
3642
3643 5. A negated single-character class was not being recognized as fixed-length
3644 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3645 compile error "lookbehind assertion is not fixed length".
3646
3647 6. The RunPerlTest auxiliary script was showing an unexpected difference
3648 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3649 write a Perl script that can interpret lines of an input file either as
3650 byte characters or as UTF-8, which is what "perltest" was being required to
3651 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3652 can't do is switch easily at run time between having the "use utf8;" pragma
3653 or not. In the end, I fudged it by using the RunPerlTest script to insert
3654 "use utf8;" explicitly for the UTF-8 tests.
3655
3656 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3657 the end of the subject string, contrary to the documentation and to what
3658 Perl does. This was true of both matching functions. Now it matches only at
3659 the start of the subject and immediately after *internal* newlines.
3660
3661 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3662 a pointer to an int instead of a pointer to an unsigned long int. This
3663 caused problems on 64-bit systems.
3664
3665 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3666 instance of the 'standard' template library not being so standard".
3667
3668 10. There was no check on the number of named subpatterns nor the maximum
3669 length of a subpattern name. The product of these values is used to compute
3670 the size of the memory block for a compiled pattern. By supplying a very
3671 long subpattern name and a large number of named subpatterns, the size
3672 computation could be caused to overflow. This is now prevented by limiting
3673 the length of names to 32 characters, and the number of named subpatterns
3674 to 10,000.
3675
3676 11. Subpatterns that are repeated with specific counts have to be replicated in
3677 the compiled pattern. The size of memory for this was computed from the
3678 length of the subpattern and the repeat count. The latter is limited to
3679 65535, but there was no limit on the former, meaning that integer overflow
3680 could in principle occur. The compiled length of a repeated subpattern is
3681 now limited to 30,000 bytes in order to prevent this.
3682
3683 12. Added the optional facility to have named substrings with the same name.
3684
3685 13. Added the ability to use a named substring as a condition, using the
3686 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3687 are numbers (not recommended). Forward references are permitted.
3688
3689 14. Added forward references in named backreferences (if you see what I mean).
3690
3691 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3692 pattern could run off the end of the subject. For example, the pattern
3693 "(?s)(.{1,5})"8 did this with the subject "ab".
3694
3695 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3696 PCRE_CASELESS was set when matching characters that were quantified with ?
3697 or *.
3698
3699 17. A character class other than a single negated character that had a minimum
3700 but no maximum quantifier - for example [ab]{6,} - was not handled
3701 correctly by pce_dfa_exec(). It would match only one character.
3702
3703 18. A valid (though odd) pattern that looked like a POSIX character
3704 class but used an invalid character after [ (for example [[,abc,]]) caused
3705 pcre_compile() to give the error "Failed: internal error: code overflow" or
3706 in some cases to crash with a glibc free() error. This could even happen if
3707 the pattern terminated after [[ but there just happened to be a sequence of
3708 letters, a binary zero, and a closing ] in the memory that followed.
3709
3710 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3711 over the years. Originally (before any Unicode support), just the bottom 8
3712 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3713 output from "man perlunicode" includes this:
3714
3715 The regular expression compiler produces polymorphic opcodes. That
3716 is, the pattern adapts to the data and automatically switches to
3717 the Unicode character scheme when presented with Unicode data--or
3718 instead uses a traditional byte scheme when presented with byte
3719 data.
3720
3721 Sadly, a wide octal escape does not cause a switch, and in a string with
3722 no other multibyte characters, these octal escapes are treated as before.
3723 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3724 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3725 Unicode string.
3726
3727 I have not perpetrated such confusion in PCRE. Up till now, it took just
3728 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3729 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3730 translate to the appropriate multibyte character.
3731
3732 29. Applied some refactoring to reduce the number of warnings from Microsoft
3733 and Borland compilers. This has included removing the fudge introduced
3734 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3735 a warning about an unused variable.
3736
3737 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3738 characters since release 4.0, because Perl (from release 5.004) does not.
3739 [Or at least, is documented not to: some releases seem to be in conflict
3740 with the documentation.] However, when a pattern was studied with
3741 pcre_study() and all its branches started with \s, PCRE still included VT
3742 as a possible starting character. Of course, this did no harm; it just
3743 caused an unnecessary match attempt.
3744
3745 22. Removed a now-redundant internal flag bit that recorded the fact that case
3746 dependency changed within the pattern. This was once needed for "required
3747 byte" processing, but is no longer used. This recovers a now-scarce options
3748 bit. Also moved the least significant internal flag bit to the most-
3749 significant bit of the word, which was not previously used (hangover from
3750 the days when it was an int rather than a uint) to free up another bit for
3751 the future.
3752
3753 23. Added support for CRLF line endings as well as CR and LF. As well as the
3754 default being selectable at build time, it can now be changed at runtime
3755 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3756 specify that it is scanning data with non-default line endings.
3757
3758 24. Changed the definition of CXXLINK to make it agree with the definition of
3759 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3760
3761 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3762 recursions. This makes a big different to stack usage for some patterns.
3763
3764 26. If a subpattern containing a named recursion or subroutine reference such
3765 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3766 the space required for the compiled pattern went wrong and gave too small a
3767 value. Depending on the environment, this could lead to "Failed: internal
3768 error: code overflow at offset 49" or "glibc detected double free or
3769 corruption" errors.
3770
3771 27. Applied patches from Google (a) to support the new newline modes and (b) to
3772 advance over multibyte UTF-8 characters in GlobalReplace.
3773
3774 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3775 difference for some implementation of PCRE in some Windows version.
3776
3777 29. Added some extra testing facilities to pcretest:
3778
3779 \q<number> in a data line sets the "match limit" value
3780 \Q<number> in a data line sets the "match recursion limt" value
3781 -S <number> sets the stack size, where <number> is in megabytes
3782
3783 The -S option isn't available for Windows.
3784
3785
3786 Version 6.6 06-Feb-06
3787 ---------------------
3788
3789 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3790 in pcreposix.h. I have copied the definition from pcre.h.
3791
3792 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3793 because pcre.h is no longer a built file.
3794
3795 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3796 not normally included in the compiled code.
3797
3798
3799 Version 6.5 01-Feb-06
3800 ---------------------
3801
3802 1. When using the partial match feature with pcre_dfa_exec(), it was not
3803 anchoring the second and subsequent partial matches at the new starting
3804 point. This could lead to incorrect results. For example, with the pattern
3805 /1234/, partially matching against "123" and then "a4" gave a match.
3806
3807 2. Changes to pcregrep:
3808
3809 (a) All non-match returns from pcre_exec() were being treated as failures
3810 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3811 error message is output. Some extra information is given for the
3812 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3813 probably the only errors that are likely to be caused by users (by
3814 specifying a regex that has nested indefinite repeats, for instance).
3815 If there are more than 20 of these errors, pcregrep is abandoned.
3816
3817 (b) A binary zero was treated as data while matching, but terminated the
3818 output line if it was written out. This has been fixed: binary zeroes
3819 are now no different to any other data bytes.
3820
3821 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3822 used to set a locale for matching. The --locale=xxxx long option has
3823 been added (no short equivalent) to specify a locale explicitly on the
3824 pcregrep command, overriding the environment variables.
3825
3826 (d) When -B was used with -n, some line numbers in the output were one less
3827 than they should have been.
3828
3829 (e) Added the -o (--only-matching) option.
3830
3831 (f) If -A or -C was used with -c (count only), some lines of context were
3832 accidentally printed for the final match.
3833
3834 (g) Added the -H (--with-filename) option.
3835
3836 (h) The combination of options -rh failed to suppress file names for files
3837 that were found from directory arguments.
3838
3839 (i) Added the -D (--devices) and -d (--directories) options.
3840
3841 (j) Added the -F (--fixed-strings) option.
3842
3843 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3844
3845 (l) Added the --colo(u)r option.
3846
3847 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3848 is not present by default.
3849
3850 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3851 items such as (?R) or (?1), when the recursion could match a number of
3852 alternatives. If it matched one of the alternatives, but subsequently,
3853 outside the recursion, there was a failure, the code tried to back up into
3854 the recursion. However, because of the way PCRE is implemented, this is not
3855 possible, and the result was an incorrect result from the match.
3856
3857 In order to prevent this happening, the specification of recursion has
3858 been changed so that all such subpatterns are automatically treated as
3859 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3860
3861 4. I had overlooked the fact that, in some locales, there are characters for
3862 which isalpha() is true but neither isupper() nor islower() are true. In
3863 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3864 and ordfeminine) are like this. This affected the treatment of \w and \W
3865 when they appeared in character classes, but not when they appeared outside
3866 a character class. The bit map for "word" characters is now created
3867 separately from the results of isalnum() instead of just taking it from the
3868 upper, lower, and digit maps. (Plus the underscore character, of course.)
3869
3870 5. The above bug also affected the handling of POSIX character classes such as
3871 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3872 permanent tables. Instead, the bit maps for such a class were previously
3873 created as the appropriate unions of the upper, lower, and digit bitmaps.
3874 Now they are created by subtraction from the [[:word:]] class, which has
3875 its own bitmap.
3876
3877 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3878 It is created by subtracting the vertical space characters (\x09, \x0a,
3879 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3880 subtraction was done in the overall bitmap for a character class, meaning
3881 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3882 be recognized. This bug has been fixed.
3883
3884 7. Patches from the folks at Google:
3885
3886 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3887 real life, but is still worth protecting against".
3888
3889 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3890 regular expressions".
3891
3892 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3893 have it.
3894
3895 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3896 "configure" and the latter not, in order to fix a problem somebody had
3897 with compiling the Arg class on HP-UX.
3898
3899 (e) Improve the error-handling of the C++ wrapper a little bit.
3900
3901 (f) New tests for checking recursion limiting.
3902
3903 8. The pcre_memmove() function, which is used only if the environment does not
3904 have a standard memmove() function (and is therefore rarely compiled),
3905 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3906 returning a result (though PCRE never actually uses the result).
3907
3908 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3909 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3910 returned instead of calling malloc() with an overflowing number that would
3911 most likely cause subsequent chaos.
3912
3913 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3914
3915 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3916 with this option is matched, the nmatch and pmatch options of regexec() are
3917 ignored.
3918
3919 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3920 provided in case anyone wants to the the POSIX interface with UTF-8
3921 strings.
3922
3923 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3924 C++ linking (needed for some HP-UX environments).
3925
3926 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3927 (unused parameter) and in the pcre_printint() function (omitted "default"
3928 switch label when the default is to do nothing).
3929
3930 15. Added some code to make it possible, when PCRE is compiled as a C++
3931 library, to replace subject pointers for pcre_exec() with a smart pointer
3932 class, thus making it possible to process discontinuous strings.
3933
3934 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3935 much the same function. They were added by different people who were trying
3936 to make PCRE easy to compile on non-Unix systems. It has been suggested
3937 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3938 for compiling on Windows systems. I have therefore replaced it with
3939 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3940 defaults to "extern" for C or "extern C" for C++, which works fine on
3941 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3942 SCOPE with something explicit in config.h. In addition:
3943
3944 (a) pcreposix.h still had just "extern" instead of either of these macros;
3945 I have replaced it with PCRE_DATA_SCOPE.
3946
3947 (b) Functions such as _pcre_xclass(), which are internal to the library,
3948 but external in the C sense, all had PCRE_EXPORT in their definitions.
3949 This is apparently wrong for the Windows case, so I have removed it.
3950 (It makes no difference on Unix-like systems.)
3951
3952 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3953 of recursive calls to match(). This is different to MATCH_LIMIT because
3954 that limits the total number of calls to match(), not all of which increase
3955 the depth of recursion. Limiting the recursion depth limits the amount of
3956 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3957 when PCRE is compiled, and changed at run time. A patch from Google adds
3958 this functionality to the C++ interface.
3959
3960 18. Changes to the handling of Unicode character properties:
3961
3962 (a) Updated the table to Unicode 4.1.0.
3963
3964 (b) Recognize characters that are not in the table as "Cn" (undefined).
3965
3966 (c) I revised the way the table is implemented to a much improved format
3967 which includes recognition of ranges. It now supports the ranges that
3968 are defined in UnicodeData.txt, and it also amalgamates other
3969 characters into ranges. This has reduced the number of entries in the
3970 table from around 16,000 to around 3,000, thus reducing its size
3971 considerably. I realized I did not need to use a tree structure after
3972 all - a binary chop search is just as efficient. Having reduced the
3973 number of entries, I extended their size from 6 bytes to 8 bytes to
3974 allow for more data.
3975
3976 (d) Added support for Unicode script names via properties such as \p{Han}.
3977
3978 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3979 matching that character.
3980
3981 20. When matching a repeated Unicode property with a minimum greater than zero,
3982 (for example \pL{2,}), PCRE could look past the end of the subject if it
3983 reached it while seeking the minimum number of characters. This could
3984 happen only if some of the characters were more than one byte long, because
3985 there is a check for at least the minimum number of bytes.
3986
3987 21. Refactored the implementation of \p and \P so as to be more general, to
3988 allow for more different types of property in future. This has changed the
3989 compiled form incompatibly. Anybody with saved compiled patterns that use
3990 \p or \P will have to recompile them.
3991
3992 22. Added "Any" and "L&" to the supported property types.
3993
3994 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3995 but give a compile time error if the value is greater than 0xff.
3996
3997 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3998 accidentally not being installed or uninstalled.
3999
4000 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4001 made were to insert the current release number. This seemed silly, because
4002 it made things harder for people building PCRE on systems that don't run
4003 "configure". I have turned pcre.h into a distributed file, no longer built
4004 by "configure", with the version identification directly included. There is
4005 no longer a pcre.h.in file.
4006
4007 However, this change necessitated a change to the pcre-config script as
4008 well. It is built from pcre-config.in, and one of the substitutions was the
4009 release number. I have updated configure.ac so that ./configure now finds
4010 the release number by grepping pcre.h.
4011
4012 26. Added the ability to run the tests under valgrind.
4013
4014
4015 Version 6.4 05-Sep-05
4016 ---------------------
4017
4018 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4019 "--" to be printed when multiple files were scanned, even when none of the
4020 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4021 consider it to be a bug, and have restored the previous behaviour.
4022
4023 2. A couple of code tidies to get rid of compiler warnings.
4024
4025 3. The pcretest program used to cheat by referring to symbols in the library
4026 whose names begin with _pcre_. These are internal symbols that are not
4027 really supposed to be visible externally, and in some environments it is
4028 possible to suppress them. The cheating is now confined to including
4029 certain files from the library's source, which is a bit cleaner.
4030
4031 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4032 file's purpose clearer.
4033
4034 5. Reorganized pcre_ucp_findchar().
4035
4036
4037 Version 6.3 15-Aug-05
4038 ---------------------
4039
4040 1. The file libpcre.pc.in did not have general read permission in the tarball.
4041
4042 2. There were some problems when building without C++ support:
4043
4044 (a) If C++ support was not built, "make install" and "make test" still
4045 tried to test it.
4046
4047 (b) There were problems when the value of CXX was explicitly set. Some
4048 changes have been made to try to fix these, and ...
4049
4050 (c) --disable-cpp can now be used to explicitly disable C++ support.
4051
4052 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4053 backslash in a target when C++ was disabled. This confuses some
4054 versions of "make", apparently. Using an intermediate variable solves
4055 this. (Same for CPP_LOBJ.)
4056
4057 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4058 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4059 necessary on certain architectures.
4060
4061 4. Added a setting of -export-symbols-regex to the link command to remove
4062 those symbols that are exported in the C sense, but actually are local
4063 within the library, and not documented. Their names all begin with
4064 "_pcre_". This is not a perfect job, because (a) we have to except some
4065 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4066 available (and never for static libraries). I have made a note to try to
4067 find a way round (a) in the future.
4068
4069
4070 Version 6.2 01-Aug-05
4071 ---------------------
4072
4073 1. There was no test for integer overflow of quantifier values. A construction
4074 such as {1111111111111111} would give undefined results. What is worse, if
4075 a minimum quantifier for a parenthesized subpattern overflowed and became
4076 negative, the calculation of the memory size went wrong. This could have
4077 led to memory overwriting.
4078
4079 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4080
4081 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4082 operating environments where this matters.
4083
4084 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4085 PCRE options from within the C++ wrapper.
4086
4087 5. Named capturing subpatterns were not being correctly counted when a pattern
4088 was compiled. This caused two problems: (a) If there were more than 100
4089 such subpatterns, the calculation of the memory needed for the whole
4090 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4091 back references of the form \12, where the number was greater than 9, were
4092 not recognized as back references, even though there were sufficient
4093 previous subpatterns.
4094
4095 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4096 versions of gcc, e.g. 2.95.4.
4097
4098
4099 Version 6.1 21-Jun-05
4100 ---------------------
4101
4102 1. There was one reference to the variable "posix" in pcretest.c that was not
4103 surrounded by "#if !defined NOPOSIX".
4104
4105 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4106 the cross-check on the old pcre_info() function, for the benefit of the
4107 cut-down version of PCRE that is currently imported into Exim.
4108
4109 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4110 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4111 patterns that start (?i) but I don't think that matters. The use of (?i) is
4112 just an example; this all applies to the other options as well.
4113
4114 4. Since libtool seems to echo the compile commands it is issuing, the output
4115 from "make" can be reduced a bit by putting "@" in front of each libtool
4116 compile command.
4117
4118 5. Patch from the folks at Google for configure.in to be a bit more thorough
4119 in checking for a suitable C++ installation before trying to compile the
4120 C++ stuff. This should fix a reported problem when a compiler was present,
4121 but no suitable headers.
4122
4123 6. The man pages all had just "PCRE" as their title. I have changed them to
4124 be the relevant file name. I have also arranged that these names are
4125 retained in the file doc/pcre.txt, which is a concatenation in text format
4126 of all the man pages except the little individual ones for each function.
4127
4128 7. The NON-UNIX-USE file had not been updated for the different set of source
4129 files that come with release 6. I also added a few comments about the C++
4130 wrapper.
4131
4132
4133 Version 6.0 07-Jun-05
4134 ---------------------
4135
4136 1. Some minor internal re-organization to help with my DFA experiments.
4137
4138 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4139 didn't matter for the library itself when fully configured, but did matter
4140 when compiling without UCP support, or within Exim, where the ucp files are
4141 not imported.
4142
4143 3. Refactoring of the library code to split up the various functions into
4144 different source modules. The addition of the new DFA matching code (see
4145 below) to a single monolithic source would have made it really too
4146 unwieldy, quite apart from causing all the code to be include in a
4147 statically linked application, when only some functions are used. This is
4148 relevant even without the DFA addition now that patterns can be compiled in
4149 one application and matched in another.
4150
4151 The downside of splitting up is that there have to be some external
4152 functions and data tables that are used internally in different modules of
4153 the library but which are not part of the API. These have all had their
4154 names changed to start with "_pcre_" so that they are unlikely to clash
4155 with other external names.
4156
4157 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4158 a different (DFA) algorithm. Although it is slower than the original
4159 function, it does have some advantages for certain types of matching
4160 problem.
4161
4162 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4163 including restarting after a partial match.
4164
4165 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4166 defined when compiling for Windows was sent to me. I have put it into the
4167 code, though I have no means of testing or verifying it.
4168
4169 7. Added the pcre_refcount() auxiliary function.
4170
4171 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4172 match before or at the first newline in the subject string. In pcretest,
4173 the /f option on a pattern can be used to set this.
4174
4175 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4176 would behave wrongly. This has been present in PCRE since release 4.0.
4177
4178 10. A number of changes to the pcregrep command:
4179
4180 (a) Refactored how -x works; insert ^(...)$ instead of setting
4181 PCRE_ANCHORED and checking the length, in preparation for adding
4182 something similar for -w.
4183
4184 (b) Added the -w (match as a word) option.
4185
4186 (c) Refactored the way lines are read and buffered so as to have more
4187 than one at a time available.
4188
4189 (d) Implemented a pcregrep test script.
4190
4191 (e) Added the -M (multiline match) option. This allows patterns to match
4192 over several lines of the subject. The buffering ensures that at least
4193 8K, or the rest of the document (whichever is the shorter) is available
4194 for matching (and similarly the previous 8K for lookbehind assertions).
4195
4196 (f) Changed the --help output so that it now says
4197
4198 -w, --word-regex(p)
4199
4200 instead of two lines, one with "regex" and the other with "regexp"
4201 because that confused at least one person since the short forms are the
4202 same. (This required a bit of code, as the output is generated
4203 automatically from a table. It wasn't just a text change.)
4204
4205 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4206 option but starts with a hyphen. Could be a pattern or a path name
4207 starting with a hyphen, for instance.
4208
4209 (h) "-" can be given as a file name to represent stdin.
4210
4211 (i) When file names are being printed, "(standard input)" is used for
4212 the standard input, for compatibility with GNU grep. Previously
4213 "<stdin>" was used.
4214
4215 (j) The option --label=xxx can be used to supply a name to be used for
4216 stdin when file names are being printed. There is no short form.
4217
4218 (k) Re-factored the options decoding logic because we are going to add
4219 two more options that take data. Such options can now be given in four
4220 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4221
4222 (l) Added the -A, -B, and -C options for requesting that lines of context
4223 around matches be printed.
4224
4225 (m) Added the -L option to print the names of files that do not contain
4226 any matching lines, that is, the complement of -l.
4227
4228 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4229 continue to scan other files.
4230
4231 (o) The -s option was incorrectly implemented. For compatibility with other
4232 greps, it now suppresses the error message for a non-existent or non-
4233 accessible file (but not the return code). There is a new option called
4234 -q that suppresses the output of matching lines, which was what -s was
4235 previously doing.
4236
4237 (p) Added --include and --exclude options to specify files for inclusion
4238 and exclusion when recursing.
4239
4240 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4241 Hopefully, it now does.
4242
4243 12. Missing cast in pcre_study().
4244
4245 13. Added an "uninstall" target to the makefile.
4246
4247 14. Replaced "extern" in the function prototypes in Makefile.in with
4248 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4249 world, but is set differently for Windows.
4250
4251 15. Added a second compiling function called pcre_compile2(). The only
4252 difference is that it has an extra argument, which is a pointer to an
4253 integer error code. When there is a compile-time failure, this is set
4254 non-zero, in addition to the error test pointer being set to point to an
4255 error message. The new argument may be NULL if no error number is required
4256 (but then you may as well call pcre_compile(), which is now just a
4257 wrapper). This facility is provided because some applications need a
4258 numeric error indication, but it has also enabled me to tidy up the way
4259 compile-time errors are handled in the POSIX wrapper.
4260
4261 16. Added VPATH=.libs to the makefile; this should help when building with one
4262 prefix path and installing with another. (Or so I'm told by someone who
4263 knows more about this stuff than I do.)
4264
4265 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4266 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4267 match everything, including newlines. This is not POSIX-compatible, but
4268 somebody wanted the feature. From pcretest it can be activated by using
4269 both the P and the s flags.
4270
4271 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4272
4273 19. libpcre.pc was being incorrectly installed as executable.
4274
4275 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4276 it now also looks for '\r' so that it will work unmodified on Windows.
4277
4278 21. Added Google's contributed C++ wrapper to the distribution.
4279
4280 22. Added some untidy missing memory free() calls in pcretest, to keep
4281 Electric Fence happy when testing.
4282
4283
4284
4285 Version 5.0 13-Sep-04
4286 ---------------------
4287
4288 1. Internal change: literal characters are no longer packed up into items
4289 containing multiple characters in a single byte-string. Each character
4290 is now matched using a separate opcode. However, there may be more than one
4291 byte in the character in UTF-8 mode.
4292
4293 2. The pcre_callout_block structure has two new fields: pattern_position and
4294 next_item_length. These contain the offset in the pattern to the next match
4295 item, and its length, respectively.
4296
4297 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4298 insertion of callouts before each pattern item. Added the /C option to
4299 pcretest to make use of this.
4300
4301 4. On the advice of a Windows user, the lines
4302
4303 #if defined(_WIN32) || defined(WIN32)
4304 _setmode( _fileno( stdout ), 0x8000 );
4305 #endif /* defined(_WIN32) || defined(WIN32) */
4306
4307 have been added to the source of pcretest. This apparently does useful
4308 magic in relation to line terminators.
4309
4310 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4311 for the benefit of those environments where the "b" makes a difference.
4312
4313 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4314 to know about it. I have put a hack into configure.in that adds in code
4315 to set GCC=yes if CC=icc. This seems to end up at a point in the
4316 generated configure script that is early enough to affect the setting of
4317 compiler options, which is what is needed, but I have no means of testing
4318 whether it really works. (The user who reported this had patched the
4319 generated configure script, which of course I cannot do.)
4320
4321 LATER: After change 22 below (new libtool files), the configure script
4322 seems to know about icc (and also ecc). Therefore, I have commented out
4323 this hack in configure.in.
4324
4325 7. Added support for pkg-config (2 patches were sent in).
4326
4327 8. Negated POSIX character classes that used a combination of internal tables
4328 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4329 [[:^ascii]]. Typically, they would match almost any characters. The other
4330 POSIX classes were not broken in this way.
4331
4332 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4333 to find the match, as PCRE was deluded into thinking that the match had to
4334 start at the start point or following a newline. The same bug applied to
4335 patterns with negative forward assertions or any backward assertions
4336 preceding ".*" at the start, unless the pattern required a fixed first
4337 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4338
4339 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4340 starting at the last subject character, bytes beyond the end of the subject
4341 string were read.
4342
4343 11. Renamed the variable "class" as "classbits" to make life easier for C++
4344 users. (Previously there was a macro definition, but it apparently wasn't
4345 enough.)
4346
4347 12. Added the new field "tables" to the extra data so that tables can be passed
4348 in at exec time, or the internal tables can be re-selected. This allows
4349 a compiled regex to be saved and re-used at a later time by a different
4350 program that might have everything at different addresses.
4351
4352 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4353 -R library as well as a -L library.
4354
4355 14. The debugging options of pcretest (-d on the command line or D on a
4356 pattern) showed incorrect output for anything following an extended class
4357 that contained multibyte characters and which was followed by a quantifier.
4358
4359 15. Added optional support for general category Unicode character properties
4360 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4361 support. It adds about 90K to the size of the library. The meanings of the
4362 inbuilt class escapes such as \d and \s have NOT been changed.
4363
4364 16. Updated pcredemo.c to include calls to free() to release the memory for the
4365 compiled pattern.
4366
4367 17. The generated file chartables.c was being created in the source directory
4368 instead of in the building directory. This caused the build to fail if the
4369 source directory was different from the building directory, and was
4370 read-only.
4371
4372 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4373 file. No doubt somebody will tell me if they don't make sense... Also added
4374 Dan Mooney's comments about building on OpenVMS.
4375
4376 19. Added support for partial matching via the PCRE_PARTIAL option for
4377 pcre_exec() and the \P data escape in pcretest.
4378
4379 20. Extended pcretest with 3 new pattern features:
4380
4381 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4382 write the compiled pattern to the file whose name is "rest-of-line".
4383 This is a straight binary dump of the data, with the saved pointer to
4384 the character tables forced to be NULL. The study data, if any, is
4385 written too. After writing, pcretest reads a new pattern.
4386
4387 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4388 compiled pattern from the given file. There must not be any
4389 occurrences of "<" in the file name (pretty unlikely); if there are,
4390 pcretest will instead treat the initial "<" as a pattern delimiter.
4391 After reading in the pattern, pcretest goes on to read data lines as
4392 usual.
4393
4394 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4395 and 16-bit fields in a compiled pattern, to simulate a pattern that
4396 was compiled on a host of opposite endianness.
4397
4398 21. The pcre-exec() function can now cope with patterns that were compiled on
4399 hosts of opposite endianness, with this restriction:
4400
4401 As for any compiled expression that is saved and used later, the tables
4402 pointer field cannot be preserved; the extra_data field in the arguments
4403 to pcre_exec() should be used to pass in a tables address if a value
4404 other than the default internal tables were used at compile time.
4405
4406 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4407 now diagnosed as an error. Previously, most of the time, a negative number
4408 would have been treated as zero, but if in addition "ovector" was passed as
4409 NULL, a crash could occur.
4410
4411 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4412 new versions from the libtool 1.5 distribution (the last one is a copy of
4413 a file called libtool.m4). This seems to have fixed the need to patch
4414 "configure" to support Darwin 1.3 (which I used to do). However, I still
4415 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4416 workstation).
4417
4418 24. Changed the PCRE licence to be the more standard "BSD" licence.
4419
4420
4421 Version 4.5 01-Dec-03
4422 ---------------------
4423
4424 1. There has been some re-arrangement of the code for the match() function so
4425 that it can be compiled in a version that does not call itself recursively.
4426 Instead, it keeps those local variables that need separate instances for
4427 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4428 needs to "recurse". Keeping track of where control must go is done by means
4429 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4430 hide most of the details from the main code, and operates only if
4431 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4432 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4433 operating.
4434
4435 To make it easier for callers to provide specially tailored get/free
4436 functions for this usage, two new functions, pcre_stack_malloc, and
4437 pcre_stack_free, are used. They are always called in strict stacking order,
4438 and the size of block requested is always the same.
4439
4440 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4441 PCRE has been compiled to use the stack or the heap for recursion. The
4442 -C option of pcretest uses this to show which version is compiled.
4443
4444 A new data escape \S, is added to pcretest; it causes the amounts of store
4445 obtained and freed by both kinds of malloc/free at match time to be added
4446 to the output.
4447
4448 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4449 what's available on my current Linux desktop machine.
4450
4451 3. When matching a UTF-8 string, the test for a valid string at the start has
4452 been extended. If start_offset is not zero, PCRE now checks that it points
4453 to a byte that is the start of a UTF-8 character. If not, it returns
4454 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4455 this is necessary because there may be backward assertions in the pattern.
4456 When matching the same subject several times, it may save resources to use
4457 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4458
4459 4. The code for checking the validity of UTF-8 strings has been tightened so
4460 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4461 containing "overlong sequences".
4462
4463 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4464 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4465 should have been "&", but it just so happened that all the cases this let
4466 through by mistake were picked up later in the function.
4467
4468 6. I had used a variable called "isblank" - this is a C99 function, causing
4469 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4470
4471 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4472 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4473 so the version is shown; (c) stop "make test" echoing "./RunTest".
4474
4475 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4476
4477 9. The prototype for memmove() for systems that don't have it was using
4478 size_t, but the inclusion of the header that defines size_t was later. I've
4479 moved the #includes for the C headers earlier to avoid this.
4480
4481 10. Added some adjustments to the code to make it easier to compiler on certain
4482 special systems:
4483
4484 (a) Some "const" qualifiers were missing.
4485 (b) Added the macro EXPORT before all exported functions; by default this
4486 is defined to be empty.
4487 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4488 that it reads its output file name as an argument instead of writing
4489 to the standard output and assuming this can be redirected.
4490
4491 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4492 class containing characters with values greater than 255, PCRE compilation
4493 went into a loop.
4494
4495 12. A recursive reference to a subpattern that was within another subpattern
4496 that had a minimum quantifier of zero caused PCRE to crash. For example,
4497 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4498 recursion. If the recursively-called subpattern itself had a zero repeat,
4499 that was OK.
4500
4501 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4502 buffer into which it was copied (for escape processing) was still set at
4503 1024, so long lines caused crashes.
4504
4505 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4506 "internal error: code overflow...". This applied to any character class
4507 that was followed by a possessive quantifier.
4508
4509 15. Modified the Makefile to add libpcre.la as a prerequisite for
4510 libpcreposix.la because I was told this is needed for a parallel build to
4511 work.
4512
4513 16. If a pattern that contained .* following optional items at the start was
4514 studied, the wrong optimizing data was generated, leading to matching
4515 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4516 matching string must start with a or b or c. The correct conclusion for
4517 this pattern is that a match can start with any character.
4518
4519
4520 Version 4.4 13-Aug-03
4521 ---------------------
4522
4523 1. In UTF-8 mode, a character class containing characters with values between
4524 127 and 255 was not handled correctly if the compiled pattern was studied.
4525 In fixing this, I have also improved the studying algorithm for such
4526 classes (slightly).
4527
4528 2. Three internal functions had redundant arguments passed to them. Removal
4529 might give a very teeny performance improvement.
4530
4531 3. Documentation bug: the value of the capture_top field in a callout is *one
4532 more than* the number of the hightest numbered captured substring.
4533
4534 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4535 in incorrectly linking with a previously installed version. They now link
4536 explicitly with libpcre.la.
4537
4538 5. configure.in no longer needs to recognize Cygwin specially.
4539
4540 6. A problem in pcre.in for Windows platforms is fixed.
4541
4542 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4543 pcretest, it used to include the size of the study block as part of its
4544 output. Unfortunately, the structure contains a field that has a different
4545 size on different hardware architectures. This meant that the tests that
4546 showed this size failed. As the block is currently always of a fixed size,
4547 this information isn't actually particularly useful in pcretest output, so
4548 I have just removed it.
4549
4550 8. Three pre-processor statements accidentally did not start in column 1.
4551 Sadly, there are *still* compilers around that complain, even though
4552 standard C has not required this for well over a decade. Sigh.
4553
4554 9. In pcretest, the code for checking callouts passed small integers in the
4555 callout_data field, which is a void * field. However, some picky compilers
4556 complained about the casts involved for this on 64-bit systems. Now
4557 pcretest passes the address of the small integer instead, which should get
4558 rid of the warnings.
4559
4560 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4561 both compile and run time, and gives an error if an invalid UTF-8 sequence
4562 is found. There is a option for disabling this check in cases where the
4563 string is known to be correct and/or the maximum performance is wanted.
4564
4565 11. In response to a bug report, I changed one line in Makefile.in from
4566
4567 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4568 to
4569 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4570
4571 to look similar to other lines, but I have no way of telling whether this
4572 is the right thing to do, as I do not use Windows. No doubt I'll get told
4573 if it's wrong...
4574
4575
4576 Version 4.3 21-May-03
4577 ---------------------
4578
4579 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4580 Makefile.
4581
4582 2. Some refactoring to improve the quality of the code:
4583
4584 (i) The utf8_table... variables are now declared "const".
4585
4586 (ii) The code for \cx, which used the "case flipping" table to upper case
4587 lower case letters, now just substracts 32. This is ASCII-specific,
4588 but the whole concept of \cx is ASCII-specific, so it seems
4589 reasonable.
4590
4591 (iii) PCRE was using its character types table to recognize decimal and
4592 hexadecimal digits in the pattern. This is silly, because it handles
4593 only 0-9, a-f, and A-F, but the character types table is locale-
4594 specific, which means strange things might happen. A private
4595 table is now used for this - though it costs 256 bytes, a table is
4596 much faster than multiple explicit tests. Of course, the standard
4597 character types table is still used for matching digits in subject
4598 strings against \d.
4599
4600 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4601 ending in _t are). So I've renamed it as ESC_tee.
4602
4603 3. The first argument for regexec() in the POSIX wrapper should have been
4604 defined as "const".
4605
4606 4. Changed pcretest to use malloc() for its buffers so that they can be
4607 Electric Fenced for debugging.
4608
4609 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4610 to read one or more bytes before the start of the subject string. Often this
4611 had no effect on PCRE's behaviour, but in some circumstances it could
4612 provoke a segmentation fault.
4613
4614 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4615 to try to read one or more bytes before the start of the subject string.
4616
4617 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4618 UTF-8 support could misbehave in various ways if the subject string
4619 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4620 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4621 back over UTF-8 characters.)
4622
4623
4624 Version 4.2 14-Apr-03
4625 ---------------------
4626
4627 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4628
4629 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4630 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4631 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4632 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4633 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4634 and BUILD_EXEEXT
4635 Note: automatic setting of the BUILD variables is not yet working
4636 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4637 compile-time but not at link-time
4638 [LINK]: use for linking executables only
4639 make different versions for Windows and non-Windows
4640 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4641 libraries
4642 [LINK_FOR_BUILD]: new variable
4643 [OBJEXT]: use throughout
4644 [EXEEXT]: use throughout
4645 <winshared>: new target
4646 <wininstall>: new target
4647 <dftables.o>: use native compiler
4648 <dftables>: use native linker
4649 <install>: handle Windows platform correctly
4650 <clean>: ditto
4651 <check>: ditto
4652 copy DLL to top builddir before testing
4653
4654 As part of these changes, -no-undefined was removed again. This was reported
4655 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4656 in any case.
4657
4658 3. Some tidies to get rid of compiler warnings:
4659
4660 . In the match_data structure, match_limit was an unsigned long int, whereas
4661 match_call_count was an int. I've made them both unsigned long ints.
4662
4663 . In pcretest the fact that a const uschar * doesn't automatically cast to
4664 a void * provoked a warning.
4665
4666 . Turning on some more compiler warnings threw up some "shadow" variables
4667 and a few more missing casts.
4668
4669 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4670 option, a class that contained a single character with a value between 128
4671 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4672
4673 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4674 option, a class that contained several characters, but with at least one
4675 whose value was between 128 and 255 caused PCRE to crash.
4676
4677
4678 Version 4.1 12-Mar-03
4679 ---------------------
4680
4681 1. Compiling with gcc -pedantic found a couple of places where casts were
4682 needed, and a string in dftables.c that was longer than standard compilers are
4683 required to support.
4684
4685 2. Compiling with Sun's compiler found a few more places where the code could
4686 be tidied up in order to avoid warnings.
4687
4688 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4689 first of these names is deprecated in the latest Autoconf in favour of the name
4690 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4691 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4692 analogy I have changed it to CFLAGS_FOR_BUILD.
4693
4694 4. Added -no-undefined to the linking command in the Makefile, because this is
4695 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4696 linking step for the pcreposix library.
4697
4698 5. PCRE was failing to diagnose the case of two named groups with the same
4699 name.
4700
4701 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4702 literal character that is needed in the subject for a match, and scans along to
4703 ensure that it is present before embarking on the full matching process. This
4704 saves time in cases of nested unlimited repeats that are never going to match.
4705 Problem: the scan can take a lot of time if the subject is very long (e.g.
4706 megabytes), thus penalizing straightforward matches. It is now done only if the
4707 amount of subject to be scanned is less than 1000 bytes.
4708
4709 7. A lesser problem with the same optimization is that it was recording the
4710 first character of an anchored pattern as "needed", thus provoking a search
4711 right along the subject, even when the first match of the pattern was going to
4712 fail. The "needed" character is now not set for anchored patterns, unless it
4713 follows something in the pattern that is of non-fixed length. Thus, it still
4714 fulfils its original purpose of finding quick non-matches in cases of nested
4715 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4716
4717
4718 Version 4.0 17-Feb-03
4719 ---------------------
4720
4721 1. If a comment in an extended regex that started immediately after a meta-item
4722 extended to the end of string, PCRE compiled incorrect data. This could lead to
4723 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4724
4725 2. Moved to autoconf 2.53 and libtool 1.4.2.
4726
4727 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4728 the special perltest8 script is no longer needed - all the tests can be run
4729 from a single perltest script.
4730
4731 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4732 by \s. It has now been removed in PCRE. This means it isn't recognized as
4733 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4734 class [:space:] *does* include VT, thereby creating a mess.
4735
4736 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4737 space and tab.
4738
4739 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4740 its new features into the main test script, reducing the number of scripts.
4741
4742 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4743 were backward compatible, and made the (?i) apply to the whole pattern, as if
4744 /i were given. Now it behaves more logically, and applies the option setting
4745 only to what follows. PCRE has been changed to follow suit. However, if it
4746 finds options settings right at the start of the pattern, it extracts them into
4747 the global options, as before. Thus, they show up in the info data.
4748
4749 8. Added support for the \Q...\E escape sequence. Characters in between are
4750 treated as literals. This is slightly different from Perl in that $ and @ are
4751 also handled as literals inside the quotes. In Perl, they will cause variable
4752 interpolation. Note the following examples:
4753
4754 Pattern PCRE matches Perl matches
4755
4756 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4757 \Qabc\$xyz\E abc\$xyz abc\$xyz
4758 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4759
4760 For compatibility with Perl, \Q...\E sequences are recognized inside character
4761 classes as well as outside them.
4762
4763 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4764 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4765 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4766 signed/unsigned warnings.
4767
4768 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4769 option for pcretest, so I've replaced it by a simple function that does just
4770 that job.
4771
4772 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4773 "pcregrep -".
4774
4775 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4776 Java package. This provides some syntactic sugar for simple cases of what my
4777 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4778 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4779 item, you can use this simplified notation. Note that only makes sense with
4780 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4781 greediness, whatever the setting of the PCRE_UNGREEDY option.
4782
4783 13. A change of greediness default within a pattern was not taking effect at
4784 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4785 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4786 was abstracted outside.
4787
4788 14. PCRE now supports the \G assertion. It is true when the current matching
4789 position is at the start point of the match. This differs from \A when the
4790 starting offset is non-zero. Used with the /g option of pcretest (or similar
4791 code), it works in the same way as it does for Perl's /g option. If all
4792 alternatives of a regex begin with \G, the expression is anchored to the start
4793 match position, and the "anchored" flag is set in the compiled expression.
4794
4795 15. Some bugs concerning the handling of certain option changes within patterns
4796 have been fixed. These applied to options other than (?ims). For example,
4797 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4798 been the other way round. Some of this was related to change 7 above.
4799
4800 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4801 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4802 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4803 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4804
4805 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4806 mode. Unlike ".", it always matches newline, whatever the setting of
4807 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4808 assertions. Perl allows it, but it doesn't (in general) work because it can't
4809 calculate the length of the lookbehind. At least, that's the case for Perl
4810 5.8.0 - I've been told they are going to document that it doesn't work in
4811 future.
4812
4813 18. Added an error diagnosis for escapes that PCRE does not support: these are
4814 \L, \l, \N, \P, \p, \U, \u, and \X.
4815
4816 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4817 reading past the end of the pattern in cases such as /[abcd/.
4818
4819 20. PCRE was getting more memory than necessary for patterns with classes that
4820 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4821
4822 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4823 compiling PCRE for use with Virtual Pascal.
4824
4825 22. Small fix to the Makefile to make it work properly if the build is done
4826 outside the source tree.
4827
4828 23. Added a new extension: a condition to go with recursion. If a conditional
4829 subpattern starts with (?(R) the "true" branch is used if recursion has
4830 happened, whereas the "false" branch is used only at the top level.
4831
4832 24. When there was a very long string of literal characters (over 255 bytes
4833 without UTF support, over 250 bytes with UTF support), the computation of how
4834 much memory was required could be incorrect, leading to segfaults or other
4835 strange effects.
4836
4837 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4838 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4839 there was a subsequent back reference to those brackets. This meant that, for
4840 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4841 possible to check for precisely this case. All we can do is abandon the
4842 optimization if .* occurs inside capturing brackets when there are any back
4843 references whatsoever. (See below for a better fix that came later.)
4844
4845 26. The handling of the optimization for finding the first character of a
4846 non-anchored pattern, and for finding a character that is required later in the
4847 match were failing in some cases. This didn't break the matching; it just
4848 failed to optimize when it could. The way this is done has been re-implemented.
4849
4850 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4851
4852 28. Added a new feature that provides some of the functionality that Perl
4853 provides with (?{...}). The facility is termed a "callout". The way it is done
4854 in PCRE is for the caller to provide an optional function, by setting
4855 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4856 global variable. By default it is unset, which disables all calling out. To get
4857 the function called, the regex must include (?C) at appropriate points. This
4858 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4859 This provides a means of identifying different callout points. When PCRE
4860 reaches such a point in the regex, if pcre_callout has been set, the external
4861 function is called. It is provided with data in a structure called
4862 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4863 matching continues; if it returns a non-zero value, the match at the current
4864 point fails. However, backtracking will occur if possible. [This was changed
4865 later and other features added - see item 49 below.]
4866
4867 29. pcretest is upgraded to test the callout functionality. It provides a
4868 callout function that displays information. By default, it shows the start of
4869 the match and the current position in the text. There are some new data escapes
4870 to vary what happens:
4871
4872 \C+ in addition, show current contents of captured substrings
4873 \C- do not supply a callout function
4874 \C!n return 1 when callout number n is reached
4875 \C!n!m return 1 when callout number n is reached for the mth time
4876
4877 30. If pcregrep was called with the -l option and just a single file name, it
4878 output "<stdin>" if a match was found, instead of the file name.
4879
4880 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4881 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4882 pcre_exec(). This saves a malloc/free per call. The default value of
4883 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4884 when configuring.
4885
4886 32. The default maximum size of a compiled pattern is 64K. There have been a
4887 few cases of people hitting this limit. The code now uses macros to handle the
4888 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4889 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4890 configuring. Tests 2 and 5 work only with 2-byte links because they output
4891 debugging information about compiled patterns.
4892
4893 33. Internal code re-arrangements:
4894
4895 (a) Moved the debugging function for printing out a compiled regex into
4896 its own source file (printint.c) and used #include to pull it into
4897 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4898 separate copies.
4899
4900 (b) Defined the list of op-code names for debugging as a macro in
4901 internal.h so that it is next to the definition of the opcodes.
4902
4903 (c) Defined a table of op-code lengths for simpler skipping along compiled
4904 code. This is again a macro in internal.h so that it is next to the
4905 definition of the opcodes.
4906
4907 34. Added support for recursive calls to individual subpatterns, along the
4908 lines of Robin Houston's patch (but implemented somewhat differently).
4909
4910 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4911 allow it to read and process whole directories in Win32. This code was
4912 contributed by Lionel Fourquaux; it has not been tested by me.
4913
4914 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4915 used to name a group. Names consist of alphanumerics and underscores, and must
4916 be unique. Back references use the syntax (?P=name) and recursive calls use
4917 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4918 numbers. The function pcre_fullinfo() can be used after compilation to extract
4919 a name/number map. There are three relevant calls:
4920
4921 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4922 PCRE_INFO_NAMECOUNT yields the number of entries
4923 PCRE_INFO_NAMETABLE yields a pointer to the map.
4924
4925 The map is a vector of fixed-size entries. The size of each entry depends on
4926 the length of the longest name used. The first two bytes of each entry are the
4927 group number, most significant byte first. There follows the corresponding
4928 name, zero terminated. The names are in alphabetical order.
4929
4930 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4931 case instead of 255. Making it the same both with and without UTF-8 support
4932 means that the same test output works with both.
4933
4934 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4935 calling malloc() with a zero argument.
4936
4937 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4938 optimization. I've improved things by keeping a bitmap of backreferences with
4939 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4940 fact referenced, the optimization can be applied. It is unlikely that a
4941 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4942 the match to follow \n) will appear inside brackets with a number greater than
4943 31, but if it does, any back reference > 31 suppresses the optimization.
4944
4945 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4946 of disabling numbered capturing parentheses. Any opening parenthesis that is
4947 not followed by ? behaves as if it were followed by ?: but named parentheses
4948 can still be used for capturing (and they will acquire numbers in the usual
4949 way).
4950
4951 41. Redesigned the return codes from the match() function into yes/no/error so
4952 that errors can be passed back from deep inside the nested calls. A malloc
4953 failure while inside a recursive subpattern call now causes the
4954 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4955
4956 42. It is now possible to set a limit on the number of times the match()
4957 function is called in a call to pcre_exec(). This facility makes it possible to
4958 limit the amount of recursion and backtracking, though not in a directly
4959 obvious way, because the match() function is used in a number of different
4960 circumstances. The count starts from zero for each position in the subject
4961 string (for non-anchored patterns). The default limit is, for compatibility, a
4962 large number, namely 10 000 000. You can change this in two ways:
4963
4964 (a) When configuring PCRE before making, you can use --with-match-limit=n
4965 to set a default value for the compiled library.
4966
4967 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4968 a different value is set. See 45 below.
4969
4970 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4971
4972 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4973 of things that can be changed at compile time. The first argument specifies
4974 what is wanted and the second points to where the information is to be placed.
4975 The current list of available information is:
4976
4977 PCRE_CONFIG_UTF8
4978
4979 The output is an integer that is set to one if UTF-8 support is available;
4980 otherwise it is set to zero.
4981
4982 PCRE_CONFIG_NEWLINE
4983
4984 The output is an integer that it set to the value of the code that is used for
4985 newline. It is either LF (10) or CR (13).
4986
4987 PCRE_CONFIG_LINK_SIZE
4988
4989 The output is an integer that contains the number of bytes used for internal
4990 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4991
4992 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4993
4994 The output is an integer that contains the threshold above which the POSIX
4995 interface uses malloc() for output vectors. See item 31 above.
4996
4997 PCRE_CONFIG_MATCH_LIMIT
4998
4999 The output is an unsigned integer that contains the default limit of the number
5000 of match() calls in a pcre_exec() execution. See 42 above.
5001
5002 44. pcretest has been upgraded by the addition of the -C option. This causes it
5003 to extract all the available output from the new pcre_config() function, and to
5004 output it. The program then exits immediately.
5005
5006 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5007 order to support additional features. One way would have been to define
5008 pcre_exec2() (for example) with extra arguments, but this would not have been
5009 extensible, and would also have required all calls to the original function to
5010 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5011 is used for passing in "extra" data from pcre_study().
5012
5013 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5014 contains the following fields:
5015
5016 flags a bitmap indicating which of the following fields are set
5017 study_data opaque data from pcre_study()
5018 match_limit a way of specifying a limit on match() calls for a specific
5019 call to pcre_exec()
5020 callout_data data for callouts (see 49 below)
5021
5022 The flag bits are also defined in pcre.h, and are
5023
5024 PCRE_EXTRA_STUDY_DATA
5025 PCRE_EXTRA_MATCH_LIMIT
5026 PCRE_EXTRA_CALLOUT_DATA
5027
5028 The pcre_study() function now returns one of these new pcre_extra blocks, with
5029 the actual study data pointed to by the study_data field, and the
5030 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5031 before. That is, this change is entirely upwards-compatible and requires no
5032 change to existing code.
5033
5034 If you want to pass in additional data to pcre_exec(), you can either place it
5035 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5036 block.
5037
5038 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5039 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5040 times with different match limits, until it finds the minimum value needed for
5041 pcre_exec() to complete. The value is then output. This can be instructive; for
5042 most simple matches the number is quite small, but for pathological cases it
5043 gets very large very quickly.
5044
5045 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5046 returns the size of the data block pointed to by the study_data field in a
5047 pcre_extra block, that is, the value that was passed as the argument to
5048 pcre_malloc() when PCRE was getting memory in which to place the information
5049 created by pcre_study(). The fourth argument should point to a size_t variable.
5050 pcretest has been extended so that this information is shown after a successful
5051 pcre_study() call when information about the compiled regex is being displayed.
5052
5053 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5054 because what follows is always an absolute path. (Later: it turns out that this
5055 is more than cosmetic for MinGW, because it doesn't like empty path
5056 components.)
5057
5058 49. Some changes have been made to the callout feature (see 28 above):
5059
5060 (i) A callout function now has three choices for what it returns:
5061
5062 0 => success, carry on matching
5063 > 0 => failure at this point, but backtrack if possible
5064 < 0 => serious error, return this value from pcre_exec()
5065
5066 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5067 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5068 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5069 use by callout functions. It will never be used by PCRE itself.
5070
5071 (ii) The pcre_extra structure (see 45 above) has a void * field called
5072 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5073 pcre_callout_block structure has a field of the same name. The contents of
5074 the field passed in the pcre_extra structure are passed to the callout
5075 function in the corresponding field in the callout block. This makes it
5076 easier to use the same callout-containing regex from multiple threads. For
5077 testing, the pcretest program has a new data escape
5078
5079 \C*n pass the number n (may be negative) as callout_data
5080
5081 If the callout function in pcretest receives a non-zero value as
5082 callout_data, it returns that value.
5083
5084 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5085 there were some redundant $(CFLAGS) in commands that are now specified as
5086 $(LINK), which already includes $(CFLAGS).
5087
5088 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5089 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5090 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5091 one-byte characters throughout. Note that case-insensitive matching applies
5092 only to characters whose values are less than 256. PCRE doesn't support the
5093 notion of cases for higher-valued characters.
5094
5095 (i) A character class whose characters are all within 0-255 is handled as
5096 a bit map, and the map is inverted for negative classes. Previously, a
5097 character > 255 always failed to match such a class; however it should
5098 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5099
5100 (ii) A negated character class with a single character < 255 is coded as
5101 "not this character" (OP_NOT). This wasn't working properly when the test
5102 character was multibyte, either singly or repeated.
5103
5104 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5105 mode, for example: \x{100}{2,3}.
5106
5107 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5108 singly or repeated) now correctly test multibyte characters. However,
5109 PCRE doesn't recognize any characters with values greater than 255 as
5110 digits, spaces, or word characters. Such characters always match \D, \S,
5111 and \W, and never match \d, \s, or \w.
5112
5113 (v) Classes may now contain characters and character ranges with values
5114 greater than 255. For example: [ab\x{100}-\x{400}].
5115
5116 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5117 PCRE in UTF-8 mode.
5118
5119 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5120 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5121 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5122 value.)
5123
5124 53. The single man page has become too large. I have therefore split it up into
5125 a number of separate man pages. These also give rise to individual HTML pages;
5126 these are now put in a separate directory, and there is an index.html page that
5127 lists them all. Some hyperlinking between the pages has been installed.
5128
5129 54. Added convenience functions for handling named capturing parentheses.
5130
5131 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5132 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5133 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5134 are faulted.
5135
5136 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5137 calling configure. These values are used when compiling the dftables.c program
5138 which is run to generate the source of the default character tables. They
5139 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5140 you will need to set these values.
5141
5142 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5143
5144
5145 Version 3.9 02-Jan-02
5146 ---------------------
5147
5148 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5149
5150 2. If --disable-static was given, the building process failed when trying to
5151 build pcretest and pcregrep. (For some reason it was using libtool to compile
5152 them, which is not right, as they aren't part of the library.)
5153
5154
5155 Version 3.8 18-Dec-01
5156 ---------------------
5157
5158 1. The experimental UTF-8 code was completely screwed up. It was packing the
5159 bytes in the wrong order. How dumb can you get?
5160
5161
5162 Version 3.7 29-Oct-01
5163 ---------------------
5164
5165 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5166 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5167 this didn't happen under Solaris 8, where I normally test things.
5168
5169 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5170 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5171 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5172 this fix an hour or so after the initial 3.7 release.)
5173
5174
5175 Version 3.6 23-Oct-01
5176 ---------------------
5177
5178 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5179 offsets passed as NULL with zero offset count.
5180
5181 2. The config.guess and config.sub files had not been updated when I moved to
5182 the latest autoconf.
5183
5184
5185 Version 3.5 15-Aug-01
5186 ---------------------
5187
5188 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5189 had been forgotten.
5190
5191 2. By using declared but undefined structures, we can avoid using "void"
5192 definitions in pcre.h while keeping the internal definitions of the structures
5193 private.
5194
5195 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5196 user point of view, this means that both static and shared libraries are built
5197 by default, but this can be individually controlled. More of the work of
5198 handling this static/shared cases is now inside libtool instead of PCRE's make
5199 file.
5200
5201 4. The pcretest utility is now installed along with pcregrep because it is
5202 useful for users (to test regexs) and by doing this, it automatically gets
5203 relinked by libtool. The documentation has been turned into a man page, so
5204 there are now .1, .txt, and .html versions in /doc.
5205
5206 5. Upgrades to pcregrep:
5207 (i) Added long-form option names like gnu grep.
5208 (ii) Added --help to list all options with an explanatory phrase.
5209 (iii) Added -r, --recursive to recurse into sub-directories.
5210 (iv) Added -f, --file to read patterns from a file.
5211
5212 6. pcre_exec() was referring to its "code" argument before testing that
5213 argument for NULL (and giving an error if it was NULL).
5214
5215 7. Upgraded Makefile.in to allow for compiling in a different directory from
5216 the source directory.
5217
5218 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5219 options bits, the pointer it was passed was to an int instead of to an unsigned
5220 long int. This mattered only on 64-bit systems.
5221
5222 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5223 generated) instead of pcre.in, which it its source. Also made the same change
5224 in several of the .c files.
5225
5226 10. A new release of gcc defines printf() as a macro, which broke pcretest
5227 because it had an ifdef in the middle of a string argument for printf(). Fixed
5228 by using separate calls to printf().
5229
5230 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5231 script, to force use of CR or LF instead of \n in the source. On non-Unix
5232 systems, the value can be set in config.h.
5233
5234 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5235 absolute limit. Changed the text of the error message to make this clear, and
5236 likewise updated the man page.
5237
5238 13. The limit of 99 on the number of capturing subpatterns has been removed.
5239 The new limit is 65535, which I hope will not be a "real" limit.
5240
5241
5242 Version 3.4 22-Aug-00
5243 ---------------------
5244
5245 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5246
5247 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5248
5249
5250 Version 3.3 01-Aug-00
5251 ---------------------
5252
5253 1. If an octal character was given, but the value was greater than \377, it
5254 was not getting masked to the least significant bits, as documented. This could
5255 lead to crashes in some systems.
5256
5257 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5258 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5259
5260 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5261 These just pass their arguments on to (pcre_free)(), but they are provided
5262 because some uses of PCRE bind it to non-C systems that can call its functions,
5263 but cannot call free() or pcre_free() directly.
5264
5265 4. Add "make test" as a synonym for "make check". Corrected some comments in
5266 the Makefile.
5267
5268 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5269 Makefile.
5270
5271 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5272 command called pgrep for grepping around the active processes.
5273
5274 7. Added the beginnings of support for UTF-8 character strings.
5275
5276 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5277 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5278 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5279 out for the ar command.)
5280
5281
5282 Version 3.2 12-May-00
5283 ---------------------
5284
5285 This is purely a bug fixing release.
5286
5287 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5288 of ZA. This was just one example of several cases that could provoke this bug,
5289 which was introduced by change 9 of version 2.00. The code for breaking
5290 infinite loops after an iteration that matches an empty string was't working
5291 correctly.
5292
5293 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5294 when matched against abbab (for example). After matching an empty string, it
5295 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5296 caused it to match further down the string than it should.
5297
5298 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5299 was there because it doesn't seem to be needed, and it causes trouble on some
5300 systems, as it is not a Standard C header. It has been removed.
5301
5302 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5303 were reported on the Macintosh. The changes were from
5304
5305 while ((c = *(++ptr)) != 0 && c != '\n');
5306 to
5307 while ((c = *(++ptr)) != 0 && c != '\n') ;
5308
5309 Totally extraordinary, but if that's what it takes...
5310
5311 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5312 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5313 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5314 assumes the way PCRE uses memmove() (always moving upwards).
5315
5316 6. PCRE is being used in one environment where strchr() is not available. There
5317 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5318 faster code anyway.
5319
5320
5321 Version 3.1 09-Feb-00
5322 ---------------------
5323
5324 The only change in this release is the fixing of some bugs in Makefile.in for
5325 the "install" target:
5326
5327 (1) It was failing to install pcreposix.h.
5328
5329 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5330
5331
5332 Version 3.0 01-Feb-00
5333 ---------------------
5334
5335 1. Add support for the /+ modifier to perltest (to output $` like it does in
5336 pcretest).
5337
5338 2. Add support for the /g modifier to perltest.
5339
5340 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5341 matches null strings.
5342
5343 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5344 pattern. Perl treats empty patterns specially - it reuses the most recent
5345 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5346 effect.
5347
5348 5. The POSIX interface was broken in that it was just handing over the POSIX
5349 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5350 required a bigger vector, with some working space on the end. This means that
5351 the POSIX wrapper now has to get and free some memory, and copy the results.
5352
5353 6. Added some simple autoconf support, placing the test data and the
5354 documentation in separate directories, re-organizing some of the
5355 information files, and making it build pcre-config (a GNU standard). Also added
5356 libtool support for building PCRE as a shared library, which is now the
5357 default.
5358
5359 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5360 09 are not valid octal constants. Single digits will be used for minor values
5361 less than 10.
5362
5363 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5364 existing programs that set these in the POSIX interface can use PCRE without
5365 modification.
5366
5367 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5368 return all that pcre_info() returns, plus additional data. The pcre_info()
5369 function is retained for compatibility, but is considered to be obsolete.
5370
5371 10. Added experimental recursion feature (?R) to handle one common case that
5372 Perl 5.6 will be able to do with (?p{...}).
5373
5374 11. Added support for POSIX character classes like [:alpha:], which Perl is
5375 adopting.
5376
5377
5378 Version 2.08 31-Aug-99
5379 ----------------------
5380
5381 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5382 trying to match at the startoffset position, but instead was moving forward to
5383 the next newline as if a previous match had failed.
5384
5385 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5386 and could get into a loop if a null string was matched other than at the start
5387 of the subject.
5388
5389 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5390 be distinguished at compile time, and for completeness also added PCRE_DATE.
5391
5392 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5393 in GnuWin32 environments.
5394
5395
5396 Version 2.07 29-Jul-99
5397 ----------------------
5398
5399 1. The documentation is now supplied in plain text form and HTML as well as in
5400 the form of man page sources.
5401
5402 2. C++ compilers don't like assigning (void *) values to other pointer types.
5403 In particular this affects malloc(). Although there is no problem in Standard
5404 C, I've put in casts to keep C++ compilers happy.
5405
5406 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5407 should be (const char *).
5408
5409 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5410 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5411 However, I haven't made this a standard facility. The documentation doesn't
5412 mention it, and the Makefile doesn't support it.
5413
5414 5. The Makefile now contains an "install" target, with editable destinations at
5415 the top of the file. The pcretest program is not installed.
5416
5417 6. pgrep -V now gives the PCRE version number and date.
5418
5419 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5420 causing the entire string to be ignored, instead of just the last character.
5421
5422 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5423 non-matching string, it can take a very, very long time, even for strings of
5424 quite modest length, because of the nested recursion. PCRE now does better in
5425 some of these cases. It does this by remembering the last required literal
5426 character in the pattern, and pre-searching the subject to ensure it is present
5427 before running the real match. In other words, it applies a heuristic to detect
5428 some types of certain failure quickly, and in the above example, if presented
5429 with a string that has no trailing " it gives "no match" very quickly.
5430
5431 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5432 other alternatives are tried instead.
5433
5434
5435 Version 2.06 09-Jun-99
5436 ----------------------
5437
5438 1. Change pcretest's output for amount of store used to show just the code
5439 space, because the remainder (the data block) varies in size between 32-bit and
5440 64-bit systems.
5441
5442 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5443 start matching at. This allows lookbehinds to work when searching for multiple
5444 occurrences in a string.
5445
5446 3. Added additional options to pcretest for testing multiple occurrences:
5447
5448 /+ outputs the rest of the string that follows a match
5449 /g loops for multiple occurrences, using the new startoffset argument
5450 /G loops for multiple occurrences by passing an incremented pointer
5451
5452 4. PCRE wasn't doing the "first character" optimization for patterns starting
5453 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5454 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5455 the letter 'x'. On long subject strings, this gives a significant speed-up.
5456
5457
5458 Version 2.05 21-Apr-99
5459 ----------------------
5460
5461 1. Changed the type of magic_number from int to long int so that it works
5462 properly on 16-bit systems.
5463
5464 2. Fixed a bug which caused patterns starting with .* not to work correctly
5465 when the subject string contained newline characters. PCRE was assuming
5466 anchoring for such patterns in all cases, which is not correct because .* will
5467 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5468 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5469 must be retried after every newline in the subject.
5470
5471
5472 Version 2.04 18-Feb-99
5473 ----------------------
5474
5475 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5476 computation of the store needed to hold the pattern was incorrect (too large).
5477 If such patterns were nested a few deep, this could multiply and become a real
5478 problem.
5479
5480 2. Added /M option to pcretest to show the memory requirement of a specific
5481 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5482
5483 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5484 compiled in such a way that the backtracking after subsequent failure was
5485 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5486 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5487
5488
5489 Version 2.03 02-Feb-99
5490 ----------------------
5491
5492 1. Fixed typo and small mistake in man page.
5493
5494 2. Added 4th condition (GPL supersedes if conflict) and created separate
5495 LICENCE file containing the conditions.
5496
5497 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5498 Perl, that is the internal \ allows the delimiter to be included in the
5499 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5500 the final delimiter, add \ to the end of the pattern (to test the error).
5501
5502 4. Added the convenience functions for extracting substrings after a successful
5503 match. Updated pcretest to make it able to test these functions.
5504
5505
5506 Version 2.02 14-Jan-99
5507 ----------------------
5508
5509 1. Initialized the working variables associated with each extraction so that
5510 their saving and restoring doesn't refer to uninitialized store.
5511
5512 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5513 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5514 fix the problem.
5515
5516 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5517 calls, and wasn't printing the correct value for compiling calls. Increased the
5518 default value of LOOPREPEAT, and the number of significant figures in the
5519 times.
5520
5521 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5522
5523 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5524 a building problem on Windows NT with a FAT file system.
5525
5526
5527 Version 2.01 21-Oct-98
5528 ----------------------
5529
5530 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5531 to character tables built by pcre_maketables() in the current locale. If NULL
5532 is passed, the default tables are used.
5533
5534
5535 Version 2.00 24-Sep-98
5536 ----------------------
5537
5538 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5539 it any more.
5540
5541 2. Allow quantification of (?>) groups, and make it work correctly.
5542
5543 3. The first character computation wasn't working for (?>) groups.
5544
5545 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5546 end of the subject) and add 5.005's \z, which really does match only at the
5547 very end of the subject.
5548
5549 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5550
5551 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5552 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5553 localized options. All options to pcre_study() were also removed.
5554
5555 7. Add other new features from 5.005:
5556
5557 $(?<= positive lookbehind
5558 $(?<! negative lookbehind
5559 (?imsx-imsx) added the unsetting capability
5560 such a setting is global if at outer level; local otherwise
5561 (?imsx-imsx:) non-capturing groups with option setting
5562 (?(cond)re|re) conditional pattern matching
5563
5564 A backreference to itself in a repeated group matches the previous
5565 captured string.
5566
5567 8. General tidying up of studying (both automatic and via "study")
5568 consequential on the addition of new assertions.
5569
5570 9. As in 5.005, unlimited repeated groups that could match an empty substring
5571 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5572 runtime if any iteration does actually match an empty substring.
5573
5574 10. Include the RunTest script in the distribution.
5575
5576 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5577 discrepancies, some of which were old and were also with respect to 5.004. They
5578 have now been fixed.
5579
5580
5581 Version 1.09 28-Apr-98
5582 ----------------------
5583
5584 1. A negated single character class followed by a quantifier with a minimum
5585 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5586 program crashes, or just wrong answers. This did not apply to negated classes
5587 containing more than one character, or to minima other than one.
5588
5589
5590 Version 1.08 27-Mar-98
5591 ----------------------
5592
5593 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5594
5595 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5596 latter must appear before anything that relies on it in the pattern.
5597
5598
5599 Version 1.07 16-Feb-98
5600 ----------------------
5601
5602 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5603 repeat of a potentially empty string).
5604
5605
5606 Version 1.06 23-Jan-98
5607 ----------------------
5608
5609 1. Added Markus Oberhumer's little patches for C++.
5610
5611 2. Literal strings longer than 255 characters were broken.
5612
5613
5614 Version 1.05 23-Dec-97
5615 ----------------------
5616
5617 1. Negated character classes containing more than one character were failing if
5618 PCRE_CASELESS was set at run time.
5619
5620
5621 Version 1.04 19-Dec-97
5622 ----------------------
5623
5624 1. Corrected the man page, where some "const" qualifiers had been omitted.
5625
5626 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5627 input syntax.
5628
5629 3. Fixed memory leak which occurred when a regex with back references was
5630 matched with an offsets vector that wasn't big enough. The temporary memory
5631 that is used in this case wasn't being freed if the match failed.
5632
5633 4. Tidied pcretest to ensure it frees memory that it gets.
5634
5635 5. Temporary memory was being obtained in the case where the passed offsets
5636 vector was exactly big enough.
5637
5638 6. Corrected definition of offsetof() from change 5 below.
5639
5640 7. I had screwed up change 6 below and broken the rules for the use of
5641 setjmp(). Now fixed.
5642
5643
5644 Version 1.03 18-Dec-97
5645 ----------------------
5646
5647 1. A erroneous regex with a missing opening parenthesis was correctly
5648 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5649 on some systems.
5650
5651 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5652 it was reported that one broken compiler failed on the former because "code" is
5653 also an independent variable.
5654
5655 3. The erroneous regex a[]b caused an array overrun reference.
5656
5657 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5658 fail on data ending with that character. (It was going on too far, and checking
5659 the next character, typically a binary zero.) This was specific to the
5660 optimized code for single-character negative classes.
5661
5662 5. Added a contributed patch from the TIN world which does the following:
5663
5664 + Add an undef for memmove, in case the the system defines a macro for it.
5665
5666 + Add a definition of offsetof(), in case there isn't one. (I don't know
5667 the reason behind this - offsetof() is part of the ANSI standard - but
5668 it does no harm).
5669
5670 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5671 most of the places where whitespace preceded '#'. I have given up and
5672 allowed the remaining 2 cases to be at the margin.
5673
5674 + Rename some variables in pcre to eliminate shadowing. This seems very
5675 pedantic, but does no harm, of course.
5676
5677 6. Moved the call to setjmp() into its own function, to get rid of warnings
5678 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5679
5680 7. Constructs such as \d{8,} were compiling into the equivalent of
5681 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5682 outcome, but in this particular case used more store than had been allocated,
5683 which caused the bug to be discovered because it threw up an internal error.
5684
5685 8. The debugging code in both pcre and pcretest for outputting the compiled
5686 form of a regex was going wrong in the case of back references followed by
5687 curly-bracketed repeats.
5688
5689
5690 Version 1.02 12-Dec-97
5691 ----------------------
5692
5693 1. Typos in pcre.3 and comments in the source fixed.
5694
5695 2. Applied a contributed patch to get rid of places where it used to remove
5696 'const' from variables, and fixed some signed/unsigned and uninitialized
5697 variable warnings.
5698
5699 3. Added the "runtest" target to Makefile.
5700
5701 4. Set default compiler flag to -O2 rather than just -O.
5702
5703
5704 Version 1.01 19-Nov-97
5705 ----------------------
5706
5707 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5708 like /([ab]*)*/, that is, for classes with more than one character in them.
5709
5710 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5711 as /((?>a*))*/ (a PCRE_EXTRA facility).
5712
5713
5714 Version 1.00 18-Nov-97
5715 ----------------------
5716
5717 1. Added compile-time macros to support systems such as SunOS4 which don't have
5718 memmove() or strerror() but have other things that can be used instead.
5719
5720 2. Arranged that "make clean" removes the executables.
5721
5722
5723 Version 0.99 27-Oct-97
5724 ----------------------
5725
5726 1. Fixed bug in code for optimizing classes with only one character. It was
5727 initializing a 32-byte map regardless, which could cause it to run off the end
5728 of the memory it had got.
5729
5730 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5731
5732
5733 Version 0.98 22-Oct-97
5734 ----------------------
5735
5736 1. Fixed bug in code for handling temporary memory usage when there are more
5737 back references than supplied space in the ovector. This could cause segfaults.
5738
5739
5740 Version 0.97 21-Oct-97
5741 ----------------------
5742
5743 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5744
5745 2. Optimized negated single characters not to use a bit map.
5746
5747 3. Brought error texts together as macro definitions; clarified some of them;
5748 fixed one that was wrong - it said "range out of order" when it meant "invalid
5749 escape sequence".
5750
5751 4. Changed some char * arguments to const char *.
5752
5753 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5754
5755 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5756 pcretest.
5757
5758
5759 Version 0.96 16-Oct-97
5760 ----------------------
5761
5762 1. Added a simple "pgrep" utility to the distribution.
5763
5764 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5765 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5766 where "ddd" means "one or more decimal digits".
5767
5768 3. Fixed serious bug. If a pattern had a back reference, but the call to
5769 pcre_exec() didn't supply a large enough ovector to record the related
5770 identifying subpattern, the match always failed. PCRE now remembers the number
5771 of the largest back reference, and gets some temporary memory in which to save
5772 the offsets during matching if necessary, in order to ensure that
5773 backreferences always work.
5774
5775 4. Increased the compatibility with Perl in a number of ways:
5776
5777 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5778 to request this handling. The option can be set at compile or exec time.
5779
5780 (b) $ matches before a terminating newline by default; an option
5781 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5782 mode). The option can be set at compile or exec time.
5783
5784 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5785 the same as Perl's. If the decimal number it represents is less than 10
5786 or there aren't that many previous left capturing parentheses, an octal
5787 escape is read. Inside a character class, it's always an octal escape,
5788 even if it is a single digit.
5789
5790 (d) An escaped but undefined alphabetic character is taken as a literal,
5791 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5792 escapes.
5793
5794 (e) {0} is now permitted. (The previous item is removed from the compiled
5795 pattern).
5796
5797 5. Changed all the names of code files so that the basic parts are no longer
5798 than 10 characters, and abolished the teeny "globals.c" file.
5799
5800 6. Changed the handling of character classes; they are now done with a 32-byte
5801 bit map always.
5802
5803 7. Added the -d and /D options to pcretest to make it possible to look at the
5804 internals of compilation without having to recompile pcre.
5805
5806
5807 Version 0.95 23-Sep-97
5808 ----------------------
5809
5810 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5811 \x20 at the start of a run of normal characters. These were being treated as
5812 real characters, instead of the source characters being re-checked.
5813
5814
5815 Version 0.94 18-Sep-97
5816 ----------------------
5817
5818 1. The functions are now thread-safe, with the caveat that the global variables
5819 containing pointers to malloc() and free() or alternative functions are the
5820 same for all threads.
5821
5822 2. Get pcre_study() to generate a bitmap of initial characters for non-
5823 anchored patterns when this is possible, and use it if passed to pcre_exec().
5824
5825
5826 Version 0.93 15-Sep-97
5827 ----------------------
5828
5829 1. /(b)|(:+)/ was computing an incorrect first character.
5830
5831 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5832 but not actually doing anything yet.
5833
5834 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5835 as Perl does (e.g. [-az] or [az-]).
5836
5837 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5838 all possible positions.
5839
5840 5. Split up into different modules to avoid including unneeded functions in a
5841 compiled binary. However, compile and exec are still in one module. The "study"
5842 function is split off.
5843
5844 6. The character tables are now in a separate module whose source is generated
5845 by an auxiliary program - but can then be edited by hand if required. There are
5846 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5847 toupper() in the code.
5848
5849 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5850 make them global. Abolish the function for setting them, as the caller can now
5851 set them directly.
5852
5853
5854 Version 0.92 11-Sep-97
5855 ----------------------
5856
5857 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5858 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5859
5860 2. Caseless matching was not working in character classes if the characters in
5861 the pattern were in upper case.
5862
5863 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5864
5865 4. Make PCRE_ANCHORED public and accept as a compile option.
5866
5867 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5868 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5869 pass them.
5870
5871 6. Give an error if bad option bits passed at compile or run time.
5872
5873 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5874 pcretest to cause it to pass that flag.
5875
5876 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5877 options, and the first character, if set.
5878
5879 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5880
5881
5882 Version 0.91 10-Sep-97
5883 ----------------------
5884
5885 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5886 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5887
5888 2. PCRE was looping on encountering an indefinitely repeated back reference to
5889 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5890 Perl does - treats the match as successful.
5891
5892 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5