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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1596 - (show annotations)
Wed Aug 26 13:44:24 2015 UTC (4 years, 2 months ago) by zherczeg
File size: 284025 byte(s)
Error occurred while calculating annotation data.
Fixed a corner case of range optimization in JIT.
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
149 Version 8.37 28-April-2015
150 --------------------------
151
152 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
153 for those parentheses to be closed with whatever has been captured so far.
154 However, it was failing to mark any other groups between the hightest
155 capture so far and the currrent group as "unset". Thus, the ovector for
156 those groups contained whatever was previously there. An example is the
157 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
158
159 2. If an assertion condition was quantified with a minimum of zero (an odd
160 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
161
162 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
163 unrecognized modifier, a crash could occur.
164
165 4. An attempt to do global matching in pcretest with a zero-length ovector
166 caused a crash.
167
168 5. Fixed a memory leak during matching that could occur for a subpattern
169 subroutine call (recursive or otherwise) if the number of captured groups
170 that had to be saved was greater than ten.
171
172 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
173 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
174 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
175
176 7. A UTF pattern containing a "not" match of a non-ASCII character and a
177 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
178
179 8. When a pattern is compiled, it remembers the highest back reference so that
180 when matching, if the ovector is too small, extra memory can be obtained to
181 use instead. A conditional subpattern whose condition is a check on a
182 capture having happened, such as, for example in the pattern
183 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
184 setting the highest backreference number. This mattered only if pcre_exec()
185 was called with an ovector that was too small to hold the capture, and there
186 was no other kind of back reference (a situation which is probably quite
187 rare). The effect of the bug was that the condition was always treated as
188 FALSE when the capture could not be consulted, leading to a incorrect
189 behaviour by pcre_exec(). This bug has been fixed.
190
191 9. A reference to a duplicated named group (either a back reference or a test
192 for being set in a conditional) that occurred in a part of the pattern where
193 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
194 to be incorrectly calculated, leading to overwriting.
195
196 10. A mutually recursive set of back references such as (\2)(\1) caused a
197 segfault at study time (while trying to find the minimum matching length).
198 The infinite loop is now broken (with the minimum length unset, that is,
199 zero).
200
201 11. If an assertion that was used as a condition was quantified with a minimum
202 of zero, matching went wrong. In particular, if the whole group had
203 unlimited repetition and could match an empty string, a segfault was
204 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
205 assertions to be quantified, but not if they are being used as conditions,
206 so the above pattern is faulted by Perl. PCRE has now been changed so that
207 it also rejects such patterns.
208
209 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
210 failed to allow the zero-repeat case if pcre2_exec() was called with an
211 ovector too small to capture the group.
212
213 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
214 Red Hat Product Security:
215
216 (a) A crash if /K and /F were both set with the option to save the compiled
217 pattern.
218
219 (b) Another crash if the option to print captured substrings in a callout
220 was combined with setting a null ovector, for example \O\C+ as a subject
221 string.
222
223 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
224 forward reference repeated a large (but limited) number of times within a
225 repeated outer group that has a zero minimum quantifier, caused incorrect
226 code to be compiled, leading to the error "internal error:
227 previously-checked referenced subpattern not found" when an incorrect
228 memory address was read. This bug was reported as "heap overflow",
229 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
230 CVE-2015-2325.
231
232 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
233 call within a group that also contained a recursive back reference caused
234 incorrect code to be compiled. This bug was reported as "heap overflow",
235 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
236 number CVE-2015-2326.
237
238 24. Computing the size of the JIT read-only data in advance has been a source
239 of various issues, and new ones are still appear unfortunately. To fix
240 existing and future issues, size computation is eliminated from the code,
241 and replaced by on-demand memory allocation.
242
243 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
244 adjacent to the end of the range, and the range contained characters with
245 more than one other case, caused incorrect behaviour when compiled in UTF
246 mode. In that example, the range a-j was left out of the class.
247
248 26. Fix JIT compilation of conditional blocks, which assertion
249 is converted to (*FAIL). E.g: /(?(?!))/.
250
251 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
252 discovered by the LLVM fuzzer.
253
254 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
255 when this assertion was used as a condition, for example (?(?!)a|b). In
256 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
257 error about an unsupported item.
258
259 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
260 possessification code could take exponential time to complete. A recursion
261 depth limit of 1000 has been imposed to limit the resources used by this
262 optimization.
263
264 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
265 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
266 because \S ensures they are all in the class. The code for doing this was
267 interacting badly with the code for computing the amount of space needed to
268 compile the pattern, leading to a buffer overflow. This bug was discovered
269 by the LLVM fuzzer.
270
271 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
272 other kinds of group caused stack overflow at compile time. This bug was
273 discovered by the LLVM fuzzer.
274
275 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
276 between a subroutine call and its quantifier was incorrectly compiled,
277 leading to buffer overflow or other errors. This bug was discovered by the
278 LLVM fuzzer.
279
280 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
281 assertion after (?(. The code was failing to check the character after
282 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
283 was discovered by the LLVM fuzzer.
284
285 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
286 a fixed maximum following a group that contains a subroutine reference was
287 incorrectly compiled and could trigger buffer overflow. This bug was
288 discovered by the LLVM fuzzer.
289
290 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
291 caused a stack overflow instead of the diagnosis of a non-fixed length
292 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
293
294 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
295 (e.g. /(?<=\Ka)/) could make pcregrep loop.
296
297 37. There was a similar problem to 36 in pcretest for global matches.
298
299 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
300 and a subsequent item in the pattern caused a non-match, backtracking over
301 the repeated \X did not stop, but carried on past the start of the subject,
302 causing reference to random memory and/or a segfault. There were also some
303 other cases where backtracking after \C could crash. This set of bugs was
304 discovered by the LLVM fuzzer.
305
306 39. The function for finding the minimum length of a matching string could take
307 a very long time if mutual recursion was present many times in a pattern,
308 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
309 method has been implemented. This infelicity was discovered by the LLVM
310 fuzzer.
311
312 40. Static linking against the PCRE library using the pkg-config module was
313 failing on missing pthread symbols.
314
315
316 Version 8.36 26-September-2014
317 ------------------------------
318
319 1. Got rid of some compiler warnings in the C++ modules that were shown up by
320 -Wmissing-field-initializers and -Wunused-parameter.
321
322 2. The tests for quantifiers being too big (greater than 65535) were being
323 applied after reading the number, and stupidly assuming that integer
324 overflow would give a negative number. The tests are now applied as the
325 numbers are read.
326
327 3. Tidy code in pcre_exec.c where two branches that used to be different are
328 now the same.
329
330 4. The JIT compiler did not generate match limit checks for certain
331 bracketed expressions with quantifiers. This may lead to exponential
332 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
333 issue should be resolved now.
334
335 5. Fixed an issue, which occures when nested alternatives are optimized
336 with table jumps.
337
338 6. Inserted two casts and changed some ints to size_t in the light of some
339 reported 64-bit compiler warnings (Bugzilla 1477).
340
341 7. Fixed a bug concerned with zero-minimum possessive groups that could match
342 an empty string, which sometimes were behaving incorrectly in the
343 interpreter (though correctly in the JIT matcher). This pcretest input is
344 an example:
345
346 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
347 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
348
349 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
350 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
351 for an empty string was breaking the inner loop and carrying on at a lower
352 level, when possessive repeated groups should always return to a higher
353 level as they have no backtrack points in them. The empty string test now
354 occurs at the outer level.
355
356 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
357 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
358
359 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
360 doesn't).
361
362 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
363 to hold a pattern to be incorrectly computed (too small) when there were
364 named back references to duplicated names. This could cause "internal
365 error: code overflow" or "double free or corruption" or other memory
366 handling errors.
367
368 11. When named subpatterns had the same prefixes, back references could be
369 confused. For example, in this pattern:
370
371 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
372
373 the reference to 'Name' was incorrectly treated as a reference to a
374 duplicate name.
375
376 12. A pattern such as /^s?c/mi8 where the optional character has more than
377 one "other case" was incorrectly compiled such that it would only try to
378 match starting at "c".
379
380 13. When a pattern starting with \s was studied, VT was not included in the
381 list of possible starting characters; this should have been part of the
382 8.34/18 patch.
383
384 14. If a character class started [\Qx]... where x is any character, the class
385 was incorrectly terminated at the ].
386
387 15. If a pattern that started with a caseless match for a character with more
388 than one "other case" was studied, PCRE did not set up the starting code
389 unit bit map for the list of possible characters. Now it does. This is an
390 optimization improvement, not a bug fix.
391
392 16. The Unicode data tables have been updated to Unicode 7.0.0.
393
394 17. Fixed a number of memory leaks in pcregrep.
395
396 18. Avoid a compiler warning (from some compilers) for a function call with
397 a cast that removes "const" from an lvalue by using an intermediate
398 variable (to which the compiler does not object).
399
400 19. Incorrect code was compiled if a group that contained an internal recursive
401 back reference was optional (had quantifier with a minimum of zero). This
402 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
403 caused segmentation faults because of stack overflows at compile time.
404
405 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
406 group that is quantified with an indefinite repeat, caused a compile-time
407 loop which used up all the system stack and provoked a segmentation fault.
408 This was not the same bug as 19 above.
409
410 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
411 Patch by Mike Frysinger.
412
413
414 Version 8.35 04-April-2014
415 --------------------------
416
417 1. A new flag is set, when property checks are present in an XCLASS.
418 When this flag is not set, PCRE can perform certain optimizations
419 such as studying these XCLASS-es.
420
421 2. The auto-possessification of character sets were improved: a normal
422 and an extended character set can be compared now. Furthermore
423 the JIT compiler optimizes more character set checks.
424
425 3. Got rid of some compiler warnings for potentially uninitialized variables
426 that show up only when compiled with -O2.
427
428 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
429 of a match later then the end of the match. The pcretest program was not
430 handling the case sensibly - it was outputting from the start to the next
431 binary zero. It now reports this situation in a message, and outputs the
432 text from the end to the start.
433
434 5. Fast forward search is improved in JIT. Instead of the first three
435 characters, any three characters with fixed position can be searched.
436 Search order: first, last, middle.
437
438 6. Improve character range checks in JIT. Characters are read by an inprecise
439 function now, which returns with an unknown value if the character code is
440 above a certain threshold (e.g: 256). The only limitation is that the value
441 must be bigger than the threshold as well. This function is useful when
442 the characters above the threshold are handled in the same way.
443
444 7. The macros whose names start with RAWUCHAR are placeholders for a future
445 mode in which only the bottom 21 bits of 32-bit data items are used. To
446 make this more memorable for those maintaining the code, the names have
447 been changed to start with UCHAR21, and an extensive comment has been added
448 to their definition.
449
450 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
451 to the export list in Makefile.am (they were accidentally omitted from the
452 8.34 tarball).
453
454 9. The informational output from pcretest used the phrase "starting byte set"
455 which is inappropriate for the 16-bit and 32-bit libraries. As the output
456 for "first char" and "need char" really means "non-UTF-char", I've changed
457 "byte" to "char", and slightly reworded the output. The documentation about
458 these values has also been (I hope) clarified.
459
460 10. Another JIT related optimization: use table jumps for selecting the correct
461 backtracking path, when more than four alternatives are present inside a
462 bracket.
463
464 11. Empty match is not possible, when the minimum length is greater than zero,
465 and there is no \K in the pattern. JIT should avoid empty match checks in
466 such cases.
467
468 12. In a caseless character class with UCP support, when a character with more
469 than one alternative case was not the first character of a range, not all
470 the alternative cases were added to the class. For example, s and \x{17f}
471 are both alternative cases for S: the class [RST] was handled correctly,
472 but [R-T] was not.
473
474 13. The configure.ac file always checked for pthread support when JIT was
475 enabled. This is not used in Windows, so I have put this test inside a
476 check for the presence of windows.h (which was already tested for).
477
478 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
479 The algorithm provides a way to skip certain starting offsets, and usually
480 faster than linear prefix searches.
481
482 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
483 as for 'fr_FR' and 'french'. For some reason, however, it then used the
484 Windows-specific input and output files, which have 'french' screwed in.
485 So this could never have worked. One of the problems with locales is that
486 they aren't always the same. I have now updated RunTest so that it checks
487 the output of the locale test (test 3) against three different output
488 files, and it allows the test to pass if any one of them matches. With luck
489 this should make the test pass on some versions of Solaris where it was
490 failing. Because of the uncertainty, the script did not used to stop if
491 test 3 failed; it now does. If further versions of a French locale ever
492 come to light, they can now easily be added.
493
494 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
495 there was a message during ./configure, but it did not stop. This now
496 provokes an error. The invalid example in README has been corrected.
497 If a value less than the minimum is given, the minimum value has always
498 been used, but now a warning is given.
499
500 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
501 was a bug in the test system, which is now fixed. Also, the list of various
502 configurations that are tested for each release did not have one with both
503 16/32 bits and --enable-bar-anycrlf. It now does.
504
505 18. pcretest was missing "-C bsr" for displaying the \R default setting.
506
507 19. Little endian PowerPC systems are supported now by the JIT compiler.
508
509 20. The fast forward newline mechanism could enter to an infinite loop on
510 certain invalid UTF-8 input. Although we don't support these cases
511 this issue can be fixed by a performance optimization.
512
513 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
514 not take account if existing stack usage. There is now a new global
515 variable called pcre_stack_guard that can be set to point to an external
516 function to check stack availability. It is called at the start of
517 processing every parenthesized group.
518
519 22. A typo in the code meant that in ungreedy mode the max/min qualifier
520 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
521 not match "ab".
522
523 23. When UTF was disabled, the JIT program reported some incorrect compile
524 errors. These messages are silenced now.
525
526 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
527 compiler.
528
529 25. Change all the temporary files used in RunGrepTest to be different to those
530 used by RunTest so that the tests can be run simultaneously, for example by
531 "make -j check".
532
533
534 Version 8.34 15-December-2013
535 -----------------------------
536
537 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
538 executable memory. Patch inspired by Carsten Klein.
539
540 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
541 this macro is never tested and has no effect, because the work to support
542 coverage involves only compiling and linking options and special targets in
543 the Makefile. The comment in config.h implied that defining the macro would
544 enable coverage support, which is totally false. There was also support for
545 setting this macro in the CMake files (my fault, I just copied it from
546 configure). SUPPORT_GCOV has now been removed.
547
548 3. Make a small performance improvement in strlen16() and strlen32() in
549 pcretest.
550
551 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
552 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
553 these cases). There was also one in pcretest.c.
554
555 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
556
557 6. In UTF mode, the code for checking whether a group could match an empty
558 string (which is used for indefinitely repeated groups to allow for
559 breaking an infinite loop) was broken when the group contained a repeated
560 negated single-character class with a character that occupied more than one
561 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
562 UTF-8 mode). The effect was undefined: the group might or might not be
563 deemed as matching an empty string, or the program might have crashed.
564
565 7. The code for checking whether a group could match an empty string was not
566 recognizing that \h, \H, \v, \V, and \R must match a character.
567
568 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
569 an empty string. If it can, pcretest shows this in its information output.
570
571 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
572 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
573 matched by pcre_exec() without using JIT:
574
575 (a) If the rest of the pattern did not match after a maximal run of
576 grapheme clusters, the code for backing up to try with fewer of them
577 did not always back up over a full grapheme when characters that do not
578 have the modifier quality were involved, e.g. Hangul syllables.
579
580 (b) If the match point in a subject started with modifier character, and
581 there was no match, the code could incorrectly back up beyond the match
582 point, and potentially beyond the first character in the subject,
583 leading to a segfault or an incorrect match result.
584
585 10. A conditional group with an assertion condition could lead to PCRE
586 recording an incorrect first data item for a match if no other first data
587 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
588 first data item, and therefore matched "ca" after "c" instead of at the
589 start.
590
591 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
592 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
593
594 12. The source of pcregrep now includes z/OS-specific code so that it can be
595 compiled for z/OS as part of the special z/OS distribution.
596
597 13. Added the -T and -TM options to pcretest.
598
599 14. The code in pcre_compile.c for creating the table of named capturing groups
600 has been refactored. Instead of creating the table dynamically during the
601 actual compiling pass, the information is remembered during the pre-compile
602 pass (on the stack unless there are more than 20 named groups, in which
603 case malloc() is used) and the whole table is created before the actual
604 compile happens. This has simplified the code (it is now nearly 150 lines
605 shorter) and prepared the way for better handling of references to groups
606 with duplicate names.
607
608 15. A back reference to a named subpattern when there is more than one of the
609 same name now checks them in the order in which they appear in the pattern.
610 The first one that is set is used for the reference. Previously only the
611 first one was inspected. This change makes PCRE more compatible with Perl.
612
613 16. Unicode character properties were updated from Unicode 6.3.0.
614
615 17. The compile-time code for auto-possessification has been refactored, based
616 on a patch by Zoltan Herczeg. It now happens after instead of during
617 compilation. The code is cleaner, and more cases are handled. The option
618 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
619 options in pcretest are provided to set it. It can also be set by
620 (*NO_AUTO_POSSESS) at the start of a pattern.
621
622 18. The character VT has been added to the default ("C" locale) set of
623 characters that match \s and are generally treated as white space,
624 following this same change in Perl 5.18. There is now no difference between
625 "Perl space" and "POSIX space". Whether VT is treated as white space in
626 other locales depends on the locale.
627
628 19. The code for checking named groups as conditions, either for being set or
629 for being recursed, has been refactored (this is related to 14 and 15
630 above). Processing unduplicated named groups should now be as fast at
631 numerical groups, and processing duplicated groups should be faster than
632 before.
633
634 20. Two patches to the CMake build system, by Alexander Barkov:
635
636 (1) Replace the "source" command by "." in CMakeLists.txt because
637 "source" is a bash-ism.
638
639 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
640 without these the CMake build does not work on Solaris.
641
642 21. Perl has changed its handling of \8 and \9. If there is no previously
643 encountered capturing group of those numbers, they are treated as the
644 literal characters 8 and 9 instead of a binary zero followed by the
645 literals. PCRE now does the same.
646
647 22. Following Perl, added \o{} to specify codepoints in octal, making it
648 possible to specify values greater than 0777 and also making them
649 unambiguous.
650
651 23. Perl now gives an error for missing closing braces after \x{... instead of
652 treating the string as literal. PCRE now does the same.
653
654 24. RunTest used to grumble if an inappropriate test was selected explicitly,
655 but just skip it when running all tests. This make it awkward to run ranges
656 of tests when one of them was inappropriate. Now it just skips any
657 inappropriate tests, as it always did when running all tests.
658
659 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
660 character types such as \d or \w, too many callouts were inserted, and the
661 data that they returned was rubbish.
662
663 26. In UCP mode, \s was not matching two of the characters that Perl matches,
664 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
665 were matched by \h. The code has now been refactored so that the lists of
666 the horizontal and vertical whitespace characters used for \h and \v (which
667 are defined only in one place) are now also used for \s.
668
669 27. Add JIT support for the 64 bit TileGX architecture.
670 Patch by Jiong Wang (Tilera Corporation).
671
672 28. Possessive quantifiers for classes (both explicit and automatically
673 generated) now use special opcodes instead of wrapping in ONCE brackets.
674
675 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
676 (because it's meaningless), this was not happening when PCRE_CASELESS was
677 set. Not wrong, but inefficient.
678
679 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
680 properties for \w, \d, etc) is present in a test regex. Otherwise if the
681 test contains no characters greater than 255, Perl doesn't realise it
682 should be using Unicode semantics.
683
684 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
685 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
686 does in Unicode mode.
687
688 32. Added the "forbid" facility to pcretest so that putting tests into the
689 wrong test files can sometimes be quickly detected.
690
691 33. There is now a limit (default 250) on the depth of nesting of parentheses.
692 This limit is imposed to control the amount of system stack used at compile
693 time. It can be changed at build time by --with-parens-nest-limit=xxx or
694 the equivalent in CMake.
695
696 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
697 errors. Perl warns for these when in warning mode, but PCRE has no facility
698 for giving warnings.
699
700 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
701 However, this was not working for (?!) because it is optimized to (*FAIL),
702 for which PCRE does not allow quantifiers. The optimization is now disabled
703 when a quantifier follows (?!). I can't see any use for this, but it makes
704 things uniform.
705
706 36. Perl no longer allows group names to start with digits, so I have made this
707 change also in PCRE. It simplifies the code a bit.
708
709 37. In extended mode, Perl ignores spaces before a + that indicates a
710 possessive quantifier. PCRE allowed a space before the quantifier, but not
711 before the possessive +. It now does.
712
713 38. The use of \K (reset reported match start) within a repeated possessive
714 group such as (a\Kb)*+ was not working.
715
716 40. Document that the same character tables must be used at compile time and
717 run time, and that the facility to pass tables to pcre_exec() and
718 pcre_dfa_exec() is for use only with saved/restored patterns.
719
720 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
721 features for Builds with MSVC:
722
723 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
724 builds). The .rc files can be used to set FileDescription and many other
725 attributes.
726
727 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
728 This allows higher-level build scripts which want .pdb files to avoid
729 hard-coding the exact files needed."
730
731 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
732 mean "start of word" and "end of word", respectively, as a transition aid.
733
734 43. A minimizing repeat of a class containing codepoints greater than 255 in
735 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
736 compiled to use the heap for recursion.
737
738 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
739 is configured.
740
741
742 Version 8.33 28-May-2013
743 ------------------------
744
745 1. Added 'U' to some constants that are compared to unsigned integers, to
746 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
747 variables that are added to signed variables, to ensure the result is
748 signed and can be negated.
749
750 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
751 CMake config files.
752
753 3. Revise the creation of config.h.generic so that all boolean macros are
754 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
755 overriding via -D on the command line possible.
756
757 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
758 to unsigned int is reported to make a quite noticeable speed difference in
759 a specific Windows environment. Testing on Linux did also appear to show
760 some benefit (and it is clearly not harmful). Also fixed the definition of
761 Xop which should be unsigned.
762
763 5. Related to (4), changing the definition of the intermediate variable cc
764 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
765 improvements.
766
767 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
768 unnecessary spaces.
769
770 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
771 and later.
772
773 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
774
775 9. Optimizing fast_forward_start_bits in JIT.
776
777 10. Adding support for callouts in JIT, and fixing some issues revealed
778 during this work. Namely:
779
780 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
781
782 (b) Minimum length was not checked before the matching is started.
783
784 11. The value of capture_last that is passed to callouts was incorrect in some
785 cases when there was a capture on one path that was subsequently abandoned
786 after a backtrack. Also, the capture_last value is now reset after a
787 recursion, since all captures are also reset in this case.
788
789 12. The interpreter no longer returns the "too many substrings" error in the
790 case when an overflowing capture is in a branch that is subsequently
791 abandoned after a backtrack.
792
793 13. In the pathological case when an offset vector of size 2 is used, pcretest
794 now prints out the matched string after a yield of 0 or 1.
795
796 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
797 Only supported by the JIT compiler at the moment.
798
799 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
800
801 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
802 offset of the starting point of the matching process, provided the offsets
803 vector is large enough.
804
805 17. The \A escape now records a lookbehind value of 1, though its execution
806 does not actually inspect the previous character. This is to ensure that,
807 in partial multi-segment matching, at least one character from the old
808 segment is retained when a new segment is processed. Otherwise, if there
809 are no lookbehinds in the pattern, \A might match incorrectly at the start
810 of a new segment.
811
812 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
813
814 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
815 gives some modest performance improvement in 8-bit mode.
816
817 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
818 be expressed in certain programming languages using Universal Character
819 Names.
820
821 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
822 which points out that "non characters" are not "characters that may not
823 appear in Unicode strings" but rather "characters that are reserved for
824 internal use and have only local meaning".
825
826 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
827 there was a conditional group that depended on an assertion, if the
828 assertion was false, the callout that immediately followed the alternation
829 in the condition was skipped when pcre_exec() was used for matching.
830
831 23. Allow an explicit callout to be inserted before an assertion that is the
832 condition for a conditional group, for compatibility with automatic
833 callouts, which always insert a callout at this point.
834
835 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
836 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
837
838 25. (*PRUNE) is now supported by the JIT compiler.
839
840 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
841
842 27. Fix the case where there are two or more SKIPs with arguments that may be
843 ignored.
844
845 28. (*SKIP) is now supported by the JIT compiler.
846
847 29. (*THEN) is now supported by the JIT compiler.
848
849 30. Update RunTest with additional test selector options.
850
851 31. The way PCRE handles backtracking verbs has been changed in two ways.
852
853 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
854 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
855 backtracking. In some cases this makes it more Perl-compatible, but Perl's
856 rather obscure rules do not always do the same thing.
857
858 (2) Previously, backtracking verbs were confined within assertions. This is
859 no longer the case for positive assertions, except for (*ACCEPT). Again,
860 this sometimes improves Perl compatibility, and sometimes does not.
861
862 32. A number of tests that were in test 2 because Perl did things differently
863 have been moved to test 1, because either Perl or PCRE has changed, and
864 these tests are now compatible.
865
866 32. Backtracking control verbs are now handled in the same way in JIT and
867 interpreter.
868
869 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
870 contained a forward subroutine reference caused a compile error.
871
872 34. Auto-detect and optimize limited repetitions in JIT.
873
874 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
875 blocking (*UTF) etc.
876
877 36. In the interpreter, maximizing pattern repetitions for characters and
878 character types now use tail recursion, which reduces stack usage.
879
880 37. The value of the max lookbehind was not correctly preserved if a compiled
881 and saved regex was reloaded on a host of different endianness.
882
883 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
884 of the compiled pattern block, expand the flags field from 16 to 32 bits
885 because it was almost full.
886
887 39. Try madvise first before posix_madvise.
888
889 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
890 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
891 original change locked out all matches of empty strings. This has been
892 changed so that one match of an empty string per line is recognized.
893 Subsequent searches on the same line (for colouring or for --only-matching,
894 for example) do not recognize empty strings.
895
896 41. Applied a user patch to fix a number of spelling mistakes in comments.
897
898 42. Data lines longer than 65536 caused pcretest to crash.
899
900 43. Clarified the data type for length and startoffset arguments for pcre_exec
901 and pcre_dfa_exec in the function-specific man pages, where they were
902 explicitly stated to be in bytes, never having been updated. I also added
903 some clarification to the pcreapi man page.
904
905 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
906 a segmentation fault.
907
908
909 Version 8.32 30-November-2012
910 -----------------------------
911
912 1. Improved JIT compiler optimizations for first character search and single
913 character iterators.
914
915 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
916 Patch by Daniel Richard G.
917
918 3. Single character iterator optimizations in the JIT compiler.
919
920 4. Improved JIT compiler optimizations for character ranges.
921
922 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
923 Reported by Giuseppe D'Angelo.
924
925 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
926 of a line, was being set incorrectly in cases where .* appeared inside
927 atomic brackets at the start of a pattern, or where there was a subsequent
928 *PRUNE or *SKIP.
929
930 7. Improved instruction cache flush for POWER/PowerPC.
931 Patch by Daniel Richard G.
932
933 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
934 grep:
935
936 (a) There is now no limit to the number of patterns to be matched.
937
938 (b) An error is given if a pattern is too long.
939
940 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
941 are now supported.
942
943 (d) --exclude-from and --include-from (multiple use) have been added.
944
945 (e) Exclusions and inclusions now apply to all files and directories, not
946 just to those obtained from scanning a directory recursively.
947
948 (f) Multiple uses of -f and --file-list are now supported.
949
950 (g) In a Windows environment, the default for -d has been changed from
951 "read" (the GNU grep default) to "skip", because otherwise the presence
952 of a directory in the file list provokes an error.
953
954 (h) The documentation has been revised and clarified in places.
955
956 9. Improve the matching speed of capturing brackets.
957
958 10. Changed the meaning of \X so that it now matches a Unicode extended
959 grapheme cluster.
960
961 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
962 out POSIX threads when JIT support is configured.
963
964 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
965
966 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
967 structure could go wrong in environments where size_t is not the same size
968 as int.
969
970 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
971 set.
972
973 15. The EBCDIC support had decayed; later updates to the code had included
974 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
975 general tidy up of EBCDIC-related issues, and the documentation was also
976 not quite right. There is now a test that can be run on ASCII systems to
977 check some of the EBCDIC-related things (but is it not a full test).
978
979 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
980 in a small tidy to the code.
981
982 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
983
984 18. If the --only-matching (-o) option in pcregrep is specified multiple
985 times, each one causes appropriate output. For example, -o1 -o2 outputs the
986 substrings matched by the 1st and 2nd capturing parentheses. A separating
987 string can be specified by --om-separator (default empty).
988
989 19. Improving the first n character searches.
990
991 20. Turn case lists for horizontal and vertical white space into macros so that
992 they are defined only once.
993
994 21. This set of changes together give more compatible Unicode case-folding
995 behaviour for characters that have more than one other case when UCP
996 support is available.
997
998 (a) The Unicode property table now has offsets into a new table of sets of
999 three or more characters that are case-equivalent. The MultiStage2.py
1000 script that generates these tables (the pcre_ucd.c file) now scans
1001 CaseFolding.txt instead of UnicodeData.txt for character case
1002 information.
1003
1004 (b) The code for adding characters or ranges of characters to a character
1005 class has been abstracted into a generalized function that also handles
1006 case-independence. In UTF-mode with UCP support, this uses the new data
1007 to handle characters with more than one other case.
1008
1009 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1010 whose other case is greater than 256 are now correctly matched
1011 caselessly. Previously, the high codepoint matched the low one, but not
1012 vice versa.
1013
1014 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1015 of the new class addition function, using character lists defined as
1016 macros alongside the case definitions of 20 above.
1017
1018 (e) Caseless back references now work with characters that have more than
1019 one other case.
1020
1021 (f) General caseless matching of characters with more than one other case
1022 is supported.
1023
1024 22. Unicode character properties were updated from Unicode 6.2.0
1025
1026 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1027
1028 24. Add support for 32-bit character strings, and UTF-32
1029
1030 25. Major JIT compiler update (code refactoring and bugfixing).
1031 Experimental Sparc 32 support is added.
1032
1033 26. Applied a modified version of Daniel Richard G's patch to create
1034 pcre.h.generic and config.h.generic by "make" instead of in the
1035 PrepareRelease script.
1036
1037 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1038 pcre_compile.c when checking for a zero character.
1039
1040 28. Introducing a native interface for JIT. Through this interface, the compiled
1041 machine code can be directly executed. The purpose of this interface is to
1042 provide fast pattern matching, so several sanity checks are not performed.
1043 However, feature tests are still performed. The new interface provides
1044 1.4x speedup compared to the old one.
1045
1046 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1047 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1048 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1049
1050 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1051 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1052 the "old" RFC 2279). Instead, it ended up passing a negative length to
1053 pcre_exec().
1054
1055 31. Add support for GCC's visibility feature to hide internal functions.
1056
1057 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1058 "unknown -C option" after outputting 0 or 1.
1059
1060 33. There is now support for generating a code coverage report for the test
1061 suite in environments where gcc is the compiler and lcov is installed. This
1062 is mainly for the benefit of the developers.
1063
1064 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1065 unaddressable using valgrind annotations, allowing valgrind to detect
1066 invalid memory accesses. This is mainly for the benefit of the developers.
1067
1068 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1069
1070 26. Give configure error if --enable-cpp but no C++ compiler found.
1071
1072
1073 Version 8.31 06-July-2012
1074 -------------------------
1075
1076 1. Fixing a wrong JIT test case and some compiler warnings.
1077
1078 2. Removed a bashism from the RunTest script.
1079
1080 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1081 to unsigned type, result still unsigned" that was given by an MS compiler
1082 on encountering the code "-sizeof(xxx)".
1083
1084 4. Partial matching support is added to the JIT compiler.
1085
1086 5. Fixed several bugs concerned with partial matching of items that consist
1087 of more than one character:
1088
1089 (a) /^(..)\1/ did not partially match "aba" because checking references was
1090 done on an "all or nothing" basis. This also applied to repeated
1091 references.
1092
1093 (b) \R did not give a hard partial match if \r was found at the end of the
1094 subject.
1095
1096 (c) \X did not give a hard partial match after matching one or more
1097 characters at the end of the subject.
1098
1099 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1100 a partial match for the string "\r".
1101
1102 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1103 a partial match for a CR character at the end of the subject string.
1104
1105 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1106 running pcretest, the text "(JIT)" added to the output whenever JIT is
1107 actually used to run the match.
1108
1109 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1110 or /S+[+] with a digit between 1 and 7.
1111
1112 8. OP_NOT now supports any UTF character not just single-byte ones.
1113
1114 9. (*MARK) control verb is now supported by the JIT compiler.
1115
1116 10. The command "./RunTest list" lists the available tests without actually
1117 running any of them. (Because I keep forgetting what they all are.)
1118
1119 11. Add PCRE_INFO_MAXLOOKBEHIND.
1120
1121 12. Applied a (slightly modified) user-supplied patch that improves performance
1122 when the heap is used for recursion (compiled with --disable-stack-for-
1123 recursion). Instead of malloc and free for each heap frame each time a
1124 logical recursion happens, frames are retained on a chain and re-used where
1125 possible. This sometimes gives as much as 30% improvement.
1126
1127 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1128 call.
1129
1130 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1131
1132 15. It is now possible to link pcretest with libedit as an alternative to
1133 libreadline.
1134
1135 16. (*COMMIT) control verb is now supported by the JIT compiler.
1136
1137 17. The Unicode data tables have been updated to Unicode 6.1.0.
1138
1139 18. Added --file-list option to pcregrep.
1140
1141 19. Added binary file support to pcregrep, including the -a, --binary-files,
1142 -I, and --text options.
1143
1144 20. The madvise function is renamed for posix_madvise for QNX compatibility
1145 reasons. Fixed by Giuseppe D'Angelo.
1146
1147 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1148
1149 22. Changed the option for creating symbolic links for 16-bit man pages from
1150 -s to -sf so that re-installing does not cause issues.
1151
1152 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1153
1154 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1155 matches in certain environments (the workspace was not being correctly
1156 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1157 some of the workspace data at the beginning of a restart.
1158
1159 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1160 was not doing so when it should - probably a typo introduced by SVN 528
1161 (change 8.10/14).
1162
1163 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1164 \w+ when the character tables indicated that \x{c4} was a word character.
1165 There were several related cases, all because the tests for doing a table
1166 lookup were testing for characters less than 127 instead of 255.
1167
1168 27. If a pattern contains capturing parentheses that are not used in a match,
1169 their slots in the ovector are set to -1. For those that are higher than
1170 any matched groups, this happens at the end of processing. In the case when
1171 there were back references that the ovector was too small to contain
1172 (causing temporary malloc'd memory to be used during matching), and the
1173 highest capturing number was not used, memory off the end of the ovector
1174 was incorrectly being set to -1. (It was using the size of the temporary
1175 memory instead of the true size.)
1176
1177 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1178 ovector size, it uses memory at the end of the block that it has got.
1179
1180 29. Check for an overlong MARK name and give an error at compile time. The
1181 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1182
1183 30. JIT compiler update.
1184
1185 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1186 Rill for the patch.
1187
1188 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1189
1190 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1191
1192 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1193 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1194 was enabled.
1195
1196 35. Improve JIT code generation for greedy plus quantifier.
1197
1198 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1199 1 to "aa" instead of to an empty string. The bug affected repeated groups
1200 that could potentially match an empty string.
1201
1202 37. Optimizing single character iterators in JIT.
1203
1204 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1205 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1206 codepoints that are too big for the mode are faulted, and in a UTF mode,
1207 disallowed codepoints are also faulted.
1208
1209 39. If PCRE was compiled with UTF support, in three places in the DFA
1210 matcher there was code that should only have been obeyed in UTF mode, but
1211 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1212 processing when bytes with values greater than 127 were present. In 16-bit
1213 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1214 both cases the values are those that cannot be the first data item in a UTF
1215 character. The three items that might have provoked this were recursions,
1216 possessively repeated groups, and atomic groups.
1217
1218 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1219 and pcregrep, because some OS require shared objects to be explicitly
1220 passed to ld, causing the link step to fail if they are not.
1221
1222 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1223 mode, patterns that started with \h* or \R* might be incorrectly matched.
1224
1225
1226 Version 8.30 04-February-2012
1227 -----------------------------
1228
1229 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1230 name is defined in ctype.h.
1231
1232 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1233 only in quite long subpatterns.
1234
1235 3. Removed the function pcre_info(), which has been obsolete and deprecated
1236 since it was replaced by pcre_fullinfo() in February 2000.
1237
1238 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1239 match a (*MARK), and the match failed at the start of the subject, a
1240 reference to memory before the start of the subject could occur. This bug
1241 was introduced by fix 17 of release 8.21.
1242
1243 5. A reference to an unset group with zero minimum repetition was giving
1244 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1245 /(another)?(\1?)test/ matched against "hello world test". This bug was
1246 introduced in release 8.13.
1247
1248 6. Add support for 16-bit character strings (a large amount of work involving
1249 many changes and refactorings).
1250
1251 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1252 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1253 from a file.
1254
1255 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1256 rounding is not applied in this particular case).
1257
1258 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1259 if they appear, or are escaped, in patterns.
1260
1261 10. Get rid of a number of -Wunused-but-set-variable warnings.
1262
1263 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1264 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1265 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1266 also returns the mark "x". This bug applied to capturing parentheses,
1267 non-capturing parentheses, and atomic parentheses. It also applied to some
1268 assertions.
1269
1270 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1271 information out of configure.ac instead of relying on pcre.h.generic, which
1272 is not stored in the repository.
1273
1274 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1275 -lreadline.
1276
1277 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1278
1279 15. Applied Graycode's patch to put the top-level frame on the stack rather
1280 than the heap when not using the stack for recursion. This gives a
1281 performance improvement in many cases when recursion is not deep.
1282
1283 16. Experimental code added to "pcretest -C" to output the stack frame size.
1284
1285
1286 Version 8.21 12-Dec-2011
1287 ------------------------
1288
1289 1. Updating the JIT compiler.
1290
1291 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1292 are added as well.
1293
1294 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1295 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1296 calling _pcre_jit_exec. Some extra comments are added.
1297
1298 4. (*MARK) settings inside atomic groups that do not contain any capturing
1299 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1300 was introduced by change 18 for 8.20.
1301
1302 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1303 ECMA-262 standard.
1304
1305 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1306 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1307 This bug was probably introduced by change 9 of 8.13.
1308
1309 7. While fixing 6 above, I noticed that a number of other items were being
1310 incorrectly rejected as "not fixed length". This arose partly because newer
1311 opcodes had not been added to the fixed-length checking code. I have (a)
1312 corrected the bug and added tests for these items, and (b) arranged for an
1313 error to occur if an unknown opcode is encountered while checking for fixed
1314 length instead of just assuming "not fixed length". The items that were
1315 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1316 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1317 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1318
1319 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1320 being incorrectly compiled and would have given unpredicatble results.
1321
1322 9. A possessively repeated subpattern with minimum repeat count greater than
1323 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1324 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1325 the first (A) could occur when it should not.
1326
1327 10. Add a cast and remove a redundant test from the code.
1328
1329 11. JIT should use pcre_malloc/pcre_free for allocation.
1330
1331 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1332 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1333 is anything other than /usr, -L is still shown).
1334
1335 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1336
1337 14. Perl does not support \N without a following name in a [] class; PCRE now
1338 also gives an error.
1339
1340 15. If a forward reference was repeated with an upper limit of around 2000,
1341 it caused the error "internal error: overran compiling workspace". The
1342 maximum number of forward references (including repeats) was limited by the
1343 internal workspace, and dependent on the LINK_SIZE. The code has been
1344 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1345 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1346 of around 200,000 forward references. While doing this, I also speeded up
1347 the filling in of repeated forward references.
1348
1349 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1350 incorrectly expecting the subject to contain another "a" after the start.
1351
1352 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1353 in the match, the SKIP should be ignored. This was not happening; instead
1354 the SKIP was being treated as NOMATCH. For patterns such as
1355 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1356 tested.
1357
1358 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1359 now much more compatible with Perl, in particular in cases where the result
1360 is a non-match for a non-anchored pattern. For example, if
1361 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1362 "m", where previously it did not return a name. A side effect of this
1363 change is that for partial matches, the last encountered mark name is
1364 returned, as for non matches. A number of tests that were previously not
1365 Perl-compatible have been moved into the Perl-compatible test files. The
1366 refactoring has had the pleasing side effect of removing one argument from
1367 the match() function, thus reducing its stack requirements.
1368
1369 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1370 subsequent uses of /S (without +) incorrectly behaved like /S+.
1371
1372 21. Retrieve executable code size support for the JIT compiler and fixing
1373 some warnings.
1374
1375 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1376 not work when the shorter character appeared right at the end of the
1377 subject string.
1378
1379 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1380 systems.
1381
1382 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1383 output it when the /M option is used in pcretest.
1384
1385 25. The CheckMan script was not being included in the distribution. Also, added
1386 an explicit "perl" to run Perl scripts from the PrepareRelease script
1387 because this is reportedly needed in Windows.
1388
1389 26. If study data was being save in a file and studying had not found a set of
1390 "starts with" bytes for the pattern, the data written to the file (though
1391 never used) was taken from uninitialized memory and so caused valgrind to
1392 complain.
1393
1394 27. Updated RunTest.bat as provided by Sheri Pierce.
1395
1396 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1397
1398 29. Computation of memory usage for the table of capturing group names was
1399 giving an unnecessarily large value.
1400
1401
1402 Version 8.20 21-Oct-2011
1403 ------------------------
1404
1405 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1406 a POSIX class. After further experiments with Perl, which convinced me that
1407 Perl has bugs and confusions, a closing square bracket is no longer allowed
1408 in a POSIX name. This bug also affected patterns with classes that started
1409 with full stops.
1410
1411 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1412 captured substring, but while checking the failing first alternative,
1413 substring 1 is temporarily captured. If the output vector supplied to
1414 pcre_exec() was not big enough for this capture, the yield of the function
1415 was still zero ("insufficient space for captured substrings"). This cannot
1416 be totally fixed without adding another stack variable, which seems a lot
1417 of expense for a edge case. However, I have improved the situation in cases
1418 such as /(a)(b)x|abc/ matched against "abc", where the return code
1419 indicates that fewer than the maximum number of slots in the ovector have
1420 been set.
1421
1422 3. Related to (2) above: when there are more back references in a pattern than
1423 slots in the output vector, pcre_exec() uses temporary memory during
1424 matching, and copies in the captures as far as possible afterwards. It was
1425 using the entire output vector, but this conflicts with the specification
1426 that only 2/3 is used for passing back captured substrings. Now it uses
1427 only the first 2/3, for compatibility. This is, of course, another edge
1428 case.
1429
1430 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1431 main code base, and can be used by building with --enable-jit. When this is
1432 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1433 runtime --no-jit option is given.
1434
1435 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1436 ovector, the return from the function was zero, implying that there were
1437 other matches that did not fit. The correct "exactly full" value is now
1438 returned.
1439
1440 6. If a subpattern that was called recursively or as a subroutine contained
1441 (*PRUNE) or any other control that caused it to give a non-standard return,
1442 invalid errors such as "Error -26 (nested recursion at the same subject
1443 position)" or even infinite loops could occur.
1444
1445 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1446 computing the minimum length on reaching *ACCEPT, and so ended up with the
1447 wrong value of 1 rather than 0. Further investigation indicates that
1448 computing a minimum subject length in the presence of *ACCEPT is difficult
1449 (think back references, subroutine calls), and so I have changed the code
1450 so that no minimum is registered for a pattern that contains *ACCEPT.
1451
1452 8. If (*THEN) was present in the first (true) branch of a conditional group,
1453 it was not handled as intended. [But see 16 below.]
1454
1455 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1456 Sheri Pierce.
1457
1458 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1459 the first byte in a match must be "a".
1460
1461 11. Change 17 for 8.13 increased the recursion depth for patterns like
1462 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1463 pattern contains any instances of (*THEN). If it does not, the old
1464 optimizations are restored. It would be nice to do this on a per-group
1465 basis, but at the moment that is not feasible.
1466
1467 12. In some environments, the output of pcretest -C is CRLF terminated. This
1468 broke RunTest's code that checks for the link size. A single white space
1469 character after the value is now allowed for.
1470
1471 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1472 For "fr", it uses the Windows-specific input and output files.
1473
1474 14. If (*THEN) appeared in a group that was called recursively or as a
1475 subroutine, it did not work as intended. [But see next item.]
1476
1477 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1478 pattern fragments (but not containing any | characters). If A and B are
1479 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1480 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1481 D. In other words, Perl considers parentheses that do not contain any |
1482 characters to be part of a surrounding alternative, whereas PCRE was
1483 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1484 differently. PCRE now behaves in the same way as Perl, except in the case
1485 of subroutine/recursion calls such as (?1) which have in any case always
1486 been different (but PCRE had them first :-).
1487
1488 16. Related to 15 above: Perl does not treat the | in a conditional group as
1489 creating alternatives. Such a group is treated in the same way as an
1490 ordinary group without any | characters when processing (*THEN). PCRE has
1491 been changed to match Perl's behaviour.
1492
1493 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1494 RunGrepTest script failed.
1495
1496 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1497 inevitable for groups that contain captures, but it can lead to a lot of
1498 stack use in large patterns. The old behaviour has been restored for atomic
1499 groups that do not contain any capturing parentheses.
1500
1501 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1502 suppress the check for a minimum subject length at run time. (If it was
1503 given to pcre_exec() or pcre_dfa_exec() it did work.)
1504
1505 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1506 fail to work when decoding hex characters in data strings in EBCDIC
1507 environments.
1508
1509 21. It appears that in at least one Mac OS environment, the isxdigit() function
1510 is implemented as a macro that evaluates to its argument more than once,
1511 contravening the C 90 Standard (I haven't checked a later standard). There
1512 was an instance in pcretest which caused it to go wrong when processing
1513 \x{...} escapes in subject strings. The has been rewritten to avoid using
1514 things like p++ in the argument of isxdigit().
1515
1516
1517 Version 8.13 16-Aug-2011
1518 ------------------------
1519
1520 1. The Unicode data tables have been updated to Unicode 6.0.0.
1521
1522 2. Two minor typos in pcre_internal.h have been fixed.
1523
1524 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1525 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1526 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1527
1528 4. There were a number of related bugs in the code for matching backrefences
1529 caselessly in UTF-8 mode when codes for the characters concerned were
1530 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1531 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1532 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1533 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1534 2-byte code at the end of the subject (it thought there wasn't enough data
1535 left).
1536
1537 5. Comprehensive information about what went wrong is now returned by
1538 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1539 as the output vector has at least 2 elements. The offset of the start of
1540 the failing character and a reason code are placed in the vector.
1541
1542 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1543 now returned is for the first byte of the failing character, instead of the
1544 last byte inspected. This is an incompatible change, but I hope it is small
1545 enough not to be a problem. It makes the returned offset consistent with
1546 pcre_exec() and pcre_dfa_exec().
1547
1548 7. pcretest now gives a text phrase as well as the error number when
1549 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1550 failure, the offset and reason code are output.
1551
1552 8. When \R was used with a maximizing quantifier it failed to skip backwards
1553 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1554 back over a single character (\n). This seems wrong (because it treated the
1555 two characters as a single entity when going forwards), conflicts with the
1556 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1557 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1558 has been changed.
1559
1560 9. Some internal refactoring has changed the processing so that the handling
1561 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1562 time (the PCRE_DOTALL option was changed this way some time ago: version
1563 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1564 which was always a bit of a fudge. It also means that there is one less
1565 argument for the match() function, which reduces its stack requirements
1566 slightly. This change also fixes an incompatibility with Perl: the pattern
1567 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1568
1569 10. More internal refactoring has drastically reduced the number of recursive
1570 calls to match() for possessively repeated groups such as (abc)++ when
1571 using pcre_exec().
1572
1573 11. While implementing 10, a number of bugs in the handling of groups were
1574 discovered and fixed:
1575
1576 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1577 (a|)*(?1) gave a compile-time internal error.
1578 ((a|)+)+ did not notice that the outer group could match an empty string.
1579 (^a|^)+ was not marked as anchored.
1580 (.*a|.*)+ was not marked as matching at start or after a newline.
1581
1582 12. Yet more internal refactoring has removed another argument from the match()
1583 function. Special calls to this function are now indicated by setting a
1584 value in a variable in the "match data" data block.
1585
1586 13. Be more explicit in pcre_study() instead of relying on "default" for
1587 opcodes that mean there is no starting character; this means that when new
1588 ones are added and accidentally left out of pcre_study(), testing should
1589 pick them up.
1590
1591 14. The -s option of pcretest has been documented for ages as being an old
1592 synonym of -m (show memory usage). I have changed it to mean "force study
1593 for every regex", that is, assume /S for every regex. This is similar to -i
1594 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1595 using it. It makes it easier to run collections of tests with and without
1596 study enabled, and thereby test pcre_study() more easily. All the standard
1597 tests are now run with and without -s (but some patterns can be marked as
1598 "never study" - see 20 below).
1599
1600 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1601 restoration of the capturing data to the outer values was not happening
1602 correctly.
1603
1604 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1605 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1606 pattern had matched an empty string, and so incorrectly returned a no
1607 match.
1608
1609 17. There was optimizing code for the last branch of non-capturing parentheses,
1610 and also for the obeyed branch of a conditional subexpression, which used
1611 tail recursion to cut down on stack usage. Unfortunately, now that there is
1612 the possibility of (*THEN) occurring in these branches, tail recursion is
1613 no longer possible because the return has to be checked for (*THEN). These
1614 two optimizations have therefore been removed. [But see 8.20/11 above.]
1615
1616 18. If a pattern containing \R was studied, it was assumed that \R always
1617 matched two bytes, thus causing the minimum subject length to be
1618 incorrectly computed because \R can also match just one byte.
1619
1620 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1621 was incorrectly computed.
1622
1623 20. If /S is present twice on a test pattern in pcretest input, it now
1624 *disables* studying, thereby overriding the use of -s on the command line
1625 (see 14 above). This is necessary for one or two tests to keep the output
1626 identical in both cases.
1627
1628 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1629 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1630
1631 22. When an atomic group that contained a capturing parenthesis was
1632 successfully matched, but the branch in which it appeared failed, the
1633 capturing was not being forgotten if a higher numbered group was later
1634 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1635 group 1 to "a", when in fact it should be unset. This applied to multi-
1636 branched capturing and non-capturing groups, repeated or not, and also to
1637 positive assertions (capturing in negative assertions does not happen
1638 in PCRE) and also to nested atomic groups.
1639
1640 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1641 subject after a captured substring, to make it easier to tell which of a
1642 number of identical substrings has been captured.
1643
1644 24. The way atomic groups are processed by pcre_exec() has been changed so that
1645 if they are repeated, backtracking one repetition now resets captured
1646 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1647 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1648 "aaa". Previously, it would have been "a". As part of this code
1649 refactoring, the way recursive calls are handled has also been changed.
1650
1651 25. If an assertion condition captured any substrings, they were not passed
1652 back unless some other capturing happened later. For example, if
1653 (?(?=(a))a) was matched against "a", no capturing was returned.
1654
1655 26. When studying a pattern that contained subroutine calls or assertions,
1656 the code for finding the minimum length of a possible match was handling
1657 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1658 group 1 called group 2 while simultaneously a separate group 2 called group
1659 1). A stack overflow occurred in this case. I have fixed this by limiting
1660 the recursion depth to 10.
1661
1662 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1663 Fortmann. This supports explicit test numbers on the command line, and has
1664 argument validation and error reporting.
1665
1666 28. An instance of \X with an unlimited repeat could fail if at any point the
1667 first character it looked at was a mark character.
1668
1669 29. Some minor code refactoring concerning Unicode properties and scripts
1670 should reduce the stack requirement of match() slightly.
1671
1672 30. Added the '=' option to pcretest to check the setting of unused capturing
1673 slots at the end of the pattern, which are documented as being -1, but are
1674 not included in the return count.
1675
1676 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1677 compiled something random. Now it gives a compile-time error (as does
1678 Perl).
1679
1680 32. A *MARK encountered during the processing of a positive assertion is now
1681 recorded and passed back (compatible with Perl).
1682
1683 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1684 had alternative anchored branches, the search for a second match in a line
1685 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1686 matched the line "0102" twice. The same bug affected patterns that started
1687 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1688 twice.
1689
1690 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1691 does, and because of capturing effects, quantifying parenthesized
1692 assertions may at times be useful. Quantifiers are now allowed for
1693 parenthesized assertions.
1694
1695 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1696
1697 36. \g was being checked for fancy things in a character class, when it should
1698 just be a literal "g".
1699
1700 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1701 appearance of a nested POSIX class supersedes an apparent external class.
1702 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1703 unescaped square brackets may also appear as part of class names. For
1704 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1705 more like Perl. (But see 8.20/1 above.)
1706
1707 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1708 was because it thought it was \N{name}, which is not supported).
1709
1710 39. Add minix to OS list not supporting the -S option in pcretest.
1711
1712 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1713 cannot analyze patterns in sufficient detail to catch mutual recursions
1714 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1715 subgroup is called recursively as a subpattern for a second time at the
1716 same position in the subject string. In previous releases this might have
1717 been caught by the recursion limit, or it might have run out of stack.
1718
1719 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1720 happen only once. PCRE was, however incorrectly giving a compile time error
1721 "recursive call could loop indefinitely" because it cannot analyze the
1722 pattern in sufficient detail. The compile time test no longer happens when
1723 PCRE is compiling a conditional subpattern, but actual runaway loops are
1724 now caught at runtime (see 40 above).
1725
1726 42. It seems that Perl allows any characters other than a closing parenthesis
1727 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1728 has been changed to be the same.
1729
1730 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1731 as not to get warnings when autogen.sh is called. Also changed
1732 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1733
1734 44. To help people who use pcregrep to scan files containing exceedingly long
1735 lines, the following changes have been made:
1736
1737 (a) The default value of the buffer size parameter has been increased from
1738 8K to 20K. (The actual buffer used is three times this size.)
1739
1740 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1741 PCRE is built.
1742
1743 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1744 to be set at run time.
1745
1746 (d) Numerical values in pcregrep options can be followed by K or M, for
1747 example --buffer-size=50K.
1748
1749 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1750 given and the return code is set to 2.
1751
1752 45. Add a pointer to the latest mark to the callout data block.
1753
1754 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1755 partial match of an empty string instead of no match. This was specific to
1756 the use of ".".
1757
1758 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1759 complete match instead of a partial match. This bug was dependent on both
1760 the PCRE_UTF8 and PCRE_DOTALL options being set.
1761
1762 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1763 starting byte set, because \b was not being ignored.
1764
1765
1766 Version 8.12 15-Jan-2011
1767 ------------------------
1768
1769 1. Fixed some typos in the markup of the man pages, and wrote a script that
1770 checks for such things as part of the documentation building process.
1771
1772 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1773 --match-limit and --recursion-limit options (added for 8.11). In
1774 particular, this made one of the standard tests fail. (The integer value
1775 went into the wrong half of a long int.)
1776
1777 3. If the --colour option was given to pcregrep with -v (invert match), it
1778 did strange things, either producing crazy output, or crashing. It should,
1779 of course, ignore a request for colour when reporting lines that do not
1780 match.
1781
1782 4. Another pcregrep bug caused similar problems if --colour was specified with
1783 -M (multiline) and the pattern match finished with a line ending.
1784
1785 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1786 matched in multiline mode, the following line was shown as part of the
1787 match. This seems wrong, so I have changed it.
1788
1789 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1790 the check for further matches in the same line (so they could be coloured)
1791 to overrun the end of the current line. If another match was found, it was
1792 incorrectly shown (and then shown again when found in the next line).
1793
1794 7. If pcregrep was compiled under Windows, there was a reference to the
1795 function pcregrep_exit() before it was defined. I am assuming this was
1796 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1797 reported by a user. I've moved the definition above the reference.
1798
1799
1800 Version 8.11 10-Dec-2010
1801 ------------------------
1802
1803 1. (*THEN) was not working properly if there were untried alternatives prior
1804 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1805 backtracked to try for "b" instead of moving to the next alternative branch
1806 at the same level (in this case, to look for "c"). The Perl documentation
1807 is clear that when (*THEN) is backtracked onto, it goes to the "next
1808 alternative in the innermost enclosing group".
1809
1810 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1811 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1812 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1813 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1814 (*THEN).
1815
1816 3. If \s appeared in a character class, it removed the VT character from
1817 the class, even if it had been included by some previous item, for example
1818 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1819 of \s, but is part of the POSIX "space" class.)
1820
1821 4. A partial match never returns an empty string (because you can always
1822 match an empty string at the end of the subject); however the checking for
1823 an empty string was starting at the "start of match" point. This has been
1824 changed to the "earliest inspected character" point, because the returned
1825 data for a partial match starts at this character. This means that, for
1826 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1827 (previously it gave "no match").
1828
1829 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1830 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1831 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1832 has an implication that the given string is incomplete (because a partial
1833 match is preferred over a full match). For this reason, these items now
1834 give a partial match in this situation. [Aside: previously, the one case
1835 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1836 match rather than a full match, which was wrong by the old rules, but is
1837 now correct.]
1838
1839 6. There was a bug in the handling of #-introduced comments, recognized when
1840 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1841 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1842 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1843 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1844 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1845 places in pcre_compile().
1846
1847 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1848 comments when looking ahead for named forward references to subpatterns,
1849 the only newline sequence it recognized was NL. It now handles newlines
1850 according to the set newline convention.
1851
1852 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1853 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1854 cater for a lack of strerror(). These oversights have been fixed.
1855
1856 9. Added --match-limit and --recursion-limit to pcregrep.
1857
1858 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1859
1860 11. When the -o option was used, pcregrep was setting a return code of 1, even
1861 when matches were found, and --line-buffered was not being honoured.
1862
1863 12. Added an optional parentheses number to the -o and --only-matching options
1864 of pcregrep.
1865
1866 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1867 can match an empty string. The code to do it in pcretest and pcredemo
1868 needed fixing:
1869
1870 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1871 only one byte after an empty string match just before CRLF (this case
1872 just got forgotten; "any" and "anycrlf" were OK).
1873
1874 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1875 mode when an empty string match preceded an ASCII character followed by
1876 a non-ASCII character. (The code for advancing by one character rather
1877 than one byte was nonsense.)
1878
1879 (c) The pcredemo.c sample program did not have any code at all to handle
1880 the cases when CRLF is a valid newline sequence.
1881
1882 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1883 as a starting offset was within the subject string. There is now a new
1884 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1885 negative or greater than the length of the string. In order to test this,
1886 pcretest is extended to allow the setting of negative starting offsets.
1887
1888 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1889 starting offset points to the beginning of a UTF-8 character was
1890 unnecessarily clumsy. I tidied it up.
1891
1892 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1893 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1894
1895 17. Nobody had reported that the --include_dir option, which was added in
1896 release 7.7 should have been called --include-dir (hyphen, not underscore)
1897 for compatibility with GNU grep. I have changed it to --include-dir, but
1898 left --include_dir as an undocumented synonym, and the same for
1899 --exclude-dir, though that is not available in GNU grep, at least as of
1900 release 2.5.4.
1901
1902 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1903 characters from a string of bytes) have been redefined so as not to use
1904 loops, in order to improve performance in some environments. At the same
1905 time, I abstracted some of the common code into auxiliary macros to save
1906 repetition (this should not affect the compiled code).
1907
1908 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1909 compile-time error is now given if \c is not followed by an ASCII
1910 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1911 different, and any byte value is allowed.)
1912
1913 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1914 START_OPTIMIZE option, which is now allowed at compile time - but just
1915 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1916 to pcregrep and other applications that have no direct access to PCRE
1917 options. The new /Y option in pcretest sets this option when calling
1918 pcre_compile().
1919
1920 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1921 back references. Groups containing recursive back references were forced to
1922 be atomic by that change, but in the case of named groups, the amount of
1923 memory required was incorrectly computed, leading to "Failed: internal
1924 error: code overflow". This has been fixed.
1925
1926 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1927 pcretest.c, to avoid build problems in some Borland environments.
1928
1929
1930 Version 8.10 25-Jun-2010
1931 ------------------------
1932
1933 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1934 THEN.
1935
1936 2. (*ACCEPT) was not working when inside an atomic group.
1937
1938 3. Inside a character class, \B is treated as a literal by default, but
1939 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1940 causes the error). The code is unchanged, but I tidied the documentation.
1941
1942 4. Inside a character class, PCRE always treated \R and \X as literals,
1943 whereas Perl faults them if its -w option is set. I have changed PCRE so
1944 that it faults them when PCRE_EXTRA is set.
1945
1946 5. Added support for \N, which always matches any character other than
1947 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1948
1949 6. When compiling pcregrep with newer versions of gcc which may have
1950 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1951 declared with attribute warn_unused_result" were given. Just casting the
1952 result to (void) does not stop the warnings; a more elaborate fudge is
1953 needed. I've used a macro to implement this.
1954
1955 7. Minor change to pcretest.c to avoid a compiler warning.
1956
1957 8. Added four artifical Unicode properties to help with an option to make
1958 \s etc use properties (see next item). The new properties are: Xan
1959 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1960
1961 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1962 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1963 this option. Modified pcretest to add /W to test this facility. Added
1964 REG_UCP to make it available via the POSIX interface.
1965
1966 10. Added --line-buffered to pcregrep.
1967
1968 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1969 studied, and the match started with a letter with a code point greater than
1970 127 whose first byte was different to the first byte of the other case of
1971 the letter, the other case of this starting letter was not recognized
1972 (#976).
1973
1974 12. If a pattern that was studied started with a repeated Unicode property
1975 test, for example, \p{Nd}+, there was the theoretical possibility of
1976 setting up an incorrect bitmap of starting bytes, but fortunately it could
1977 not have actually happened in practice until change 8 above was made (it
1978 added property types that matched character-matching opcodes).
1979
1980 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1981 possible starting bytes for non-anchored patterns.
1982
1983 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1984 \R, and also a number of cases that involve Unicode properties, both
1985 explicit and implicit when PCRE_UCP is set.
1986
1987 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1988 input, it could crash or give wrong results if characters with values
1989 greater than 0xc0 were present in the subject string. (Detail: it assumed
1990 UTF-8 input when processing these items.)
1991
1992 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1993 size_t is 64-bit (#991).
1994
1995 17. Added a check for running out of memory when PCRE is compiled with
1996 --disable-stack-for-recursion (#990).
1997
1998 18. If the last data line in a file for pcretest does not have a newline on
1999 the end, a newline was missing in the output.
2000
2001 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2002 less than 128) in its various bitmaps. However, there is a facility for
2003 generating tables according to the current locale when PCRE is compiled. It
2004 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2005 characters, are recognized by isspace() and therefore were getting set in
2006 these tables, and indeed these tables seem to approximate to ISO 8859. This
2007 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2008 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2009 which of course cannot start UTF-8 characters. I have changed the code so
2010 that only real ASCII characters (less than 128) and the correct starting
2011 bytes for UTF-8 encodings are set for characters greater than 127 when in
2012 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2013 altogether.)
2014
2015 20. Added the /T option to pcretest so as to be able to run tests with non-
2016 standard character tables, thus making it possible to include the tests
2017 used for 19 above in the standard set of tests.
2018
2019 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2020 reference to a subpattern the other side of a comment that contains an
2021 opening parenthesis caused either an internal compiling error, or a
2022 reference to the wrong subpattern.
2023
2024
2025 Version 8.02 19-Mar-2010
2026 ------------------------
2027
2028 1. The Unicode data tables have been updated to Unicode 5.2.0.
2029
2030 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2031 configured.
2032
2033 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2034 original author of that file, following a query about its status.
2035
2036 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2037 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2038
2039 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2040 quantifier applied to a forward-referencing subroutine call, could compile
2041 incorrect code or give the error "internal error: previously-checked
2042 referenced subpattern not found".
2043
2044 6. Both MS Visual Studio and Symbian OS have problems with initializing
2045 variables to point to external functions. For these systems, therefore,
2046 pcre_malloc etc. are now initialized to local functions that call the
2047 relevant global functions.
2048
2049 7. There were two entries missing in the vectors called coptable and poptable
2050 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2051 I've fixed the data, and added a kludgy way of testing at compile time that
2052 the lengths are correct (equal to the number of opcodes).
2053
2054 8. Following on from 7, I added a similar kludge to check the length of the
2055 eint vector in pcreposix.c.
2056
2057 9. Error texts for pcre_compile() are held as one long string to avoid too
2058 much relocation at load time. To find a text, the string is searched,
2059 counting zeros. There was no check for running off the end of the string,
2060 which could happen if a new error number was added without updating the
2061 string.
2062
2063 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2064
2065 11. \K was not working if it appeared in an atomic group or in a group that
2066 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2067 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2068 the assertion is positive, but not if it is negative.
2069
2070 12. Change 11 fortuitously reduced the size of the stack frame used in the
2071 "match()" function of pcre_exec.c by one pointer. Forthcoming
2072 implementation of support for (*MARK) will need an extra pointer on the
2073 stack; I have reserved it now, so that the stack frame size does not
2074 decrease.
2075
2076 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2077 item in branch that calls a recursion is a subroutine call - as in the
2078 second branch in the above example - was incorrectly given the compile-
2079 time error "recursive call could loop indefinitely" because pcre_compile()
2080 was not correctly checking the subroutine for matching a non-empty string.
2081
2082 14. The checks for overrunning compiling workspace could trigger after an
2083 overrun had occurred. This is a "should never occur" error, but it can be
2084 triggered by pathological patterns such as hundreds of nested parentheses.
2085 The checks now trigger 100 bytes before the end of the workspace.
2086
2087 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2088
2089
2090 Version 8.01 19-Jan-2010
2091 ------------------------
2092
2093 1. If a pattern contained a conditional subpattern with only one branch (in
2094 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2095 computed the wrong minimum data length (which is of course zero for such
2096 subpatterns). This could cause incorrect "no match" results.
2097
2098 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2099 the pattern is reset in the first branch, pcre_compile() failed with
2100 "internal error: code overflow at offset...". This happened only when
2101 the reset was to the original external option setting. (An optimization
2102 abstracts leading options settings into an external setting, which was the
2103 cause of this.)
2104
2105 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2106 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2107 assertion pattern did not match (meaning that the assertion was true), it
2108 was incorrectly treated as false if the SKIP had been reached during the
2109 matching. This also applied to assertions used as conditions.
2110
2111 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2112 assertion subpattern, including such a pattern used as a condition,
2113 unpredictable results occurred, instead of the error return
2114 PCRE_ERROR_DFA_UITEM.
2115
2116 5. The C++ GlobalReplace function was not working like Perl for the special
2117 situation when an empty string is matched. It now does the fancy magic
2118 stuff that is necessary.
2119
2120 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2121 removed. (These were left over from very, very early versions of PCRE.)
2122
2123 7. Some cosmetic changes to the code to make life easier when compiling it
2124 as part of something else:
2125
2126 (a) Change DEBUG to PCRE_DEBUG.
2127
2128 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2129 called "current" as "current_branch", to prevent a collision with the
2130 Linux macro when compiled as a kernel module.
2131
2132 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2133 prevent a collision with the Linux macro when compiled as a kernel
2134 module.
2135
2136 8. In pcre_compile() there are some checks for integer overflows that used to
2137 cast potentially large values to (double). This has been changed to that
2138 when building, a check for int64_t is made, and if it is found, it is used
2139 instead, thus avoiding the use of floating point arithmetic. (There is no
2140 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2141 double.
2142
2143 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2144 2005 (difference between two addresses compared to an unsigned value).
2145
2146 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2147 custom one, because of the following reported problem in Windows:
2148
2149 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2150 under Win32.
2151 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2152 therefore missing the function definition.
2153 - The compiler thus generates a "C" signature for the test function.
2154 - The linker fails to find the "C" function.
2155 - PCRE fails to configure if asked to do so against libbz2.
2156
2157 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2158 messages were output:
2159
2160 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2161 rerunning libtoolize, to keep the correct libtool macros in-tree.
2162 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2163
2164 I have done both of these things.
2165
2166 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2167 most of the time, it *can* run out if it is given a pattern that contains a
2168 runaway infinite recursion. I updated the discussion in the pcrestack man
2169 page.
2170
2171 13. Now that we have gone to the x.xx style of version numbers, the minor
2172 version may start with zero. Using 08 or 09 is a bad idea because users
2173 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2174 interpreted as invalid octal numbers. I've updated the previous comment in
2175 configure.ac, and also added a check that gives an error if 08 or 09 are
2176 used.
2177
2178 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2179 causing partial matching to fail when the end of the subject matched \W
2180 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2181
2182 15. There were some discrepancies between the declarations in pcre_internal.h
2183 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2184 their definitions. The declarations used "const uschar *" and the
2185 definitions used USPTR. Even though USPTR is normally defined as "const
2186 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2187 reported that: "This difference in casting confuses some C++ compilers, for
2188 example, SunCC recognizes above declarations as different functions and
2189 generates broken code for hbpcre." I have changed the declarations to use
2190 USPTR.
2191
2192 16. GNU libtool is named differently on some systems. The autogen.sh script now
2193 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2194 (FreeBSD).
2195
2196 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2197 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2198 comment: "Figure out how to create a longlong from a string: strtoll and
2199 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2200 instance, but it only takes 2 args instead of 3!"
2201
2202 18. A subtle bug concerned with back references has been fixed by a change of
2203 specification, with a corresponding code fix. A pattern such as
2204 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2205 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2206 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2207 same bug. Such groups have to be quantified to be useful, or contained
2208 inside another quantified group. (If there's no repetition, the reference
2209 can never match.) The problem arises because, having left the group and
2210 moved on to the rest of the pattern, a later failure that backtracks into
2211 the group uses the captured value from the final iteration of the group
2212 rather than the correct earlier one. I have fixed this in PCRE by forcing
2213 any group that contains a reference to itself to be an atomic group; that
2214 is, there cannot be any backtracking into it once it has completed. This is
2215 similar to recursive and subroutine calls.
2216
2217
2218 Version 8.00 19-Oct-09
2219 ----------------------
2220
2221 1. The table for translating pcre_compile() error codes into POSIX error codes
2222 was out-of-date, and there was no check on the pcre_compile() error code
2223 being within the table. This could lead to an OK return being given in
2224 error.
2225
2226 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2227 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2228 in a Windows environment.
2229
2230 3. The pcregrep --count option prints the count for each file even when it is
2231 zero, as does GNU grep. However, pcregrep was also printing all files when
2232 --files-with-matches was added. Now, when both options are given, it prints
2233 counts only for those files that have at least one match. (GNU grep just
2234 prints the file name in this circumstance, but including the count seems
2235 more useful - otherwise, why use --count?) Also ensured that the
2236 combination -clh just lists non-zero counts, with no names.
2237
2238 4. The long form of the pcregrep -F option was incorrectly implemented as
2239 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2240 but it seems right to fix it, and I didn't think it was worth preserving
2241 the old behaviour.
2242
2243 5. The command line items --regex=pattern and --regexp=pattern were not
2244 recognized by pcregrep, which required --regex pattern or --regexp pattern
2245 (with a space rather than an '='). The man page documented the '=' forms,
2246 which are compatible with GNU grep; these now work.
2247
2248 6. No libpcreposix.pc file was created for pkg-config; there was just
2249 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2250
2251 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2252 when UCP support is not needed, by modifying the Python script that
2253 generates it from Unicode data files. This should not matter if the module
2254 is correctly used as a library, but I received one complaint about 50K of
2255 unwanted data. My guess is that the person linked everything into his
2256 program rather than using a library. Anyway, it does no harm.
2257
2258 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2259 was a minimum greater than 1 for a wide character in a possessive
2260 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2261 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2262 character. Chaos in the form of incorrect output or a compiling loop could
2263 result.
2264
2265 9. The restrictions on what a pattern can contain when partial matching is
2266 requested for pcre_exec() have been removed. All patterns can now be
2267 partially matched by this function. In addition, if there are at least two
2268 slots in the offset vector, the offset of the earliest inspected character
2269 for the match and the offset of the end of the subject are set in them when
2270 PCRE_ERROR_PARTIAL is returned.
2271
2272 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2273 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2274 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2275 and may be more useful for multi-segment matching.
2276
2277 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2278 used to be given if ever the end of the subject was reached; now it is
2279 given only if matching could not proceed because another character was
2280 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2281 string "Z", which now yields "no match" instead of "partial match". In the
2282 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2283 final character ended with (*FAIL).
2284
2285 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2286 if the pattern had a "must contain" character that was already found in the
2287 earlier partial match, unless partial matching was again requested. For
2288 example, with the pattern /dog.(body)?/, the "must contain" character is
2289 "g". If the first part-match was for the string "dog", restarting with
2290 "sbody" failed. This bug has been fixed.
2291
2292 13. The string returned by pcre_dfa_exec() after a partial match has been
2293 changed so that it starts at the first inspected character rather than the
2294 first character of the match. This makes a difference only if the pattern
2295 starts with a lookbehind assertion or \b or \B (\K is not supported by
2296 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2297 matching functions compatible, and I think it's the right thing to do.
2298
2299 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2300 so that the demonstration program is easily available in environments where
2301 PCRE has not been installed from source.
2302
2303 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2304 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2305 library.
2306
2307 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2308 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2309 is not the first non-POSIX option to be added. Clearly some people find
2310 these options useful.
2311
2312 17. If a caller to the POSIX matching function regexec() passes a non-zero
2313 value for nmatch with a NULL value for pmatch, the value of
2314 nmatch is forced to zero.
2315
2316 18. RunGrepTest did not have a test for the availability of the -u option of
2317 the diff command, as RunTest does. It now checks in the same way as
2318 RunTest, and also checks for the -b option.
2319
2320 19. If an odd number of negated classes containing just a single character
2321 interposed, within parentheses, between a forward reference to a named
2322 subpattern and the definition of the subpattern, compilation crashed with
2323 an internal error, complaining that it could not find the referenced
2324 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2325 [The bug was that it was starting one character too far in when skipping
2326 over the character class, thus treating the ] as data rather than
2327 terminating the class. This meant it could skip too much.]
2328
2329 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2330 /g option in pcretest when the pattern contains \K, which makes it possible
2331 to have an empty string match not at the start, even when the pattern is
2332 anchored. Updated pcretest and pcredemo to use this option.
2333
2334 21. If the maximum number of capturing subpatterns in a recursion was greater
2335 than the maximum at the outer level, the higher number was returned, but
2336 with unset values at the outer level. The correct (outer level) value is
2337 now given.
2338
2339 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2340 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2341 make it do so. The string so far is captured, making this feature
2342 compatible with Perl.
2343
2344 23. The tests have been re-organized, adding tests 11 and 12, to make it
2345 possible to check the Perl 5.10 features against Perl 5.10.
2346
2347 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2348 pattern matches a fixed length string. PCRE did not allow this; now it
2349 does. Neither allows recursion.
2350
2351 25. I finally figured out how to implement a request to provide the minimum
2352 length of subject string that was needed in order to match a given pattern.
2353 (It was back references and recursion that I had previously got hung up
2354 on.) This code has now been added to pcre_study(); it finds a lower bound
2355 to the length of subject needed. It is not necessarily the greatest lower
2356 bound, but using it to avoid searching strings that are too short does give
2357 some useful speed-ups. The value is available to calling programs via
2358 pcre_fullinfo().
2359
2360 26. While implementing 25, I discovered to my embarrassment that pcretest had
2361 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2362 study optimizations had never been tested with that matching function.
2363 Oops. What is worse, even when it was passed study data, there was a bug in
2364 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2365 were also very few tests of studied patterns with pcre_dfa_exec().
2366
2367 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2368 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2369 on the other side of the coin, they are no longer allowed to have different
2370 names, because these cannot be distinguished in PCRE, and this has caused
2371 confusion. (This is a difference from Perl.)
2372
2373 28. When duplicate subpattern names are present (necessarily with different
2374 numbers, as required by 27 above), and a test is made by name in a
2375 conditional pattern, either for a subpattern having been matched, or for
2376 recursion in such a pattern, all the associated numbered subpatterns are
2377 tested, and the overall condition is true if the condition is true for any
2378 one of them. This is the way Perl works, and is also more like the way
2379 testing by number works.
2380
2381
2382 Version 7.9 11-Apr-09
2383 ---------------------
2384
2385 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2386 (pcretest), all targets were linked against these libraries. This included
2387 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2388 libraries. This caused unwanted dependencies to be created. This problem
2389 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2390 pcretest is linked with readline.
2391
2392 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2393 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2394 moved outside it again, because FALSE and TRUE are already defined in AIX,
2395 but BOOL is not.
2396
2397 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2398 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2399
2400 4. The pcregrep documentation said spaces were inserted as well as colons (or
2401 hyphens) following file names and line numbers when outputting matching
2402 lines. This is not true; no spaces are inserted. I have also clarified the
2403 wording for the --colour (or --color) option.
2404
2405 5. In pcregrep, when --colour was used with -o, the list of matching strings
2406 was not coloured; this is different to GNU grep, so I have changed it to be
2407 the same.
2408
2409 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2410 each matching line was coloured. Now it goes on to look for further matches
2411 of any of the test patterns, which is the same behaviour as GNU grep.
2412
2413 7. A pattern that could match an empty string could cause pcregrep to loop; it
2414 doesn't make sense to accept an empty string match in pcregrep, so I have
2415 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2416 seems to be how GNU grep behaves. [But see later change 40 for release
2417 8.33.]
2418
2419 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2420 start or after a newline", because the conditional assertion was not being
2421 correctly handled. The rule now is that both the assertion and what follows
2422 in the first alternative must satisfy the test.
2423
2424 9. If auto-callout was enabled in a pattern with a conditional group whose
2425 condition was an assertion, PCRE could crash during matching, both with
2426 pcre_exec() and pcre_dfa_exec().
2427
2428 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2429 used for matching.
2430
2431 11. Unicode property support in character classes was not working for
2432 characters (bytes) greater than 127 when not in UTF-8 mode.
2433
2434 12. Added the -M command line option to pcretest.
2435
2436 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2437
2438 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2439
2440 16. Added comments and documentation about mis-use of no_arg in the C++
2441 wrapper.
2442
2443 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2444 from Martin Jerabek that uses macro names for all relevant character and
2445 string constants.
2446
2447 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2448 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2449 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2450 these, but not everybody uses configure.
2451
2452 19. A conditional group that had only one branch was not being correctly
2453 recognized as an item that could match an empty string. This meant that an
2454 enclosing group might also not be so recognized, causing infinite looping
2455 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2456 with the subject "ab", where knowledge that the repeated group can match
2457 nothing is needed in order to break the loop.
2458
2459 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2460 exec(), but without supplying a callout function, matching went wrong.
2461
2462 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2463 leak if the size of the offset vector was greater than 30. When the vector
2464 is smaller, the saved offsets during recursion go onto a local stack
2465 vector, but for larger vectors malloc() is used. It was failing to free
2466 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2467 error, in fact).
2468
2469 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2470 heapframe that is used only when UTF-8 support is enabled. This caused no
2471 problem, but was untidy.
2472
2473 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2474 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2475 included within another project.
2476
2477 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2478 slightly modified by me:
2479
2480 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2481 not building pcregrep.
2482
2483 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2484 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2485
2486 25. Forward references, both numeric and by name, in patterns that made use of
2487 duplicate group numbers, could behave incorrectly or give incorrect errors,
2488 because when scanning forward to find the reference group, PCRE was not
2489 taking into account the duplicate group numbers. A pattern such as
2490 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2491
2492 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2493 the feature of a custom pointer more persuasive (as requested by a user).
2494
2495 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2496 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2497 pre-defined.
2498
2499 28. Added support for (*UTF8) at the start of a pattern.
2500
2501 29. Arrange for flags added by the "release type" setting in CMake to be shown
2502 in the configuration summary.
2503
2504
2505 Version 7.8 05-Sep-08
2506 ---------------------
2507
2508 1. Replaced UCP searching code with optimized version as implemented for Ad
2509 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2510 stage table and inline lookup instead of a function, giving speed ups of 2
2511 to 5 times on some simple patterns that I tested. Permission was given to
2512 distribute the MultiStage2.py script that generates the tables (it's not in
2513 the tarball, but is in the Subversion repository).
2514
2515 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2516 scripts.
2517
2518 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2519 a group with a zero qualifier. The result of the study could be incorrect,
2520 or the function might crash, depending on the pattern.
2521
2522 4. Caseless matching was not working for non-ASCII characters in back
2523 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2524 It now works when Unicode Property Support is available.
2525
2526 5. In pcretest, an escape such as \x{de} in the data was always generating
2527 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2528 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2529 truncation.
2530
2531 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2532
2533 7. Added two (int) casts to pcregrep when printing the difference of two
2534 pointers, in case they are 64-bit values.
2535
2536 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2537 test 2 if it fails.
2538
2539 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2540 and a #define of that name to empty if it is not externally set. This is to
2541 allow users of MSVC to set it if necessary.
2542
2543 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2544 the convenience functions in the pcre_get.c source file.
2545
2546 11. An option change at the start of a pattern that had top-level alternatives
2547 could cause overwriting and/or a crash. This command provoked a crash in
2548 some environments:
2549
2550 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2551
2552 This potential security problem was recorded as CVE-2008-2371.
2553
2554 12. For a pattern where the match had to start at the beginning or immediately
2555 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2556 pcre_dfa_exec() could read past the end of the passed subject if there was
2557 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2558 pcretest so that it places the subject at the end of its malloc-ed buffer.
2559
2560 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2561 exec() might read past the end of the data buffer in UTF-8 mode.
2562
2563 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2564 the data contained the byte 0x85 as part of a UTF-8 character within its
2565 first line. This applied both to normal and DFA matching.
2566
2567 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2568 /^[^d]*?$/8 failed to match "abc".
2569
2570 16. Added a missing copyright notice to pcrecpp_internal.h.
2571
2572 17. Make it more clear in the documentation that values returned from
2573 pcre_exec() in ovector are byte offsets, not character counts.
2574
2575 18. Tidied a few places to stop certain compilers from issuing warnings.
2576
2577 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2578 supplied by Stefan Weber. I made a further small update for 7.8 because
2579 there is a change of source arrangements: the pcre_searchfuncs.c module is
2580 replaced by pcre_ucd.c.
2581
2582
2583 Version 7.7 07-May-08
2584 ---------------------
2585
2586 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2587 a string to a long long, pretend we don't even have a long long." This is
2588 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2589
2590 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2591 pre-7.6 versions, which defined a global no_arg variable instead of putting
2592 it in the RE class. (See also #8 below.)
2593
2594 3. Remove a line of dead code, identified by coverity and reported by Nuno
2595 Lopes.
2596
2597 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2598
2599 (1) The include/exclude patterns were being applied to the whole pathnames
2600 of files, instead of just to the final components.
2601
2602 (2) If there was more than one level of directory, the subdirectories were
2603 skipped unless they satisfied the include/exclude conditions. This is
2604 inconsistent with GNU grep (and could even be seen as contrary to the
2605 pcregrep specification - which I improved to make it absolutely clear).
2606 The action now is always to scan all levels of directory, and just
2607 apply the include/exclude patterns to regular files.
2608
2609 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2610 --exclude_dir in the tests to avoid scanning .svn directories.
2611
2612 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2613 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2614 doesn't support NULs in patterns.
2615
2616 7. Added some missing "const"s to declarations of static tables in
2617 pcre_compile.c and pcre_dfa_exec.c.
2618
2619 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2620 caused by fix #2 above. (Subsequently also a second patch to fix the
2621 first patch. And a third patch - this was a messy problem.)
2622
2623 9. Applied Craig's patch to remove the use of push_back().
2624
2625 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2626 matching function regexec().
2627
2628 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2629 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2630 references. PCRE supports relative numbers with this syntax (I don't think
2631 Oniguruma does).
2632
2633 12. Previously, a group with a zero repeat such as (...){0} was completely
2634 omitted from the compiled regex. However, this means that if the group
2635 was called as a subroutine from elsewhere in the pattern, things went wrong
2636 (an internal error was given). Such groups are now left in the compiled
2637 pattern, with a new opcode that causes them to be skipped at execution
2638 time.
2639
2640 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2641 to the way PCRE behaves:
2642
2643 (a) A lone ] character is dis-allowed (Perl treats it as data).
2644
2645 (b) A back reference to an unmatched subpattern matches an empty string
2646 (Perl fails the current match path).
2647
2648 (c) A data ] in a character class must be notated as \] because if the
2649 first data character in a class is ], it defines an empty class. (In
2650 Perl it is not possible to have an empty class.) The empty class []
2651 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2652 The negative empty class [^] matches any one character, independently
2653 of the DOTALL setting.
2654
2655 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2656 non-existent subpattern following a character class starting with ']' and
2657 containing () gave an internal compiling error instead of "reference to
2658 non-existent subpattern". Fortunately, when the pattern did exist, the
2659 compiled code was correct. (When scanning forwards to check for the
2660 existence of the subpattern, it was treating the data ']' as terminating
2661 the class, so got the count wrong. When actually compiling, the reference
2662 was subsequently set up correctly.)
2663
2664 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2665 it was being rejected as not supported by pcre_dfa_exec(), even though
2666 other assertions are supported. I have made pcre_dfa_exec() support
2667 (*FAIL).
2668
2669 16. The implementation of 13c above involved the invention of a new opcode,
2670 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2671 cannot be changed at match time, I realized I could make a small
2672 improvement to matching performance by compiling OP_ALLANY instead of
2673 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2674 on the OP_ANY path.
2675
2676 17. Compiling pcretest on Windows with readline support failed without the
2677 following two fixes: (1) Make the unistd.h include conditional on
2678 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2679
2680 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2681 ncurses library to be included for pcretest when ReadLine support is
2682 requested, but also to allow for it to be overridden. This patch came from
2683 Daniel Bergström.
2684
2685 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2686 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2687 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2688 spotting this.
2689
2690
2691 Version 7.6 28-Jan-08
2692 ---------------------
2693
2694 1. A character class containing a very large number of characters with
2695 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2696 overflow.
2697
2698 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2699 HAVE_LONG_LONG is not defined.
2700
2701 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2702 bring them up to date and include new features. This patch includes:
2703
2704 - Fixed PH's badly added libz and libbz2 support.
2705 - Fixed a problem with static linking.
2706 - Added pcredemo. [But later removed - see 7 below.]
2707 - Fixed dftables problem and added an option.
2708 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2709 HAVE_LONG_LONG.
2710 - Added readline support for pcretest.
2711 - Added an listing of the option settings after cmake has run.
2712
2713 4. A user submitted a patch to Makefile that makes it easy to create
2714 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2715 Makefile.am that cause it to include this special target, without
2716 affecting anything else. Note that the same mingw target plus all
2717 the other distribution libraries and programs are now supported
2718 when configuring with CMake (see 6 below) instead of with
2719 Configure/Make.
2720
2721 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2722 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2723 exported in the Windows port". It has not yet been confirmed that the patch
2724 solves the problem, but it does no harm.
2725
2726 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2727 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2728 with CMake, and also correct the comment about stack recursion.
2729
2730 7. Remove the automatic building of pcredemo from the ./configure system and
2731 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2732 of a program that users should build themselves after PCRE is installed, so
2733 building it automatically is not really right. What is more, it gave
2734 trouble in some build environments.
2735
2736 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2737
2738
2739 Version 7.5 10-Jan-08
2740 ---------------------
2741
2742 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2743 values in parens when parsing an RE using the C++ wrapper."
2744
2745 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2746 Characters greater than 255 were excluded from the class instead of being
2747 included.
2748
2749 3. The same bug as (2) above applied to negated POSIX classes such as
2750 [:^space:].
2751
2752 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2753 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2754 I have changed it.
2755
2756 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2757 first named subpattern) and a construct such as (?&a) would reference the
2758 first named subpattern whose name started with "a" (in other words, the
2759 length check was missing). Both these problems are fixed. "Subpattern name
2760 expected" is now given for (?&) (a zero-length name), and this patch also
2761 makes it give the same error for \k'' (previously it complained that that
2762 was a reference to a non-existent subpattern).
2763
2764 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2765 this is right because (?- can be followed by option settings as well as by
2766 digits. I have, however, made the messages clearer.
2767
2768 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2769 than the number used in the conditional) now cause a compile-time error.
2770 This is actually not compatible with Perl, which accepts such patterns, but
2771 treats the conditional as always being FALSE (as PCRE used to), but it
2772 seems to me that giving a diagnostic is better.
2773
2774 8. Change "alphameric" to the more common word "alphanumeric" in comments
2775 and messages.
2776
2777 9. Fix two occurrences of "backslash" in comments that should have been
2778 "backspace".
2779
2780 10. Remove two redundant lines of code that can never be obeyed (their function
2781 was moved elsewhere).
2782
2783 11. The program that makes PCRE's Unicode character property table had a bug
2784 which caused it to generate incorrect table entries for sequences of
2785 characters that have the same character type, but are in different scripts.
2786 It amalgamated them into a single range, with the script of the first of
2787 them. In other words, some characters were in the wrong script. There were
2788 thirteen such cases, affecting characters in the following ranges:
2789
2790 U+002b0 - U+002c1
2791 U+0060c - U+0060d
2792 U+0061e - U+00612
2793 U+0064b - U+0065e
2794 U+0074d - U+0076d
2795 U+01800 - U+01805
2796 U+01d00 - U+01d77
2797 U+01d9b - U+01dbf
2798 U+0200b - U+0200f
2799 U+030fc - U+030fe
2800 U+03260 - U+0327f
2801 U+0fb46 - U+0fbb1
2802 U+10450 - U+1049d
2803
2804 12. The -o option (show only the matching part of a line) for pcregrep was not
2805 compatible with GNU grep in that, if there was more than one match in a
2806 line, it showed only the first of them. It now behaves in the same way as
2807 GNU grep.
2808
2809 13. If the -o and -v options were combined for pcregrep, it printed a blank
2810 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2811 does the same. The return code can be used to tell if there were any
2812 non-matching lines.
2813
2814 14. Added --file-offsets and --line-offsets to pcregrep.
2815
2816 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2817 infinitely looping recursion. The bug was that positive lookaheads were not
2818 being skipped when checking for a possible empty match (negative lookaheads
2819 and both kinds of lookbehind were skipped).
2820
2821 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2822 inclusion of <windows.h> to before rather than after the definition of
2823 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2824
2825 17. Specifying a possessive quantifier with a specific limit for a Unicode
2826 character property caused pcre_compile() to compile bad code, which led at
2827 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2828 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2829 caused the error; without that there was no problem.
2830
2831 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2832
2833 19. Added --enable-pcretest-libreadline.
2834
2835 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2836 RE::GlobalReplace(). As a result, the number of replacements returned was
2837 double what it should be. I removed one of the increments, but Craig sent a
2838 later patch that removed the other one (the right fix) and added unit tests
2839 that check the return values (which was not done before).
2840
2841 21. Several CMake things:
2842
2843 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2844 the names libpcre and libpcreposix, not just pcre and pcreposix.
2845
2846 (2) The above change means that pcretest and pcregrep are now correctly
2847 linked with the newly-built libraries, not previously installed ones.
2848
2849 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2850
2851 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2852 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2853 UTF-8 newline character). The key issue is that the pattern starts .*;
2854 this means that the match must be either at the beginning, or after a
2855 newline. The bug was in the code for advancing after a failed match and
2856 checking that the new position followed a newline. It was not taking
2857 account of UTF-8 characters correctly.
2858
2859 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2860 character classes. PCRE was not treating the sequence [:...:] as a
2861 character class unless the ... were all letters. Perl, however, seems to
2862 allow any characters between [: and :], though of course it rejects as
2863 unknown any "names" that contain non-letters, because all the known class
2864 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2865 for example, whereas PCRE did not - it did not recognize a POSIX character
2866 class. This seemed a bit dangerous, so the code has been changed to be
2867 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2868 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2869 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2870 Perl does, and where it didn't before.
2871
2872 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2873 Windows environments %n is disabled by default.
2874
2875
2876 Version 7.4 21-Sep-07
2877 ---------------------
2878
2879 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2880 means that a class such as [\s] counted as "explicit reference to CR or
2881 LF". That isn't really right - the whole point of the change was to try to
2882 help when there was an actual mention of one of the two characters. So now
2883 the change happens only if \r or \n (or a literal CR or LF) character is
2884 encountered.
2885
2886 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2887 of both had grown to the point where there were only 3 bits left.
2888 Fortunately, there was spare space in the data structure, and so I have
2889 moved the internal flags into a new 16-bit field to free up more option
2890 bits.
2891
2892 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2893 but did not set the internal JCHANGED flag - either of these is enough to
2894 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2895 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2896 start sets both bits.
2897
2898 4. Added options (at build time, compile time, exec time) to change \R from
2899 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2900
2901 5. doc/pcresyntax.html was missing from the distribution.
2902
2903 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2904 compatibility, even though it is no longer used.
2905
2906 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2907 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2908 windows.h file is present (where different names are used). [This was
2909 reversed later after testing - see 16 below.]
2910
2911 8. Changed all #include <config.h> to #include "config.h". There were also
2912 some further <pcre.h> cases that I changed to "pcre.h".
2913
2914 9. When pcregrep was used with the --colour option, it missed the line ending
2915 sequence off the lines that it output.
2916
2917 10. It was pointed out to me that arrays of string pointers cause lots of
2918 relocations when a shared library is dynamically loaded. A technique of
2919 using a single long string with a table of offsets can drastically reduce
2920 these. I have refactored PCRE in four places to do this. The result is
2921 dramatic:
2922
2923 Originally: 290
2924 After changing UCP table: 187
2925 After changing error message table: 43
2926 After changing table of "verbs" 36
2927 After changing table of Posix names 22
2928
2929 Thanks to the folks working on Gregex for glib for this insight.
2930
2931 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2932 unicode-properties was also set.
2933
2934 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2935
2936 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2937 checked only for CRLF.
2938
2939 14. Added casts to pcretest.c to avoid compiler warnings.
2940
2941 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2942
2943 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2944 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2945 entirely. This removes changes made in 7 above.
2946
2947 17. The CMake files have been updated, and there is now more information about
2948 building with CMake in the NON-UNIX-USE document.
2949
2950
2951 Version 7.3 28-Aug-07
2952 ---------------------
2953
2954 1. In the rejigging of the build system that eventually resulted in 7.1, the
2955 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2956 brackets there is not right, since it causes compilers to look for an
2957 installed pcre.h, not the version that is in the source that is being
2958 compiled (which of course may be different). I have changed it back to:
2959
2960 #include "pcre.h"
2961
2962 I have a vague recollection that the change was concerned with compiling in
2963 different directories, but in the new build system, that is taken care of
2964 by the VPATH setting the Makefile.
2965
2966 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2967 when the subject happened to end in the byte 0x85 (e.g. if the last
2968 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2969 characters but of course it shouldn't be taken as a newline when it is part
2970 of another character. The bug was that, for an unlimited repeat of . in
2971 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2972 characters when looking for a newline.
2973
2974 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2975
2976 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2977 in debug output.
2978
2979 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2980 long printing in the pcrecpp unittest when running under MinGW.
2981
2982 6. ESC_K was left out of the EBCDIC table.
2983
2984 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2985 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2986 limit also applies to "virtual nesting" when a pattern is recursive, and in
2987 this case 1000 isn't so big. I have been able to remove this limit at the
2988 expense of backing off one optimization in certain circumstances. Normally,
2989 when pcre_exec() would call its internal match() function recursively and
2990 immediately return the result unconditionally, it uses a "tail recursion"
2991 feature to save stack. However, when a subpattern that can match an empty
2992 string has an unlimited repetition quantifier, it no longer makes this
2993 optimization. That gives it a stack frame in which to save the data for
2994 checking that an empty string has been matched. Previously this was taken
2995 from the 1000-entry workspace that had been reserved. So now there is no
2996 explicit limit, but more stack is used.
2997
2998 8. Applied Daniel's patches to solve problems with the import/export magic
2999 syntax that is required for Windows, and which was going wrong for the
3000 pcreposix and pcrecpp parts of the library. These were overlooked when this
3001 problem was solved for the main library.
3002
3003 9. There were some crude static tests to avoid integer overflow when computing
3004 the size of patterns that contain repeated groups with explicit upper
3005 limits. As the maximum quantifier is 65535, the maximum group length was
3006 set at 30,000 so that the product of these two numbers did not overflow a
3007 32-bit integer. However, it turns out that people want to use groups that
3008 are longer than 30,000 bytes (though not repeat them that many times).
3009 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3010 made it possible to implement the integer overflow checks in a much more
3011 dynamic way, which I have now done. The artificial limitation on group
3012 length has been removed - we now have only the limit on the total length of
3013 the compiled pattern, which depends on the LINK_SIZE setting.
3014
3015 10. Fixed a bug in the documentation for get/copy named substring when
3016 duplicate names are permitted. If none of the named substrings are set, the
3017 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3018 empty string.
3019
3020 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3021 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3022 because the ] is interpreted as the first data character and the
3023 terminating ] is not found. PCRE has been made compatible with Perl in this
3024 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3025 cause memory overwriting.
3026
3027 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3028 string has been matched (to stop an infinite loop). It was not recognizing
3029 a conditional subpattern that could match an empty string if that
3030 subpattern was within another subpattern. For example, it looped when
3031 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3032 condition was not nested. This bug has been fixed.
3033
3034 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3035 past the start of the subject in the presence of bytes with the top bit
3036 set, for example "\x8aBCD".
3037
3038 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3039 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3040
3041 14. Optimized (?!) to (*FAIL).
3042
3043 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3044 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3045 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3046 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3047 does: it's just the validity check that is more restrictive.
3048
3049 16. Inserted checks for integer overflows during escape sequence (backslash)
3050 processing, and also fixed erroneous offset values for syntax errors during
3051 backslash processing.
3052
3053 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3054 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3055
3056 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3057 caused an overrun.
3058
3059 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3060 something other than just ASCII characters) inside a group that had an
3061 unlimited repeat caused a loop at compile time (while checking to see
3062 whether the group could match an empty string).
3063
3064 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3065 [\P{Any}] did so. (Error in the code for printing property names.)
3066
3067 21. An orphan \E inside a character class could cause a crash.
3068
3069 22. A repeated capturing bracket such as (A)? could cause a wild memory
3070 reference during compilation.
3071
3072 23. There are several functions in pcre_compile() that scan along a compiled
3073 expression for various reasons (e.g. to see if it's fixed length for look
3074 behind). There were bugs in these functions when a repeated \p or \P was
3075 present in the pattern. These operators have additional parameters compared
3076 with \d, etc, and these were not being taken into account when moving along
3077 the compiled data. Specifically:
3078
3079 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3080 length.
3081
3082 (b) An item such as \pL+ within a repeated group could cause crashes or
3083 loops.
3084
3085 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3086 "reference to non-existent subpattern" error.
3087
3088 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3089
3090 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3091 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3092
3093 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3094
3095 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3096 character were causing crashes (broken optimization).
3097
3098 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3099 \p or \P) caused a compile-time loop.
3100
3101 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3102 break. For example, the unstudied pattern [\r\n]A does not match the string
3103 "\r\nA" because change 7.0/46 below moves the current point on by two
3104 characters after failing to match at the start. However, the pattern \nA
3105 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3106 the same is true. There doesn't seem any very clean way out of this, but
3107 what I have chosen to do makes the common cases work: PCRE now takes note
3108 of whether there can be an explicit match for \r or \n anywhere in the
3109 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3110 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3111 pattern has explicit CR or LF references.
3112
3113 29. Added (*CR) etc for changing newline setting at start of pattern.
3114
3115
3116 Version 7.2 19-Jun-07
3117 ---------------------
3118
3119 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3120 which is apparently normally available under Windows.
3121
3122 2. Re-jig the pcregrep tests with different newline settings in an attempt
3123 to make them independent of the local environment's newline setting.
3124
3125 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3126
3127 4. Some of the "internals" tests were previously cut out when the link size
3128 was not 2, because the output contained actual offsets. The recent new
3129 "Z" feature of pcretest means that these can be cut out, making the tests
3130 usable with all link sizes.
3131
3132 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3133 stack recursion. This gives a massive performance boost under BSD, but just
3134 a small improvement under Linux. However, it saves one field in the frame
3135 in all cases.
3136
3137 6. Added more features from the forthcoming Perl 5.10:
3138
3139 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3140 recursion call. It refers to the nth most recently opened parentheses.
3141
3142 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3143 to be opened parentheses.
3144
3145 (c) Conditions that refer to capturing parentheses can be specified
3146 relatively, for example, (?(-2)... or (?(+3)...
3147
3148 (d) \K resets the start of the current match so that everything before
3149 is not part of it.
3150
3151 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3152
3153 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3154 reference syntax.
3155
3156 (g) (?| introduces a group in which the numbering of parentheses in each
3157 alternative starts with the same number.
3158
3159 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3160
3161 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3162 PCRE_INFO_JCHANGED.
3163
3164 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3165 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3166 for detecting groups that can match an empty string.
3167
3168 9. A pattern with a very large number of alternatives (more than several
3169 hundred) was running out of internal workspace during the pre-compile
3170 phase, where pcre_compile() figures out how much memory will be needed. A
3171 bit of new cunning has reduced the workspace needed for groups with
3172 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3173 workspace instead of running out of the 4096 that are available.
3174
3175 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3176
3177 11. Applied patch from Google to remove an optimization that didn't quite work.
3178 The report of the bug said:
3179
3180 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3181 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3182 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3183
3184 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3185 it matched the wrong number of bytes.
3186
3187
3188 Version 7.1 24-Apr-07
3189 ---------------------
3190
3191 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3192 that is more "standard", making use of automake and other Autotools. There
3193 is some re-arrangement of the files and adjustment of comments consequent
3194 on this.
3195
3196 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3197 for recursive directory scanning broke on some systems because the files
3198 are not scanned in any specific order and on different systems the order
3199 was different. A call to "sort" has been inserted into RunGrepTest for the
3200 approprate test as a short-term fix. In the longer term there may be an
3201 alternative.
3202
3203 3. I had an email from Eric Raymond about problems translating some of PCRE's
3204 man pages to HTML (despite the fact that I distribute HTML pages, some
3205 people do their own conversions for various reasons). The problems
3206 concerned the use of low-level troff macros .br and .in. I have therefore
3207 removed all such uses from the man pages (some were redundant, some could
3208 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3209 HTML has been updated to handle .nf/.fi and to complain if it encounters
3210 .br or .in.
3211
3212 4. Updated comments in configure.ac that get placed in config.h.in and also
3213 arranged for config.h to be included in the distribution, with the name
3214 config.h.generic, for the benefit of those who have to compile without
3215 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3216
3217 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3218 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3219 makevp.bat for the current PCRE, using the additional files
3220 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3221
3222 6. A Windows user reported a minor discrepancy with test 2, which turned out
3223 to be caused by a trailing space on an input line that had got lost in his
3224 copy. The trailing space was an accident, so I've just removed it.
3225
3226 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3227 that is needed.
3228
3229 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3230 as "const" (a) because they are and (b) because it helps the PHP
3231 maintainers who have recently made a script to detect big data structures
3232 in the php code that should be moved to the .rodata section. I remembered
3233 to update Builducptable as well, so it won't revert if ucptable.h is ever
3234 re-created.
3235
3236 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3237 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3238 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3239 support is not required. This saves 1.5-2K of code, which is important in
3240 some applications.
3241
3242 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3243 so as not to refer to the tables, even though these functions will never be
3244 called when UTF-8 support is disabled. Otherwise there are problems with a
3245 shared library.
3246
3247 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3248
3249 (a) It was defining its arguments as char * instead of void *.
3250
3251 (b) It was assuming that all moves were upwards in memory; this was true
3252 a long time ago when I wrote it, but is no longer the case.
3253
3254 The emulated memove() is provided for those environments that have neither
3255 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3256 is clearly not the case, as these two bugs were recently reported.
3257
3258 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3259 and Detrail to create the HTML documentation, the .txt form of the man
3260 pages, and it removes trailing spaces from listed files. It also creates
3261 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3262 case, it wraps all the #defines with #ifndefs. This script should be run
3263 before "make dist".
3264
3265 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3266 with Unicode property support.
3267
3268 (a) For a maximizing quantifier, if the two different cases of the
3269 character were of different lengths in their UTF-8 codings (there are
3270 some cases like this - I found 11), and the matching function had to
3271 back up over a mixture of the two cases, it incorrectly assumed they
3272 were both the same length.
3273
3274 (b) When PCRE was configured to use the heap rather than the stack for
3275 recursion during matching, it was not correctly preserving the data for
3276 the other case of a UTF-8 character when checking ahead for a match
3277 while processing a minimizing repeat. If the check also involved
3278 matching a wide character, but failed, corruption could cause an
3279 erroneous result when trying to check for a repeat of the original
3280 character.
3281
3282 13. Some tidying changes to the testing mechanism:
3283
3284 (a) The RunTest script now detects the internal link size and whether there
3285 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3286 values substituted by "configure". (The RunGrepTest script already did
3287 this for UTF-8.) The configure.ac script no longer substitutes the
3288 relevant variables.
3289
3290 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3291 with length and offset values. This means that the output is different
3292 for different internal link sizes. Test 2 is skipped for link sizes
3293 other than 2 because of this, bypassing the problem. Unfortunately,
3294 there was also a test in test 3 (the locale tests) that used /B and
3295 failed for link sizes other than 2. Rather than cut the whole test out,
3296 I have added a new /Z option to pcretest that replaces the length and
3297 offset values with spaces. This is now used to make test 3 independent
3298 of link size. (Test 2 will be tidied up later.)
3299
3300 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3301 segmentation fault instead of returning the appropriate error message.
3302
3303 15. In multiline mode when the newline sequence was set to "any", the pattern
3304 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3305 This doesn't seem right; it now treats the CRLF combination as the line
3306 ending, and so does not match in that case. It's only a pattern such as ^$
3307 that would hit this one: something like ^ABC$ would have failed after \r
3308 and then tried again after \r\n.
3309
3310 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3311 in an attempt to make files that differ only in their line terminators
3312 compare equal. This works on Linux.
3313
3314 17. Under certain error circumstances pcregrep might try to free random memory
3315 as it exited. This is now fixed, thanks to valgrind.
3316
3317 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3318 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3319 was because its rules for how to advance for /g after matching an empty
3320 string at the end of a line did not allow for this case. They now check for
3321 it specially.
3322
3323 20. pcretest is supposed to handle patterns and data of any length, by
3324 extending its buffers when necessary. It was getting this wrong when the
3325 buffer for a data line had to be extended.
3326
3327 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3328 CRLF as a newline sequence.
3329
3330 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3331 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3332 I have nevertheless tidied it up.
3333
3334 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3335
3336 24. Added a man page for pcre-config.
3337
3338
3339 Version 7.0 19-Dec-06
3340 ---------------------
3341
3342 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3343 moving to gcc 4.1.1.
3344
3345 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3346 sys/time.h, which is documented as needed for this function. It doesn't
3347 seem to matter on Linux, but it showed up on some releases of OS X.
3348
3349 3. It seems that there are systems where bytes whose values are greater than
3350 127 match isprint() in the "C" locale. The "C" locale should be the
3351 default when a C program starts up. In most systems, only ASCII printing
3352 characters match isprint(). This difference caused the output from pcretest
3353 to vary, making some of the tests fail. I have changed pcretest so that:
3354
3355 (a) When it is outputting text in the compiled version of a pattern, bytes
3356 other than 32-126 are always shown as hex escapes.
3357
3358 (b) When it is outputting text that is a matched part of a subject string,
3359 it does the same, unless a different locale has been set for the match
3360 (using the /L modifier). In this case, it uses isprint() to decide.
3361
3362 4. Fixed a major bug that caused incorrect computation of the amount of memory
3363 required for a compiled pattern when options that changed within the
3364 pattern affected the logic of the preliminary scan that determines the
3365 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3366 that the computed length was too small. The symptoms of this bug were
3367 either the PCRE error "internal error: code overflow" from pcre_compile(),
3368 or a glibc crash with a message such as "pcretest: free(): invalid next
3369 size (fast)". Examples of patterns that provoked this bug (shown in
3370 pcretest format) are:
3371
3372 /(?-x: )/x
3373 /(?x)(?-x: \s*#\s*)/
3374 /((?i)[\x{c0}])/8
3375 /(?i:[\x{c0}])/8
3376
3377 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3378 is now done differently.
3379
3380 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3381 wrapper classes; (b) implement a new function in the C++ scanner that is
3382 more efficient than the old way of doing things because it avoids levels of
3383 recursion in the regex matching; (c) add a paragraph to the documentation
3384 for the FullMatch() function.
3385
3386 6. The escape sequence \n was being treated as whatever was defined as
3387 "newline". Not only was this contrary to the documentation, which states
3388 that \n is character 10 (hex 0A), but it also went horribly wrong when
3389 "newline" was defined as CRLF. This has been fixed.
3390
3391 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3392 was being set to -1 for the "end of line" case (supposedly a value that no
3393 character can have). Though this value is never used (the check for end of
3394 line is "zero bytes in current character"), it caused compiler complaints.
3395 I've changed it to 0xffffffff.
3396
3397 8. In pcre_version.c, the version string was being built by a sequence of
3398 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3399 string (as it is for production releases) called a macro with an empty
3400 argument. The C standard says the result of this is undefined. The gcc
3401 compiler treats it as an empty string (which was what was wanted) but it is
3402 reported that Visual C gives an error. The source has been hacked around to
3403 avoid this problem.
3404
3405 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3406 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3407 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3408 of them did).
3409
3410 10. Originally, pcretest opened its input and output without "b"; then I was
3411 told that "b" was needed in some environments, so it was added for release
3412 5.0 to both the input and output. (It makes no difference on Unix-like
3413 systems.) Later I was told that it is wrong for the input on Windows. I've
3414 now abstracted the modes into two macros, to make it easier to fiddle with
3415 them, and removed "b" from the input mode under Windows.
3416
3417 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3418
3419 12. Added -help and --help to pcretest as an official way of being reminded
3420 of the options.
3421
3422 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3423 and pcrecpp.cc because they annoy compilers at high warning levels.
3424
3425 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3426
3427 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3428 scripts are not C programs :-) and which was not noticed because it works
3429 on Linux.
3430
3431 16. pcretest is supposed to handle any length of pattern and data line (as one
3432 line or as a continued sequence of lines) by extending its input buffer if
3433 necessary. This feature was broken for very long pattern lines, leading to
3434 a string of junk being passed to pcre_compile() if the pattern was longer
3435 than about 50K.
3436
3437 17. I have done a major re-factoring of the way pcre_compile() computes the
3438 amount of memory needed for a compiled pattern. Previously, there was code
3439 that made a preliminary scan of the pattern in order to do this. That was
3440 OK when PCRE was new, but as the facilities have expanded, it has become
3441 harder and harder to keep it in step with the real compile phase, and there
3442 have been a number of bugs (see for example, 4 above). I have now found a
3443 cunning way of running the real compile function in a "fake" mode that
3444 enables it to compute how much memory it would need, while actually only
3445 ever using a few hundred bytes of working memory and without too many
3446 tests of the mode. This should make future maintenance and development
3447 easier. A side effect of this work is that the limit of 200 on the nesting
3448 depth of parentheses has been removed (though this was never a serious
3449 limitation, I suspect). However, there is a downside: pcre_compile() now
3450 runs more slowly than before (30% or more, depending on the pattern). I
3451 hope this isn't a big issue. There is no effect on runtime performance.
3452
3453 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3454 newline (only possible for the last line of a file) and it was a
3455 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3456
3457 19. Added additional timing features to pcretest. (1) The -tm option now times
3458 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3459 separate command line item, by a number that specifies the number of
3460 repeats to use when timing. The default is 50000; this gives better
3461 precision, but takes uncomfortably long for very large patterns.
3462
3463 20. Extended pcre_study() to be more clever in cases where a branch of a
3464 subpattern has no definite first character. For example, (a*|b*)[cd] would
3465 previously give no result from pcre_study(). Now it recognizes that the
3466 first character must be a, b, c, or d.
3467
3468 21. There was an incorrect error "recursive call could loop indefinitely" if
3469 a subpattern (or the entire pattern) that was being tested for matching an
3470 empty string contained only one non-empty item after a nested subpattern.
3471 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3472 incorrectly, because the \d was being skipped in the check.
3473
3474 22. The pcretest program now has a new pattern option /B and a command line
3475 option -b, which is equivalent to adding /B to every pattern. This causes
3476 it to show the compiled bytecode, without the additional information that
3477 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3478 is the same as /B/I).
3479
3480 23. A new optimization is now able automatically to treat some sequences such
3481 as a*b as a*+b. More specifically, if something simple (such as a character
3482 or a simple class like \d) has an unlimited quantifier, and is followed by
3483 something that cannot possibly match the quantified thing, the quantifier
3484 is automatically "possessified".
3485
3486 24. A recursive reference to a subpattern whose number was greater than 39
3487 went wrong under certain circumstances in UTF-8 mode. This bug could also
3488 have affected the operation of pcre_study().
3489
3490 25. Realized that a little bit of performance could be had by replacing
3491 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3492
3493 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3494
3495 27. Possessive quantifiers such as a++ were previously implemented by turning
3496 them into atomic groups such as ($>a+). Now they have their own opcodes,
3497 which improves performance. This includes the automatically created ones
3498 from 23 above.
3499
3500 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3501 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3502 the first matched character to be a colon. This applied both to named and
3503 numbered groups.
3504
3505 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3506
3507 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3508 building PCRE on Windows easier, so I have included it in the distribution.
3509
3510 31. There is now a check in pcretest against a ridiculously large number being
3511 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3512 loop, the loop is abandoned.
3513
3514 32. Forward references to subpatterns in conditions such as (?(2)...) where
3515 subpattern 2 is defined later cause pcre_compile() to search forwards in
3516 the pattern for the relevant set of parentheses. This search went wrong
3517 when there were unescaped parentheses in a character class, parentheses
3518 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3519
3520 33. "Subroutine" calls and backreferences were previously restricted to
3521 referencing subpatterns earlier in the regex. This restriction has now
3522 been removed.
3523
3524 34. Added a number of extra features that are going to be in Perl 5.10. On the
3525 whole, these are just syntactic alternatives for features that PCRE had
3526 previously implemented using the Python syntax or my own invention. The
3527 other formats are all retained for compatibility.
3528
3529 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3530 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3531 also .NET compatible.
3532
3533 (b) A recursion or subroutine call to a named group can now be defined as
3534 (?&name) as well as (?P>name).
3535
3536 (c) A backreference to a named group can now be defined as \k<name> or
3537 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3538 5.10, are also .NET compatible.
3539
3540 (d) A conditional reference to a named group can now use the syntax
3541 (?(<name>) or (?('name') as well as (?(name).
3542
3543 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3544 groups (named and numbered) that are never evaluated inline, but can be
3545 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3546 is always false. There may be only one alternative in such a group.
3547
3548 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3549 as the simple (?(R). The condition is true only if the most recent
3550 recursion is that of the given number or name. It does not search out
3551 through the entire recursion stack.
3552
3553 (g) The escape \gN or \g{N} has been added, where N is a positive or
3554 negative number, specifying an absolute or relative reference.
3555
3556 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3557 some "unreachable code" warnings.
3558
3559 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3560 things, this adds five new scripts.
3561
3562 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3563 There were also incompatibilities regarding the handling of \Q..\E inside
3564 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3565 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3566
3567 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3568 matches an empty string, and forcibly breaks the loop. There were bugs in
3569 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3570 against aaaa the result was just "a" rather than "aaaa", for example. Two
3571 separate and independent bugs (that affected different cases) have been
3572 fixed.
3573
3574 39. Refactored the code to abolish the use of different opcodes for small
3575 capturing bracket numbers. This is a tidy that I avoided doing when I
3576 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3577 The new approach is not only tidier, it makes it possible to reduce the
3578 memory needed to fix the previous bug (38).
3579
3580 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3581 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3582 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3583 mode.
3584
3585 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3586 report.
3587
3588 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3589 copy construction and assignment in the C++ wrapper.
3590
3591 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3592 couple of bugs that could have given wrong results in the "--newline=crlf"
3593 case.
3594
3595 44. Added a number of casts and did some reorganization of signed/unsigned int
3596 variables following suggestions from Dair Grant. Also renamed the variable
3597 "this" as "item" because it is a C++ keyword.
3598
3599 45. Arranged for dftables to add
3600
3601 #include "pcre_internal.h"
3602
3603 to pcre_chartables.c because without it, gcc 4.x may remove the array
3604 definition from the final binary if PCRE is built into a static library and
3605 dead code stripping is activated.
3606
3607 46. For an unanchored pattern, if a match attempt fails at the start of a
3608 newline sequence, and the newline setting is CRLF or ANY, and the next two
3609 characters are CRLF, advance by two characters instead of one.
3610
3611
3612 Version 6.7 04-Jul-06
3613 ---------------------
3614
3615 1. In order to handle tests when input lines are enormously long, pcretest has
3616 been re-factored so that it automatically extends its buffers when
3617 necessary. The code is crude, but this _is_ just a test program. The
3618 default size has been increased from 32K to 50K.
3619
3620 2. The code in pcre_study() was using the value of the re argument before
3621 testing it for NULL. (Of course, in any sensible call of the function, it
3622 won't be NULL.)
3623
3624 3. The memmove() emulation function in pcre_internal.h, which is used on
3625 systems that lack both memmove() and bcopy() - that is, hardly ever -
3626 was missing a "static" storage class specifier.
3627
3628 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3629 containing an extended class (one that cannot be represented by a bitmap
3630 because it contains high-valued characters or Unicode property items, e.g.
3631 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3632 pattern, but PCRE should not loop if you do not (it no longer does).
3633 [Detail: two cases were found: (a) a repeated subpattern containing an
3634 extended class; (b) a recursive reference to a subpattern that followed a
3635 previous extended class. It wasn't skipping over the extended class
3636 correctly when UTF-8 mode was not set.]
3637
3638 5. A negated single-character class was not being recognized as fixed-length
3639 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3640 compile error "lookbehind assertion is not fixed length".
3641
3642 6. The RunPerlTest auxiliary script was showing an unexpected difference
3643 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3644 write a Perl script that can interpret lines of an input file either as
3645 byte characters or as UTF-8, which is what "perltest" was being required to
3646 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3647 can't do is switch easily at run time between having the "use utf8;" pragma
3648 or not. In the end, I fudged it by using the RunPerlTest script to insert
3649 "use utf8;" explicitly for the UTF-8 tests.
3650
3651 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3652 the end of the subject string, contrary to the documentation and to what
3653 Perl does. This was true of both matching functions. Now it matches only at
3654 the start of the subject and immediately after *internal* newlines.
3655
3656 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3657 a pointer to an int instead of a pointer to an unsigned long int. This
3658 caused problems on 64-bit systems.
3659
3660 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3661 instance of the 'standard' template library not being so standard".
3662
3663 10. There was no check on the number of named subpatterns nor the maximum
3664 length of a subpattern name. The product of these values is used to compute
3665 the size of the memory block for a compiled pattern. By supplying a very
3666 long subpattern name and a large number of named subpatterns, the size
3667 computation could be caused to overflow. This is now prevented by limiting
3668 the length of names to 32 characters, and the number of named subpatterns
3669 to 10,000.
3670
3671 11. Subpatterns that are repeated with specific counts have to be replicated in
3672 the compiled pattern. The size of memory for this was computed from the
3673 length of the subpattern and the repeat count. The latter is limited to
3674 65535, but there was no limit on the former, meaning that integer overflow
3675 could in principle occur. The compiled length of a repeated subpattern is
3676 now limited to 30,000 bytes in order to prevent this.
3677
3678 12. Added the optional facility to have named substrings with the same name.
3679
3680 13. Added the ability to use a named substring as a condition, using the
3681 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3682 are numbers (not recommended). Forward references are permitted.
3683
3684 14. Added forward references in named backreferences (if you see what I mean).
3685
3686 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3687 pattern could run off the end of the subject. For example, the pattern
3688 "(?s)(.{1,5})"8 did this with the subject "ab".
3689
3690 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3691 PCRE_CASELESS was set when matching characters that were quantified with ?
3692 or *.
3693
3694 17. A character class other than a single negated character that had a minimum
3695 but no maximum quantifier - for example [ab]{6,} - was not handled
3696 correctly by pce_dfa_exec(). It would match only one character.
3697
3698 18. A valid (though odd) pattern that looked like a POSIX character
3699 class but used an invalid character after [ (for example [[,abc,]]) caused
3700 pcre_compile() to give the error "Failed: internal error: code overflow" or
3701 in some cases to crash with a glibc free() error. This could even happen if
3702 the pattern terminated after [[ but there just happened to be a sequence of
3703 letters, a binary zero, and a closing ] in the memory that followed.
3704
3705 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3706 over the years. Originally (before any Unicode support), just the bottom 8
3707 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3708 output from "man perlunicode" includes this:
3709
3710 The regular expression compiler produces polymorphic opcodes. That
3711 is, the pattern adapts to the data and automatically switches to
3712 the Unicode character scheme when presented with Unicode data--or
3713 instead uses a traditional byte scheme when presented with byte
3714 data.
3715
3716 Sadly, a wide octal escape does not cause a switch, and in a string with
3717 no other multibyte characters, these octal escapes are treated as before.
3718 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3719 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3720 Unicode string.
3721
3722 I have not perpetrated such confusion in PCRE. Up till now, it took just
3723 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3724 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3725 translate to the appropriate multibyte character.
3726
3727 29. Applied some refactoring to reduce the number of warnings from Microsoft
3728 and Borland compilers. This has included removing the fudge introduced
3729 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3730 a warning about an unused variable.
3731
3732 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3733 characters since release 4.0, because Perl (from release 5.004) does not.
3734 [Or at least, is documented not to: some releases seem to be in conflict
3735 with the documentation.] However, when a pattern was studied with
3736 pcre_study() and all its branches started with \s, PCRE still included VT
3737 as a possible starting character. Of course, this did no harm; it just
3738 caused an unnecessary match attempt.
3739
3740 22. Removed a now-redundant internal flag bit that recorded the fact that case
3741 dependency changed within the pattern. This was once needed for "required
3742 byte" processing, but is no longer used. This recovers a now-scarce options
3743 bit. Also moved the least significant internal flag bit to the most-
3744 significant bit of the word, which was not previously used (hangover from
3745 the days when it was an int rather than a uint) to free up another bit for
3746 the future.
3747
3748 23. Added support for CRLF line endings as well as CR and LF. As well as the
3749 default being selectable at build time, it can now be changed at runtime
3750 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3751 specify that it is scanning data with non-default line endings.
3752
3753 24. Changed the definition of CXXLINK to make it agree with the definition of
3754 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3755
3756 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3757 recursions. This makes a big different to stack usage for some patterns.
3758
3759 26. If a subpattern containing a named recursion or subroutine reference such
3760 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3761 the space required for the compiled pattern went wrong and gave too small a
3762 value. Depending on the environment, this could lead to "Failed: internal
3763 error: code overflow at offset 49" or "glibc detected double free or
3764 corruption" errors.
3765
3766 27. Applied patches from Google (a) to support the new newline modes and (b) to
3767 advance over multibyte UTF-8 characters in GlobalReplace.
3768
3769 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3770 difference for some implementation of PCRE in some Windows version.
3771
3772 29. Added some extra testing facilities to pcretest:
3773
3774 \q<number> in a data line sets the "match limit" value
3775 \Q<number> in a data line sets the "match recursion limt" value
3776 -S <number> sets the stack size, where <number> is in megabytes
3777
3778 The -S option isn't available for Windows.
3779
3780
3781 Version 6.6 06-Feb-06
3782 ---------------------
3783
3784 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3785 in pcreposix.h. I have copied the definition from pcre.h.
3786
3787 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3788 because pcre.h is no longer a built file.
3789
3790 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3791 not normally included in the compiled code.
3792
3793
3794 Version 6.5 01-Feb-06
3795 ---------------------
3796
3797 1. When using the partial match feature with pcre_dfa_exec(), it was not
3798 anchoring the second and subsequent partial matches at the new starting
3799 point. This could lead to incorrect results. For example, with the pattern
3800 /1234/, partially matching against "123" and then "a4" gave a match.
3801
3802 2. Changes to pcregrep:
3803
3804 (a) All non-match returns from pcre_exec() were being treated as failures
3805 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3806 error message is output. Some extra information is given for the
3807 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3808 probably the only errors that are likely to be caused by users (by
3809 specifying a regex that has nested indefinite repeats, for instance).
3810 If there are more than 20 of these errors, pcregrep is abandoned.
3811
3812 (b) A binary zero was treated as data while matching, but terminated the
3813 output line if it was written out. This has been fixed: binary zeroes
3814 are now no different to any other data bytes.
3815
3816 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3817 used to set a locale for matching. The --locale=xxxx long option has
3818 been added (no short equivalent) to specify a locale explicitly on the
3819 pcregrep command, overriding the environment variables.
3820
3821 (d) When -B was used with -n, some line numbers in the output were one less
3822 than they should have been.
3823
3824 (e) Added the -o (--only-matching) option.
3825
3826 (f) If -A or -C was used with -c (count only), some lines of context were
3827 accidentally printed for the final match.
3828
3829 (g) Added the -H (--with-filename) option.
3830
3831 (h) The combination of options -rh failed to suppress file names for files
3832 that were found from directory arguments.
3833
3834 (i) Added the -D (--devices) and -d (--directories) options.
3835
3836 (j) Added the -F (--fixed-strings) option.
3837
3838 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3839
3840 (l) Added the --colo(u)r option.
3841
3842 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3843 is not present by default.
3844
3845 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3846 items such as (?R) or (?1), when the recursion could match a number of
3847 alternatives. If it matched one of the alternatives, but subsequently,
3848 outside the recursion, there was a failure, the code tried to back up into
3849 the recursion. However, because of the way PCRE is implemented, this is not
3850 possible, and the result was an incorrect result from the match.
3851
3852 In order to prevent this happening, the specification of recursion has
3853 been changed so that all such subpatterns are automatically treated as
3854 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3855
3856 4. I had overlooked the fact that, in some locales, there are characters for
3857 which isalpha() is true but neither isupper() nor islower() are true. In
3858 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3859 and ordfeminine) are like this. This affected the treatment of \w and \W
3860 when they appeared in character classes, but not when they appeared outside
3861 a character class. The bit map for "word" characters is now created
3862 separately from the results of isalnum() instead of just taking it from the
3863 upper, lower, and digit maps. (Plus the underscore character, of course.)
3864
3865 5. The above bug also affected the handling of POSIX character classes such as
3866 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3867 permanent tables. Instead, the bit maps for such a class were previously
3868 created as the appropriate unions of the upper, lower, and digit bitmaps.
3869 Now they are created by subtraction from the [[:word:]] class, which has
3870 its own bitmap.
3871
3872 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3873 It is created by subtracting the vertical space characters (\x09, \x0a,
3874 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3875 subtraction was done in the overall bitmap for a character class, meaning
3876 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3877 be recognized. This bug has been fixed.
3878
3879 7. Patches from the folks at Google:
3880
3881 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3882 real life, but is still worth protecting against".
3883
3884 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3885 regular expressions".
3886
3887 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3888 have it.
3889
3890 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3891 "configure" and the latter not, in order to fix a problem somebody had
3892 with compiling the Arg class on HP-UX.
3893
3894 (e) Improve the error-handling of the C++ wrapper a little bit.
3895
3896 (f) New tests for checking recursion limiting.
3897
3898 8. The pcre_memmove() function, which is used only if the environment does not
3899 have a standard memmove() function (and is therefore rarely compiled),
3900 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3901 returning a result (though PCRE never actually uses the result).
3902
3903 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3904 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3905 returned instead of calling malloc() with an overflowing number that would
3906 most likely cause subsequent chaos.
3907
3908 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3909
3910 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3911 with this option is matched, the nmatch and pmatch options of regexec() are
3912 ignored.
3913
3914 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3915 provided in case anyone wants to the the POSIX interface with UTF-8
3916 strings.
3917
3918 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3919 C++ linking (needed for some HP-UX environments).
3920
3921 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3922 (unused parameter) and in the pcre_printint() function (omitted "default"
3923 switch label when the default is to do nothing).
3924
3925 15. Added some code to make it possible, when PCRE is compiled as a C++
3926 library, to replace subject pointers for pcre_exec() with a smart pointer
3927 class, thus making it possible to process discontinuous strings.
3928
3929 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3930 much the same function. They were added by different people who were trying
3931 to make PCRE easy to compile on non-Unix systems. It has been suggested
3932 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3933 for compiling on Windows systems. I have therefore replaced it with
3934 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3935 defaults to "extern" for C or "extern C" for C++, which works fine on
3936 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3937 SCOPE with something explicit in config.h. In addition:
3938
3939 (a) pcreposix.h still had just "extern" instead of either of these macros;
3940 I have replaced it with PCRE_DATA_SCOPE.
3941
3942 (b) Functions such as _pcre_xclass(), which are internal to the library,
3943 but external in the C sense, all had PCRE_EXPORT in their definitions.
3944 This is apparently wrong for the Windows case, so I have removed it.
3945 (It makes no difference on Unix-like systems.)
3946
3947 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3948 of recursive calls to match(). This is different to MATCH_LIMIT because
3949 that limits the total number of calls to match(), not all of which increase
3950 the depth of recursion. Limiting the recursion depth limits the amount of
3951 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3952 when PCRE is compiled, and changed at run time. A patch from Google adds
3953 this functionality to the C++ interface.
3954
3955 18. Changes to the handling of Unicode character properties:
3956
3957 (a) Updated the table to Unicode 4.1.0.
3958
3959 (b) Recognize characters that are not in the table as "Cn" (undefined).
3960
3961 (c) I revised the way the table is implemented to a much improved format
3962 which includes recognition of ranges. It now supports the ranges that
3963 are defined in UnicodeData.txt, and it also amalgamates other
3964 characters into ranges. This has reduced the number of entries in the
3965 table from around 16,000 to around 3,000, thus reducing its size
3966 considerably. I realized I did not need to use a tree structure after
3967 all - a binary chop search is just as efficient. Having reduced the
3968 number of entries, I extended their size from 6 bytes to 8 bytes to
3969 allow for more data.
3970
3971 (d) Added support for Unicode script names via properties such as \p{Han}.
3972
3973 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3974 matching that character.
3975
3976 20. When matching a repeated Unicode property with a minimum greater than zero,
3977 (for example \pL{2,}), PCRE could look past the end of the subject if it
3978 reached it while seeking the minimum number of characters. This could
3979 happen only if some of the characters were more than one byte long, because
3980 there is a check for at least the minimum number of bytes.
3981
3982 21. Refactored the implementation of \p and \P so as to be more general, to
3983 allow for more different types of property in future. This has changed the
3984 compiled form incompatibly. Anybody with saved compiled patterns that use
3985 \p or \P will have to recompile them.
3986
3987 22. Added "Any" and "L&" to the supported property types.
3988
3989 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3990 but give a compile time error if the value is greater than 0xff.
3991
3992 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3993 accidentally not being installed or uninstalled.
3994
3995 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3996 made were to insert the current release number. This seemed silly, because
3997 it made things harder for people building PCRE on systems that don't run
3998 "configure". I have turned pcre.h into a distributed file, no longer built
3999 by "configure", with the version identification directly included. There is
4000 no longer a pcre.h.in file.
4001
4002 However, this change necessitated a change to the pcre-config script as
4003 well. It is built from pcre-config.in, and one of the substitutions was the
4004 release number. I have updated configure.ac so that ./configure now finds
4005 the release number by grepping pcre.h.
4006
4007 26. Added the ability to run the tests under valgrind.
4008
4009
4010 Version 6.4 05-Sep-05
4011 ---------------------
4012
4013 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4014 "--" to be printed when multiple files were scanned, even when none of the
4015 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4016 consider it to be a bug, and have restored the previous behaviour.
4017
4018 2. A couple of code tidies to get rid of compiler warnings.
4019
4020 3. The pcretest program used to cheat by referring to symbols in the library
4021 whose names begin with _pcre_. These are internal symbols that are not
4022 really supposed to be visible externally, and in some environments it is
4023 possible to suppress them. The cheating is now confined to including
4024 certain files from the library's source, which is a bit cleaner.
4025
4026 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4027 file's purpose clearer.
4028
4029 5. Reorganized pcre_ucp_findchar().
4030
4031
4032 Version 6.3 15-Aug-05
4033 ---------------------
4034
4035 1. The file libpcre.pc.in did not have general read permission in the tarball.
4036
4037 2. There were some problems when building without C++ support:
4038
4039 (a) If C++ support was not built, "make install" and "make test" still
4040 tried to test it.
4041
4042 (b) There were problems when the value of CXX was explicitly set. Some
4043 changes have been made to try to fix these, and ...
4044
4045 (c) --disable-cpp can now be used to explicitly disable C++ support.
4046
4047 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4048 backslash in a target when C++ was disabled. This confuses some
4049 versions of "make", apparently. Using an intermediate variable solves
4050 this. (Same for CPP_LOBJ.)
4051
4052 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4053 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4054 necessary on certain architectures.
4055
4056 4. Added a setting of -export-symbols-regex to the link command to remove
4057 those symbols that are exported in the C sense, but actually are local
4058 within the library, and not documented. Their names all begin with
4059 "_pcre_". This is not a perfect job, because (a) we have to except some
4060 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4061 available (and never for static libraries). I have made a note to try to
4062 find a way round (a) in the future.
4063
4064
4065 Version 6.2 01-Aug-05
4066 ---------------------
4067
4068 1. There was no test for integer overflow of quantifier values. A construction
4069 such as {1111111111111111} would give undefined results. What is worse, if
4070 a minimum quantifier for a parenthesized subpattern overflowed and became
4071 negative, the calculation of the memory size went wrong. This could have
4072 led to memory overwriting.
4073
4074 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4075
4076 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4077 operating environments where this matters.
4078
4079 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4080 PCRE options from within the C++ wrapper.
4081
4082 5. Named capturing subpatterns were not being correctly counted when a pattern
4083 was compiled. This caused two problems: (a) If there were more than 100
4084 such subpatterns, the calculation of the memory needed for the whole
4085 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4086 back references of the form \12, where the number was greater than 9, were
4087 not recognized as back references, even though there were sufficient
4088 previous subpatterns.
4089
4090 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4091 versions of gcc, e.g. 2.95.4.
4092
4093
4094 Version 6.1 21-Jun-05
4095 ---------------------
4096
4097 1. There was one reference to the variable "posix" in pcretest.c that was not
4098 surrounded by "#if !defined NOPOSIX".
4099
4100 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4101 the cross-check on the old pcre_info() function, for the benefit of the
4102 cut-down version of PCRE that is currently imported into Exim.
4103
4104 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4105 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4106 patterns that start (?i) but I don't think that matters. The use of (?i) is
4107 just an example; this all applies to the other options as well.
4108
4109 4. Since libtool seems to echo the compile commands it is issuing, the output
4110 from "make" can be reduced a bit by putting "@" in front of each libtool
4111 compile command.
4112
4113 5. Patch from the folks at Google for configure.in to be a bit more thorough
4114 in checking for a suitable C++ installation before trying to compile the
4115 C++ stuff. This should fix a reported problem when a compiler was present,
4116 but no suitable headers.
4117
4118 6. The man pages all had just "PCRE" as their title. I have changed them to
4119 be the relevant file name. I have also arranged that these names are
4120 retained in the file doc/pcre.txt, which is a concatenation in text format
4121 of all the man pages except the little individual ones for each function.
4122
4123 7. The NON-UNIX-USE file had not been updated for the different set of source
4124 files that come with release 6. I also added a few comments about the C++
4125 wrapper.
4126
4127
4128 Version 6.0 07-Jun-05
4129 ---------------------
4130
4131 1. Some minor internal re-organization to help with my DFA experiments.
4132
4133 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4134 didn't matter for the library itself when fully configured, but did matter
4135 when compiling without UCP support, or within Exim, where the ucp files are
4136 not imported.
4137
4138 3. Refactoring of the library code to split up the various functions into
4139 different source modules. The addition of the new DFA matching code (see
4140 below) to a single monolithic source would have made it really too
4141 unwieldy, quite apart from causing all the code to be include in a
4142 statically linked application, when only some functions are used. This is
4143 relevant even without the DFA addition now that patterns can be compiled in
4144 one application and matched in another.
4145
4146 The downside of splitting up is that there have to be some external
4147 functions and data tables that are used internally in different modules of
4148 the library but which are not part of the API. These have all had their
4149 names changed to start with "_pcre_" so that they are unlikely to clash
4150 with other external names.
4151
4152 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4153 a different (DFA) algorithm. Although it is slower than the original
4154 function, it does have some advantages for certain types of matching
4155 problem.
4156
4157 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4158 including restarting after a partial match.
4159
4160 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4161 defined when compiling for Windows was sent to me. I have put it into the
4162 code, though I have no means of testing or verifying it.
4163
4164 7. Added the pcre_refcount() auxiliary function.
4165
4166 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4167 match before or at the first newline in the subject string. In pcretest,
4168 the /f option on a pattern can be used to set this.
4169
4170 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4171 would behave wrongly. This has been present in PCRE since release 4.0.
4172
4173 10. A number of changes to the pcregrep command:
4174
4175 (a) Refactored how -x works; insert ^(...)$ instead of setting
4176 PCRE_ANCHORED and checking the length, in preparation for adding
4177 something similar for -w.
4178
4179 (b) Added the -w (match as a word) option.
4180
4181 (c) Refactored the way lines are read and buffered so as to have more
4182 than one at a time available.
4183
4184 (d) Implemented a pcregrep test script.
4185
4186 (e) Added the -M (multiline match) option. This allows patterns to match
4187 over several lines of the subject. The buffering ensures that at least
4188 8K, or the rest of the document (whichever is the shorter) is available
4189 for matching (and similarly the previous 8K for lookbehind assertions).
4190
4191 (f) Changed the --help output so that it now says
4192
4193 -w, --word-regex(p)
4194
4195 instead of two lines, one with "regex" and the other with "regexp"
4196 because that confused at least one person since the short forms are the
4197 same. (This required a bit of code, as the output is generated
4198 automatically from a table. It wasn't just a text change.)
4199
4200 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4201 option but starts with a hyphen. Could be a pattern or a path name
4202 starting with a hyphen, for instance.
4203
4204 (h) "-" can be given as a file name to represent stdin.
4205
4206 (i) When file names are being printed, "(standard input)" is used for
4207 the standard input, for compatibility with GNU grep. Previously
4208 "<stdin>" was used.
4209
4210 (j) The option --label=xxx can be used to supply a name to be used for
4211 stdin when file names are being printed. There is no short form.
4212
4213 (k) Re-factored the options decoding logic because we are going to add
4214 two more options that take data. Such options can now be given in four
4215 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4216
4217 (l) Added the -A, -B, and -C options for requesting that lines of context
4218 around matches be printed.
4219
4220 (m) Added the -L option to print the names of files that do not contain
4221 any matching lines, that is, the complement of -l.
4222
4223 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4224 continue to scan other files.
4225
4226 (o) The -s option was incorrectly implemented. For compatibility with other
4227 greps, it now suppresses the error message for a non-existent or non-
4228 accessible file (but not the return code). There is a new option called
4229 -q that suppresses the output of matching lines, which was what -s was
4230 previously doing.
4231
4232 (p) Added --include and --exclude options to specify files for inclusion
4233 and exclusion when recursing.
4234
4235 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4236 Hopefully, it now does.
4237
4238 12. Missing cast in pcre_study().
4239
4240 13. Added an "uninstall" target to the makefile.
4241
4242 14. Replaced "extern" in the function prototypes in Makefile.in with
4243 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4244 world, but is set differently for Windows.
4245
4246 15. Added a second compiling function called pcre_compile2(). The only
4247 difference is that it has an extra argument, which is a pointer to an
4248 integer error code. When there is a compile-time failure, this is set
4249 non-zero, in addition to the error test pointer being set to point to an
4250 error message. The new argument may be NULL if no error number is required
4251 (but then you may as well call pcre_compile(), which is now just a
4252 wrapper). This facility is provided because some applications need a
4253 numeric error indication, but it has also enabled me to tidy up the way
4254 compile-time errors are handled in the POSIX wrapper.
4255
4256 16. Added VPATH=.libs to the makefile; this should help when building with one
4257 prefix path and installing with another. (Or so I'm told by someone who
4258 knows more about this stuff than I do.)
4259
4260 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4261 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4262 match everything, including newlines. This is not POSIX-compatible, but
4263 somebody wanted the feature. From pcretest it can be activated by using
4264 both the P and the s flags.
4265
4266 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4267
4268 19. libpcre.pc was being incorrectly installed as executable.
4269
4270 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4271 it now also looks for '\r' so that it will work unmodified on Windows.
4272
4273 21. Added Google's contributed C++ wrapper to the distribution.
4274
4275 22. Added some untidy missing memory free() calls in pcretest, to keep
4276 Electric Fence happy when testing.
4277
4278
4279
4280 Version 5.0 13-Sep-04
4281 ---------------------
4282
4283 1. Internal change: literal characters are no longer packed up into items
4284 containing multiple characters in a single byte-string. Each character
4285 is now matched using a separate opcode. However, there may be more than one
4286 byte in the character in UTF-8 mode.
4287
4288 2. The pcre_callout_block structure has two new fields: pattern_position and
4289 next_item_length. These contain the offset in the pattern to the next match
4290 item, and its length, respectively.
4291
4292 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4293 insertion of callouts before each pattern item. Added the /C option to
4294 pcretest to make use of this.
4295
4296 4. On the advice of a Windows user, the lines
4297
4298 #if defined(_WIN32) || defined(WIN32)
4299 _setmode( _fileno( stdout ), 0x8000 );
4300 #endif /* defined(_WIN32) || defined(WIN32) */
4301
4302 have been added to the source of pcretest. This apparently does useful
4303 magic in relation to line terminators.
4304
4305 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4306 for the benefit of those environments where the "b" makes a difference.
4307
4308 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4309 to know about it. I have put a hack into configure.in that adds in code
4310 to set GCC=yes if CC=icc. This seems to end up at a point in the
4311 generated configure script that is early enough to affect the setting of
4312 compiler options, which is what is needed, but I have no means of testing
4313 whether it really works. (The user who reported this had patched the
4314 generated configure script, which of course I cannot do.)
4315
4316 LATER: After change 22 below (new libtool files), the configure script
4317 seems to know about icc (and also ecc). Therefore, I have commented out
4318 this hack in configure.in.
4319
4320 7. Added support for pkg-config (2 patches were sent in).
4321
4322 8. Negated POSIX character classes that used a combination of internal tables
4323 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4324 [[:^ascii]]. Typically, they would match almost any characters. The other
4325 POSIX classes were not broken in this way.
4326
4327 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4328 to find the match, as PCRE was deluded into thinking that the match had to
4329 start at the start point or following a newline. The same bug applied to
4330 patterns with negative forward assertions or any backward assertions
4331 preceding ".*" at the start, unless the pattern required a fixed first
4332 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4333
4334 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4335 starting at the last subject character, bytes beyond the end of the subject
4336 string were read.
4337
4338 11. Renamed the variable "class" as "classbits" to make life easier for C++
4339 users. (Previously there was a macro definition, but it apparently wasn't
4340 enough.)
4341
4342 12. Added the new field "tables" to the extra data so that tables can be passed
4343 in at exec time, or the internal tables can be re-selected. This allows
4344 a compiled regex to be saved and re-used at a later time by a different
4345 program that might have everything at different addresses.
4346
4347 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4348 -R library as well as a -L library.
4349
4350 14. The debugging options of pcretest (-d on the command line or D on a
4351 pattern) showed incorrect output for anything following an extended class
4352 that contained multibyte characters and which was followed by a quantifier.
4353
4354 15. Added optional support for general category Unicode character properties
4355 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4356 support. It adds about 90K to the size of the library. The meanings of the
4357 inbuilt class escapes such as \d and \s have NOT been changed.
4358
4359 16. Updated pcredemo.c to include calls to free() to release the memory for the
4360 compiled pattern.
4361
4362 17. The generated file chartables.c was being created in the source directory
4363 instead of in the building directory. This caused the build to fail if the
4364 source directory was different from the building directory, and was
4365 read-only.
4366
4367 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4368 file. No doubt somebody will tell me if they don't make sense... Also added
4369 Dan Mooney's comments about building on OpenVMS.
4370
4371 19. Added support for partial matching via the PCRE_PARTIAL option for
4372 pcre_exec() and the \P data escape in pcretest.
4373
4374 20. Extended pcretest with 3 new pattern features:
4375
4376 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4377 write the compiled pattern to the file whose name is "rest-of-line".
4378 This is a straight binary dump of the data, with the saved pointer to
4379 the character tables forced to be NULL. The study data, if any, is
4380 written too. After writing, pcretest reads a new pattern.
4381
4382 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4383 compiled pattern from the given file. There must not be any
4384 occurrences of "<" in the file name (pretty unlikely); if there are,
4385 pcretest will instead treat the initial "<" as a pattern delimiter.
4386 After reading in the pattern, pcretest goes on to read data lines as
4387 usual.
4388
4389 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4390 and 16-bit fields in a compiled pattern, to simulate a pattern that
4391 was compiled on a host of opposite endianness.
4392
4393 21. The pcre-exec() function can now cope with patterns that were compiled on
4394 hosts of opposite endianness, with this restriction:
4395
4396 As for any compiled expression that is saved and used later, the tables
4397 pointer field cannot be preserved; the extra_data field in the arguments
4398 to pcre_exec() should be used to pass in a tables address if a value
4399 other than the default internal tables were used at compile time.
4400
4401 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4402 now diagnosed as an error. Previously, most of the time, a negative number
4403 would have been treated as zero, but if in addition "ovector" was passed as
4404 NULL, a crash could occur.
4405
4406 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4407 new versions from the libtool 1.5 distribution (the last one is a copy of
4408 a file called libtool.m4). This seems to have fixed the need to patch
4409 "configure" to support Darwin 1.3 (which I used to do). However, I still
4410 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4411 workstation).
4412
4413 24. Changed the PCRE licence to be the more standard "BSD" licence.
4414
4415
4416 Version 4.5 01-Dec-03
4417 ---------------------
4418
4419 1. There has been some re-arrangement of the code for the match() function so
4420 that it can be compiled in a version that does not call itself recursively.
4421 Instead, it keeps those local variables that need separate instances for
4422 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4423 needs to "recurse". Keeping track of where control must go is done by means
4424 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4425 hide most of the details from the main code, and operates only if
4426 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4427 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4428 operating.
4429
4430 To make it easier for callers to provide specially tailored get/free
4431 functions for this usage, two new functions, pcre_stack_malloc, and
4432 pcre_stack_free, are used. They are always called in strict stacking order,
4433 and the size of block requested is always the same.
4434
4435 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4436 PCRE has been compiled to use the stack or the heap for recursion. The
4437 -C option of pcretest uses this to show which version is compiled.
4438
4439 A new data escape \S, is added to pcretest; it causes the amounts of store
4440 obtained and freed by both kinds of malloc/free at match time to be added
4441 to the output.
4442
4443 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4444 what's available on my current Linux desktop machine.
4445
4446 3. When matching a UTF-8 string, the test for a valid string at the start has
4447 been extended. If start_offset is not zero, PCRE now checks that it points
4448 to a byte that is the start of a UTF-8 character. If not, it returns
4449 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4450 this is necessary because there may be backward assertions in the pattern.
4451 When matching the same subject several times, it may save resources to use
4452 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4453
4454 4. The code for checking the validity of UTF-8 strings has been tightened so
4455 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4456 containing "overlong sequences".
4457
4458 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4459 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4460 should have been "&", but it just so happened that all the cases this let
4461 through by mistake were picked up later in the function.
4462
4463 6. I had used a variable called "isblank" - this is a C99 function, causing
4464 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4465
4466 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4467 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4468 so the version is shown; (c) stop "make test" echoing "./RunTest".
4469
4470 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4471
4472 9. The prototype for memmove() for systems that don't have it was using
4473 size_t, but the inclusion of the header that defines size_t was later. I've
4474 moved the #includes for the C headers earlier to avoid this.
4475
4476 10. Added some adjustments to the code to make it easier to compiler on certain
4477 special systems:
4478
4479 (a) Some "const" qualifiers were missing.
4480 (b) Added the macro EXPORT before all exported functions; by default this
4481 is defined to be empty.
4482 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4483 that it reads its output file name as an argument instead of writing
4484 to the standard output and assuming this can be redirected.
4485
4486 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4487 class containing characters with values greater than 255, PCRE compilation
4488 went into a loop.
4489
4490 12. A recursive reference to a subpattern that was within another subpattern
4491 that had a minimum quantifier of zero caused PCRE to crash. For example,
4492 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4493 recursion. If the recursively-called subpattern itself had a zero repeat,
4494 that was OK.
4495
4496 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4497 buffer into which it was copied (for escape processing) was still set at
4498 1024, so long lines caused crashes.
4499
4500 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4501 "internal error: code overflow...". This applied to any character class
4502 that was followed by a possessive quantifier.
4503
4504 15. Modified the Makefile to add libpcre.la as a prerequisite for
4505 libpcreposix.la because I was told this is needed for a parallel build to
4506 work.
4507
4508 16. If a pattern that contained .* following optional items at the start was
4509 studied, the wrong optimizing data was generated, leading to matching
4510 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4511 matching string must start with a or b or c. The correct conclusion for
4512 this pattern is that a match can start with any character.
4513
4514
4515 Version 4.4 13-Aug-03
4516 ---------------------
4517
4518 1. In UTF-8 mode, a character class containing characters with values between
4519 127 and 255 was not handled correctly if the compiled pattern was studied.
4520 In fixing this, I have also improved the studying algorithm for such
4521 classes (slightly).
4522
4523 2. Three internal functions had redundant arguments passed to them. Removal
4524 might give a very teeny performance improvement.
4525
4526 3. Documentation bug: the value of the capture_top field in a callout is *one
4527 more than* the number of the hightest numbered captured substring.
4528
4529 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4530 in incorrectly linking with a previously installed version. They now link
4531 explicitly with libpcre.la.
4532
4533 5. configure.in no longer needs to recognize Cygwin specially.
4534
4535 6. A problem in pcre.in for Windows platforms is fixed.
4536
4537 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4538 pcretest, it used to include the size of the study block as part of its
4539 output. Unfortunately, the structure contains a field that has a different
4540 size on different hardware architectures. This meant that the tests that
4541 showed this size failed. As the block is currently always of a fixed size,
4542 this information isn't actually particularly useful in pcretest output, so
4543 I have just removed it.
4544
4545 8. Three pre-processor statements accidentally did not start in column 1.
4546 Sadly, there are *still* compilers around that complain, even though
4547 standard C has not required this for well over a decade. Sigh.
4548
4549 9. In pcretest, the code for checking callouts passed small integers in the
4550 callout_data field, which is a void * field. However, some picky compilers
4551 complained about the casts involved for this on 64-bit systems. Now
4552 pcretest passes the address of the small integer instead, which should get
4553 rid of the warnings.
4554
4555 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4556 both compile and run time, and gives an error if an invalid UTF-8 sequence
4557 is found. There is a option for disabling this check in cases where the
4558 string is known to be correct and/or the maximum performance is wanted.
4559
4560 11. In response to a bug report, I changed one line in Makefile.in from
4561
4562 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4563 to
4564 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4565
4566 to look similar to other lines, but I have no way of telling whether this
4567 is the right thing to do, as I do not use Windows. No doubt I'll get told
4568 if it's wrong...
4569
4570
4571 Version 4.3 21-May-03
4572 ---------------------
4573
4574 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4575 Makefile.
4576
4577 2. Some refactoring to improve the quality of the code:
4578
4579 (i) The utf8_table... variables are now declared "const".
4580
4581 (ii) The code for \cx, which used the "case flipping" table to upper case
4582 lower case letters, now just substracts 32. This is ASCII-specific,
4583 but the whole concept of \cx is ASCII-specific, so it seems
4584 reasonable.
4585
4586 (iii) PCRE was using its character types table to recognize decimal and
4587 hexadecimal digits in the pattern. This is silly, because it handles
4588 only 0-9, a-f, and A-F, but the character types table is locale-
4589 specific, which means strange things might happen. A private
4590 table is now used for this - though it costs 256 bytes, a table is
4591 much faster than multiple explicit tests. Of course, the standard
4592 character types table is still used for matching digits in subject
4593 strings against \d.
4594
4595 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4596 ending in _t are). So I've renamed it as ESC_tee.
4597
4598 3. The first argument for regexec() in the POSIX wrapper should have been
4599 defined as "const".
4600
4601 4. Changed pcretest to use malloc() for its buffers so that they can be
4602 Electric Fenced for debugging.
4603
4604 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4605 to read one or more bytes before the start of the subject string. Often this
4606 had no effect on PCRE's behaviour, but in some circumstances it could
4607 provoke a segmentation fault.
4608
4609 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4610 to try to read one or more bytes before the start of the subject string.
4611
4612 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4613 UTF-8 support could misbehave in various ways if the subject string
4614 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4615 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4616 back over UTF-8 characters.)
4617
4618
4619 Version 4.2 14-Apr-03
4620 ---------------------
4621
4622 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4623
4624 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4625 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4626 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4627 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4628 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4629 and BUILD_EXEEXT
4630 Note: automatic setting of the BUILD variables is not yet working
4631 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4632 compile-time but not at link-time
4633 [LINK]: use for linking executables only
4634 make different versions for Windows and non-Windows
4635 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4636 libraries
4637 [LINK_FOR_BUILD]: new variable
4638 [OBJEXT]: use throughout
4639 [EXEEXT]: use throughout
4640 <winshared>: new target
4641 <wininstall>: new target
4642 <dftables.o>: use native compiler
4643 <dftables>: use native linker
4644 <install>: handle Windows platform correctly
4645 <clean>: ditto
4646 <check>: ditto
4647 copy DLL to top builddir before testing
4648
4649 As part of these changes, -no-undefined was removed again. This was reported
4650 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4651 in any case.
4652
4653 3. Some tidies to get rid of compiler warnings:
4654
4655 . In the match_data structure, match_limit was an unsigned long int, whereas
4656 match_call_count was an int. I've made them both unsigned long ints.
4657
4658 . In pcretest the fact that a const uschar * doesn't automatically cast to
4659 a void * provoked a warning.
4660
4661 . Turning on some more compiler warnings threw up some "shadow" variables
4662 and a few more missing casts.
4663
4664 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4665 option, a class that contained a single character with a value between 128
4666 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4667
4668 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4669 option, a class that contained several characters, but with at least one
4670 whose value was between 128 and 255 caused PCRE to crash.
4671
4672
4673 Version 4.1 12-Mar-03
4674 ---------------------
4675
4676 1. Compiling with gcc -pedantic found a couple of places where casts were
4677 needed, and a string in dftables.c that was longer than standard compilers are
4678 required to support.
4679
4680 2. Compiling with Sun's compiler found a few more places where the code could
4681 be tidied up in order to avoid warnings.
4682
4683 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4684 first of these names is deprecated in the latest Autoconf in favour of the name
4685 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4686 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4687 analogy I have changed it to CFLAGS_FOR_BUILD.
4688
4689 4. Added -no-undefined to the linking command in the Makefile, because this is
4690 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4691 linking step for the pcreposix library.
4692
4693 5. PCRE was failing to diagnose the case of two named groups with the same
4694 name.
4695
4696 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4697 literal character that is needed in the subject for a match, and scans along to
4698 ensure that it is present before embarking on the full matching process. This
4699 saves time in cases of nested unlimited repeats that are never going to match.
4700 Problem: the scan can take a lot of time if the subject is very long (e.g.
4701 megabytes), thus penalizing straightforward matches. It is now done only if the
4702 amount of subject to be scanned is less than 1000 bytes.
4703
4704 7. A lesser problem with the same optimization is that it was recording the
4705 first character of an anchored pattern as "needed", thus provoking a search
4706 right along the subject, even when the first match of the pattern was going to
4707 fail. The "needed" character is now not set for anchored patterns, unless it
4708 follows something in the pattern that is of non-fixed length. Thus, it still
4709 fulfils its original purpose of finding quick non-matches in cases of nested
4710 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4711
4712
4713 Version 4.0 17-Feb-03
4714 ---------------------
4715
4716 1. If a comment in an extended regex that started immediately after a meta-item
4717 extended to the end of string, PCRE compiled incorrect data. This could lead to
4718 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4719
4720 2. Moved to autoconf 2.53 and libtool 1.4.2.
4721
4722 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4723 the special perltest8 script is no longer needed - all the tests can be run
4724 from a single perltest script.
4725
4726 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4727 by \s. It has now been removed in PCRE. This means it isn't recognized as
4728 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4729 class [:space:] *does* include VT, thereby creating a mess.
4730
4731 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4732 space and tab.
4733
4734 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4735 its new features into the main test script, reducing the number of scripts.
4736
4737 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4738 were backward compatible, and made the (?i) apply to the whole pattern, as if
4739 /i were given. Now it behaves more logically, and applies the option setting
4740 only to what follows. PCRE has been changed to follow suit. However, if it
4741 finds options settings right at the start of the pattern, it extracts them into
4742 the global options, as before. Thus, they show up in the info data.
4743
4744 8. Added support for the \Q...\E escape sequence. Characters in between are
4745 treated as literals. This is slightly different from Perl in that $ and @ are
4746 also handled as literals inside the quotes. In Perl, they will cause variable
4747 interpolation. Note the following examples:
4748
4749 Pattern PCRE matches Perl matches
4750
4751 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4752 \Qabc\$xyz\E abc\$xyz abc\$xyz
4753 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4754
4755 For compatibility with Perl, \Q...\E sequences are recognized inside character
4756 classes as well as outside them.
4757
4758 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4759 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4760 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4761 signed/unsigned warnings.
4762
4763 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4764 option for pcretest, so I've replaced it by a simple function that does just
4765 that job.
4766
4767 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4768 "pcregrep -".
4769
4770 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4771 Java package. This provides some syntactic sugar for simple cases of what my
4772 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4773 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4774 item, you can use this simplified notation. Note that only makes sense with
4775 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4776 greediness, whatever the setting of the PCRE_UNGREEDY option.
4777
4778 13. A change of greediness default within a pattern was not taking effect at
4779 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4780 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4781 was abstracted outside.
4782
4783 14. PCRE now supports the \G assertion. It is true when the current matching
4784 position is at the start point of the match. This differs from \A when the
4785 starting offset is non-zero. Used with the /g option of pcretest (or similar
4786 code), it works in the same way as it does for Perl's /g option. If all
4787 alternatives of a regex begin with \G, the expression is anchored to the start
4788 match position, and the "anchored" flag is set in the compiled expression.
4789
4790 15. Some bugs concerning the handling of certain option changes within patterns
4791 have been fixed. These applied to options other than (?ims). For example,
4792 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4793 been the other way round. Some of this was related to change 7 above.
4794
4795 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4796 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4797 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4798 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4799
4800 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4801 mode. Unlike ".", it always matches newline, whatever the setting of
4802 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4803 assertions. Perl allows it, but it doesn't (in general) work because it can't
4804 calculate the length of the lookbehind. At least, that's the case for Perl
4805 5.8.0 - I've been told they are going to document that it doesn't work in
4806 future.
4807
4808 18. Added an error diagnosis for escapes that PCRE does not support: these are
4809 \L, \l, \N, \P, \p, \U, \u, and \X.
4810
4811 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4812 reading past the end of the pattern in cases such as /[abcd/.
4813
4814 20. PCRE was getting more memory than necessary for patterns with classes that
4815 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4816
4817 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4818 compiling PCRE for use with Virtual Pascal.
4819
4820 22. Small fix to the Makefile to make it work properly if the build is done
4821 outside the source tree.
4822
4823 23. Added a new extension: a condition to go with recursion. If a conditional
4824 subpattern starts with (?(R) the "true" branch is used if recursion has
4825 happened, whereas the "false" branch is used only at the top level.
4826
4827 24. When there was a very long string of literal characters (over 255 bytes
4828 without UTF support, over 250 bytes with UTF support), the computation of how
4829 much memory was required could be incorrect, leading to segfaults or other
4830 strange effects.
4831
4832 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4833 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4834 there was a subsequent back reference to those brackets. This meant that, for
4835 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4836 possible to check for precisely this case. All we can do is abandon the
4837 optimization if .* occurs inside capturing brackets when there are any back
4838 references whatsoever. (See below for a better fix that came later.)
4839
4840 26. The handling of the optimization for finding the first character of a
4841 non-anchored pattern, and for finding a character that is required later in the
4842 match were failing in some cases. This didn't break the matching; it just
4843 failed to optimize when it could. The way this is done has been re-implemented.
4844
4845 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4846
4847 28. Added a new feature that provides some of the functionality that Perl
4848 provides with (?{...}). The facility is termed a "callout". The way it is done
4849 in PCRE is for the caller to provide an optional function, by setting
4850 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4851 global variable. By default it is unset, which disables all calling out. To get
4852 the function called, the regex must include (?C) at appropriate points. This
4853 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4854 This provides a means of identifying different callout points. When PCRE
4855 reaches such a point in the regex, if pcre_callout has been set, the external
4856 function is called. It is provided with data in a structure called
4857 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4858 matching continues; if it returns a non-zero value, the match at the current
4859 point fails. However, backtracking will occur if possible. [This was changed
4860 later and other features added - see item 49 below.]
4861
4862 29. pcretest is upgraded to test the callout functionality. It provides a
4863 callout function that displays information. By default, it shows the start of
4864 the match and the current position in the text. There are some new data escapes
4865 to vary what happens:
4866
4867 \C+ in addition, show current contents of captured substrings
4868 \C- do not supply a callout function
4869 \C!n return 1 when callout number n is reached
4870 \C!n!m return 1 when callout number n is reached for the mth time
4871
4872 30. If pcregrep was called with the -l option and just a single file name, it
4873 output "<stdin>" if a match was found, instead of the file name.
4874
4875 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4876 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4877 pcre_exec(). This saves a malloc/free per call. The default value of
4878 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4879 when configuring.
4880
4881 32. The default maximum size of a compiled pattern is 64K. There have been a
4882 few cases of people hitting this limit. The code now uses macros to handle the
4883 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4884 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4885 configuring. Tests 2 and 5 work only with 2-byte links because they output
4886 debugging information about compiled patterns.
4887
4888 33. Internal code re-arrangements:
4889
4890 (a) Moved the debugging function for printing out a compiled regex into
4891 its own source file (printint.c) and used #include to pull it into
4892 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4893 separate copies.
4894
4895 (b) Defined the list of op-code names for debugging as a macro in
4896 internal.h so that it is next to the definition of the opcodes.
4897
4898 (c) Defined a table of op-code lengths for simpler skipping along compiled
4899 code. This is again a macro in internal.h so that it is next to the
4900 definition of the opcodes.
4901
4902 34. Added support for recursive calls to individual subpatterns, along the
4903 lines of Robin Houston's patch (but implemented somewhat differently).
4904
4905 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4906 allow it to read and process whole directories in Win32. This code was
4907 contributed by Lionel Fourquaux; it has not been tested by me.
4908
4909 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4910 used to name a group. Names consist of alphanumerics and underscores, and must
4911 be unique. Back references use the syntax (?P=name) and recursive calls use
4912 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4913 numbers. The function pcre_fullinfo() can be used after compilation to extract
4914 a name/number map. There are three relevant calls:
4915
4916 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4917 PCRE_INFO_NAMECOUNT yields the number of entries
4918 PCRE_INFO_NAMETABLE yields a pointer to the map.
4919
4920 The map is a vector of fixed-size entries. The size of each entry depends on
4921 the length of the longest name used. The first two bytes of each entry are the
4922 group number, most significant byte first. There follows the corresponding
4923 name, zero terminated. The names are in alphabetical order.
4924
4925 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4926 case instead of 255. Making it the same both with and without UTF-8 support
4927 means that the same test output works with both.
4928
4929 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4930 calling malloc() with a zero argument.
4931
4932 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4933 optimization. I've improved things by keeping a bitmap of backreferences with
4934 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4935 fact referenced, the optimization can be applied. It is unlikely that a
4936 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4937 the match to follow \n) will appear inside brackets with a number greater than
4938 31, but if it does, any back reference > 31 suppresses the optimization.
4939
4940 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4941 of disabling numbered capturing parentheses. Any opening parenthesis that is
4942 not followed by ? behaves as if it were followed by ?: but named parentheses
4943 can still be used for capturing (and they will acquire numbers in the usual
4944 way).
4945
4946 41. Redesigned the return codes from the match() function into yes/no/error so
4947 that errors can be passed back from deep inside the nested calls. A malloc
4948 failure while inside a recursive subpattern call now causes the
4949 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4950
4951 42. It is now possible to set a limit on the number of times the match()
4952 function is called in a call to pcre_exec(). This facility makes it possible to
4953 limit the amount of recursion and backtracking, though not in a directly
4954 obvious way, because the match() function is used in a number of different
4955 circumstances. The count starts from zero for each position in the subject
4956 string (for non-anchored patterns). The default limit is, for compatibility, a
4957 large number, namely 10 000 000. You can change this in two ways:
4958
4959 (a) When configuring PCRE before making, you can use --with-match-limit=n
4960 to set a default value for the compiled library.
4961
4962 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4963 a different value is set. See 45 below.
4964
4965 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4966
4967 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4968 of things that can be changed at compile time. The first argument specifies
4969 what is wanted and the second points to where the information is to be placed.
4970 The current list of available information is:
4971
4972 PCRE_CONFIG_UTF8
4973
4974 The output is an integer that is set to one if UTF-8 support is available;
4975 otherwise it is set to zero.
4976
4977 PCRE_CONFIG_NEWLINE
4978
4979 The output is an integer that it set to the value of the code that is used for
4980 newline. It is either LF (10) or CR (13).
4981
4982 PCRE_CONFIG_LINK_SIZE
4983
4984 The output is an integer that contains the number of bytes used for internal
4985 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4986
4987 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4988
4989 The output is an integer that contains the threshold above which the POSIX
4990 interface uses malloc() for output vectors. See item 31 above.
4991
4992 PCRE_CONFIG_MATCH_LIMIT
4993
4994 The output is an unsigned integer that contains the default limit of the number
4995 of match() calls in a pcre_exec() execution. See 42 above.
4996
4997 44. pcretest has been upgraded by the addition of the -C option. This causes it
4998 to extract all the available output from the new pcre_config() function, and to
4999 output it. The program then exits immediately.
5000
5001 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5002 order to support additional features. One way would have been to define
5003 pcre_exec2() (for example) with extra arguments, but this would not have been
5004 extensible, and would also have required all calls to the original function to
5005 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5006 is used for passing in "extra" data from pcre_study().
5007
5008 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5009 contains the following fields:
5010
5011 flags a bitmap indicating which of the following fields are set
5012 study_data opaque data from pcre_study()
5013 match_limit a way of specifying a limit on match() calls for a specific
5014 call to pcre_exec()
5015 callout_data data for callouts (see 49 below)
5016
5017 The flag bits are also defined in pcre.h, and are
5018
5019 PCRE_EXTRA_STUDY_DATA
5020 PCRE_EXTRA_MATCH_LIMIT
5021 PCRE_EXTRA_CALLOUT_DATA
5022
5023 The pcre_study() function now returns one of these new pcre_extra blocks, with
5024 the actual study data pointed to by the study_data field, and the
5025 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5026 before. That is, this change is entirely upwards-compatible and requires no
5027 change to existing code.
5028
5029 If you want to pass in additional data to pcre_exec(), you can either place it
5030 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5031 block.
5032
5033 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5034 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5035 times with different match limits, until it finds the minimum value needed for
5036 pcre_exec() to complete. The value is then output. This can be instructive; for
5037 most simple matches the number is quite small, but for pathological cases it
5038 gets very large very quickly.
5039
5040 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5041 returns the size of the data block pointed to by the study_data field in a
5042 pcre_extra block, that is, the value that was passed as the argument to
5043 pcre_malloc() when PCRE was getting memory in which to place the information
5044 created by pcre_study(). The fourth argument should point to a size_t variable.
5045 pcretest has been extended so that this information is shown after a successful
5046 pcre_study() call when information about the compiled regex is being displayed.
5047
5048 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5049 because what follows is always an absolute path. (Later: it turns out that this
5050 is more than cosmetic for MinGW, because it doesn't like empty path
5051 components.)
5052
5053 49. Some changes have been made to the callout feature (see 28 above):
5054
5055 (i) A callout function now has three choices for what it returns:
5056
5057 0 => success, carry on matching
5058 > 0 => failure at this point, but backtrack if possible
5059 < 0 => serious error, return this value from pcre_exec()
5060
5061 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5062 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5063 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5064 use by callout functions. It will never be used by PCRE itself.
5065
5066 (ii) The pcre_extra structure (see 45 above) has a void * field called
5067 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5068 pcre_callout_block structure has a field of the same name. The contents of
5069 the field passed in the pcre_extra structure are passed to the callout
5070 function in the corresponding field in the callout block. This makes it
5071 easier to use the same callout-containing regex from multiple threads. For
5072 testing, the pcretest program has a new data escape
5073
5074 \C*n pass the number n (may be negative) as callout_data
5075
5076 If the callout function in pcretest receives a non-zero value as
5077 callout_data, it returns that value.
5078
5079 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5080 there were some redundant $(CFLAGS) in commands that are now specified as
5081 $(LINK), which already includes $(CFLAGS).
5082
5083 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5084 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5085 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5086 one-byte characters throughout. Note that case-insensitive matching applies
5087 only to characters whose values are less than 256. PCRE doesn't support the
5088 notion of cases for higher-valued characters.
5089
5090 (i) A character class whose characters are all within 0-255 is handled as
5091 a bit map, and the map is inverted for negative classes. Previously, a
5092 character > 255 always failed to match such a class; however it should
5093 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5094
5095 (ii) A negated character class with a single character < 255 is coded as
5096 "not this character" (OP_NOT). This wasn't working properly when the test
5097 character was multibyte, either singly or repeated.
5098
5099 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5100 mode, for example: \x{100}{2,3}.
5101
5102 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5103 singly or repeated) now correctly test multibyte characters. However,
5104 PCRE doesn't recognize any characters with values greater than 255 as
5105 digits, spaces, or word characters. Such characters always match \D, \S,
5106 and \W, and never match \d, \s, or \w.
5107
5108 (v) Classes may now contain characters and character ranges with values
5109 greater than 255. For example: [ab\x{100}-\x{400}].
5110
5111 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5112 PCRE in UTF-8 mode.
5113
5114 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5115 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5116 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5117 value.)
5118
5119 53. The single man page has become too large. I have therefore split it up into
5120 a number of separate man pages. These also give rise to individual HTML pages;
5121 these are now put in a separate directory, and there is an index.html page that
5122 lists them all. Some hyperlinking between the pages has been installed.
5123
5124 54. Added convenience functions for handling named capturing parentheses.
5125
5126 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5127 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5128 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5129 are faulted.
5130
5131 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5132 calling configure. These values are used when compiling the dftables.c program
5133 which is run to generate the source of the default character tables. They
5134 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5135 you will need to set these values.
5136
5137 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5138
5139
5140 Version 3.9 02-Jan-02
5141 ---------------------
5142
5143 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5144
5145 2. If --disable-static was given, the building process failed when trying to
5146 build pcretest and pcregrep. (For some reason it was using libtool to compile
5147 them, which is not right, as they aren't part of the library.)
5148
5149
5150 Version 3.8 18-Dec-01
5151 ---------------------
5152
5153 1. The experimental UTF-8 code was completely screwed up. It was packing the
5154 bytes in the wrong order. How dumb can you get?
5155
5156
5157 Version 3.7 29-Oct-01
5158 ---------------------
5159
5160 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5161 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5162 this didn't happen under Solaris 8, where I normally test things.
5163
5164 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5165 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5166 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5167 this fix an hour or so after the initial 3.7 release.)
5168
5169
5170 Version 3.6 23-Oct-01
5171 ---------------------
5172
5173 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5174 offsets passed as NULL with zero offset count.
5175
5176 2. The config.guess and config.sub files had not been updated when I moved to
5177 the latest autoconf.
5178
5179
5180 Version 3.5 15-Aug-01
5181 ---------------------
5182
5183 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5184 had been forgotten.
5185
5186 2. By using declared but undefined structures, we can avoid using "void"
5187 definitions in pcre.h while keeping the internal definitions of the structures
5188 private.
5189
5190 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5191 user point of view, this means that both static and shared libraries are built
5192 by default, but this can be individually controlled. More of the work of
5193 handling this static/shared cases is now inside libtool instead of PCRE's make
5194 file.
5195
5196 4. The pcretest utility is now installed along with pcregrep because it is
5197 useful for users (to test regexs) and by doing this, it automatically gets
5198 relinked by libtool. The documentation has been turned into a man page, so
5199 there are now .1, .txt, and .html versions in /doc.
5200
5201 5. Upgrades to pcregrep:
5202 (i) Added long-form option names like gnu grep.
5203 (ii) Added --help to list all options with an explanatory phrase.
5204 (iii) Added -r, --recursive to recurse into sub-directories.
5205 (iv) Added -f, --file to read patterns from a file.
5206
5207 6. pcre_exec() was referring to its "code" argument before testing that
5208 argument for NULL (and giving an error if it was NULL).
5209
5210 7. Upgraded Makefile.in to allow for compiling in a different directory from
5211 the source directory.
5212
5213 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5214 options bits, the pointer it was passed was to an int instead of to an unsigned
5215 long int. This mattered only on 64-bit systems.
5216
5217 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5218 generated) instead of pcre.in, which it its source. Also made the same change
5219 in several of the .c files.
5220
5221 10. A new release of gcc defines printf() as a macro, which broke pcretest
5222 because it had an ifdef in the middle of a string argument for printf(). Fixed
5223 by using separate calls to printf().
5224
5225 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5226 script, to force use of CR or LF instead of \n in the source. On non-Unix
5227 systems, the value can be set in config.h.
5228
5229 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5230 absolute limit. Changed the text of the error message to make this clear, and
5231 likewise updated the man page.
5232
5233 13. The limit of 99 on the number of capturing subpatterns has been removed.
5234 The new limit is 65535, which I hope will not be a "real" limit.
5235
5236
5237 Version 3.4 22-Aug-00
5238 ---------------------
5239
5240 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5241
5242 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5243
5244
5245 Version 3.3 01-Aug-00
5246 ---------------------
5247
5248 1. If an octal character was given, but the value was greater than \377, it
5249 was not getting masked to the least significant bits, as documented. This could
5250 lead to crashes in some systems.
5251
5252 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5253 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5254
5255 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5256 These just pass their arguments on to (pcre_free)(), but they are provided
5257 because some uses of PCRE bind it to non-C systems that can call its functions,
5258 but cannot call free() or pcre_free() directly.
5259
5260 4. Add "make test" as a synonym for "make check". Corrected some comments in
5261 the Makefile.
5262
5263 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5264 Makefile.
5265
5266 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5267 command called pgrep for grepping around the active processes.
5268
5269 7. Added the beginnings of support for UTF-8 character strings.
5270
5271 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5272 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5273 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5274 out for the ar command.)
5275
5276
5277 Version 3.2 12-May-00
5278 ---------------------
5279
5280 This is purely a bug fixing release.
5281
5282 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5283 of ZA. This was just one example of several cases that could provoke this bug,
5284 which was introduced by change 9 of version 2.00. The code for breaking
5285 infinite loops after an iteration that matches an empty string was't working
5286 correctly.
5287
5288 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5289 when matched against abbab (for example). After matching an empty string, it
5290 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5291 caused it to match further down the string than it should.
5292
5293 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5294 was there because it doesn't seem to be needed, and it causes trouble on some
5295 systems, as it is not a Standard C header. It has been removed.
5296
5297 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5298 were reported on the Macintosh. The changes were from
5299
5300 while ((c = *(++ptr)) != 0 && c != '\n');
5301 to
5302 while ((c = *(++ptr)) != 0 && c != '\n') ;
5303
5304 Totally extraordinary, but if that's what it takes...
5305
5306 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5307 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5308 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5309 assumes the way PCRE uses memmove() (always moving upwards).
5310
5311 6. PCRE is being used in one environment where strchr() is not available. There
5312 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5313 faster code anyway.
5314
5315
5316 Version 3.1 09-Feb-00
5317 ---------------------
5318
5319 The only change in this release is the fixing of some bugs in Makefile.in for
5320 the "install" target:
5321
5322 (1) It was failing to install pcreposix.h.
5323
5324 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5325
5326
5327 Version 3.0 01-Feb-00
5328 ---------------------
5329
5330 1. Add support for the /+ modifier to perltest (to output $` like it does in
5331 pcretest).
5332
5333 2. Add support for the /g modifier to perltest.
5334
5335 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5336 matches null strings.
5337
5338 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5339 pattern. Perl treats empty patterns specially - it reuses the most recent
5340 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5341 effect.
5342
5343 5. The POSIX interface was broken in that it was just handing over the POSIX
5344 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5345 required a bigger vector, with some working space on the end. This means that
5346 the POSIX wrapper now has to get and free some memory, and copy the results.
5347
5348 6. Added some simple autoconf support, placing the test data and the
5349 documentation in separate directories, re-organizing some of the
5350 information files, and making it build pcre-config (a GNU standard). Also added
5351 libtool support for building PCRE as a shared library, which is now the
5352 default.
5353
5354 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5355 09 are not valid octal constants. Single digits will be used for minor values
5356 less than 10.
5357
5358 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5359 existing programs that set these in the POSIX interface can use PCRE without
5360 modification.
5361
5362 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5363 return all that pcre_info() returns, plus additional data. The pcre_info()
5364 function is retained for compatibility, but is considered to be obsolete.
5365
5366 10. Added experimental recursion feature (?R) to handle one common case that
5367 Perl 5.6 will be able to do with (?p{...}).
5368
5369 11. Added support for POSIX character classes like [:alpha:], which Perl is
5370 adopting.
5371
5372
5373 Version 2.08 31-Aug-99
5374 ----------------------
5375
5376 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5377 trying to match at the startoffset position, but instead was moving forward to
5378 the next newline as if a previous match had failed.
5379
5380 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5381 and could get into a loop if a null string was matched other than at the start
5382 of the subject.
5383
5384 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5385 be distinguished at compile time, and for completeness also added PCRE_DATE.
5386
5387 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5388 in GnuWin32 environments.
5389
5390
5391 Version 2.07 29-Jul-99
5392 ----------------------
5393
5394 1. The documentation is now supplied in plain text form and HTML as well as in
5395 the form of man page sources.
5396
5397 2. C++ compilers don't like assigning (void *) values to other pointer types.
5398 In particular this affects malloc(). Although there is no problem in Standard
5399 C, I've put in casts to keep C++ compilers happy.
5400
5401 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5402 should be (const char *).
5403
5404 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5405 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5406 However, I haven't made this a standard facility. The documentation doesn't
5407 mention it, and the Makefile doesn't support it.
5408
5409 5. The Makefile now contains an "install" target, with editable destinations at
5410 the top of the file. The pcretest program is not installed.
5411
5412 6. pgrep -V now gives the PCRE version number and date.
5413
5414 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5415 causing the entire string to be ignored, instead of just the last character.
5416
5417 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5418 non-matching string, it can take a very, very long time, even for strings of
5419 quite modest length, because of the nested recursion. PCRE now does better in
5420 some of these cases. It does this by remembering the last required literal
5421 character in the pattern, and pre-searching the subject to ensure it is present
5422 before running the real match. In other words, it applies a heuristic to detect
5423 some types of certain failure quickly, and in the above example, if presented
5424 with a string that has no trailing " it gives "no match" very quickly.
5425
5426 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5427 other alternatives are tried instead.
5428
5429
5430 Version 2.06 09-Jun-99
5431 ----------------------
5432
5433 1. Change pcretest's output for amount of store used to show just the code
5434 space, because the remainder (the data block) varies in size between 32-bit and
5435 64-bit systems.
5436
5437 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5438 start matching at. This allows lookbehinds to work when searching for multiple
5439 occurrences in a string.
5440
5441 3. Added additional options to pcretest for testing multiple occurrences:
5442
5443 /+ outputs the rest of the string that follows a match
5444 /g loops for multiple occurrences, using the new startoffset argument
5445 /G loops for multiple occurrences by passing an incremented pointer
5446
5447 4. PCRE wasn't doing the "first character" optimization for patterns starting
5448 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5449 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5450 the letter 'x'. On long subject strings, this gives a significant speed-up.
5451
5452
5453 Version 2.05 21-Apr-99
5454 ----------------------
5455
5456 1. Changed the type of magic_number from int to long int so that it works
5457 properly on 16-bit systems.
5458
5459 2. Fixed a bug which caused patterns starting with .* not to work correctly
5460 when the subject string contained newline characters. PCRE was assuming
5461 anchoring for such patterns in all cases, which is not correct because .* will
5462 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5463 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5464 must be retried after every newline in the subject.
5465
5466
5467 Version 2.04 18-Feb-99
5468 ----------------------
5469
5470 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5471 computation of the store needed to hold the pattern was incorrect (too large).
5472 If such patterns were nested a few deep, this could multiply and become a real
5473 problem.
5474
5475 2. Added /M option to pcretest to show the memory requirement of a specific
5476 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5477
5478 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5479 compiled in such a way that the backtracking after subsequent failure was
5480 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5481 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5482
5483
5484 Version 2.03 02-Feb-99
5485 ----------------------
5486
5487 1. Fixed typo and small mistake in man page.
5488
5489 2. Added 4th condition (GPL supersedes if conflict) and created separate
5490 LICENCE file containing the conditions.
5491
5492 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5493 Perl, that is the internal \ allows the delimiter to be included in the
5494 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5495 the final delimiter, add \ to the end of the pattern (to test the error).
5496
5497 4. Added the convenience functions for extracting substrings after a successful
5498 match. Updated pcretest to make it able to test these functions.
5499
5500
5501 Version 2.02 14-Jan-99
5502 ----------------------
5503
5504 1. Initialized the working variables associated with each extraction so that
5505 their saving and restoring doesn't refer to uninitialized store.
5506
5507 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5508 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5509 fix the problem.
5510
5511 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5512 calls, and wasn't printing the correct value for compiling calls. Increased the
5513 default value of LOOPREPEAT, and the number of significant figures in the
5514 times.
5515
5516 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5517
5518 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5519 a building problem on Windows NT with a FAT file system.
5520
5521
5522 Version 2.01 21-Oct-98
5523 ----------------------
5524
5525 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5526 to character tables built by pcre_maketables() in the current locale. If NULL
5527 is passed, the default tables are used.
5528
5529
5530 Version 2.00 24-Sep-98
5531 ----------------------
5532
5533 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5534 it any more.
5535
5536 2. Allow quantification of (?>) groups, and make it work correctly.
5537
5538 3. The first character computation wasn't working for (?>) groups.
5539
5540 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5541 end of the subject) and add 5.005's \z, which really does match only at the
5542 very end of the subject.
5543
5544 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5545
5546 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5547 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5548 localized options. All options to pcre_study() were also removed.
5549
5550 7. Add other new features from 5.005:
5551
5552 $(?<= positive lookbehind
5553 $(?<! negative lookbehind
5554 (?imsx-imsx) added the unsetting capability
5555 such a setting is global if at outer level; local otherwise
5556 (?imsx-imsx:) non-capturing groups with option setting
5557 (?(cond)re|re) conditional pattern matching
5558
5559 A backreference to itself in a repeated group matches the previous
5560 captured string.
5561
5562 8. General tidying up of studying (both automatic and via "study")
5563 consequential on the addition of new assertions.
5564
5565 9. As in 5.005, unlimited repeated groups that could match an empty substring
5566 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5567 runtime if any iteration does actually match an empty substring.
5568
5569 10. Include the RunTest script in the distribution.
5570
5571 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5572 discrepancies, some of which were old and were also with respect to 5.004. They
5573 have now been fixed.
5574
5575
5576 Version 1.09 28-Apr-98
5577 ----------------------
5578
5579 1. A negated single character class followed by a quantifier with a minimum
5580 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5581 program crashes, or just wrong answers. This did not apply to negated classes
5582 containing more than one character, or to minima other than one.
5583
5584
5585 Version 1.08 27-Mar-98
5586 ----------------------
5587
5588 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5589
5590 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5591 latter must appear before anything that relies on it in the pattern.
5592
5593
5594 Version 1.07 16-Feb-98
5595 ----------------------
5596
5597 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5598 repeat of a potentially empty string).
5599
5600
5601 Version 1.06 23-Jan-98
5602 ----------------------
5603
5604 1. Added Markus Oberhumer's little patches for C++.
5605
5606 2. Literal strings longer than 255 characters were broken.
5607
5608
5609 Version 1.05 23-Dec-97
5610 ----------------------
5611
5612 1. Negated character classes containing more than one character were failing if
5613 PCRE_CASELESS was set at run time.
5614
5615
5616 Version 1.04 19-Dec-97
5617 ----------------------
5618
5619 1. Corrected the man page, where some "const" qualifiers had been omitted.
5620
5621 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5622 input syntax.
5623
5624 3. Fixed memory leak which occurred when a regex with back references was
5625 matched with an offsets vector that wasn't big enough. The temporary memory
5626 that is used in this case wasn't being freed if the match failed.
5627
5628 4. Tidied pcretest to ensure it frees memory that it gets.
5629
5630 5. Temporary memory was being obtained in the case where the passed offsets
5631 vector was exactly big enough.
5632
5633 6. Corrected definition of offsetof() from change 5 below.
5634
5635 7. I had screwed up change 6 below and broken the rules for the use of
5636 setjmp(). Now fixed.
5637
5638
5639 Version 1.03 18-Dec-97
5640 ----------------------
5641
5642 1. A erroneous regex with a missing opening parenthesis was correctly
5643 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5644 on some systems.
5645
5646 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5647 it was reported that one broken compiler failed on the former because "code" is
5648 also an independent variable.
5649
5650 3. The erroneous regex a[]b caused an array overrun reference.
5651
5652 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5653 fail on data ending with that character. (It was going on too far, and checking
5654 the next character, typically a binary zero.) This was specific to the
5655 optimized code for single-character negative classes.
5656
5657 5. Added a contributed patch from the TIN world which does the following:
5658
5659 + Add an undef for memmove, in case the the system defines a macro for it.
5660
5661 + Add a definition of offsetof(), in case there isn't one. (I don't know
5662 the reason behind this - offsetof() is part of the ANSI standard - but
5663 it does no harm).
5664
5665 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5666 most of the places where whitespace preceded '#'. I have given up and
5667 allowed the remaining 2 cases to be at the margin.
5668
5669 + Rename some variables in pcre to eliminate shadowing. This seems very
5670 pedantic, but does no harm, of course.
5671
5672 6. Moved the call to setjmp() into its own function, to get rid of warnings
5673 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5674
5675 7. Constructs such as \d{8,} were compiling into the equivalent of
5676 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5677 outcome, but in this particular case used more store than had been allocated,
5678 which caused the bug to be discovered because it threw up an internal error.
5679
5680 8. The debugging code in both pcre and pcretest for outputting the compiled
5681 form of a regex was going wrong in the case of back references followed by
5682 curly-bracketed repeats.
5683
5684
5685 Version 1.02 12-Dec-97
5686 ----------------------
5687
5688 1. Typos in pcre.3 and comments in the source fixed.
5689
5690 2. Applied a contributed patch to get rid of places where it used to remove
5691 'const' from variables, and fixed some signed/unsigned and uninitialized
5692 variable warnings.
5693
5694 3. Added the "runtest" target to Makefile.
5695
5696 4. Set default compiler flag to -O2 rather than just -O.
5697
5698
5699 Version 1.01 19-Nov-97
5700 ----------------------
5701
5702 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5703 like /([ab]*)*/, that is, for classes with more than one character in them.
5704
5705 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5706 as /((?>a*))*/ (a PCRE_EXTRA facility).
5707
5708
5709 Version 1.00 18-Nov-97
5710 ----------------------
5711
5712 1. Added compile-time macros to support systems such as SunOS4 which don't have
5713 memmove() or strerror() but have other things that can be used instead.
5714
5715 2. Arranged that "make clean" removes the executables.
5716
5717
5718 Version 0.99 27-Oct-97
5719 ----------------------
5720
5721 1. Fixed bug in code for optimizing classes with only one character. It was
5722 initializing a 32-byte map regardless, which could cause it to run off the end
5723 of the memory it had got.
5724
5725 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5726
5727
5728 Version 0.98 22-Oct-97
5729 ----------------------
5730
5731 1. Fixed bug in code for handling temporary memory usage when there are more
5732 back references than supplied space in the ovector. This could cause segfaults.
5733
5734
5735 Version 0.97 21-Oct-97
5736 ----------------------
5737
5738 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5739
5740 2. Optimized negated single characters not to use a bit map.
5741
5742 3. Brought error texts together as macro definitions; clarified some of them;
5743 fixed one that was wrong - it said "range out of order" when it meant "invalid
5744 escape sequence".
5745
5746 4. Changed some char * arguments to const char *.
5747
5748 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5749
5750 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5751 pcretest.
5752
5753
5754 Version 0.96 16-Oct-97
5755 ----------------------
5756
5757 1. Added a simple "pgrep" utility to the distribution.
5758
5759 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5760 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5761 where "ddd" means "one or more decimal digits".
5762
5763 3. Fixed serious bug. If a pattern had a back reference, but the call to
5764 pcre_exec() didn't supply a large enough ovector to record the related
5765 identifying subpattern, the match always failed. PCRE now remembers the number
5766 of the largest back reference, and gets some temporary memory in which to save
5767 the offsets during matching if necessary, in order to ensure that
5768 backreferences always work.
5769
5770 4. Increased the compatibility with Perl in a number of ways:
5771
5772 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5773 to request this handling. The option can be set at compile or exec time.
5774
5775 (b) $ matches before a terminating newline by default; an option
5776 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5777 mode). The option can be set at compile or exec time.
5778
5779 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5780 the same as Perl's. If the decimal number it represents is less than 10
5781 or there aren't that many previous left capturing parentheses, an octal
5782 escape is read. Inside a character class, it's always an octal escape,
5783 even if it is a single digit.
5784
5785 (d) An escaped but undefined alphabetic character is taken as a literal,
5786 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5787 escapes.
5788
5789 (e) {0} is now permitted. (The previous item is removed from the compiled
5790 pattern).
5791
5792 5. Changed all the names of code files so that the basic parts are no longer
5793 than 10 characters, and abolished the teeny "globals.c" file.
5794
5795 6. Changed the handling of character classes; they are now done with a 32-byte
5796 bit map always.
5797
5798 7. Added the -d and /D options to pcretest to make it possible to look at the
5799 internals of compilation without having to recompile pcre.
5800
5801
5802 Version 0.95 23-Sep-97
5803 ----------------------
5804
5805 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5806 \x20 at the start of a run of normal characters. These were being treated as
5807 real characters, instead of the source characters being re-checked.
5808
5809
5810 Version 0.94 18-Sep-97
5811 ----------------------
5812
5813 1. The functions are now thread-safe, with the caveat that the global variables
5814 containing pointers to malloc() and free() or alternative functions are the
5815 same for all threads.
5816
5817 2. Get pcre_study() to generate a bitmap of initial characters for non-
5818 anchored patterns when this is possible, and use it if passed to pcre_exec().
5819
5820
5821 Version 0.93 15-Sep-97
5822 ----------------------
5823
5824 1. /(b)|(:+)/ was computing an incorrect first character.
5825
5826 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5827 but not actually doing anything yet.
5828
5829 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5830 as Perl does (e.g. [-az] or [az-]).
5831
5832 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5833 all possible positions.
5834
5835 5. Split up into different modules to avoid including unneeded functions in a
5836 compiled binary. However, compile and exec are still in one module. The "study"
5837 function is split off.
5838
5839 6. The character tables are now in a separate module whose source is generated
5840 by an auxiliary program - but can then be edited by hand if required. There are
5841 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5842 toupper() in the code.
5843
5844 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5845 make them global. Abolish the function for setting them, as the caller can now
5846 set them directly.
5847
5848
5849 Version 0.92 11-Sep-97
5850 ----------------------
5851
5852 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5853 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5854
5855 2. Caseless matching was not working in character classes if the characters in
5856 the pattern were in upper case.
5857
5858 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5859
5860 4. Make PCRE_ANCHORED public and accept as a compile option.
5861
5862 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5863 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5864 pass them.
5865
5866 6. Give an error if bad option bits passed at compile or run time.
5867
5868 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5869 pcretest to cause it to pass that flag.
5870
5871 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5872 options, and the first character, if set.
5873
5874 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5875
5876
5877 Version 0.91 10-Sep-97
5878 ----------------------
5879
5880 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5881 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5882
5883 2. PCRE was looping on encountering an indefinitely repeated back reference to
5884 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5885 Perl does - treats the match as successful.
5886
5887 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5