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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1691 - (show annotations)
Mon Mar 27 16:00:16 2017 UTC (2 years, 10 months ago) by ph10
File size: 293249 byte(s)
Error occurred while calculating annotation data.
Fix typo (leading to possible buffer overflow in pcre_copy_substring()) in 
pcretest.
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 8.41
8 ------------
9
10 1. Fixed typo in CMakeLists.txt (wrong number of arguments for
11 PCRE_STATIC_RUNTIME (affects MSVC only).
12
13 2. Issue 1 for 8.40 below was not correctly fixed. If pcregrep in multiline
14 mode with --only-matching matched several lines, it restarted scanning at the
15 next line instead of moving on to the end of the matched string, which can be
16 several lines after the start.
17
18 3. Fix a missing else in the JIT compiler reported by 'idaifish'.
19
20 4. A (?# style comment is now ignored between a basic quantifier and a
21 following '+' or '?' (example: /X+(?#comment)?Y/.
22
23 5. Avoid use of a potentially overflowing buffer in pcregrep (patch by Petr
24 Pisar).
25
26 6. Fuzzers have reported issues in pcretest. These are NOT serious (it is,
27 after all, just a test program). However, to stop the reports, some easy ones
28 are fixed:
29
30 (a) Check for values < 256 when calling isprint() in pcretest.
31 (b) Give an error for too big a number after \O.
32
33 7. In the 32-bit library in non-UTF mode, an attempt to find a Unicode
34 property for a character with a code point greater than 0x10ffff (the Unicode
35 maximum) caused a crash.
36
37 8. The alternative matching function, pcre_dfa_exec() misbehaved if it
38 encountered a character class with a possessive repeat, for example [a-f]{3}+.
39
40 9. When pcretest called pcre_copy_substring() in 32-bit mode, it set the buffer
41 length incorrectly, which could result in buffer overflow.
42
43
44 Version 8.40 11-January-2017
45 ----------------------------
46
47 1. Using -o with -M in pcregrep could cause unnecessary repeated output when
48 the match extended over a line boundary.
49
50 2. Applied Chris Wilson's second patch (Bugzilla #1681) to CMakeLists.txt for
51 MSVC static compilation, putting the first patch under a new option.
52
53 3. Fix register overwite in JIT when SSE2 acceleration is enabled.
54
55 4. Ignore "show all captures" (/=) for DFA matching.
56
57 5. Fix JIT unaligned accesses on x86. Patch by Marc Mutz.
58
59 6. In any wide-character mode (8-bit UTF or any 16-bit or 32-bit mode),
60 without PCRE_UCP set, a negative character type such as \D in a positive
61 class should cause all characters greater than 255 to match, whatever else
62 is in the class. There was a bug that caused this not to happen if a
63 Unicode property item was added to such a class, for example [\D\P{Nd}] or
64 [\W\pL].
65
66 7. When pcretest was outputing information from a callout, the caret indicator
67 for the current position in the subject line was incorrect if it was after
68 an escape sequence for a character whose code point was greater than
69 \x{ff}.
70
71 8. A pattern such as (?<RA>abc)(?(R)xyz) was incorrectly compiled such that
72 the conditional was interpreted as a reference to capturing group 1 instead
73 of a test for recursion. Any group whose name began with R was
74 misinterpreted in this way. (The reference interpretation should only
75 happen if the group's name is precisely "R".)
76
77 9. A number of bugs have been mended relating to match start-up optimizations
78 when the first thing in a pattern is a positive lookahead. These all
79 applied only when PCRE_NO_START_OPTIMIZE was *not* set:
80
81 (a) A pattern such as (?=.*X)X$ was incorrectly optimized as if it needed
82 both an initial 'X' and a following 'X'.
83 (b) Some patterns starting with an assertion that started with .* were
84 incorrectly optimized as having to match at the start of the subject or
85 after a newline. There are cases where this is not true, for example,
86 (?=.*[A-Z])(?=.{8,16})(?!.*[\s]) matches after the start in lines that
87 start with spaces. Starting .* in an assertion is no longer taken as an
88 indication of matching at the start (or after a newline).
89
90
91 Version 8.39 14-June-2016
92 -------------------------
93
94 1. If PCRE_AUTO_CALLOUT was set on a pattern that had a (?# comment between
95 an item and its qualifier (for example, A(?#comment)?B) pcre_compile()
96 misbehaved. This bug was found by the LLVM fuzzer.
97
98 2. Similar to the above, if an isolated \E was present between an item and its
99 qualifier when PCRE_AUTO_CALLOUT was set, pcre_compile() misbehaved. This
100 bug was found by the LLVM fuzzer.
101
102 3. Further to 8.38/46, negated classes such as [^[:^ascii:]\d] were also not
103 working correctly in UCP mode.
104
105 4. The POSIX wrapper function regexec() crashed if the option REG_STARTEND
106 was set when the pmatch argument was NULL. It now returns REG_INVARG.
107
108 5. Allow for up to 32-bit numbers in the ordin() function in pcregrep.
109
110 6. An empty \Q\E sequence between an item and its qualifier caused
111 pcre_compile() to misbehave when auto callouts were enabled. This bug was
112 found by the LLVM fuzzer.
113
114 7. If a pattern that was compiled with PCRE_EXTENDED started with white
115 space or a #-type comment that was followed by (?-x), which turns off
116 PCRE_EXTENDED, and there was no subsequent (?x) to turn it on again,
117 pcre_compile() assumed that (?-x) applied to the whole pattern and
118 consequently mis-compiled it. This bug was found by the LLVM fuzzer.
119
120 8. A call of pcre_copy_named_substring() for a named substring whose number
121 was greater than the space in the ovector could cause a crash.
122
123 9. Yet another buffer overflow bug involved duplicate named groups with a
124 group that reset capture numbers (compare 8.38/7 below). Once again, I have
125 just allowed for more memory, even if not needed. (A proper fix is
126 implemented in PCRE2, but it involves a lot of refactoring.)
127
128 10. pcre_get_substring_list() crashed if the use of \K in a match caused the
129 start of the match to be earlier than the end.
130
131 11. Migrating appropriate PCRE2 JIT improvements to PCRE.
132
133 12. A pattern such as /(?<=((?C)0))/, which has a callout inside a lookbehind
134 assertion, caused pcretest to generate incorrect output, and also to read
135 uninitialized memory (detected by ASAN or valgrind).
136
137 13. A pattern that included (*ACCEPT) in the middle of a sufficiently deeply
138 nested set of parentheses of sufficient size caused an overflow of the
139 compiling workspace (which was diagnosed, but of course is not desirable).
140
141 14. And yet another buffer overflow bug involving duplicate named groups, this
142 time nested, with a nested back reference. Yet again, I have just allowed
143 for more memory, because anything more needs all the refactoring that has
144 been done for PCRE2. An example pattern that provoked this bug is:
145 /((?J)(?'R'(?'R'(?'R'(?'R'(?'R'(?|(\k'R'))))))))/ and the bug was
146 registered as CVE-2016-1283.
147
148 15. pcretest went into a loop if global matching was requested with an ovector
149 size less than 2. It now gives an error message. This bug was found by
150 afl-fuzz.
151
152 16. An invalid pattern fragment such as (?(?C)0 was not diagnosing an error
153 ("assertion expected") when (?(?C) was not followed by an opening
154 parenthesis.
155
156 17. Fixed typo ("&&" for "&") in pcre_study(). Fortunately, this could not
157 actually affect anything, by sheer luck.
158
159 18. Applied Chris Wilson's patch (Bugzilla #1681) to CMakeLists.txt for MSVC
160 static compilation.
161
162 19. Modified the RunTest script to incorporate a valgrind suppressions file so
163 that certain errors, provoked by the SSE2 instruction set when JIT is used,
164 are ignored.
165
166 20. A racing condition is fixed in JIT reported by Mozilla.
167
168 21. Minor code refactor to avoid "array subscript is below array bounds"
169 compiler warning.
170
171 22. Minor code refactor to avoid "left shift of negative number" warning.
172
173 23. Fix typo causing compile error when 16- or 32-bit JIT is compiled without
174 UCP support.
175
176 24. Refactor to avoid compiler warnings in pcrecpp.cc.
177
178 25. Refactor to fix a typo in pcre_jit_test.c
179
180 26. Patch to support compiling pcrecpp.cc with Intel compiler.
181
182
183 Version 8.38 23-November-2015
184 -----------------------------
185
186 1. If a group that contained a recursive back reference also contained a
187 forward reference subroutine call followed by a non-forward-reference
188 subroutine call, for example /.((?2)(?R)\1)()/, pcre_compile() failed to
189 compile correct code, leading to undefined behaviour or an internally
190 detected error. This bug was discovered by the LLVM fuzzer.
191
192 2. Quantification of certain items (e.g. atomic back references) could cause
193 incorrect code to be compiled when recursive forward references were
194 involved. For example, in this pattern: /(?1)()((((((\1++))\x85)+)|))/.
195 This bug was discovered by the LLVM fuzzer.
196
197 3. A repeated conditional group whose condition was a reference by name caused
198 a buffer overflow if there was more than one group with the given name.
199 This bug was discovered by the LLVM fuzzer.
200
201 4. A recursive back reference by name within a group that had the same name as
202 another group caused a buffer overflow. For example:
203 /(?J)(?'d'(?'d'\g{d}))/. This bug was discovered by the LLVM fuzzer.
204
205 5. A forward reference by name to a group whose number is the same as the
206 current group, for example in this pattern: /(?|(\k'Pm')|(?'Pm'))/, caused
207 a buffer overflow at compile time. This bug was discovered by the LLVM
208 fuzzer.
209
210 6. A lookbehind assertion within a set of mutually recursive subpatterns could
211 provoke a buffer overflow. This bug was discovered by the LLVM fuzzer.
212
213 7. Another buffer overflow bug involved duplicate named groups with a
214 reference between their definition, with a group that reset capture
215 numbers, for example: /(?J:(?|(?'R')(\k'R')|((?'R'))))/. This has been
216 fixed by always allowing for more memory, even if not needed. (A proper fix
217 is implemented in PCRE2, but it involves more refactoring.)
218
219 8. There was no check for integer overflow in subroutine calls such as (?123).
220
221 9. The table entry for \l in EBCDIC environments was incorrect, leading to its
222 being treated as a literal 'l' instead of causing an error.
223
224 10. There was a buffer overflow if pcre_exec() was called with an ovector of
225 size 1. This bug was found by american fuzzy lop.
226
227 11. If a non-capturing group containing a conditional group that could match
228 an empty string was repeated, it was not identified as matching an empty
229 string itself. For example: /^(?:(?(1)x|)+)+$()/.
230
231 12. In an EBCDIC environment, pcretest was mishandling the escape sequences
232 \a and \e in test subject lines.
233
234 13. In an EBCDIC environment, \a in a pattern was converted to the ASCII
235 instead of the EBCDIC value.
236
237 14. The handling of \c in an EBCDIC environment has been revised so that it is
238 now compatible with the specification in Perl's perlebcdic page.
239
240 15. The EBCDIC character 0x41 is a non-breaking space, equivalent to 0xa0 in
241 ASCII/Unicode. This has now been added to the list of characters that are
242 recognized as white space in EBCDIC.
243
244 16. When PCRE was compiled without UCP support, the use of \p and \P gave an
245 error (correctly) when used outside a class, but did not give an error
246 within a class.
247
248 17. \h within a class was incorrectly compiled in EBCDIC environments.
249
250 18. A pattern with an unmatched closing parenthesis that contained a backward
251 assertion which itself contained a forward reference caused buffer
252 overflow. And example pattern is: /(?=di(?<=(?1))|(?=(.))))/.
253
254 19. JIT should return with error when the compiled pattern requires more stack
255 space than the maximum.
256
257 20. A possessively repeated conditional group that could match an empty string,
258 for example, /(?(R))*+/, was incorrectly compiled.
259
260 21. Fix infinite recursion in the JIT compiler when certain patterns such as
261 /(?:|a|){100}x/ are analysed.
262
263 22. Some patterns with character classes involving [: and \\ were incorrectly
264 compiled and could cause reading from uninitialized memory or an incorrect
265 error diagnosis.
266
267 23. Pathological patterns containing many nested occurrences of [: caused
268 pcre_compile() to run for a very long time.
269
270 24. A conditional group with only one branch has an implicit empty alternative
271 branch and must therefore be treated as potentially matching an empty
272 string.
273
274 25. If (?R was followed by - or + incorrect behaviour happened instead of a
275 diagnostic.
276
277 26. Arrange to give up on finding the minimum matching length for overly
278 complex patterns.
279
280 27. Similar to (4) above: in a pattern with duplicated named groups and an
281 occurrence of (?| it is possible for an apparently non-recursive back
282 reference to become recursive if a later named group with the relevant
283 number is encountered. This could lead to a buffer overflow. Wen Guanxing
284 from Venustech ADLAB discovered this bug.
285
286 28. If pcregrep was given the -q option with -c or -l, or when handling a
287 binary file, it incorrectly wrote output to stdout.
288
289 29. The JIT compiler did not restore the control verb head in case of *THEN
290 control verbs. This issue was found by Karl Skomski with a custom LLVM
291 fuzzer.
292
293 30. Error messages for syntax errors following \g and \k were giving inaccurate
294 offsets in the pattern.
295
296 31. Added a check for integer overflow in conditions (?(<digits>) and
297 (?(R<digits>). This omission was discovered by Karl Skomski with the LLVM
298 fuzzer.
299
300 32. Handling recursive references such as (?2) when the reference is to a group
301 later in the pattern uses code that is very hacked about and error-prone.
302 It has been re-written for PCRE2. Here in PCRE1, a check has been added to
303 give an internal error if it is obvious that compiling has gone wrong.
304
305 33. The JIT compiler should not check repeats after a {0,1} repeat byte code.
306 This issue was found by Karl Skomski with a custom LLVM fuzzer.
307
308 34. The JIT compiler should restore the control chain for empty possessive
309 repeats. This issue was found by Karl Skomski with a custom LLVM fuzzer.
310
311 35. Match limit check added to JIT recursion. This issue was found by Karl
312 Skomski with a custom LLVM fuzzer.
313
314 36. Yet another case similar to 27 above has been circumvented by an
315 unconditional allocation of extra memory. This issue is fixed "properly" in
316 PCRE2 by refactoring the way references are handled. Wen Guanxing
317 from Venustech ADLAB discovered this bug.
318
319 37. Fix two assertion fails in JIT. These issues were found by Karl Skomski
320 with a custom LLVM fuzzer.
321
322 38. Fixed a corner case of range optimization in JIT.
323
324 39. An incorrect error "overran compiling workspace" was given if there were
325 exactly enough group forward references such that the last one extended
326 into the workspace safety margin. The next one would have expanded the
327 workspace. The test for overflow was not including the safety margin.
328
329 40. A match limit issue is fixed in JIT which was found by Karl Skomski
330 with a custom LLVM fuzzer.
331
332 41. Remove the use of /dev/null in testdata/testinput2, because it doesn't
333 work under Windows. (Why has it taken so long for anyone to notice?)
334
335 42. In a character class such as [\W\p{Any}] where both a negative-type escape
336 ("not a word character") and a property escape were present, the property
337 escape was being ignored.
338
339 43. Fix crash caused by very long (*MARK) or (*THEN) names.
340
341 44. A sequence such as [[:punct:]b] that is, a POSIX character class followed
342 by a single ASCII character in a class item, was incorrectly compiled in
343 UCP mode. The POSIX class got lost, but only if the single character
344 followed it.
345
346 45. [:punct:] in UCP mode was matching some characters in the range 128-255
347 that should not have been matched.
348
349 46. If [:^ascii:] or [:^xdigit:] or [:^cntrl:] are present in a non-negated
350 class, all characters with code points greater than 255 are in the class.
351 When a Unicode property was also in the class (if PCRE_UCP is set, escapes
352 such as \w are turned into Unicode properties), wide characters were not
353 correctly handled, and could fail to match.
354
355
356 Version 8.37 28-April-2015
357 --------------------------
358
359 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
360 for those parentheses to be closed with whatever has been captured so far.
361 However, it was failing to mark any other groups between the hightest
362 capture so far and the currrent group as "unset". Thus, the ovector for
363 those groups contained whatever was previously there. An example is the
364 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
365
366 2. If an assertion condition was quantified with a minimum of zero (an odd
367 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
368
369 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
370 unrecognized modifier, a crash could occur.
371
372 4. An attempt to do global matching in pcretest with a zero-length ovector
373 caused a crash.
374
375 5. Fixed a memory leak during matching that could occur for a subpattern
376 subroutine call (recursive or otherwise) if the number of captured groups
377 that had to be saved was greater than ten.
378
379 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
380 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
381 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
382
383 7. A UTF pattern containing a "not" match of a non-ASCII character and a
384 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
385
386 8. When a pattern is compiled, it remembers the highest back reference so that
387 when matching, if the ovector is too small, extra memory can be obtained to
388 use instead. A conditional subpattern whose condition is a check on a
389 capture having happened, such as, for example in the pattern
390 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
391 setting the highest backreference number. This mattered only if pcre_exec()
392 was called with an ovector that was too small to hold the capture, and there
393 was no other kind of back reference (a situation which is probably quite
394 rare). The effect of the bug was that the condition was always treated as
395 FALSE when the capture could not be consulted, leading to a incorrect
396 behaviour by pcre_exec(). This bug has been fixed.
397
398 9. A reference to a duplicated named group (either a back reference or a test
399 for being set in a conditional) that occurred in a part of the pattern where
400 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
401 to be incorrectly calculated, leading to overwriting.
402
403 10. A mutually recursive set of back references such as (\2)(\1) caused a
404 segfault at study time (while trying to find the minimum matching length).
405 The infinite loop is now broken (with the minimum length unset, that is,
406 zero).
407
408 11. If an assertion that was used as a condition was quantified with a minimum
409 of zero, matching went wrong. In particular, if the whole group had
410 unlimited repetition and could match an empty string, a segfault was
411 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
412 assertions to be quantified, but not if they are being used as conditions,
413 so the above pattern is faulted by Perl. PCRE has now been changed so that
414 it also rejects such patterns.
415
416 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
417 failed to allow the zero-repeat case if pcre2_exec() was called with an
418 ovector too small to capture the group.
419
420 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
421 Red Hat Product Security:
422
423 (a) A crash if /K and /F were both set with the option to save the compiled
424 pattern.
425
426 (b) Another crash if the option to print captured substrings in a callout
427 was combined with setting a null ovector, for example \O\C+ as a subject
428 string.
429
430 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
431 forward reference repeated a large (but limited) number of times within a
432 repeated outer group that has a zero minimum quantifier, caused incorrect
433 code to be compiled, leading to the error "internal error:
434 previously-checked referenced subpattern not found" when an incorrect
435 memory address was read. This bug was reported as "heap overflow",
436 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
437 CVE-2015-2325.
438
439 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
440 call within a group that also contained a recursive back reference caused
441 incorrect code to be compiled. This bug was reported as "heap overflow",
442 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
443 number CVE-2015-2326.
444
445 24. Computing the size of the JIT read-only data in advance has been a source
446 of various issues, and new ones are still appear unfortunately. To fix
447 existing and future issues, size computation is eliminated from the code,
448 and replaced by on-demand memory allocation.
449
450 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
451 adjacent to the end of the range, and the range contained characters with
452 more than one other case, caused incorrect behaviour when compiled in UTF
453 mode. In that example, the range a-j was left out of the class.
454
455 26. Fix JIT compilation of conditional blocks, which assertion
456 is converted to (*FAIL). E.g: /(?(?!))/.
457
458 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
459 discovered by the LLVM fuzzer.
460
461 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
462 when this assertion was used as a condition, for example (?(?!)a|b). In
463 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
464 error about an unsupported item.
465
466 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
467 possessification code could take exponential time to complete. A recursion
468 depth limit of 1000 has been imposed to limit the resources used by this
469 optimization.
470
471 30. A pattern such as /(*UTF)[\S\V\H]/, which contains a negated special class
472 such as \S in non-UCP mode, explicit wide characters (> 255) can be ignored
473 because \S ensures they are all in the class. The code for doing this was
474 interacting badly with the code for computing the amount of space needed to
475 compile the pattern, leading to a buffer overflow. This bug was discovered
476 by the LLVM fuzzer.
477
478 31. A pattern such as /((?2)+)((?1))/ which has mutual recursion nested inside
479 other kinds of group caused stack overflow at compile time. This bug was
480 discovered by the LLVM fuzzer.
481
482 32. A pattern such as /(?1)(?#?'){8}(a)/ which had a parenthesized comment
483 between a subroutine call and its quantifier was incorrectly compiled,
484 leading to buffer overflow or other errors. This bug was discovered by the
485 LLVM fuzzer.
486
487 33. The illegal pattern /(?(?<E>.*!.*)?)/ was not being diagnosed as missing an
488 assertion after (?(. The code was failing to check the character after
489 (?(?< for the ! or = that would indicate a lookbehind assertion. This bug
490 was discovered by the LLVM fuzzer.
491
492 34. A pattern such as /X((?2)()*+){2}+/ which has a possessive quantifier with
493 a fixed maximum following a group that contains a subroutine reference was
494 incorrectly compiled and could trigger buffer overflow. This bug was
495 discovered by the LLVM fuzzer.
496
497 35. A mutual recursion within a lookbehind assertion such as (?<=((?2))((?1)))
498 caused a stack overflow instead of the diagnosis of a non-fixed length
499 lookbehind assertion. This bug was discovered by the LLVM fuzzer.
500
501 36. The use of \K in a positive lookbehind assertion in a non-anchored pattern
502 (e.g. /(?<=\Ka)/) could make pcregrep loop.
503
504 37. There was a similar problem to 36 in pcretest for global matches.
505
506 38. If a greedy quantified \X was preceded by \C in UTF mode (e.g. \C\X*),
507 and a subsequent item in the pattern caused a non-match, backtracking over
508 the repeated \X did not stop, but carried on past the start of the subject,
509 causing reference to random memory and/or a segfault. There were also some
510 other cases where backtracking after \C could crash. This set of bugs was
511 discovered by the LLVM fuzzer.
512
513 39. The function for finding the minimum length of a matching string could take
514 a very long time if mutual recursion was present many times in a pattern,
515 for example, /((?2){73}(?2))((?1))/. A better mutual recursion detection
516 method has been implemented. This infelicity was discovered by the LLVM
517 fuzzer.
518
519 40. Static linking against the PCRE library using the pkg-config module was
520 failing on missing pthread symbols.
521
522
523 Version 8.36 26-September-2014
524 ------------------------------
525
526 1. Got rid of some compiler warnings in the C++ modules that were shown up by
527 -Wmissing-field-initializers and -Wunused-parameter.
528
529 2. The tests for quantifiers being too big (greater than 65535) were being
530 applied after reading the number, and stupidly assuming that integer
531 overflow would give a negative number. The tests are now applied as the
532 numbers are read.
533
534 3. Tidy code in pcre_exec.c where two branches that used to be different are
535 now the same.
536
537 4. The JIT compiler did not generate match limit checks for certain
538 bracketed expressions with quantifiers. This may lead to exponential
539 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
540 issue should be resolved now.
541
542 5. Fixed an issue, which occures when nested alternatives are optimized
543 with table jumps.
544
545 6. Inserted two casts and changed some ints to size_t in the light of some
546 reported 64-bit compiler warnings (Bugzilla 1477).
547
548 7. Fixed a bug concerned with zero-minimum possessive groups that could match
549 an empty string, which sometimes were behaving incorrectly in the
550 interpreter (though correctly in the JIT matcher). This pcretest input is
551 an example:
552
553 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
554 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
555
556 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
557 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
558 for an empty string was breaking the inner loop and carrying on at a lower
559 level, when possessive repeated groups should always return to a higher
560 level as they have no backtrack points in them. The empty string test now
561 occurs at the outer level.
562
563 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
564 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
565
566 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
567 doesn't).
568
569 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
570 to hold a pattern to be incorrectly computed (too small) when there were
571 named back references to duplicated names. This could cause "internal
572 error: code overflow" or "double free or corruption" or other memory
573 handling errors.
574
575 11. When named subpatterns had the same prefixes, back references could be
576 confused. For example, in this pattern:
577
578 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
579
580 the reference to 'Name' was incorrectly treated as a reference to a
581 duplicate name.
582
583 12. A pattern such as /^s?c/mi8 where the optional character has more than
584 one "other case" was incorrectly compiled such that it would only try to
585 match starting at "c".
586
587 13. When a pattern starting with \s was studied, VT was not included in the
588 list of possible starting characters; this should have been part of the
589 8.34/18 patch.
590
591 14. If a character class started [\Qx]... where x is any character, the class
592 was incorrectly terminated at the ].
593
594 15. If a pattern that started with a caseless match for a character with more
595 than one "other case" was studied, PCRE did not set up the starting code
596 unit bit map for the list of possible characters. Now it does. This is an
597 optimization improvement, not a bug fix.
598
599 16. The Unicode data tables have been updated to Unicode 7.0.0.
600
601 17. Fixed a number of memory leaks in pcregrep.
602
603 18. Avoid a compiler warning (from some compilers) for a function call with
604 a cast that removes "const" from an lvalue by using an intermediate
605 variable (to which the compiler does not object).
606
607 19. Incorrect code was compiled if a group that contained an internal recursive
608 back reference was optional (had quantifier with a minimum of zero). This
609 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
610 caused segmentation faults because of stack overflows at compile time.
611
612 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
613 group that is quantified with an indefinite repeat, caused a compile-time
614 loop which used up all the system stack and provoked a segmentation fault.
615 This was not the same bug as 19 above.
616
617 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
618 Patch by Mike Frysinger.
619
620
621 Version 8.35 04-April-2014
622 --------------------------
623
624 1. A new flag is set, when property checks are present in an XCLASS.
625 When this flag is not set, PCRE can perform certain optimizations
626 such as studying these XCLASS-es.
627
628 2. The auto-possessification of character sets were improved: a normal
629 and an extended character set can be compared now. Furthermore
630 the JIT compiler optimizes more character set checks.
631
632 3. Got rid of some compiler warnings for potentially uninitialized variables
633 that show up only when compiled with -O2.
634
635 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
636 of a match later then the end of the match. The pcretest program was not
637 handling the case sensibly - it was outputting from the start to the next
638 binary zero. It now reports this situation in a message, and outputs the
639 text from the end to the start.
640
641 5. Fast forward search is improved in JIT. Instead of the first three
642 characters, any three characters with fixed position can be searched.
643 Search order: first, last, middle.
644
645 6. Improve character range checks in JIT. Characters are read by an inprecise
646 function now, which returns with an unknown value if the character code is
647 above a certain threshold (e.g: 256). The only limitation is that the value
648 must be bigger than the threshold as well. This function is useful when
649 the characters above the threshold are handled in the same way.
650
651 7. The macros whose names start with RAWUCHAR are placeholders for a future
652 mode in which only the bottom 21 bits of 32-bit data items are used. To
653 make this more memorable for those maintaining the code, the names have
654 been changed to start with UCHAR21, and an extensive comment has been added
655 to their definition.
656
657 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
658 to the export list in Makefile.am (they were accidentally omitted from the
659 8.34 tarball).
660
661 9. The informational output from pcretest used the phrase "starting byte set"
662 which is inappropriate for the 16-bit and 32-bit libraries. As the output
663 for "first char" and "need char" really means "non-UTF-char", I've changed
664 "byte" to "char", and slightly reworded the output. The documentation about
665 these values has also been (I hope) clarified.
666
667 10. Another JIT related optimization: use table jumps for selecting the correct
668 backtracking path, when more than four alternatives are present inside a
669 bracket.
670
671 11. Empty match is not possible, when the minimum length is greater than zero,
672 and there is no \K in the pattern. JIT should avoid empty match checks in
673 such cases.
674
675 12. In a caseless character class with UCP support, when a character with more
676 than one alternative case was not the first character of a range, not all
677 the alternative cases were added to the class. For example, s and \x{17f}
678 are both alternative cases for S: the class [RST] was handled correctly,
679 but [R-T] was not.
680
681 13. The configure.ac file always checked for pthread support when JIT was
682 enabled. This is not used in Windows, so I have put this test inside a
683 check for the presence of windows.h (which was already tested for).
684
685 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
686 The algorithm provides a way to skip certain starting offsets, and usually
687 faster than linear prefix searches.
688
689 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
690 as for 'fr_FR' and 'french'. For some reason, however, it then used the
691 Windows-specific input and output files, which have 'french' screwed in.
692 So this could never have worked. One of the problems with locales is that
693 they aren't always the same. I have now updated RunTest so that it checks
694 the output of the locale test (test 3) against three different output
695 files, and it allows the test to pass if any one of them matches. With luck
696 this should make the test pass on some versions of Solaris where it was
697 failing. Because of the uncertainty, the script did not used to stop if
698 test 3 failed; it now does. If further versions of a French locale ever
699 come to light, they can now easily be added.
700
701 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
702 there was a message during ./configure, but it did not stop. This now
703 provokes an error. The invalid example in README has been corrected.
704 If a value less than the minimum is given, the minimum value has always
705 been used, but now a warning is given.
706
707 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
708 was a bug in the test system, which is now fixed. Also, the list of various
709 configurations that are tested for each release did not have one with both
710 16/32 bits and --enable-bar-anycrlf. It now does.
711
712 18. pcretest was missing "-C bsr" for displaying the \R default setting.
713
714 19. Little endian PowerPC systems are supported now by the JIT compiler.
715
716 20. The fast forward newline mechanism could enter to an infinite loop on
717 certain invalid UTF-8 input. Although we don't support these cases
718 this issue can be fixed by a performance optimization.
719
720 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
721 not take account if existing stack usage. There is now a new global
722 variable called pcre_stack_guard that can be set to point to an external
723 function to check stack availability. It is called at the start of
724 processing every parenthesized group.
725
726 22. A typo in the code meant that in ungreedy mode the max/min qualifier
727 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
728 not match "ab".
729
730 23. When UTF was disabled, the JIT program reported some incorrect compile
731 errors. These messages are silenced now.
732
733 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
734 compiler.
735
736 25. Change all the temporary files used in RunGrepTest to be different to those
737 used by RunTest so that the tests can be run simultaneously, for example by
738 "make -j check".
739
740
741 Version 8.34 15-December-2013
742 -----------------------------
743
744 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
745 executable memory. Patch inspired by Carsten Klein.
746
747 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
748 this macro is never tested and has no effect, because the work to support
749 coverage involves only compiling and linking options and special targets in
750 the Makefile. The comment in config.h implied that defining the macro would
751 enable coverage support, which is totally false. There was also support for
752 setting this macro in the CMake files (my fault, I just copied it from
753 configure). SUPPORT_GCOV has now been removed.
754
755 3. Make a small performance improvement in strlen16() and strlen32() in
756 pcretest.
757
758 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
759 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
760 these cases). There was also one in pcretest.c.
761
762 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
763
764 6. In UTF mode, the code for checking whether a group could match an empty
765 string (which is used for indefinitely repeated groups to allow for
766 breaking an infinite loop) was broken when the group contained a repeated
767 negated single-character class with a character that occupied more than one
768 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
769 UTF-8 mode). The effect was undefined: the group might or might not be
770 deemed as matching an empty string, or the program might have crashed.
771
772 7. The code for checking whether a group could match an empty string was not
773 recognizing that \h, \H, \v, \V, and \R must match a character.
774
775 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
776 an empty string. If it can, pcretest shows this in its information output.
777
778 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
779 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
780 matched by pcre_exec() without using JIT:
781
782 (a) If the rest of the pattern did not match after a maximal run of
783 grapheme clusters, the code for backing up to try with fewer of them
784 did not always back up over a full grapheme when characters that do not
785 have the modifier quality were involved, e.g. Hangul syllables.
786
787 (b) If the match point in a subject started with modifier character, and
788 there was no match, the code could incorrectly back up beyond the match
789 point, and potentially beyond the first character in the subject,
790 leading to a segfault or an incorrect match result.
791
792 10. A conditional group with an assertion condition could lead to PCRE
793 recording an incorrect first data item for a match if no other first data
794 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
795 first data item, and therefore matched "ca" after "c" instead of at the
796 start.
797
798 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
799 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
800
801 12. The source of pcregrep now includes z/OS-specific code so that it can be
802 compiled for z/OS as part of the special z/OS distribution.
803
804 13. Added the -T and -TM options to pcretest.
805
806 14. The code in pcre_compile.c for creating the table of named capturing groups
807 has been refactored. Instead of creating the table dynamically during the
808 actual compiling pass, the information is remembered during the pre-compile
809 pass (on the stack unless there are more than 20 named groups, in which
810 case malloc() is used) and the whole table is created before the actual
811 compile happens. This has simplified the code (it is now nearly 150 lines
812 shorter) and prepared the way for better handling of references to groups
813 with duplicate names.
814
815 15. A back reference to a named subpattern when there is more than one of the
816 same name now checks them in the order in which they appear in the pattern.
817 The first one that is set is used for the reference. Previously only the
818 first one was inspected. This change makes PCRE more compatible with Perl.
819
820 16. Unicode character properties were updated from Unicode 6.3.0.
821
822 17. The compile-time code for auto-possessification has been refactored, based
823 on a patch by Zoltan Herczeg. It now happens after instead of during
824 compilation. The code is cleaner, and more cases are handled. The option
825 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
826 options in pcretest are provided to set it. It can also be set by
827 (*NO_AUTO_POSSESS) at the start of a pattern.
828
829 18. The character VT has been added to the default ("C" locale) set of
830 characters that match \s and are generally treated as white space,
831 following this same change in Perl 5.18. There is now no difference between
832 "Perl space" and "POSIX space". Whether VT is treated as white space in
833 other locales depends on the locale.
834
835 19. The code for checking named groups as conditions, either for being set or
836 for being recursed, has been refactored (this is related to 14 and 15
837 above). Processing unduplicated named groups should now be as fast at
838 numerical groups, and processing duplicated groups should be faster than
839 before.
840
841 20. Two patches to the CMake build system, by Alexander Barkov:
842
843 (1) Replace the "source" command by "." in CMakeLists.txt because
844 "source" is a bash-ism.
845
846 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
847 without these the CMake build does not work on Solaris.
848
849 21. Perl has changed its handling of \8 and \9. If there is no previously
850 encountered capturing group of those numbers, they are treated as the
851 literal characters 8 and 9 instead of a binary zero followed by the
852 literals. PCRE now does the same.
853
854 22. Following Perl, added \o{} to specify codepoints in octal, making it
855 possible to specify values greater than 0777 and also making them
856 unambiguous.
857
858 23. Perl now gives an error for missing closing braces after \x{... instead of
859 treating the string as literal. PCRE now does the same.
860
861 24. RunTest used to grumble if an inappropriate test was selected explicitly,
862 but just skip it when running all tests. This make it awkward to run ranges
863 of tests when one of them was inappropriate. Now it just skips any
864 inappropriate tests, as it always did when running all tests.
865
866 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
867 character types such as \d or \w, too many callouts were inserted, and the
868 data that they returned was rubbish.
869
870 26. In UCP mode, \s was not matching two of the characters that Perl matches,
871 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
872 were matched by \h. The code has now been refactored so that the lists of
873 the horizontal and vertical whitespace characters used for \h and \v (which
874 are defined only in one place) are now also used for \s.
875
876 27. Add JIT support for the 64 bit TileGX architecture.
877 Patch by Jiong Wang (Tilera Corporation).
878
879 28. Possessive quantifiers for classes (both explicit and automatically
880 generated) now use special opcodes instead of wrapping in ONCE brackets.
881
882 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
883 (because it's meaningless), this was not happening when PCRE_CASELESS was
884 set. Not wrong, but inefficient.
885
886 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
887 properties for \w, \d, etc) is present in a test regex. Otherwise if the
888 test contains no characters greater than 255, Perl doesn't realise it
889 should be using Unicode semantics.
890
891 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
892 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
893 does in Unicode mode.
894
895 32. Added the "forbid" facility to pcretest so that putting tests into the
896 wrong test files can sometimes be quickly detected.
897
898 33. There is now a limit (default 250) on the depth of nesting of parentheses.
899 This limit is imposed to control the amount of system stack used at compile
900 time. It can be changed at build time by --with-parens-nest-limit=xxx or
901 the equivalent in CMake.
902
903 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
904 errors. Perl warns for these when in warning mode, but PCRE has no facility
905 for giving warnings.
906
907 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
908 However, this was not working for (?!) because it is optimized to (*FAIL),
909 for which PCRE does not allow quantifiers. The optimization is now disabled
910 when a quantifier follows (?!). I can't see any use for this, but it makes
911 things uniform.
912
913 36. Perl no longer allows group names to start with digits, so I have made this
914 change also in PCRE. It simplifies the code a bit.
915
916 37. In extended mode, Perl ignores spaces before a + that indicates a
917 possessive quantifier. PCRE allowed a space before the quantifier, but not
918 before the possessive +. It now does.
919
920 38. The use of \K (reset reported match start) within a repeated possessive
921 group such as (a\Kb)*+ was not working.
922
923 40. Document that the same character tables must be used at compile time and
924 run time, and that the facility to pass tables to pcre_exec() and
925 pcre_dfa_exec() is for use only with saved/restored patterns.
926
927 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
928 features for Builds with MSVC:
929
930 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
931 builds). The .rc files can be used to set FileDescription and many other
932 attributes.
933
934 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
935 This allows higher-level build scripts which want .pdb files to avoid
936 hard-coding the exact files needed."
937
938 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
939 mean "start of word" and "end of word", respectively, as a transition aid.
940
941 43. A minimizing repeat of a class containing codepoints greater than 255 in
942 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
943 compiled to use the heap for recursion.
944
945 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
946 is configured.
947
948
949 Version 8.33 28-May-2013
950 ------------------------
951
952 1. Added 'U' to some constants that are compared to unsigned integers, to
953 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
954 variables that are added to signed variables, to ensure the result is
955 signed and can be negated.
956
957 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
958 CMake config files.
959
960 3. Revise the creation of config.h.generic so that all boolean macros are
961 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
962 overriding via -D on the command line possible.
963
964 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
965 to unsigned int is reported to make a quite noticeable speed difference in
966 a specific Windows environment. Testing on Linux did also appear to show
967 some benefit (and it is clearly not harmful). Also fixed the definition of
968 Xop which should be unsigned.
969
970 5. Related to (4), changing the definition of the intermediate variable cc
971 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
972 improvements.
973
974 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
975 unnecessary spaces.
976
977 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
978 and later.
979
980 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
981
982 9. Optimizing fast_forward_start_bits in JIT.
983
984 10. Adding support for callouts in JIT, and fixing some issues revealed
985 during this work. Namely:
986
987 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
988
989 (b) Minimum length was not checked before the matching is started.
990
991 11. The value of capture_last that is passed to callouts was incorrect in some
992 cases when there was a capture on one path that was subsequently abandoned
993 after a backtrack. Also, the capture_last value is now reset after a
994 recursion, since all captures are also reset in this case.
995
996 12. The interpreter no longer returns the "too many substrings" error in the
997 case when an overflowing capture is in a branch that is subsequently
998 abandoned after a backtrack.
999
1000 13. In the pathological case when an offset vector of size 2 is used, pcretest
1001 now prints out the matched string after a yield of 0 or 1.
1002
1003 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
1004 Only supported by the JIT compiler at the moment.
1005
1006 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
1007
1008 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
1009 offset of the starting point of the matching process, provided the offsets
1010 vector is large enough.
1011
1012 17. The \A escape now records a lookbehind value of 1, though its execution
1013 does not actually inspect the previous character. This is to ensure that,
1014 in partial multi-segment matching, at least one character from the old
1015 segment is retained when a new segment is processed. Otherwise, if there
1016 are no lookbehinds in the pattern, \A might match incorrectly at the start
1017 of a new segment.
1018
1019 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
1020
1021 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
1022 gives some modest performance improvement in 8-bit mode.
1023
1024 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
1025 be expressed in certain programming languages using Universal Character
1026 Names.
1027
1028 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
1029 which points out that "non characters" are not "characters that may not
1030 appear in Unicode strings" but rather "characters that are reserved for
1031 internal use and have only local meaning".
1032
1033 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
1034 there was a conditional group that depended on an assertion, if the
1035 assertion was false, the callout that immediately followed the alternation
1036 in the condition was skipped when pcre_exec() was used for matching.
1037
1038 23. Allow an explicit callout to be inserted before an assertion that is the
1039 condition for a conditional group, for compatibility with automatic
1040 callouts, which always insert a callout at this point.
1041
1042 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
1043 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
1044
1045 25. (*PRUNE) is now supported by the JIT compiler.
1046
1047 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
1048
1049 27. Fix the case where there are two or more SKIPs with arguments that may be
1050 ignored.
1051
1052 28. (*SKIP) is now supported by the JIT compiler.
1053
1054 29. (*THEN) is now supported by the JIT compiler.
1055
1056 30. Update RunTest with additional test selector options.
1057
1058 31. The way PCRE handles backtracking verbs has been changed in two ways.
1059
1060 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
1061 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
1062 backtracking. In some cases this makes it more Perl-compatible, but Perl's
1063 rather obscure rules do not always do the same thing.
1064
1065 (2) Previously, backtracking verbs were confined within assertions. This is
1066 no longer the case for positive assertions, except for (*ACCEPT). Again,
1067 this sometimes improves Perl compatibility, and sometimes does not.
1068
1069 32. A number of tests that were in test 2 because Perl did things differently
1070 have been moved to test 1, because either Perl or PCRE has changed, and
1071 these tests are now compatible.
1072
1073 32. Backtracking control verbs are now handled in the same way in JIT and
1074 interpreter.
1075
1076 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
1077 contained a forward subroutine reference caused a compile error.
1078
1079 34. Auto-detect and optimize limited repetitions in JIT.
1080
1081 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
1082 blocking (*UTF) etc.
1083
1084 36. In the interpreter, maximizing pattern repetitions for characters and
1085 character types now use tail recursion, which reduces stack usage.
1086
1087 37. The value of the max lookbehind was not correctly preserved if a compiled
1088 and saved regex was reloaded on a host of different endianness.
1089
1090 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
1091 of the compiled pattern block, expand the flags field from 16 to 32 bits
1092 because it was almost full.
1093
1094 39. Try madvise first before posix_madvise.
1095
1096 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
1097 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
1098 original change locked out all matches of empty strings. This has been
1099 changed so that one match of an empty string per line is recognized.
1100 Subsequent searches on the same line (for colouring or for --only-matching,
1101 for example) do not recognize empty strings.
1102
1103 41. Applied a user patch to fix a number of spelling mistakes in comments.
1104
1105 42. Data lines longer than 65536 caused pcretest to crash.
1106
1107 43. Clarified the data type for length and startoffset arguments for pcre_exec
1108 and pcre_dfa_exec in the function-specific man pages, where they were
1109 explicitly stated to be in bytes, never having been updated. I also added
1110 some clarification to the pcreapi man page.
1111
1112 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
1113 a segmentation fault.
1114
1115
1116 Version 8.32 30-November-2012
1117 -----------------------------
1118
1119 1. Improved JIT compiler optimizations for first character search and single
1120 character iterators.
1121
1122 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
1123 Patch by Daniel Richard G.
1124
1125 3. Single character iterator optimizations in the JIT compiler.
1126
1127 4. Improved JIT compiler optimizations for character ranges.
1128
1129 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
1130 Reported by Giuseppe D'Angelo.
1131
1132 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
1133 of a line, was being set incorrectly in cases where .* appeared inside
1134 atomic brackets at the start of a pattern, or where there was a subsequent
1135 *PRUNE or *SKIP.
1136
1137 7. Improved instruction cache flush for POWER/PowerPC.
1138 Patch by Daniel Richard G.
1139
1140 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
1141 grep:
1142
1143 (a) There is now no limit to the number of patterns to be matched.
1144
1145 (b) An error is given if a pattern is too long.
1146
1147 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
1148 are now supported.
1149
1150 (d) --exclude-from and --include-from (multiple use) have been added.
1151
1152 (e) Exclusions and inclusions now apply to all files and directories, not
1153 just to those obtained from scanning a directory recursively.
1154
1155 (f) Multiple uses of -f and --file-list are now supported.
1156
1157 (g) In a Windows environment, the default for -d has been changed from
1158 "read" (the GNU grep default) to "skip", because otherwise the presence
1159 of a directory in the file list provokes an error.
1160
1161 (h) The documentation has been revised and clarified in places.
1162
1163 9. Improve the matching speed of capturing brackets.
1164
1165 10. Changed the meaning of \X so that it now matches a Unicode extended
1166 grapheme cluster.
1167
1168 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
1169 out POSIX threads when JIT support is configured.
1170
1171 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
1172
1173 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
1174 structure could go wrong in environments where size_t is not the same size
1175 as int.
1176
1177 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
1178 set.
1179
1180 15. The EBCDIC support had decayed; later updates to the code had included
1181 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
1182 general tidy up of EBCDIC-related issues, and the documentation was also
1183 not quite right. There is now a test that can be run on ASCII systems to
1184 check some of the EBCDIC-related things (but is it not a full test).
1185
1186 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
1187 in a small tidy to the code.
1188
1189 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
1190
1191 18. If the --only-matching (-o) option in pcregrep is specified multiple
1192 times, each one causes appropriate output. For example, -o1 -o2 outputs the
1193 substrings matched by the 1st and 2nd capturing parentheses. A separating
1194 string can be specified by --om-separator (default empty).
1195
1196 19. Improving the first n character searches.
1197
1198 20. Turn case lists for horizontal and vertical white space into macros so that
1199 they are defined only once.
1200
1201 21. This set of changes together give more compatible Unicode case-folding
1202 behaviour for characters that have more than one other case when UCP
1203 support is available.
1204
1205 (a) The Unicode property table now has offsets into a new table of sets of
1206 three or more characters that are case-equivalent. The MultiStage2.py
1207 script that generates these tables (the pcre_ucd.c file) now scans
1208 CaseFolding.txt instead of UnicodeData.txt for character case
1209 information.
1210
1211 (b) The code for adding characters or ranges of characters to a character
1212 class has been abstracted into a generalized function that also handles
1213 case-independence. In UTF-mode with UCP support, this uses the new data
1214 to handle characters with more than one other case.
1215
1216 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
1217 whose other case is greater than 256 are now correctly matched
1218 caselessly. Previously, the high codepoint matched the low one, but not
1219 vice versa.
1220
1221 (d) The processing of \h, \H, \v, and \ in character classes now makes use
1222 of the new class addition function, using character lists defined as
1223 macros alongside the case definitions of 20 above.
1224
1225 (e) Caseless back references now work with characters that have more than
1226 one other case.
1227
1228 (f) General caseless matching of characters with more than one other case
1229 is supported.
1230
1231 22. Unicode character properties were updated from Unicode 6.2.0
1232
1233 23. Improved CMake support under Windows. Patch by Daniel Richard G.
1234
1235 24. Add support for 32-bit character strings, and UTF-32
1236
1237 25. Major JIT compiler update (code refactoring and bugfixing).
1238 Experimental Sparc 32 support is added.
1239
1240 26. Applied a modified version of Daniel Richard G's patch to create
1241 pcre.h.generic and config.h.generic by "make" instead of in the
1242 PrepareRelease script.
1243
1244 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
1245 pcre_compile.c when checking for a zero character.
1246
1247 28. Introducing a native interface for JIT. Through this interface, the compiled
1248 machine code can be directly executed. The purpose of this interface is to
1249 provide fast pattern matching, so several sanity checks are not performed.
1250 However, feature tests are still performed. The new interface provides
1251 1.4x speedup compared to the old one.
1252
1253 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
1254 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
1255 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
1256
1257 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
1258 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
1259 the "old" RFC 2279). Instead, it ended up passing a negative length to
1260 pcre_exec().
1261
1262 31. Add support for GCC's visibility feature to hide internal functions.
1263
1264 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
1265 "unknown -C option" after outputting 0 or 1.
1266
1267 33. There is now support for generating a code coverage report for the test
1268 suite in environments where gcc is the compiler and lcov is installed. This
1269 is mainly for the benefit of the developers.
1270
1271 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
1272 unaddressable using valgrind annotations, allowing valgrind to detect
1273 invalid memory accesses. This is mainly for the benefit of the developers.
1274
1275 25. (*UTF) can now be used to start a pattern in any of the three libraries.
1276
1277 26. Give configure error if --enable-cpp but no C++ compiler found.
1278
1279
1280 Version 8.31 06-July-2012
1281 -------------------------
1282
1283 1. Fixing a wrong JIT test case and some compiler warnings.
1284
1285 2. Removed a bashism from the RunTest script.
1286
1287 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
1288 to unsigned type, result still unsigned" that was given by an MS compiler
1289 on encountering the code "-sizeof(xxx)".
1290
1291 4. Partial matching support is added to the JIT compiler.
1292
1293 5. Fixed several bugs concerned with partial matching of items that consist
1294 of more than one character:
1295
1296 (a) /^(..)\1/ did not partially match "aba" because checking references was
1297 done on an "all or nothing" basis. This also applied to repeated
1298 references.
1299
1300 (b) \R did not give a hard partial match if \r was found at the end of the
1301 subject.
1302
1303 (c) \X did not give a hard partial match after matching one or more
1304 characters at the end of the subject.
1305
1306 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
1307 a partial match for the string "\r".
1308
1309 (e) When newline was set to CRLF, the metacharacter "." did not recognize
1310 a partial match for a CR character at the end of the subject string.
1311
1312 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
1313 running pcretest, the text "(JIT)" added to the output whenever JIT is
1314 actually used to run the match.
1315
1316 7. Individual JIT compile options can be set in pcretest by following -s+[+]
1317 or /S+[+] with a digit between 1 and 7.
1318
1319 8. OP_NOT now supports any UTF character not just single-byte ones.
1320
1321 9. (*MARK) control verb is now supported by the JIT compiler.
1322
1323 10. The command "./RunTest list" lists the available tests without actually
1324 running any of them. (Because I keep forgetting what they all are.)
1325
1326 11. Add PCRE_INFO_MAXLOOKBEHIND.
1327
1328 12. Applied a (slightly modified) user-supplied patch that improves performance
1329 when the heap is used for recursion (compiled with --disable-stack-for-
1330 recursion). Instead of malloc and free for each heap frame each time a
1331 logical recursion happens, frames are retained on a chain and re-used where
1332 possible. This sometimes gives as much as 30% improvement.
1333
1334 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
1335 call.
1336
1337 14. As documented, (*COMMIT) is now confined to within a positive assertion.
1338
1339 15. It is now possible to link pcretest with libedit as an alternative to
1340 libreadline.
1341
1342 16. (*COMMIT) control verb is now supported by the JIT compiler.
1343
1344 17. The Unicode data tables have been updated to Unicode 6.1.0.
1345
1346 18. Added --file-list option to pcregrep.
1347
1348 19. Added binary file support to pcregrep, including the -a, --binary-files,
1349 -I, and --text options.
1350
1351 20. The madvise function is renamed for posix_madvise for QNX compatibility
1352 reasons. Fixed by Giuseppe D'Angelo.
1353
1354 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
1355
1356 22. Changed the option for creating symbolic links for 16-bit man pages from
1357 -s to -sf so that re-installing does not cause issues.
1358
1359 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
1360
1361 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
1362 matches in certain environments (the workspace was not being correctly
1363 retained). Also added to pcre_dfa_exec() a simple plausibility check on
1364 some of the workspace data at the beginning of a restart.
1365
1366 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
1367 was not doing so when it should - probably a typo introduced by SVN 528
1368 (change 8.10/14).
1369
1370 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
1371 \w+ when the character tables indicated that \x{c4} was a word character.
1372 There were several related cases, all because the tests for doing a table
1373 lookup were testing for characters less than 127 instead of 255.
1374
1375 27. If a pattern contains capturing parentheses that are not used in a match,
1376 their slots in the ovector are set to -1. For those that are higher than
1377 any matched groups, this happens at the end of processing. In the case when
1378 there were back references that the ovector was too small to contain
1379 (causing temporary malloc'd memory to be used during matching), and the
1380 highest capturing number was not used, memory off the end of the ovector
1381 was incorrectly being set to -1. (It was using the size of the temporary
1382 memory instead of the true size.)
1383
1384 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
1385 ovector size, it uses memory at the end of the block that it has got.
1386
1387 29. Check for an overlong MARK name and give an error at compile time. The
1388 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
1389
1390 30. JIT compiler update.
1391
1392 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
1393 Rill for the patch.
1394
1395 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
1396
1397 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
1398
1399 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
1400 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
1401 was enabled.
1402
1403 35. Improve JIT code generation for greedy plus quantifier.
1404
1405 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1406 1 to "aa" instead of to an empty string. The bug affected repeated groups
1407 that could potentially match an empty string.
1408
1409 37. Optimizing single character iterators in JIT.
1410
1411 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1412 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1413 codepoints that are too big for the mode are faulted, and in a UTF mode,
1414 disallowed codepoints are also faulted.
1415
1416 39. If PCRE was compiled with UTF support, in three places in the DFA
1417 matcher there was code that should only have been obeyed in UTF mode, but
1418 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1419 processing when bytes with values greater than 127 were present. In 16-bit
1420 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1421 both cases the values are those that cannot be the first data item in a UTF
1422 character. The three items that might have provoked this were recursions,
1423 possessively repeated groups, and atomic groups.
1424
1425 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1426 and pcregrep, because some OS require shared objects to be explicitly
1427 passed to ld, causing the link step to fail if they are not.
1428
1429 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1430 mode, patterns that started with \h* or \R* might be incorrectly matched.
1431
1432
1433 Version 8.30 04-February-2012
1434 -----------------------------
1435
1436 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1437 name is defined in ctype.h.
1438
1439 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1440 only in quite long subpatterns.
1441
1442 3. Removed the function pcre_info(), which has been obsolete and deprecated
1443 since it was replaced by pcre_fullinfo() in February 2000.
1444
1445 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1446 match a (*MARK), and the match failed at the start of the subject, a
1447 reference to memory before the start of the subject could occur. This bug
1448 was introduced by fix 17 of release 8.21.
1449
1450 5. A reference to an unset group with zero minimum repetition was giving
1451 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1452 /(another)?(\1?)test/ matched against "hello world test". This bug was
1453 introduced in release 8.13.
1454
1455 6. Add support for 16-bit character strings (a large amount of work involving
1456 many changes and refactorings).
1457
1458 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1459 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1460 from a file.
1461
1462 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1463 rounding is not applied in this particular case).
1464
1465 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1466 if they appear, or are escaped, in patterns.
1467
1468 10. Get rid of a number of -Wunused-but-set-variable warnings.
1469
1470 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1471 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1472 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1473 also returns the mark "x". This bug applied to capturing parentheses,
1474 non-capturing parentheses, and atomic parentheses. It also applied to some
1475 assertions.
1476
1477 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1478 information out of configure.ac instead of relying on pcre.h.generic, which
1479 is not stored in the repository.
1480
1481 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1482 -lreadline.
1483
1484 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1485
1486 15. Applied Graycode's patch to put the top-level frame on the stack rather
1487 than the heap when not using the stack for recursion. This gives a
1488 performance improvement in many cases when recursion is not deep.
1489
1490 16. Experimental code added to "pcretest -C" to output the stack frame size.
1491
1492
1493 Version 8.21 12-Dec-2011
1494 ------------------------
1495
1496 1. Updating the JIT compiler.
1497
1498 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1499 are added as well.
1500
1501 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1502 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1503 calling _pcre_jit_exec. Some extra comments are added.
1504
1505 4. (*MARK) settings inside atomic groups that do not contain any capturing
1506 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1507 was introduced by change 18 for 8.20.
1508
1509 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1510 ECMA-262 standard.
1511
1512 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1513 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1514 This bug was probably introduced by change 9 of 8.13.
1515
1516 7. While fixing 6 above, I noticed that a number of other items were being
1517 incorrectly rejected as "not fixed length". This arose partly because newer
1518 opcodes had not been added to the fixed-length checking code. I have (a)
1519 corrected the bug and added tests for these items, and (b) arranged for an
1520 error to occur if an unknown opcode is encountered while checking for fixed
1521 length instead of just assuming "not fixed length". The items that were
1522 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1523 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1524 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1525
1526 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1527 being incorrectly compiled and would have given unpredicatble results.
1528
1529 9. A possessively repeated subpattern with minimum repeat count greater than
1530 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1531 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1532 the first (A) could occur when it should not.
1533
1534 10. Add a cast and remove a redundant test from the code.
1535
1536 11. JIT should use pcre_malloc/pcre_free for allocation.
1537
1538 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1539 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1540 is anything other than /usr, -L is still shown).
1541
1542 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1543
1544 14. Perl does not support \N without a following name in a [] class; PCRE now
1545 also gives an error.
1546
1547 15. If a forward reference was repeated with an upper limit of around 2000,
1548 it caused the error "internal error: overran compiling workspace". The
1549 maximum number of forward references (including repeats) was limited by the
1550 internal workspace, and dependent on the LINK_SIZE. The code has been
1551 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1552 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1553 of around 200,000 forward references. While doing this, I also speeded up
1554 the filling in of repeated forward references.
1555
1556 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1557 incorrectly expecting the subject to contain another "a" after the start.
1558
1559 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1560 in the match, the SKIP should be ignored. This was not happening; instead
1561 the SKIP was being treated as NOMATCH. For patterns such as
1562 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1563 tested.
1564
1565 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1566 now much more compatible with Perl, in particular in cases where the result
1567 is a non-match for a non-anchored pattern. For example, if
1568 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1569 "m", where previously it did not return a name. A side effect of this
1570 change is that for partial matches, the last encountered mark name is
1571 returned, as for non matches. A number of tests that were previously not
1572 Perl-compatible have been moved into the Perl-compatible test files. The
1573 refactoring has had the pleasing side effect of removing one argument from
1574 the match() function, thus reducing its stack requirements.
1575
1576 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1577 subsequent uses of /S (without +) incorrectly behaved like /S+.
1578
1579 21. Retrieve executable code size support for the JIT compiler and fixing
1580 some warnings.
1581
1582 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1583 not work when the shorter character appeared right at the end of the
1584 subject string.
1585
1586 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1587 systems.
1588
1589 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1590 output it when the /M option is used in pcretest.
1591
1592 25. The CheckMan script was not being included in the distribution. Also, added
1593 an explicit "perl" to run Perl scripts from the PrepareRelease script
1594 because this is reportedly needed in Windows.
1595
1596 26. If study data was being save in a file and studying had not found a set of
1597 "starts with" bytes for the pattern, the data written to the file (though
1598 never used) was taken from uninitialized memory and so caused valgrind to
1599 complain.
1600
1601 27. Updated RunTest.bat as provided by Sheri Pierce.
1602
1603 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1604
1605 29. Computation of memory usage for the table of capturing group names was
1606 giving an unnecessarily large value.
1607
1608
1609 Version 8.20 21-Oct-2011
1610 ------------------------
1611
1612 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1613 a POSIX class. After further experiments with Perl, which convinced me that
1614 Perl has bugs and confusions, a closing square bracket is no longer allowed
1615 in a POSIX name. This bug also affected patterns with classes that started
1616 with full stops.
1617
1618 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1619 captured substring, but while checking the failing first alternative,
1620 substring 1 is temporarily captured. If the output vector supplied to
1621 pcre_exec() was not big enough for this capture, the yield of the function
1622 was still zero ("insufficient space for captured substrings"). This cannot
1623 be totally fixed without adding another stack variable, which seems a lot
1624 of expense for a edge case. However, I have improved the situation in cases
1625 such as /(a)(b)x|abc/ matched against "abc", where the return code
1626 indicates that fewer than the maximum number of slots in the ovector have
1627 been set.
1628
1629 3. Related to (2) above: when there are more back references in a pattern than
1630 slots in the output vector, pcre_exec() uses temporary memory during
1631 matching, and copies in the captures as far as possible afterwards. It was
1632 using the entire output vector, but this conflicts with the specification
1633 that only 2/3 is used for passing back captured substrings. Now it uses
1634 only the first 2/3, for compatibility. This is, of course, another edge
1635 case.
1636
1637 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1638 main code base, and can be used by building with --enable-jit. When this is
1639 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1640 runtime --no-jit option is given.
1641
1642 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1643 ovector, the return from the function was zero, implying that there were
1644 other matches that did not fit. The correct "exactly full" value is now
1645 returned.
1646
1647 6. If a subpattern that was called recursively or as a subroutine contained
1648 (*PRUNE) or any other control that caused it to give a non-standard return,
1649 invalid errors such as "Error -26 (nested recursion at the same subject
1650 position)" or even infinite loops could occur.
1651
1652 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1653 computing the minimum length on reaching *ACCEPT, and so ended up with the
1654 wrong value of 1 rather than 0. Further investigation indicates that
1655 computing a minimum subject length in the presence of *ACCEPT is difficult
1656 (think back references, subroutine calls), and so I have changed the code
1657 so that no minimum is registered for a pattern that contains *ACCEPT.
1658
1659 8. If (*THEN) was present in the first (true) branch of a conditional group,
1660 it was not handled as intended. [But see 16 below.]
1661
1662 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1663 Sheri Pierce.
1664
1665 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1666 the first byte in a match must be "a".
1667
1668 11. Change 17 for 8.13 increased the recursion depth for patterns like
1669 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1670 pattern contains any instances of (*THEN). If it does not, the old
1671 optimizations are restored. It would be nice to do this on a per-group
1672 basis, but at the moment that is not feasible.
1673
1674 12. In some environments, the output of pcretest -C is CRLF terminated. This
1675 broke RunTest's code that checks for the link size. A single white space
1676 character after the value is now allowed for.
1677
1678 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1679 For "fr", it uses the Windows-specific input and output files.
1680
1681 14. If (*THEN) appeared in a group that was called recursively or as a
1682 subroutine, it did not work as intended. [But see next item.]
1683
1684 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1685 pattern fragments (but not containing any | characters). If A and B are
1686 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1687 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1688 D. In other words, Perl considers parentheses that do not contain any |
1689 characters to be part of a surrounding alternative, whereas PCRE was
1690 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1691 differently. PCRE now behaves in the same way as Perl, except in the case
1692 of subroutine/recursion calls such as (?1) which have in any case always
1693 been different (but PCRE had them first :-).
1694
1695 16. Related to 15 above: Perl does not treat the | in a conditional group as
1696 creating alternatives. Such a group is treated in the same way as an
1697 ordinary group without any | characters when processing (*THEN). PCRE has
1698 been changed to match Perl's behaviour.
1699
1700 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1701 RunGrepTest script failed.
1702
1703 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1704 inevitable for groups that contain captures, but it can lead to a lot of
1705 stack use in large patterns. The old behaviour has been restored for atomic
1706 groups that do not contain any capturing parentheses.
1707
1708 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1709 suppress the check for a minimum subject length at run time. (If it was
1710 given to pcre_exec() or pcre_dfa_exec() it did work.)
1711
1712 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1713 fail to work when decoding hex characters in data strings in EBCDIC
1714 environments.
1715
1716 21. It appears that in at least one Mac OS environment, the isxdigit() function
1717 is implemented as a macro that evaluates to its argument more than once,
1718 contravening the C 90 Standard (I haven't checked a later standard). There
1719 was an instance in pcretest which caused it to go wrong when processing
1720 \x{...} escapes in subject strings. The has been rewritten to avoid using
1721 things like p++ in the argument of isxdigit().
1722
1723
1724 Version 8.13 16-Aug-2011
1725 ------------------------
1726
1727 1. The Unicode data tables have been updated to Unicode 6.0.0.
1728
1729 2. Two minor typos in pcre_internal.h have been fixed.
1730
1731 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1732 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1733 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1734
1735 4. There were a number of related bugs in the code for matching backrefences
1736 caselessly in UTF-8 mode when codes for the characters concerned were
1737 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1738 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1739 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1740 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1741 2-byte code at the end of the subject (it thought there wasn't enough data
1742 left).
1743
1744 5. Comprehensive information about what went wrong is now returned by
1745 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1746 as the output vector has at least 2 elements. The offset of the start of
1747 the failing character and a reason code are placed in the vector.
1748
1749 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1750 now returned is for the first byte of the failing character, instead of the
1751 last byte inspected. This is an incompatible change, but I hope it is small
1752 enough not to be a problem. It makes the returned offset consistent with
1753 pcre_exec() and pcre_dfa_exec().
1754
1755 7. pcretest now gives a text phrase as well as the error number when
1756 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1757 failure, the offset and reason code are output.
1758
1759 8. When \R was used with a maximizing quantifier it failed to skip backwards
1760 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1761 back over a single character (\n). This seems wrong (because it treated the
1762 two characters as a single entity when going forwards), conflicts with the
1763 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1764 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1765 has been changed.
1766
1767 9. Some internal refactoring has changed the processing so that the handling
1768 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1769 time (the PCRE_DOTALL option was changed this way some time ago: version
1770 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1771 which was always a bit of a fudge. It also means that there is one less
1772 argument for the match() function, which reduces its stack requirements
1773 slightly. This change also fixes an incompatibility with Perl: the pattern
1774 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1775
1776 10. More internal refactoring has drastically reduced the number of recursive
1777 calls to match() for possessively repeated groups such as (abc)++ when
1778 using pcre_exec().
1779
1780 11. While implementing 10, a number of bugs in the handling of groups were
1781 discovered and fixed:
1782
1783 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1784 (a|)*(?1) gave a compile-time internal error.
1785 ((a|)+)+ did not notice that the outer group could match an empty string.
1786 (^a|^)+ was not marked as anchored.
1787 (.*a|.*)+ was not marked as matching at start or after a newline.
1788
1789 12. Yet more internal refactoring has removed another argument from the match()
1790 function. Special calls to this function are now indicated by setting a
1791 value in a variable in the "match data" data block.
1792
1793 13. Be more explicit in pcre_study() instead of relying on "default" for
1794 opcodes that mean there is no starting character; this means that when new
1795 ones are added and accidentally left out of pcre_study(), testing should
1796 pick them up.
1797
1798 14. The -s option of pcretest has been documented for ages as being an old
1799 synonym of -m (show memory usage). I have changed it to mean "force study
1800 for every regex", that is, assume /S for every regex. This is similar to -i
1801 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1802 using it. It makes it easier to run collections of tests with and without
1803 study enabled, and thereby test pcre_study() more easily. All the standard
1804 tests are now run with and without -s (but some patterns can be marked as
1805 "never study" - see 20 below).
1806
1807 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1808 restoration of the capturing data to the outer values was not happening
1809 correctly.
1810
1811 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1812 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1813 pattern had matched an empty string, and so incorrectly returned a no
1814 match.
1815
1816 17. There was optimizing code for the last branch of non-capturing parentheses,
1817 and also for the obeyed branch of a conditional subexpression, which used
1818 tail recursion to cut down on stack usage. Unfortunately, now that there is
1819 the possibility of (*THEN) occurring in these branches, tail recursion is
1820 no longer possible because the return has to be checked for (*THEN). These
1821 two optimizations have therefore been removed. [But see 8.20/11 above.]
1822
1823 18. If a pattern containing \R was studied, it was assumed that \R always
1824 matched two bytes, thus causing the minimum subject length to be
1825 incorrectly computed because \R can also match just one byte.
1826
1827 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1828 was incorrectly computed.
1829
1830 20. If /S is present twice on a test pattern in pcretest input, it now
1831 *disables* studying, thereby overriding the use of -s on the command line
1832 (see 14 above). This is necessary for one or two tests to keep the output
1833 identical in both cases.
1834
1835 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1836 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1837
1838 22. When an atomic group that contained a capturing parenthesis was
1839 successfully matched, but the branch in which it appeared failed, the
1840 capturing was not being forgotten if a higher numbered group was later
1841 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1842 group 1 to "a", when in fact it should be unset. This applied to multi-
1843 branched capturing and non-capturing groups, repeated or not, and also to
1844 positive assertions (capturing in negative assertions does not happen
1845 in PCRE) and also to nested atomic groups.
1846
1847 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1848 subject after a captured substring, to make it easier to tell which of a
1849 number of identical substrings has been captured.
1850
1851 24. The way atomic groups are processed by pcre_exec() has been changed so that
1852 if they are repeated, backtracking one repetition now resets captured
1853 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1854 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1855 "aaa". Previously, it would have been "a". As part of this code
1856 refactoring, the way recursive calls are handled has also been changed.
1857
1858 25. If an assertion condition captured any substrings, they were not passed
1859 back unless some other capturing happened later. For example, if
1860 (?(?=(a))a) was matched against "a", no capturing was returned.
1861
1862 26. When studying a pattern that contained subroutine calls or assertions,
1863 the code for finding the minimum length of a possible match was handling
1864 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1865 group 1 called group 2 while simultaneously a separate group 2 called group
1866 1). A stack overflow occurred in this case. I have fixed this by limiting
1867 the recursion depth to 10.
1868
1869 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1870 Fortmann. This supports explicit test numbers on the command line, and has
1871 argument validation and error reporting.
1872
1873 28. An instance of \X with an unlimited repeat could fail if at any point the
1874 first character it looked at was a mark character.
1875
1876 29. Some minor code refactoring concerning Unicode properties and scripts
1877 should reduce the stack requirement of match() slightly.
1878
1879 30. Added the '=' option to pcretest to check the setting of unused capturing
1880 slots at the end of the pattern, which are documented as being -1, but are
1881 not included in the return count.
1882
1883 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1884 compiled something random. Now it gives a compile-time error (as does
1885 Perl).
1886
1887 32. A *MARK encountered during the processing of a positive assertion is now
1888 recorded and passed back (compatible with Perl).
1889
1890 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1891 had alternative anchored branches, the search for a second match in a line
1892 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1893 matched the line "0102" twice. The same bug affected patterns that started
1894 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1895 twice.
1896
1897 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1898 does, and because of capturing effects, quantifying parenthesized
1899 assertions may at times be useful. Quantifiers are now allowed for
1900 parenthesized assertions.
1901
1902 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1903
1904 36. \g was being checked for fancy things in a character class, when it should
1905 just be a literal "g".
1906
1907 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1908 appearance of a nested POSIX class supersedes an apparent external class.
1909 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1910 unescaped square brackets may also appear as part of class names. For
1911 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1912 more like Perl. (But see 8.20/1 above.)
1913
1914 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1915 was because it thought it was \N{name}, which is not supported).
1916
1917 39. Add minix to OS list not supporting the -S option in pcretest.
1918
1919 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1920 cannot analyze patterns in sufficient detail to catch mutual recursions
1921 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1922 subgroup is called recursively as a subpattern for a second time at the
1923 same position in the subject string. In previous releases this might have
1924 been caught by the recursion limit, or it might have run out of stack.
1925
1926 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1927 happen only once. PCRE was, however incorrectly giving a compile time error
1928 "recursive call could loop indefinitely" because it cannot analyze the
1929 pattern in sufficient detail. The compile time test no longer happens when
1930 PCRE is compiling a conditional subpattern, but actual runaway loops are
1931 now caught at runtime (see 40 above).
1932
1933 42. It seems that Perl allows any characters other than a closing parenthesis
1934 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1935 has been changed to be the same.
1936
1937 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1938 as not to get warnings when autogen.sh is called. Also changed
1939 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1940
1941 44. To help people who use pcregrep to scan files containing exceedingly long
1942 lines, the following changes have been made:
1943
1944 (a) The default value of the buffer size parameter has been increased from
1945 8K to 20K. (The actual buffer used is three times this size.)
1946
1947 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1948 PCRE is built.
1949
1950 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1951 to be set at run time.
1952
1953 (d) Numerical values in pcregrep options can be followed by K or M, for
1954 example --buffer-size=50K.
1955
1956 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1957 given and the return code is set to 2.
1958
1959 45. Add a pointer to the latest mark to the callout data block.
1960
1961 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1962 partial match of an empty string instead of no match. This was specific to
1963 the use of ".".
1964
1965 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1966 complete match instead of a partial match. This bug was dependent on both
1967 the PCRE_UTF8 and PCRE_DOTALL options being set.
1968
1969 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1970 starting byte set, because \b was not being ignored.
1971
1972
1973 Version 8.12 15-Jan-2011
1974 ------------------------
1975
1976 1. Fixed some typos in the markup of the man pages, and wrote a script that
1977 checks for such things as part of the documentation building process.
1978
1979 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1980 --match-limit and --recursion-limit options (added for 8.11). In
1981 particular, this made one of the standard tests fail. (The integer value
1982 went into the wrong half of a long int.)
1983
1984 3. If the --colour option was given to pcregrep with -v (invert match), it
1985 did strange things, either producing crazy output, or crashing. It should,
1986 of course, ignore a request for colour when reporting lines that do not
1987 match.
1988
1989 4. Another pcregrep bug caused similar problems if --colour was specified with
1990 -M (multiline) and the pattern match finished with a line ending.
1991
1992 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1993 matched in multiline mode, the following line was shown as part of the
1994 match. This seems wrong, so I have changed it.
1995
1996 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1997 the check for further matches in the same line (so they could be coloured)
1998 to overrun the end of the current line. If another match was found, it was
1999 incorrectly shown (and then shown again when found in the next line).
2000
2001 7. If pcregrep was compiled under Windows, there was a reference to the
2002 function pcregrep_exit() before it was defined. I am assuming this was
2003 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
2004 reported by a user. I've moved the definition above the reference.
2005
2006
2007 Version 8.11 10-Dec-2010
2008 ------------------------
2009
2010 1. (*THEN) was not working properly if there were untried alternatives prior
2011 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
2012 backtracked to try for "b" instead of moving to the next alternative branch
2013 at the same level (in this case, to look for "c"). The Perl documentation
2014 is clear that when (*THEN) is backtracked onto, it goes to the "next
2015 alternative in the innermost enclosing group".
2016
2017 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
2018 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
2019 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
2020 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
2021 (*THEN).
2022
2023 3. If \s appeared in a character class, it removed the VT character from
2024 the class, even if it had been included by some previous item, for example
2025 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
2026 of \s, but is part of the POSIX "space" class.)
2027
2028 4. A partial match never returns an empty string (because you can always
2029 match an empty string at the end of the subject); however the checking for
2030 an empty string was starting at the "start of match" point. This has been
2031 changed to the "earliest inspected character" point, because the returned
2032 data for a partial match starts at this character. This means that, for
2033 example, /(?<=abc)def/ gives a partial match for the subject "abc"
2034 (previously it gave "no match").
2035
2036 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
2037 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
2038 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
2039 has an implication that the given string is incomplete (because a partial
2040 match is preferred over a full match). For this reason, these items now
2041 give a partial match in this situation. [Aside: previously, the one case
2042 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
2043 match rather than a full match, which was wrong by the old rules, but is
2044 now correct.]
2045
2046 6. There was a bug in the handling of #-introduced comments, recognized when
2047 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
2048 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
2049 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
2050 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
2051 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
2052 places in pcre_compile().
2053
2054 7. Related to (6) above, when pcre_compile() was skipping #-introduced
2055 comments when looking ahead for named forward references to subpatterns,
2056 the only newline sequence it recognized was NL. It now handles newlines
2057 according to the set newline convention.
2058
2059 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
2060 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
2061 cater for a lack of strerror(). These oversights have been fixed.
2062
2063 9. Added --match-limit and --recursion-limit to pcregrep.
2064
2065 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
2066
2067 11. When the -o option was used, pcregrep was setting a return code of 1, even
2068 when matches were found, and --line-buffered was not being honoured.
2069
2070 12. Added an optional parentheses number to the -o and --only-matching options
2071 of pcregrep.
2072
2073 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
2074 can match an empty string. The code to do it in pcretest and pcredemo
2075 needed fixing:
2076
2077 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
2078 only one byte after an empty string match just before CRLF (this case
2079 just got forgotten; "any" and "anycrlf" were OK).
2080
2081 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
2082 mode when an empty string match preceded an ASCII character followed by
2083 a non-ASCII character. (The code for advancing by one character rather
2084 than one byte was nonsense.)
2085
2086 (c) The pcredemo.c sample program did not have any code at all to handle
2087 the cases when CRLF is a valid newline sequence.
2088
2089 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
2090 as a starting offset was within the subject string. There is now a new
2091 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
2092 negative or greater than the length of the string. In order to test this,
2093 pcretest is extended to allow the setting of negative starting offsets.
2094
2095 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
2096 starting offset points to the beginning of a UTF-8 character was
2097 unnecessarily clumsy. I tidied it up.
2098
2099 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
2100 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
2101
2102 17. Nobody had reported that the --include_dir option, which was added in
2103 release 7.7 should have been called --include-dir (hyphen, not underscore)
2104 for compatibility with GNU grep. I have changed it to --include-dir, but
2105 left --include_dir as an undocumented synonym, and the same for
2106 --exclude-dir, though that is not available in GNU grep, at least as of
2107 release 2.5.4.
2108
2109 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
2110 characters from a string of bytes) have been redefined so as not to use
2111 loops, in order to improve performance in some environments. At the same
2112 time, I abstracted some of the common code into auxiliary macros to save
2113 repetition (this should not affect the compiled code).
2114
2115 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
2116 compile-time error is now given if \c is not followed by an ASCII
2117 character, that is, a byte less than 128. (In EBCDIC mode, the code is
2118 different, and any byte value is allowed.)
2119
2120 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
2121 START_OPTIMIZE option, which is now allowed at compile time - but just
2122 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
2123 to pcregrep and other applications that have no direct access to PCRE
2124 options. The new /Y option in pcretest sets this option when calling
2125 pcre_compile().
2126
2127 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
2128 back references. Groups containing recursive back references were forced to
2129 be atomic by that change, but in the case of named groups, the amount of
2130 memory required was incorrectly computed, leading to "Failed: internal
2131 error: code overflow". This has been fixed.
2132
2133 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
2134 pcretest.c, to avoid build problems in some Borland environments.
2135
2136
2137 Version 8.10 25-Jun-2010
2138 ------------------------
2139
2140 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
2141 THEN.
2142
2143 2. (*ACCEPT) was not working when inside an atomic group.
2144
2145 3. Inside a character class, \B is treated as a literal by default, but
2146 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
2147 causes the error). The code is unchanged, but I tidied the documentation.
2148
2149 4. Inside a character class, PCRE always treated \R and \X as literals,
2150 whereas Perl faults them if its -w option is set. I have changed PCRE so
2151 that it faults them when PCRE_EXTRA is set.
2152
2153 5. Added support for \N, which always matches any character other than
2154 newline. (It is the same as "." when PCRE_DOTALL is not set.)
2155
2156 6. When compiling pcregrep with newer versions of gcc which may have
2157 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
2158 declared with attribute warn_unused_result" were given. Just casting the
2159 result to (void) does not stop the warnings; a more elaborate fudge is
2160 needed. I've used a macro to implement this.
2161
2162 7. Minor change to pcretest.c to avoid a compiler warning.
2163
2164 8. Added four artifical Unicode properties to help with an option to make
2165 \s etc use properties (see next item). The new properties are: Xan
2166 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
2167
2168 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
2169 use Unicode properties. (*UCP) at the start of a pattern can be used to set
2170 this option. Modified pcretest to add /W to test this facility. Added
2171 REG_UCP to make it available via the POSIX interface.
2172
2173 10. Added --line-buffered to pcregrep.
2174
2175 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
2176 studied, and the match started with a letter with a code point greater than
2177 127 whose first byte was different to the first byte of the other case of
2178 the letter, the other case of this starting letter was not recognized
2179 (#976).
2180
2181 12. If a pattern that was studied started with a repeated Unicode property
2182 test, for example, \p{Nd}+, there was the theoretical possibility of
2183 setting up an incorrect bitmap of starting bytes, but fortunately it could
2184 not have actually happened in practice until change 8 above was made (it
2185 added property types that matched character-matching opcodes).
2186
2187 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
2188 possible starting bytes for non-anchored patterns.
2189
2190 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
2191 \R, and also a number of cases that involve Unicode properties, both
2192 explicit and implicit when PCRE_UCP is set.
2193
2194 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
2195 input, it could crash or give wrong results if characters with values
2196 greater than 0xc0 were present in the subject string. (Detail: it assumed
2197 UTF-8 input when processing these items.)
2198
2199 16. Added a lot of (int) casts to avoid compiler warnings in systems where
2200 size_t is 64-bit (#991).
2201
2202 17. Added a check for running out of memory when PCRE is compiled with
2203 --disable-stack-for-recursion (#990).
2204
2205 18. If the last data line in a file for pcretest does not have a newline on
2206 the end, a newline was missing in the output.
2207
2208 19. The default pcre_chartables.c file recognizes only ASCII characters (values
2209 less than 128) in its various bitmaps. However, there is a facility for
2210 generating tables according to the current locale when PCRE is compiled. It
2211 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
2212 characters, are recognized by isspace() and therefore were getting set in
2213 these tables, and indeed these tables seem to approximate to ISO 8859. This
2214 caused a problem in UTF-8 mode when pcre_study() was used to create a list
2215 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
2216 which of course cannot start UTF-8 characters. I have changed the code so
2217 that only real ASCII characters (less than 128) and the correct starting
2218 bytes for UTF-8 encodings are set for characters greater than 127 when in
2219 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
2220 altogether.)
2221
2222 20. Added the /T option to pcretest so as to be able to run tests with non-
2223 standard character tables, thus making it possible to include the tests
2224 used for 19 above in the standard set of tests.
2225
2226 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
2227 reference to a subpattern the other side of a comment that contains an
2228 opening parenthesis caused either an internal compiling error, or a
2229 reference to the wrong subpattern.
2230
2231
2232 Version 8.02 19-Mar-2010
2233 ------------------------
2234
2235 1. The Unicode data tables have been updated to Unicode 5.2.0.
2236
2237 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
2238 configured.
2239
2240 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
2241 original author of that file, following a query about its status.
2242
2243 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
2244 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
2245
2246 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
2247 quantifier applied to a forward-referencing subroutine call, could compile
2248 incorrect code or give the error "internal error: previously-checked
2249 referenced subpattern not found".
2250
2251 6. Both MS Visual Studio and Symbian OS have problems with initializing
2252 variables to point to external functions. For these systems, therefore,
2253 pcre_malloc etc. are now initialized to local functions that call the
2254 relevant global functions.
2255
2256 7. There were two entries missing in the vectors called coptable and poptable
2257 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
2258 I've fixed the data, and added a kludgy way of testing at compile time that
2259 the lengths are correct (equal to the number of opcodes).
2260
2261 8. Following on from 7, I added a similar kludge to check the length of the
2262 eint vector in pcreposix.c.
2263
2264 9. Error texts for pcre_compile() are held as one long string to avoid too
2265 much relocation at load time. To find a text, the string is searched,
2266 counting zeros. There was no check for running off the end of the string,
2267 which could happen if a new error number was added without updating the
2268 string.
2269
2270 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
2271
2272 11. \K was not working if it appeared in an atomic group or in a group that
2273 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
2274 \K is "not well defined" if used in an assertion. PCRE now accepts it if
2275 the assertion is positive, but not if it is negative.
2276
2277 12. Change 11 fortuitously reduced the size of the stack frame used in the
2278 "match()" function of pcre_exec.c by one pointer. Forthcoming
2279 implementation of support for (*MARK) will need an extra pointer on the
2280 stack; I have reserved it now, so that the stack frame size does not
2281 decrease.
2282
2283 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
2284 item in branch that calls a recursion is a subroutine call - as in the
2285 second branch in the above example - was incorrectly given the compile-
2286 time error "recursive call could loop indefinitely" because pcre_compile()
2287 was not correctly checking the subroutine for matching a non-empty string.
2288
2289 14. The checks for overrunning compiling workspace could trigger after an
2290 overrun had occurred. This is a "should never occur" error, but it can be
2291 triggered by pathological patterns such as hundreds of nested parentheses.
2292 The checks now trigger 100 bytes before the end of the workspace.
2293
2294 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
2295
2296
2297 Version 8.01 19-Jan-2010
2298 ------------------------
2299
2300 1. If a pattern contained a conditional subpattern with only one branch (in
2301 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
2302 computed the wrong minimum data length (which is of course zero for such
2303 subpatterns). This could cause incorrect "no match" results.
2304
2305 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
2306 the pattern is reset in the first branch, pcre_compile() failed with
2307 "internal error: code overflow at offset...". This happened only when
2308 the reset was to the original external option setting. (An optimization
2309 abstracts leading options settings into an external setting, which was the
2310 cause of this.)
2311
2312 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
2313 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
2314 assertion pattern did not match (meaning that the assertion was true), it
2315 was incorrectly treated as false if the SKIP had been reached during the
2316 matching. This also applied to assertions used as conditions.
2317
2318 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
2319 assertion subpattern, including such a pattern used as a condition,
2320 unpredictable results occurred, instead of the error return
2321 PCRE_ERROR_DFA_UITEM.
2322
2323 5. The C++ GlobalReplace function was not working like Perl for the special
2324 situation when an empty string is matched. It now does the fancy magic
2325 stuff that is necessary.
2326
2327 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
2328 removed. (These were left over from very, very early versions of PCRE.)
2329
2330 7. Some cosmetic changes to the code to make life easier when compiling it
2331 as part of something else:
2332
2333 (a) Change DEBUG to PCRE_DEBUG.
2334
2335 (b) In pcre_compile(), rename the member of the "branch_chain" structure
2336 called "current" as "current_branch", to prevent a collision with the
2337 Linux macro when compiled as a kernel module.
2338
2339 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
2340 prevent a collision with the Linux macro when compiled as a kernel
2341 module.
2342
2343 8. In pcre_compile() there are some checks for integer overflows that used to
2344 cast potentially large values to (double). This has been changed to that
2345 when building, a check for int64_t is made, and if it is found, it is used
2346 instead, thus avoiding the use of floating point arithmetic. (There is no
2347 other use of FP in PCRE.) If int64_t is not found, the fallback is to
2348 double.
2349
2350 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
2351 2005 (difference between two addresses compared to an unsigned value).
2352
2353 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
2354 custom one, because of the following reported problem in Windows:
2355
2356 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
2357 under Win32.
2358 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
2359 therefore missing the function definition.
2360 - The compiler thus generates a "C" signature for the test function.
2361 - The linker fails to find the "C" function.
2362 - PCRE fails to configure if asked to do so against libbz2.
2363
2364 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
2365 messages were output:
2366
2367 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
2368 rerunning libtoolize, to keep the correct libtool macros in-tree.
2369 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
2370
2371 I have done both of these things.
2372
2373 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
2374 most of the time, it *can* run out if it is given a pattern that contains a
2375 runaway infinite recursion. I updated the discussion in the pcrestack man
2376 page.
2377
2378 13. Now that we have gone to the x.xx style of version numbers, the minor
2379 version may start with zero. Using 08 or 09 is a bad idea because users
2380 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
2381 interpreted as invalid octal numbers. I've updated the previous comment in
2382 configure.ac, and also added a check that gives an error if 08 or 09 are
2383 used.
2384
2385 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
2386 causing partial matching to fail when the end of the subject matched \W
2387 in a UTF-8 pattern where \W was quantified with a minimum of 3.
2388
2389 15. There were some discrepancies between the declarations in pcre_internal.h
2390 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
2391 their definitions. The declarations used "const uschar *" and the
2392 definitions used USPTR. Even though USPTR is normally defined as "const
2393 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
2394 reported that: "This difference in casting confuses some C++ compilers, for
2395 example, SunCC recognizes above declarations as different functions and
2396 generates broken code for hbpcre." I have changed the declarations to use
2397 USPTR.
2398
2399 16. GNU libtool is named differently on some systems. The autogen.sh script now
2400 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
2401 (FreeBSD).
2402
2403 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2404 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2405 comment: "Figure out how to create a longlong from a string: strtoll and
2406 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2407 instance, but it only takes 2 args instead of 3!"
2408
2409 18. A subtle bug concerned with back references has been fixed by a change of
2410 specification, with a corresponding code fix. A pattern such as
2411 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2412 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2413 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2414 same bug. Such groups have to be quantified to be useful, or contained
2415 inside another quantified group. (If there's no repetition, the reference
2416 can never match.) The problem arises because, having left the group and
2417 moved on to the rest of the pattern, a later failure that backtracks into
2418 the group uses the captured value from the final iteration of the group
2419 rather than the correct earlier one. I have fixed this in PCRE by forcing
2420 any group that contains a reference to itself to be an atomic group; that
2421 is, there cannot be any backtracking into it once it has completed. This is
2422 similar to recursive and subroutine calls.
2423
2424
2425 Version 8.00 19-Oct-09
2426 ----------------------
2427
2428 1. The table for translating pcre_compile() error codes into POSIX error codes
2429 was out-of-date, and there was no check on the pcre_compile() error code
2430 being within the table. This could lead to an OK return being given in
2431 error.
2432
2433 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2434 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2435 in a Windows environment.
2436
2437 3. The pcregrep --count option prints the count for each file even when it is
2438 zero, as does GNU grep. However, pcregrep was also printing all files when
2439 --files-with-matches was added. Now, when both options are given, it prints
2440 counts only for those files that have at least one match. (GNU grep just
2441 prints the file name in this circumstance, but including the count seems
2442 more useful - otherwise, why use --count?) Also ensured that the
2443 combination -clh just lists non-zero counts, with no names.
2444
2445 4. The long form of the pcregrep -F option was incorrectly implemented as
2446 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2447 but it seems right to fix it, and I didn't think it was worth preserving
2448 the old behaviour.
2449
2450 5. The command line items --regex=pattern and --regexp=pattern were not
2451 recognized by pcregrep, which required --regex pattern or --regexp pattern
2452 (with a space rather than an '='). The man page documented the '=' forms,
2453 which are compatible with GNU grep; these now work.
2454
2455 6. No libpcreposix.pc file was created for pkg-config; there was just
2456 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2457
2458 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2459 when UCP support is not needed, by modifying the Python script that
2460 generates it from Unicode data files. This should not matter if the module
2461 is correctly used as a library, but I received one complaint about 50K of
2462 unwanted data. My guess is that the person linked everything into his
2463 program rather than using a library. Anyway, it does no harm.
2464
2465 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2466 was a minimum greater than 1 for a wide character in a possessive
2467 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2468 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2469 character. Chaos in the form of incorrect output or a compiling loop could
2470 result.
2471
2472 9. The restrictions on what a pattern can contain when partial matching is
2473 requested for pcre_exec() have been removed. All patterns can now be
2474 partially matched by this function. In addition, if there are at least two
2475 slots in the offset vector, the offset of the earliest inspected character
2476 for the match and the offset of the end of the subject are set in them when
2477 PCRE_ERROR_PARTIAL is returned.
2478
2479 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2480 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2481 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2482 and may be more useful for multi-segment matching.
2483
2484 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2485 used to be given if ever the end of the subject was reached; now it is
2486 given only if matching could not proceed because another character was
2487 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2488 string "Z", which now yields "no match" instead of "partial match". In the
2489 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2490 final character ended with (*FAIL).
2491
2492 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2493 if the pattern had a "must contain" character that was already found in the
2494 earlier partial match, unless partial matching was again requested. For
2495 example, with the pattern /dog.(body)?/, the "must contain" character is
2496 "g". If the first part-match was for the string "dog", restarting with
2497 "sbody" failed. This bug has been fixed.
2498
2499 13. The string returned by pcre_dfa_exec() after a partial match has been
2500 changed so that it starts at the first inspected character rather than the
2501 first character of the match. This makes a difference only if the pattern
2502 starts with a lookbehind assertion or \b or \B (\K is not supported by
2503 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2504 matching functions compatible, and I think it's the right thing to do.
2505
2506 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2507 so that the demonstration program is easily available in environments where
2508 PCRE has not been installed from source.
2509
2510 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2511 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2512 library.
2513
2514 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2515 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2516 is not the first non-POSIX option to be added. Clearly some people find
2517 these options useful.
2518
2519 17. If a caller to the POSIX matching function regexec() passes a non-zero
2520 value for nmatch with a NULL value for pmatch, the value of
2521 nmatch is forced to zero.
2522
2523 18. RunGrepTest did not have a test for the availability of the -u option of
2524 the diff command, as RunTest does. It now checks in the same way as
2525 RunTest, and also checks for the -b option.
2526
2527 19. If an odd number of negated classes containing just a single character
2528 interposed, within parentheses, between a forward reference to a named
2529 subpattern and the definition of the subpattern, compilation crashed with
2530 an internal error, complaining that it could not find the referenced
2531 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2532 [The bug was that it was starting one character too far in when skipping
2533 over the character class, thus treating the ] as data rather than
2534 terminating the class. This meant it could skip too much.]
2535
2536 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2537 /g option in pcretest when the pattern contains \K, which makes it possible
2538 to have an empty string match not at the start, even when the pattern is
2539 anchored. Updated pcretest and pcredemo to use this option.
2540
2541 21. If the maximum number of capturing subpatterns in a recursion was greater
2542 than the maximum at the outer level, the higher number was returned, but
2543 with unset values at the outer level. The correct (outer level) value is
2544 now given.
2545
2546 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2547 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2548 make it do so. The string so far is captured, making this feature
2549 compatible with Perl.
2550
2551 23. The tests have been re-organized, adding tests 11 and 12, to make it
2552 possible to check the Perl 5.10 features against Perl 5.10.
2553
2554 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2555 pattern matches a fixed length string. PCRE did not allow this; now it
2556 does. Neither allows recursion.
2557
2558 25. I finally figured out how to implement a request to provide the minimum
2559 length of subject string that was needed in order to match a given pattern.
2560 (It was back references and recursion that I had previously got hung up
2561 on.) This code has now been added to pcre_study(); it finds a lower bound
2562 to the length of subject needed. It is not necessarily the greatest lower
2563 bound, but using it to avoid searching strings that are too short does give
2564 some useful speed-ups. The value is available to calling programs via
2565 pcre_fullinfo().
2566
2567 26. While implementing 25, I discovered to my embarrassment that pcretest had
2568 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2569 study optimizations had never been tested with that matching function.
2570 Oops. What is worse, even when it was passed study data, there was a bug in
2571 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2572 were also very few tests of studied patterns with pcre_dfa_exec().
2573
2574 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2575 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2576 on the other side of the coin, they are no longer allowed to have different
2577 names, because these cannot be distinguished in PCRE, and this has caused
2578 confusion. (This is a difference from Perl.)
2579
2580 28. When duplicate subpattern names are present (necessarily with different
2581 numbers, as required by 27 above), and a test is made by name in a
2582 conditional pattern, either for a subpattern having been matched, or for
2583 recursion in such a pattern, all the associated numbered subpatterns are
2584 tested, and the overall condition is true if the condition is true for any
2585 one of them. This is the way Perl works, and is also more like the way
2586 testing by number works.
2587
2588
2589 Version 7.9 11-Apr-09
2590 ---------------------
2591
2592 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2593 (pcretest), all targets were linked against these libraries. This included
2594 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2595 libraries. This caused unwanted dependencies to be created. This problem
2596 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2597 pcretest is linked with readline.
2598
2599 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2600 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2601 moved outside it again, because FALSE and TRUE are already defined in AIX,
2602 but BOOL is not.
2603
2604 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2605 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2606
2607 4. The pcregrep documentation said spaces were inserted as well as colons (or
2608 hyphens) following file names and line numbers when outputting matching
2609 lines. This is not true; no spaces are inserted. I have also clarified the
2610 wording for the --colour (or --color) option.
2611
2612 5. In pcregrep, when --colour was used with -o, the list of matching strings
2613 was not coloured; this is different to GNU grep, so I have changed it to be
2614 the same.
2615
2616 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2617 each matching line was coloured. Now it goes on to look for further matches
2618 of any of the test patterns, which is the same behaviour as GNU grep.
2619
2620 7. A pattern that could match an empty string could cause pcregrep to loop; it
2621 doesn't make sense to accept an empty string match in pcregrep, so I have
2622 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2623 seems to be how GNU grep behaves. [But see later change 40 for release
2624 8.33.]
2625
2626 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2627 start or after a newline", because the conditional assertion was not being
2628 correctly handled. The rule now is that both the assertion and what follows
2629 in the first alternative must satisfy the test.
2630
2631 9. If auto-callout was enabled in a pattern with a conditional group whose
2632 condition was an assertion, PCRE could crash during matching, both with
2633 pcre_exec() and pcre_dfa_exec().
2634
2635 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2636 used for matching.
2637
2638 11. Unicode property support in character classes was not working for
2639 characters (bytes) greater than 127 when not in UTF-8 mode.
2640
2641 12. Added the -M command line option to pcretest.
2642
2643 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2644
2645 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2646
2647 16. Added comments and documentation about mis-use of no_arg in the C++
2648 wrapper.
2649
2650 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2651 from Martin Jerabek that uses macro names for all relevant character and
2652 string constants.
2653
2654 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2655 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2656 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2657 these, but not everybody uses configure.
2658
2659 19. A conditional group that had only one branch was not being correctly
2660 recognized as an item that could match an empty string. This meant that an
2661 enclosing group might also not be so recognized, causing infinite looping
2662 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2663 with the subject "ab", where knowledge that the repeated group can match
2664 nothing is needed in order to break the loop.
2665
2666 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2667 exec(), but without supplying a callout function, matching went wrong.
2668
2669 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2670 leak if the size of the offset vector was greater than 30. When the vector
2671 is smaller, the saved offsets during recursion go onto a local stack
2672 vector, but for larger vectors malloc() is used. It was failing to free
2673 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2674 error, in fact).
2675
2676 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2677 heapframe that is used only when UTF-8 support is enabled. This caused no
2678 problem, but was untidy.
2679
2680 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2681 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2682 included within another project.
2683
2684 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2685 slightly modified by me:
2686
2687 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2688 not building pcregrep.
2689
2690 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2691 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2692
2693 25. Forward references, both numeric and by name, in patterns that made use of
2694 duplicate group numbers, could behave incorrectly or give incorrect errors,
2695 because when scanning forward to find the reference group, PCRE was not
2696 taking into account the duplicate group numbers. A pattern such as
2697 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2698
2699 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2700 the feature of a custom pointer more persuasive (as requested by a user).
2701
2702 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2703 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2704 pre-defined.
2705
2706 28. Added support for (*UTF8) at the start of a pattern.
2707
2708 29. Arrange for flags added by the "release type" setting in CMake to be shown
2709 in the configuration summary.
2710
2711
2712 Version 7.8 05-Sep-08
2713 ---------------------
2714
2715 1. Replaced UCP searching code with optimized version as implemented for Ad
2716 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2717 stage table and inline lookup instead of a function, giving speed ups of 2
2718 to 5 times on some simple patterns that I tested. Permission was given to
2719 distribute the MultiStage2.py script that generates the tables (it's not in
2720 the tarball, but is in the Subversion repository).
2721
2722 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2723 scripts.
2724
2725 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2726 a group with a zero qualifier. The result of the study could be incorrect,
2727 or the function might crash, depending on the pattern.
2728
2729 4. Caseless matching was not working for non-ASCII characters in back
2730 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2731 It now works when Unicode Property Support is available.
2732
2733 5. In pcretest, an escape such as \x{de} in the data was always generating
2734 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2735 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2736 truncation.
2737
2738 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2739
2740 7. Added two (int) casts to pcregrep when printing the difference of two
2741 pointers, in case they are 64-bit values.
2742
2743 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2744 test 2 if it fails.
2745
2746 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2747 and a #define of that name to empty if it is not externally set. This is to
2748 allow users of MSVC to set it if necessary.
2749
2750 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2751 the convenience functions in the pcre_get.c source file.
2752
2753 11. An option change at the start of a pattern that had top-level alternatives
2754 could cause overwriting and/or a crash. This command provoked a crash in
2755 some environments:
2756
2757 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2758
2759 This potential security problem was recorded as CVE-2008-2371.
2760
2761 12. For a pattern where the match had to start at the beginning or immediately
2762 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2763 pcre_dfa_exec() could read past the end of the passed subject if there was
2764 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2765 pcretest so that it places the subject at the end of its malloc-ed buffer.
2766
2767 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2768 exec() might read past the end of the data buffer in UTF-8 mode.
2769
2770 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2771 the data contained the byte 0x85 as part of a UTF-8 character within its
2772 first line. This applied both to normal and DFA matching.
2773
2774 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2775 /^[^d]*?$/8 failed to match "abc".
2776
2777 16. Added a missing copyright notice to pcrecpp_internal.h.
2778
2779 17. Make it more clear in the documentation that values returned from
2780 pcre_exec() in ovector are byte offsets, not character counts.
2781
2782 18. Tidied a few places to stop certain compilers from issuing warnings.
2783
2784 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2785 supplied by Stefan Weber. I made a further small update for 7.8 because
2786 there is a change of source arrangements: the pcre_searchfuncs.c module is
2787 replaced by pcre_ucd.c.
2788
2789
2790 Version 7.7 07-May-08
2791 ---------------------
2792
2793 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2794 a string to a long long, pretend we don't even have a long long." This is
2795 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2796
2797 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2798 pre-7.6 versions, which defined a global no_arg variable instead of putting
2799 it in the RE class. (See also #8 below.)
2800
2801 3. Remove a line of dead code, identified by coverity and reported by Nuno
2802 Lopes.
2803
2804 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2805
2806 (1) The include/exclude patterns were being applied to the whole pathnames
2807 of files, instead of just to the final components.
2808
2809 (2) If there was more than one level of directory, the subdirectories were
2810 skipped unless they satisfied the include/exclude conditions. This is
2811 inconsistent with GNU grep (and could even be seen as contrary to the
2812 pcregrep specification - which I improved to make it absolutely clear).
2813 The action now is always to scan all levels of directory, and just
2814 apply the include/exclude patterns to regular files.
2815
2816 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2817 --exclude_dir in the tests to avoid scanning .svn directories.
2818
2819 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2820 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2821 doesn't support NULs in patterns.
2822
2823 7. Added some missing "const"s to declarations of static tables in
2824 pcre_compile.c and pcre_dfa_exec.c.
2825
2826 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2827 caused by fix #2 above. (Subsequently also a second patch to fix the
2828 first patch. And a third patch - this was a messy problem.)
2829
2830 9. Applied Craig's patch to remove the use of push_back().
2831
2832 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2833 matching function regexec().
2834
2835 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2836 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2837 references. PCRE supports relative numbers with this syntax (I don't think
2838 Oniguruma does).
2839
2840 12. Previously, a group with a zero repeat such as (...){0} was completely
2841 omitted from the compiled regex. However, this means that if the group
2842 was called as a subroutine from elsewhere in the pattern, things went wrong
2843 (an internal error was given). Such groups are now left in the compiled
2844 pattern, with a new opcode that causes them to be skipped at execution
2845 time.
2846
2847 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2848 to the way PCRE behaves:
2849
2850 (a) A lone ] character is dis-allowed (Perl treats it as data).
2851
2852 (b) A back reference to an unmatched subpattern matches an empty string
2853 (Perl fails the current match path).
2854
2855 (c) A data ] in a character class must be notated as \] because if the
2856 first data character in a class is ], it defines an empty class. (In
2857 Perl it is not possible to have an empty class.) The empty class []
2858 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2859 The negative empty class [^] matches any one character, independently
2860 of the DOTALL setting.
2861
2862 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2863 non-existent subpattern following a character class starting with ']' and
2864 containing () gave an internal compiling error instead of "reference to
2865 non-existent subpattern". Fortunately, when the pattern did exist, the
2866 compiled code was correct. (When scanning forwards to check for the
2867 existence of the subpattern, it was treating the data ']' as terminating
2868 the class, so got the count wrong. When actually compiling, the reference
2869 was subsequently set up correctly.)
2870
2871 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2872 it was being rejected as not supported by pcre_dfa_exec(), even though
2873 other assertions are supported. I have made pcre_dfa_exec() support
2874 (*FAIL).
2875
2876 16. The implementation of 13c above involved the invention of a new opcode,
2877 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2878 cannot be changed at match time, I realized I could make a small
2879 improvement to matching performance by compiling OP_ALLANY instead of
2880 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2881 on the OP_ANY path.
2882
2883 17. Compiling pcretest on Windows with readline support failed without the
2884 following two fixes: (1) Make the unistd.h include conditional on
2885 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2886
2887 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2888 ncurses library to be included for pcretest when ReadLine support is
2889 requested, but also to allow for it to be overridden. This patch came from
2890 Daniel Bergström.
2891
2892 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2893 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2894 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2895 spotting this.
2896
2897
2898 Version 7.6 28-Jan-08
2899 ---------------------
2900
2901 1. A character class containing a very large number of characters with
2902 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2903 overflow.
2904
2905 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2906 HAVE_LONG_LONG is not defined.
2907
2908 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2909 bring them up to date and include new features. This patch includes:
2910
2911 - Fixed PH's badly added libz and libbz2 support.
2912 - Fixed a problem with static linking.
2913 - Added pcredemo. [But later removed - see 7 below.]
2914 - Fixed dftables problem and added an option.
2915 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2916 HAVE_LONG_LONG.
2917 - Added readline support for pcretest.
2918 - Added an listing of the option settings after cmake has run.
2919
2920 4. A user submitted a patch to Makefile that makes it easy to create
2921 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2922 Makefile.am that cause it to include this special target, without
2923 affecting anything else. Note that the same mingw target plus all
2924 the other distribution libraries and programs are now supported
2925 when configuring with CMake (see 6 below) instead of with
2926 Configure/Make.
2927
2928 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2929 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2930 exported in the Windows port". It has not yet been confirmed that the patch
2931 solves the problem, but it does no harm.
2932
2933 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2934 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2935 with CMake, and also correct the comment about stack recursion.
2936
2937 7. Remove the automatic building of pcredemo from the ./configure system and
2938 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2939 of a program that users should build themselves after PCRE is installed, so
2940 building it automatically is not really right. What is more, it gave
2941 trouble in some build environments.
2942
2943 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2944
2945
2946 Version 7.5 10-Jan-08
2947 ---------------------
2948
2949 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2950 values in parens when parsing an RE using the C++ wrapper."
2951
2952 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2953 Characters greater than 255 were excluded from the class instead of being
2954 included.
2955
2956 3. The same bug as (2) above applied to negated POSIX classes such as
2957 [:^space:].
2958
2959 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2960 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2961 I have changed it.
2962
2963 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2964 first named subpattern) and a construct such as (?&a) would reference the
2965 first named subpattern whose name started with "a" (in other words, the
2966 length check was missing). Both these problems are fixed. "Subpattern name
2967 expected" is now given for (?&) (a zero-length name), and this patch also
2968 makes it give the same error for \k'' (previously it complained that that
2969 was a reference to a non-existent subpattern).
2970
2971 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2972 this is right because (?- can be followed by option settings as well as by
2973 digits. I have, however, made the messages clearer.
2974
2975 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2976 than the number used in the conditional) now cause a compile-time error.
2977 This is actually not compatible with Perl, which accepts such patterns, but
2978 treats the conditional as always being FALSE (as PCRE used to), but it
2979 seems to me that giving a diagnostic is better.
2980
2981 8. Change "alphameric" to the more common word "alphanumeric" in comments
2982 and messages.
2983
2984 9. Fix two occurrences of "backslash" in comments that should have been
2985 "backspace".
2986
2987 10. Remove two redundant lines of code that can never be obeyed (their function
2988 was moved elsewhere).
2989
2990 11. The program that makes PCRE's Unicode character property table had a bug
2991 which caused it to generate incorrect table entries for sequences of
2992 characters that have the same character type, but are in different scripts.
2993 It amalgamated them into a single range, with the script of the first of
2994 them. In other words, some characters were in the wrong script. There were
2995 thirteen such cases, affecting characters in the following ranges:
2996
2997 U+002b0 - U+002c1
2998 U+0060c - U+0060d
2999 U+0061e - U+00612
3000 U+0064b - U+0065e
3001 U+0074d - U+0076d
3002 U+01800 - U+01805
3003 U+01d00 - U+01d77
3004 U+01d9b - U+01dbf
3005 U+0200b - U+0200f
3006 U+030fc - U+030fe
3007 U+03260 - U+0327f
3008 U+0fb46 - U+0fbb1
3009 U+10450 - U+1049d
3010
3011 12. The -o option (show only the matching part of a line) for pcregrep was not
3012 compatible with GNU grep in that, if there was more than one match in a
3013 line, it showed only the first of them. It now behaves in the same way as
3014 GNU grep.
3015
3016 13. If the -o and -v options were combined for pcregrep, it printed a blank
3017 line for every non-matching line. GNU grep prints nothing, and pcregrep now
3018 does the same. The return code can be used to tell if there were any
3019 non-matching lines.
3020
3021 14. Added --file-offsets and --line-offsets to pcregrep.
3022
3023 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
3024 infinitely looping recursion. The bug was that positive lookaheads were not
3025 being skipped when checking for a possible empty match (negative lookaheads
3026 and both kinds of lookbehind were skipped).
3027
3028 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
3029 inclusion of <windows.h> to before rather than after the definition of
3030 INVALID_FILE_ATTRIBUTES (patch from David Byron).
3031
3032 17. Specifying a possessive quantifier with a specific limit for a Unicode
3033 character property caused pcre_compile() to compile bad code, which led at
3034 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
3035 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
3036 caused the error; without that there was no problem.
3037
3038 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
3039
3040 19. Added --enable-pcretest-libreadline.
3041
3042 20. In pcrecpp.cc, the variable 'count' was incremented twice in
3043 RE::GlobalReplace(). As a result, the number of replacements returned was
3044 double what it should be. I removed one of the increments, but Craig sent a
3045 later patch that removed the other one (the right fix) and added unit tests
3046 that check the return values (which was not done before).
3047
3048 21. Several CMake things:
3049
3050 (1) Arranged that, when cmake is used on Unix, the libraries end up with
3051 the names libpcre and libpcreposix, not just pcre and pcreposix.
3052
3053 (2) The above change means that pcretest and pcregrep are now correctly
3054 linked with the newly-built libraries, not previously installed ones.
3055
3056 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
3057
3058 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
3059 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
3060 UTF-8 newline character). The key issue is that the pattern starts .*;
3061 this means that the match must be either at the beginning, or after a
3062 newline. The bug was in the code for advancing after a failed match and
3063 checking that the new position followed a newline. It was not taking
3064 account of UTF-8 characters correctly.
3065
3066 23. PCRE was behaving differently from Perl in the way it recognized POSIX
3067 character classes. PCRE was not treating the sequence [:...:] as a
3068 character class unless the ... were all letters. Perl, however, seems to
3069 allow any characters between [: and :], though of course it rejects as
3070 unknown any "names" that contain non-letters, because all the known class
3071 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
3072 for example, whereas PCRE did not - it did not recognize a POSIX character
3073 class. This seemed a bit dangerous, so the code has been changed to be
3074 closer to Perl. The behaviour is not identical to Perl, because PCRE will
3075 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
3076 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
3077 Perl does, and where it didn't before.
3078
3079 24. Rewrite so as to remove the single use of %n from pcregrep because in some
3080 Windows environments %n is disabled by default.
3081
3082
3083 Version 7.4 21-Sep-07
3084 ---------------------
3085
3086 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
3087 means that a class such as [\s] counted as "explicit reference to CR or
3088 LF". That isn't really right - the whole point of the change was to try to
3089 help when there was an actual mention of one of the two characters. So now
3090 the change happens only if \r or \n (or a literal CR or LF) character is
3091 encountered.
3092
3093 2. The 32-bit options word was also used for 6 internal flags, but the numbers
3094 of both had grown to the point where there were only 3 bits left.
3095 Fortunately, there was spare space in the data structure, and so I have
3096 moved the internal flags into a new 16-bit field to free up more option
3097 bits.
3098
3099 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
3100 but did not set the internal JCHANGED flag - either of these is enough to
3101 control the way the "get" function works - but the PCRE_INFO_JCHANGED
3102 facility is supposed to tell if (?J) was ever used, so now (?J) at the
3103 start sets both bits.
3104
3105 4. Added options (at build time, compile time, exec time) to change \R from
3106 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
3107
3108 5. doc/pcresyntax.html was missing from the distribution.
3109
3110 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
3111 compatibility, even though it is no longer used.
3112
3113 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
3114 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
3115 windows.h file is present (where different names are used). [This was
3116 reversed later after testing - see 16 below.]
3117
3118 8. Changed all #include <config.h> to #include "config.h". There were also
3119 some further <pcre.h> cases that I changed to "pcre.h".
3120
3121 9. When pcregrep was used with the --colour option, it missed the line ending
3122 sequence off the lines that it output.
3123
3124 10. It was pointed out to me that arrays of string pointers cause lots of
3125 relocations when a shared library is dynamically loaded. A technique of
3126 using a single long string with a table of offsets can drastically reduce
3127 these. I have refactored PCRE in four places to do this. The result is
3128 dramatic:
3129
3130 Originally: 290
3131 After changing UCP table: 187
3132 After changing error message table: 43
3133 After changing table of "verbs" 36
3134 After changing table of Posix names 22
3135
3136 Thanks to the folks working on Gregex for glib for this insight.
3137
3138 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
3139 unicode-properties was also set.
3140
3141 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
3142
3143 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
3144 checked only for CRLF.
3145
3146 14. Added casts to pcretest.c to avoid compiler warnings.
3147
3148 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
3149
3150 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
3151 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
3152 entirely. This removes changes made in 7 above.
3153
3154 17. The CMake files have been updated, and there is now more information about
3155 building with CMake in the NON-UNIX-USE document.
3156
3157
3158 Version 7.3 28-Aug-07
3159 ---------------------
3160
3161 1. In the rejigging of the build system that eventually resulted in 7.1, the
3162 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
3163 brackets there is not right, since it causes compilers to look for an
3164 installed pcre.h, not the version that is in the source that is being
3165 compiled (which of course may be different). I have changed it back to:
3166
3167 #include "pcre.h"
3168
3169 I have a vague recollection that the change was concerned with compiling in
3170 different directories, but in the new build system, that is taken care of
3171 by the VPATH setting the Makefile.
3172
3173 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
3174 when the subject happened to end in the byte 0x85 (e.g. if the last
3175 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
3176 characters but of course it shouldn't be taken as a newline when it is part
3177 of another character. The bug was that, for an unlimited repeat of . in
3178 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
3179 characters when looking for a newline.
3180
3181 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
3182
3183 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
3184 in debug output.
3185
3186 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
3187 long printing in the pcrecpp unittest when running under MinGW.
3188
3189 6. ESC_K was left out of the EBCDIC table.
3190
3191 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
3192 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
3193 limit also applies to "virtual nesting" when a pattern is recursive, and in
3194 this case 1000 isn't so big. I have been able to remove this limit at the
3195 expense of backing off one optimization in certain circumstances. Normally,
3196 when pcre_exec() would call its internal match() function recursively and
3197 immediately return the result unconditionally, it uses a "tail recursion"
3198 feature to save stack. However, when a subpattern that can match an empty
3199 string has an unlimited repetition quantifier, it no longer makes this
3200 optimization. That gives it a stack frame in which to save the data for
3201 checking that an empty string has been matched. Previously this was taken
3202 from the 1000-entry workspace that had been reserved. So now there is no
3203 explicit limit, but more stack is used.
3204
3205 8. Applied Daniel's patches to solve problems with the import/export magic
3206 syntax that is required for Windows, and which was going wrong for the
3207 pcreposix and pcrecpp parts of the library. These were overlooked when this
3208 problem was solved for the main library.
3209
3210 9. There were some crude static tests to avoid integer overflow when computing
3211 the size of patterns that contain repeated groups with explicit upper
3212 limits. As the maximum quantifier is 65535, the maximum group length was
3213 set at 30,000 so that the product of these two numbers did not overflow a
3214 32-bit integer. However, it turns out that people want to use groups that
3215 are longer than 30,000 bytes (though not repeat them that many times).
3216 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
3217 made it possible to implement the integer overflow checks in a much more
3218 dynamic way, which I have now done. The artificial limitation on group
3219 length has been removed - we now have only the limit on the total length of
3220 the compiled pattern, which depends on the LINK_SIZE setting.
3221
3222 10. Fixed a bug in the documentation for get/copy named substring when
3223 duplicate names are permitted. If none of the named substrings are set, the
3224 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
3225 empty string.
3226
3227 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
3228 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
3229 because the ] is interpreted as the first data character and the
3230 terminating ] is not found. PCRE has been made compatible with Perl in this
3231 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
3232 cause memory overwriting.
3233
3234 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
3235 string has been matched (to stop an infinite loop). It was not recognizing
3236 a conditional subpattern that could match an empty string if that
3237 subpattern was within another subpattern. For example, it looped when
3238 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
3239 condition was not nested. This bug has been fixed.
3240
3241 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
3242 past the start of the subject in the presence of bytes with the top bit
3243 set, for example "\x8aBCD".
3244
3245 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
3246 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
3247
3248 14. Optimized (?!) to (*FAIL).
3249
3250 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
3251 This restricts code points to be within the range 0 to 0x10FFFF, excluding
3252 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
3253 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
3254 does: it's just the validity check that is more restrictive.
3255
3256 16. Inserted checks for integer overflows during escape sequence (backslash)
3257 processing, and also fixed erroneous offset values for syntax errors during
3258 backslash processing.
3259
3260 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
3261 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
3262
3263 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
3264 caused an overrun.
3265
3266 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
3267 something other than just ASCII characters) inside a group that had an
3268 unlimited repeat caused a loop at compile time (while checking to see
3269 whether the group could match an empty string).
3270
3271 20. Debugging a pattern containing \p or \P could cause a crash. For example,
3272 [\P{Any}] did so. (Error in the code for printing property names.)
3273
3274 21. An orphan \E inside a character class could cause a crash.
3275
3276 22. A repeated capturing bracket such as (A)? could cause a wild memory
3277 reference during compilation.
3278
3279 23. There are several functions in pcre_compile() that scan along a compiled
3280 expression for various reasons (e.g. to see if it's fixed length for look
3281 behind). There were bugs in these functions when a repeated \p or \P was
3282 present in the pattern. These operators have additional parameters compared
3283 with \d, etc, and these were not being taken into account when moving along
3284 the compiled data. Specifically:
3285
3286 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
3287 length.
3288
3289 (b) An item such as \pL+ within a repeated group could cause crashes or
3290 loops.
3291
3292 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
3293 "reference to non-existent subpattern" error.
3294
3295 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
3296
3297 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
3298 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
3299
3300 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
3301
3302 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
3303 character were causing crashes (broken optimization).
3304
3305 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
3306 \p or \P) caused a compile-time loop.
3307
3308 28. More problems have arisen in unanchored patterns when CRLF is a valid line
3309 break. For example, the unstudied pattern [\r\n]A does not match the string
3310 "\r\nA" because change 7.0/46 below moves the current point on by two
3311 characters after failing to match at the start. However, the pattern \nA
3312 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
3313 the same is true. There doesn't seem any very clean way out of this, but
3314 what I have chosen to do makes the common cases work: PCRE now takes note
3315 of whether there can be an explicit match for \r or \n anywhere in the
3316 pattern, and if so, 7.0/46 no longer applies. As part of this change,
3317 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
3318 pattern has explicit CR or LF references.
3319
3320 29. Added (*CR) etc for changing newline setting at start of pattern.
3321
3322
3323 Version 7.2 19-Jun-07
3324 ---------------------
3325
3326 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
3327 which is apparently normally available under Windows.
3328
3329 2. Re-jig the pcregrep tests with different newline settings in an attempt
3330 to make them independent of the local environment's newline setting.
3331
3332 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
3333
3334 4. Some of the "internals" tests were previously cut out when the link size
3335 was not 2, because the output contained actual offsets. The recent new
3336 "Z" feature of pcretest means that these can be cut out, making the tests
3337 usable with all link sizes.
3338
3339 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
3340 stack recursion. This gives a massive performance boost under BSD, but just
3341 a small improvement under Linux. However, it saves one field in the frame
3342 in all cases.
3343
3344 6. Added more features from the forthcoming Perl 5.10:
3345
3346 (a) (?-n) (where n is a string of digits) is a relative subroutine or
3347 recursion call. It refers to the nth most recently opened parentheses.
3348
3349 (b) (?+n) is also a relative subroutine call; it refers to the nth next
3350 to be opened parentheses.
3351
3352 (c) Conditions that refer to capturing parentheses can be specified
3353 relatively, for example, (?(-2)... or (?(+3)...
3354
3355 (d) \K resets the start of the current match so that everything before
3356 is not part of it.
3357
3358 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
3359
3360 (f) \g{name} is another synonym - part of Perl 5.10's unification of
3361 reference syntax.
3362
3363 (g) (?| introduces a group in which the numbering of parentheses in each
3364 alternative starts with the same number.
3365
3366 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
3367
3368 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
3369 PCRE_INFO_JCHANGED.
3370
3371 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
3372 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
3373 for detecting groups that can match an empty string.
3374
3375 9. A pattern with a very large number of alternatives (more than several
3376 hundred) was running out of internal workspace during the pre-compile
3377 phase, where pcre_compile() figures out how much memory will be needed. A
3378 bit of new cunning has reduced the workspace needed for groups with
3379 alternatives. The 1000-alternative test pattern now uses 12 bytes of
3380 workspace instead of running out of the 4096 that are available.
3381
3382 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
3383
3384 11. Applied patch from Google to remove an optimization that didn't quite work.
3385 The report of the bug said:
3386
3387 pcrecpp::RE("a*").FullMatch("aaa") matches, while
3388 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
3389 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
3390
3391 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
3392 it matched the wrong number of bytes.
3393
3394
3395 Version 7.1 24-Apr-07
3396 ---------------------
3397
3398 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
3399 that is more "standard", making use of automake and other Autotools. There
3400 is some re-arrangement of the files and adjustment of comments consequent
3401 on this.
3402
3403 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3404 for recursive directory scanning broke on some systems because the files
3405 are not scanned in any specific order and on different systems the order
3406 was different. A call to "sort" has been inserted into RunGrepTest for the
3407 approprate test as a short-term fix. In the longer term there may be an
3408 alternative.
3409
3410 3. I had an email from Eric Raymond about problems translating some of PCRE's
3411 man pages to HTML (despite the fact that I distribute HTML pages, some
3412 people do their own conversions for various reasons). The problems
3413 concerned the use of low-level troff macros .br and .in. I have therefore
3414 removed all such uses from the man pages (some were redundant, some could
3415 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3416 HTML has been updated to handle .nf/.fi and to complain if it encounters
3417 .br or .in.
3418
3419 4. Updated comments in configure.ac that get placed in config.h.in and also
3420 arranged for config.h to be included in the distribution, with the name
3421 config.h.generic, for the benefit of those who have to compile without
3422 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3423
3424 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3425 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3426 makevp.bat for the current PCRE, using the additional files
3427 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3428
3429 6. A Windows user reported a minor discrepancy with test 2, which turned out
3430 to be caused by a trailing space on an input line that had got lost in his
3431 copy. The trailing space was an accident, so I've just removed it.
3432
3433 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3434 that is needed.
3435
3436 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3437 as "const" (a) because they are and (b) because it helps the PHP
3438 maintainers who have recently made a script to detect big data structures
3439 in the php code that should be moved to the .rodata section. I remembered
3440 to update Builducptable as well, so it won't revert if ucptable.h is ever
3441 re-created.
3442
3443 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3444 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3445 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3446 support is not required. This saves 1.5-2K of code, which is important in
3447 some applications.
3448
3449 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3450 so as not to refer to the tables, even though these functions will never be
3451 called when UTF-8 support is disabled. Otherwise there are problems with a
3452 shared library.
3453
3454 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3455
3456 (a) It was defining its arguments as char * instead of void *.
3457
3458 (b) It was assuming that all moves were upwards in memory; this was true
3459 a long time ago when I wrote it, but is no longer the case.
3460
3461 The emulated memove() is provided for those environments that have neither
3462 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3463 is clearly not the case, as these two bugs were recently reported.
3464
3465 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3466 and Detrail to create the HTML documentation, the .txt form of the man
3467 pages, and it removes trailing spaces from listed files. It also creates
3468 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3469 case, it wraps all the #defines with #ifndefs. This script should be run
3470 before "make dist".
3471
3472 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3473 with Unicode property support.
3474
3475 (a) For a maximizing quantifier, if the two different cases of the
3476 character were of different lengths in their UTF-8 codings (there are
3477 some cases like this - I found 11), and the matching function had to
3478 back up over a mixture of the two cases, it incorrectly assumed they
3479 were both the same length.
3480
3481 (b) When PCRE was configured to use the heap rather than the stack for
3482 recursion during matching, it was not correctly preserving the data for
3483 the other case of a UTF-8 character when checking ahead for a match
3484 while processing a minimizing repeat. If the check also involved
3485 matching a wide character, but failed, corruption could cause an
3486 erroneous result when trying to check for a repeat of the original
3487 character.
3488
3489 13. Some tidying changes to the testing mechanism:
3490
3491 (a) The RunTest script now detects the internal link size and whether there
3492 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3493 values substituted by "configure". (The RunGrepTest script already did
3494 this for UTF-8.) The configure.ac script no longer substitutes the
3495 relevant variables.
3496
3497 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3498 with length and offset values. This means that the output is different
3499 for different internal link sizes. Test 2 is skipped for link sizes
3500 other than 2 because of this, bypassing the problem. Unfortunately,
3501 there was also a test in test 3 (the locale tests) that used /B and
3502 failed for link sizes other than 2. Rather than cut the whole test out,
3503 I have added a new /Z option to pcretest that replaces the length and
3504 offset values with spaces. This is now used to make test 3 independent
3505 of link size. (Test 2 will be tidied up later.)
3506
3507 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3508 segmentation fault instead of returning the appropriate error message.
3509
3510 15. In multiline mode when the newline sequence was set to "any", the pattern
3511 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3512 This doesn't seem right; it now treats the CRLF combination as the line
3513 ending, and so does not match in that case. It's only a pattern such as ^$
3514 that would hit this one: something like ^ABC$ would have failed after \r
3515 and then tried again after \r\n.
3516
3517 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3518 in an attempt to make files that differ only in their line terminators
3519 compare equal. This works on Linux.
3520
3521 17. Under certain error circumstances pcregrep might try to free random memory
3522 as it exited. This is now fixed, thanks to valgrind.
3523
3524 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3525 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3526 was because its rules for how to advance for /g after matching an empty
3527 string at the end of a line did not allow for this case. They now check for
3528 it specially.
3529
3530 20. pcretest is supposed to handle patterns and data of any length, by
3531 extending its buffers when necessary. It was getting this wrong when the
3532 buffer for a data line had to be extended.
3533
3534 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3535 CRLF as a newline sequence.
3536
3537 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3538 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3539 I have nevertheless tidied it up.
3540
3541 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3542
3543 24. Added a man page for pcre-config.
3544
3545
3546 Version 7.0 19-Dec-06
3547 ---------------------
3548
3549 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3550 moving to gcc 4.1.1.
3551
3552 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3553 sys/time.h, which is documented as needed for this function. It doesn't
3554 seem to matter on Linux, but it showed up on some releases of OS X.
3555
3556 3. It seems that there are systems where bytes whose values are greater than
3557 127 match isprint() in the "C" locale. The "C" locale should be the
3558 default when a C program starts up. In most systems, only ASCII printing
3559 characters match isprint(). This difference caused the output from pcretest
3560 to vary, making some of the tests fail. I have changed pcretest so that:
3561
3562 (a) When it is outputting text in the compiled version of a pattern, bytes
3563 other than 32-126 are always shown as hex escapes.
3564
3565 (b) When it is outputting text that is a matched part of a subject string,
3566 it does the same, unless a different locale has been set for the match
3567 (using the /L modifier). In this case, it uses isprint() to decide.
3568
3569 4. Fixed a major bug that caused incorrect computation of the amount of memory
3570 required for a compiled pattern when options that changed within the
3571 pattern affected the logic of the preliminary scan that determines the
3572 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3573 that the computed length was too small. The symptoms of this bug were
3574 either the PCRE error "internal error: code overflow" from pcre_compile(),
3575 or a glibc crash with a message such as "pcretest: free(): invalid next
3576 size (fast)". Examples of patterns that provoked this bug (shown in
3577 pcretest format) are:
3578
3579 /(?-x: )/x
3580 /(?x)(?-x: \s*#\s*)/
3581 /((?i)[\x{c0}])/8
3582 /(?i:[\x{c0}])/8
3583
3584 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3585 is now done differently.
3586
3587 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3588 wrapper classes; (b) implement a new function in the C++ scanner that is
3589 more efficient than the old way of doing things because it avoids levels of
3590 recursion in the regex matching; (c) add a paragraph to the documentation
3591 for the FullMatch() function.
3592
3593 6. The escape sequence \n was being treated as whatever was defined as
3594 "newline". Not only was this contrary to the documentation, which states
3595 that \n is character 10 (hex 0A), but it also went horribly wrong when
3596 "newline" was defined as CRLF. This has been fixed.
3597
3598 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3599 was being set to -1 for the "end of line" case (supposedly a value that no
3600 character can have). Though this value is never used (the check for end of
3601 line is "zero bytes in current character"), it caused compiler complaints.
3602 I've changed it to 0xffffffff.
3603
3604 8. In pcre_version.c, the version string was being built by a sequence of
3605 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3606 string (as it is for production releases) called a macro with an empty
3607 argument. The C standard says the result of this is undefined. The gcc
3608 compiler treats it as an empty string (which was what was wanted) but it is
3609 reported that Visual C gives an error. The source has been hacked around to
3610 avoid this problem.
3611
3612 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3613 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3614 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3615 of them did).
3616
3617 10. Originally, pcretest opened its input and output without "b"; then I was
3618 told that "b" was needed in some environments, so it was added for release
3619 5.0 to both the input and output. (It makes no difference on Unix-like
3620 systems.) Later I was told that it is wrong for the input on Windows. I've
3621 now abstracted the modes into two macros, to make it easier to fiddle with
3622 them, and removed "b" from the input mode under Windows.
3623
3624 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3625
3626 12. Added -help and --help to pcretest as an official way of being reminded
3627 of the options.
3628
3629 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3630 and pcrecpp.cc because they annoy compilers at high warning levels.
3631
3632 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3633
3634 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3635 scripts are not C programs :-) and which was not noticed because it works
3636 on Linux.
3637
3638 16. pcretest is supposed to handle any length of pattern and data line (as one
3639 line or as a continued sequence of lines) by extending its input buffer if
3640 necessary. This feature was broken for very long pattern lines, leading to
3641 a string of junk being passed to pcre_compile() if the pattern was longer
3642 than about 50K.
3643
3644 17. I have done a major re-factoring of the way pcre_compile() computes the
3645 amount of memory needed for a compiled pattern. Previously, there was code
3646 that made a preliminary scan of the pattern in order to do this. That was
3647 OK when PCRE was new, but as the facilities have expanded, it has become
3648 harder and harder to keep it in step with the real compile phase, and there
3649 have been a number of bugs (see for example, 4 above). I have now found a
3650 cunning way of running the real compile function in a "fake" mode that
3651 enables it to compute how much memory it would need, while actually only
3652 ever using a few hundred bytes of working memory and without too many
3653 tests of the mode. This should make future maintenance and development
3654 easier. A side effect of this work is that the limit of 200 on the nesting
3655 depth of parentheses has been removed (though this was never a serious
3656 limitation, I suspect). However, there is a downside: pcre_compile() now
3657 runs more slowly than before (30% or more, depending on the pattern). I
3658 hope this isn't a big issue. There is no effect on runtime performance.
3659
3660 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3661 newline (only possible for the last line of a file) and it was a
3662 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3663
3664 19. Added additional timing features to pcretest. (1) The -tm option now times
3665 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3666 separate command line item, by a number that specifies the number of
3667 repeats to use when timing. The default is 50000; this gives better
3668 precision, but takes uncomfortably long for very large patterns.
3669
3670 20. Extended pcre_study() to be more clever in cases where a branch of a
3671 subpattern has no definite first character. For example, (a*|b*)[cd] would
3672 previously give no result from pcre_study(). Now it recognizes that the
3673 first character must be a, b, c, or d.
3674
3675 21. There was an incorrect error "recursive call could loop indefinitely" if
3676 a subpattern (or the entire pattern) that was being tested for matching an
3677 empty string contained only one non-empty item after a nested subpattern.
3678 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3679 incorrectly, because the \d was being skipped in the check.
3680
3681 22. The pcretest program now has a new pattern option /B and a command line
3682 option -b, which is equivalent to adding /B to every pattern. This causes
3683 it to show the compiled bytecode, without the additional information that
3684 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3685 is the same as /B/I).
3686
3687 23. A new optimization is now able automatically to treat some sequences such
3688 as a*b as a*+b. More specifically, if something simple (such as a character
3689 or a simple class like \d) has an unlimited quantifier, and is followed by
3690 something that cannot possibly match the quantified thing, the quantifier
3691 is automatically "possessified".
3692
3693 24. A recursive reference to a subpattern whose number was greater than 39
3694 went wrong under certain circumstances in UTF-8 mode. This bug could also
3695 have affected the operation of pcre_study().
3696
3697 25. Realized that a little bit of performance could be had by replacing
3698 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3699
3700 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3701
3702 27. Possessive quantifiers such as a++ were previously implemented by turning
3703 them into atomic groups such as ($>a+). Now they have their own opcodes,
3704 which improves performance. This includes the automatically created ones
3705 from 23 above.
3706
3707 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3708 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3709 the first matched character to be a colon. This applied both to named and
3710 numbered groups.
3711
3712 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3713
3714 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3715 building PCRE on Windows easier, so I have included it in the distribution.
3716
3717 31. There is now a check in pcretest against a ridiculously large number being
3718 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3719 loop, the loop is abandoned.
3720
3721 32. Forward references to subpatterns in conditions such as (?(2)...) where
3722 subpattern 2 is defined later cause pcre_compile() to search forwards in
3723 the pattern for the relevant set of parentheses. This search went wrong
3724 when there were unescaped parentheses in a character class, parentheses
3725 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3726
3727 33. "Subroutine" calls and backreferences were previously restricted to
3728 referencing subpatterns earlier in the regex. This restriction has now
3729 been removed.
3730
3731 34. Added a number of extra features that are going to be in Perl 5.10. On the
3732 whole, these are just syntactic alternatives for features that PCRE had
3733 previously implemented using the Python syntax or my own invention. The
3734 other formats are all retained for compatibility.
3735
3736 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3737 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3738 also .NET compatible.
3739
3740 (b) A recursion or subroutine call to a named group can now be defined as
3741 (?&name) as well as (?P>name).
3742
3743 (c) A backreference to a named group can now be defined as \k<name> or
3744 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3745 5.10, are also .NET compatible.
3746
3747 (d) A conditional reference to a named group can now use the syntax
3748 (?(<name>) or (?('name') as well as (?(name).
3749
3750 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3751 groups (named and numbered) that are never evaluated inline, but can be
3752 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3753 is always false. There may be only one alternative in such a group.
3754
3755 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3756 as the simple (?(R). The condition is true only if the most recent
3757 recursion is that of the given number or name. It does not search out
3758 through the entire recursion stack.
3759
3760 (g) The escape \gN or \g{N} has been added, where N is a positive or
3761 negative number, specifying an absolute or relative reference.
3762
3763 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3764 some "unreachable code" warnings.
3765
3766 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3767 things, this adds five new scripts.
3768
3769 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3770 There were also incompatibilities regarding the handling of \Q..\E inside
3771 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3772 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3773
3774 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3775 matches an empty string, and forcibly breaks the loop. There were bugs in
3776 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3777 against aaaa the result was just "a" rather than "aaaa", for example. Two
3778 separate and independent bugs (that affected different cases) have been
3779 fixed.
3780
3781 39. Refactored the code to abolish the use of different opcodes for small
3782 capturing bracket numbers. This is a tidy that I avoided doing when I
3783 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3784 The new approach is not only tidier, it makes it possible to reduce the
3785 memory needed to fix the previous bug (38).
3786
3787 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3788 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3789 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3790 mode.
3791
3792 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3793 report.
3794
3795 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3796 copy construction and assignment in the C++ wrapper.
3797
3798 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3799 couple of bugs that could have given wrong results in the "--newline=crlf"
3800 case.
3801
3802 44. Added a number of casts and did some reorganization of signed/unsigned int
3803 variables following suggestions from Dair Grant. Also renamed the variable
3804 "this" as "item" because it is a C++ keyword.
3805
3806 45. Arranged for dftables to add
3807
3808 #include "pcre_internal.h"
3809
3810 to pcre_chartables.c because without it, gcc 4.x may remove the array
3811 definition from the final binary if PCRE is built into a static library and
3812 dead code stripping is activated.
3813
3814 46. For an unanchored pattern, if a match attempt fails at the start of a
3815 newline sequence, and the newline setting is CRLF or ANY, and the next two
3816 characters are CRLF, advance by two characters instead of one.
3817
3818
3819 Version 6.7 04-Jul-06
3820 ---------------------
3821
3822 1. In order to handle tests when input lines are enormously long, pcretest has
3823 been re-factored so that it automatically extends its buffers when
3824 necessary. The code is crude, but this _is_ just a test program. The
3825 default size has been increased from 32K to 50K.
3826
3827 2. The code in pcre_study() was using the value of the re argument before
3828 testing it for NULL. (Of course, in any sensible call of the function, it
3829 won't be NULL.)
3830
3831 3. The memmove() emulation function in pcre_internal.h, which is used on
3832 systems that lack both memmove() and bcopy() - that is, hardly ever -
3833 was missing a "static" storage class specifier.
3834
3835 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3836 containing an extended class (one that cannot be represented by a bitmap
3837 because it contains high-valued characters or Unicode property items, e.g.
3838 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3839 pattern, but PCRE should not loop if you do not (it no longer does).
3840 [Detail: two cases were found: (a) a repeated subpattern containing an
3841 extended class; (b) a recursive reference to a subpattern that followed a
3842 previous extended class. It wasn't skipping over the extended class
3843 correctly when UTF-8 mode was not set.]
3844
3845 5. A negated single-character class was not being recognized as fixed-length
3846 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3847 compile error "lookbehind assertion is not fixed length".
3848
3849 6. The RunPerlTest auxiliary script was showing an unexpected difference
3850 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3851 write a Perl script that can interpret lines of an input file either as
3852 byte characters or as UTF-8, which is what "perltest" was being required to
3853 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3854 can't do is switch easily at run time between having the "use utf8;" pragma
3855 or not. In the end, I fudged it by using the RunPerlTest script to insert
3856 "use utf8;" explicitly for the UTF-8 tests.
3857
3858 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3859 the end of the subject string, contrary to the documentation and to what
3860 Perl does. This was true of both matching functions. Now it matches only at
3861 the start of the subject and immediately after *internal* newlines.
3862
3863 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3864 a pointer to an int instead of a pointer to an unsigned long int. This
3865 caused problems on 64-bit systems.
3866
3867 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3868 instance of the 'standard' template library not being so standard".
3869
3870 10. There was no check on the number of named subpatterns nor the maximum
3871 length of a subpattern name. The product of these values is used to compute
3872 the size of the memory block for a compiled pattern. By supplying a very
3873 long subpattern name and a large number of named subpatterns, the size
3874 computation could be caused to overflow. This is now prevented by limiting
3875 the length of names to 32 characters, and the number of named subpatterns
3876 to 10,000.
3877
3878 11. Subpatterns that are repeated with specific counts have to be replicated in
3879 the compiled pattern. The size of memory for this was computed from the
3880 length of the subpattern and the repeat count. The latter is limited to
3881 65535, but there was no limit on the former, meaning that integer overflow
3882 could in principle occur. The compiled length of a repeated subpattern is
3883 now limited to 30,000 bytes in order to prevent this.
3884
3885 12. Added the optional facility to have named substrings with the same name.
3886
3887 13. Added the ability to use a named substring as a condition, using the
3888 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3889 are numbers (not recommended). Forward references are permitted.
3890
3891 14. Added forward references in named backreferences (if you see what I mean).
3892
3893 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3894 pattern could run off the end of the subject. For example, the pattern
3895 "(?s)(.{1,5})"8 did this with the subject "ab".
3896
3897 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3898 PCRE_CASELESS was set when matching characters that were quantified with ?
3899 or *.
3900
3901 17. A character class other than a single negated character that had a minimum
3902 but no maximum quantifier - for example [ab]{6,} - was not handled
3903 correctly by pce_dfa_exec(). It would match only one character.
3904
3905 18. A valid (though odd) pattern that looked like a POSIX character
3906 class but used an invalid character after [ (for example [[,abc,]]) caused
3907 pcre_compile() to give the error "Failed: internal error: code overflow" or
3908 in some cases to crash with a glibc free() error. This could even happen if
3909 the pattern terminated after [[ but there just happened to be a sequence of
3910 letters, a binary zero, and a closing ] in the memory that followed.
3911
3912 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3913 over the years. Originally (before any Unicode support), just the bottom 8
3914 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3915 output from "man perlunicode" includes this:
3916
3917 The regular expression compiler produces polymorphic opcodes. That
3918 is, the pattern adapts to the data and automatically switches to
3919 the Unicode character scheme when presented with Unicode data--or
3920 instead uses a traditional byte scheme when presented with byte
3921 data.
3922
3923 Sadly, a wide octal escape does not cause a switch, and in a string with
3924 no other multibyte characters, these octal escapes are treated as before.
3925 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3926 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3927 Unicode string.
3928
3929 I have not perpetrated such confusion in PCRE. Up till now, it took just
3930 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3931 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3932 translate to the appropriate multibyte character.
3933
3934 29. Applied some refactoring to reduce the number of warnings from Microsoft
3935 and Borland compilers. This has included removing the fudge introduced
3936 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3937 a warning about an unused variable.
3938
3939 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3940 characters since release 4.0, because Perl (from release 5.004) does not.
3941 [Or at least, is documented not to: some releases seem to be in conflict
3942 with the documentation.] However, when a pattern was studied with
3943 pcre_study() and all its branches started with \s, PCRE still included VT
3944 as a possible starting character. Of course, this did no harm; it just
3945 caused an unnecessary match attempt.
3946
3947 22. Removed a now-redundant internal flag bit that recorded the fact that case
3948 dependency changed within the pattern. This was once needed for "required
3949 byte" processing, but is no longer used. This recovers a now-scarce options
3950 bit. Also moved the least significant internal flag bit to the most-
3951 significant bit of the word, which was not previously used (hangover from
3952 the days when it was an int rather than a uint) to free up another bit for
3953 the future.
3954
3955 23. Added support for CRLF line endings as well as CR and LF. As well as the
3956 default being selectable at build time, it can now be changed at runtime
3957 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3958 specify that it is scanning data with non-default line endings.
3959
3960 24. Changed the definition of CXXLINK to make it agree with the definition of
3961 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3962
3963 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3964 recursions. This makes a big different to stack usage for some patterns.
3965
3966 26. If a subpattern containing a named recursion or subroutine reference such
3967 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3968 the space required for the compiled pattern went wrong and gave too small a
3969 value. Depending on the environment, this could lead to "Failed: internal
3970 error: code overflow at offset 49" or "glibc detected double free or
3971 corruption" errors.
3972
3973 27. Applied patches from Google (a) to support the new newline modes and (b) to
3974 advance over multibyte UTF-8 characters in GlobalReplace.
3975
3976 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3977 difference for some implementation of PCRE in some Windows version.
3978
3979 29. Added some extra testing facilities to pcretest:
3980
3981 \q<number> in a data line sets the "match limit" value
3982 \Q<number> in a data line sets the "match recursion limt" value
3983 -S <number> sets the stack size, where <number> is in megabytes
3984
3985 The -S option isn't available for Windows.
3986
3987
3988 Version 6.6 06-Feb-06
3989 ---------------------
3990
3991 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3992 in pcreposix.h. I have copied the definition from pcre.h.
3993
3994 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3995 because pcre.h is no longer a built file.
3996
3997 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3998 not normally included in the compiled code.
3999
4000
4001 Version 6.5 01-Feb-06
4002 ---------------------
4003
4004 1. When using the partial match feature with pcre_dfa_exec(), it was not
4005 anchoring the second and subsequent partial matches at the new starting
4006 point. This could lead to incorrect results. For example, with the pattern
4007 /1234/, partially matching against "123" and then "a4" gave a match.
4008
4009 2. Changes to pcregrep:
4010
4011 (a) All non-match returns from pcre_exec() were being treated as failures
4012 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
4013 error message is output. Some extra information is given for the
4014 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
4015 probably the only errors that are likely to be caused by users (by
4016 specifying a regex that has nested indefinite repeats, for instance).
4017 If there are more than 20 of these errors, pcregrep is abandoned.
4018
4019 (b) A binary zero was treated as data while matching, but terminated the
4020 output line if it was written out. This has been fixed: binary zeroes
4021 are now no different to any other data bytes.
4022
4023 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
4024 used to set a locale for matching. The --locale=xxxx long option has
4025 been added (no short equivalent) to specify a locale explicitly on the
4026 pcregrep command, overriding the environment variables.
4027
4028 (d) When -B was used with -n, some line numbers in the output were one less
4029 than they should have been.
4030
4031 (e) Added the -o (--only-matching) option.
4032
4033 (f) If -A or -C was used with -c (count only), some lines of context were
4034 accidentally printed for the final match.
4035
4036 (g) Added the -H (--with-filename) option.
4037
4038 (h) The combination of options -rh failed to suppress file names for files
4039 that were found from directory arguments.
4040
4041 (i) Added the -D (--devices) and -d (--directories) options.
4042
4043 (j) Added the -F (--fixed-strings) option.
4044
4045 (k) Allow "-" to be used as a file name for -f as well as for a data file.
4046
4047 (l) Added the --colo(u)r option.
4048
4049 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
4050 is not present by default.
4051
4052 3. A nasty bug was discovered in the handling of recursive patterns, that is,
4053 items such as (?R) or (?1), when the recursion could match a number of
4054 alternatives. If it matched one of the alternatives, but subsequently,
4055 outside the recursion, there was a failure, the code tried to back up into
4056 the recursion. However, because of the way PCRE is implemented, this is not
4057 possible, and the result was an incorrect result from the match.
4058
4059 In order to prevent this happening, the specification of recursion has
4060 been changed so that all such subpatterns are automatically treated as
4061 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
4062
4063 4. I had overlooked the fact that, in some locales, there are characters for
4064 which isalpha() is true but neither isupper() nor islower() are true. In
4065 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
4066 and ordfeminine) are like this. This affected the treatment of \w and \W
4067 when they appeared in character classes, but not when they appeared outside
4068 a character class. The bit map for "word" characters is now created
4069 separately from the results of isalnum() instead of just taking it from the
4070 upper, lower, and digit maps. (Plus the underscore character, of course.)
4071
4072 5. The above bug also affected the handling of POSIX character classes such as
4073 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
4074 permanent tables. Instead, the bit maps for such a class were previously
4075 created as the appropriate unions of the upper, lower, and digit bitmaps.
4076 Now they are created by subtraction from the [[:word:]] class, which has
4077 its own bitmap.
4078
4079 6. The [[:blank:]] character class matches horizontal, but not vertical space.
4080 It is created by subtracting the vertical space characters (\x09, \x0a,
4081 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
4082 subtraction was done in the overall bitmap for a character class, meaning
4083 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
4084 be recognized. This bug has been fixed.
4085
4086 7. Patches from the folks at Google:
4087
4088 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
4089 real life, but is still worth protecting against".
4090
4091 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
4092 regular expressions".
4093
4094 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
4095 have it.
4096
4097 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
4098 "configure" and the latter not, in order to fix a problem somebody had
4099 with compiling the Arg class on HP-UX.
4100
4101 (e) Improve the error-handling of the C++ wrapper a little bit.
4102
4103 (f) New tests for checking recursion limiting.
4104
4105 8. The pcre_memmove() function, which is used only if the environment does not
4106 have a standard memmove() function (and is therefore rarely compiled),
4107 contained two bugs: (a) use of int instead of size_t, and (b) it was not
4108 returning a result (though PCRE never actually uses the result).
4109
4110 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
4111 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
4112 returned instead of calling malloc() with an overflowing number that would
4113 most likely cause subsequent chaos.
4114
4115 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
4116
4117 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
4118 with this option is matched, the nmatch and pmatch options of regexec() are
4119 ignored.
4120
4121 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
4122 provided in case anyone wants to the the POSIX interface with UTF-8
4123 strings.
4124
4125 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
4126 C++ linking (needed for some HP-UX environments).
4127
4128 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
4129 (unused parameter) and in the pcre_printint() function (omitted "default"
4130 switch label when the default is to do nothing).
4131
4132 15. Added some code to make it possible, when PCRE is compiled as a C++
4133 library, to replace subject pointers for pcre_exec() with a smart pointer
4134 class, thus making it possible to process discontinuous strings.
4135
4136 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
4137 much the same function. They were added by different people who were trying
4138 to make PCRE easy to compile on non-Unix systems. It has been suggested
4139 that PCRE_EXPORT be abolished now that there is more automatic apparatus
4140 for compiling on Windows systems. I have therefore replaced it with
4141 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
4142 defaults to "extern" for C or "extern C" for C++, which works fine on
4143 Unix-like systems. It is now possible to override the value of PCRE_DATA_
4144 SCOPE with something explicit in config.h. In addition:
4145
4146 (a) pcreposix.h still had just "extern" instead of either of these macros;
4147 I have replaced it with PCRE_DATA_SCOPE.
4148
4149 (b) Functions such as _pcre_xclass(), which are internal to the library,
4150 but external in the C sense, all had PCRE_EXPORT in their definitions.
4151 This is apparently wrong for the Windows case, so I have removed it.
4152 (It makes no difference on Unix-like systems.)
4153
4154 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
4155 of recursive calls to match(). This is different to MATCH_LIMIT because
4156 that limits the total number of calls to match(), not all of which increase
4157 the depth of recursion. Limiting the recursion depth limits the amount of
4158 stack (or heap if NO_RECURSE is set) that is used. The default can be set
4159 when PCRE is compiled, and changed at run time. A patch from Google adds
4160 this functionality to the C++ interface.
4161
4162 18. Changes to the handling of Unicode character properties:
4163
4164 (a) Updated the table to Unicode 4.1.0.
4165
4166 (b) Recognize characters that are not in the table as "Cn" (undefined).
4167
4168 (c) I revised the way the table is implemented to a much improved format
4169 which includes recognition of ranges. It now supports the ranges that
4170 are defined in UnicodeData.txt, and it also amalgamates other
4171 characters into ranges. This has reduced the number of entries in the
4172 table from around 16,000 to around 3,000, thus reducing its size
4173 considerably. I realized I did not need to use a tree structure after
4174 all - a binary chop search is just as efficient. Having reduced the
4175 number of entries, I extended their size from 6 bytes to 8 bytes to
4176 allow for more data.
4177
4178 (d) Added support for Unicode script names via properties such as \p{Han}.
4179
4180 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
4181 matching that character.
4182
4183 20. When matching a repeated Unicode property with a minimum greater than zero,
4184 (for example \pL{2,}), PCRE could look past the end of the subject if it
4185 reached it while seeking the minimum number of characters. This could
4186 happen only if some of the characters were more than one byte long, because
4187 there is a check for at least the minimum number of bytes.
4188
4189 21. Refactored the implementation of \p and \P so as to be more general, to
4190 allow for more different types of property in future. This has changed the
4191 compiled form incompatibly. Anybody with saved compiled patterns that use
4192 \p or \P will have to recompile them.
4193
4194 22. Added "Any" and "L&" to the supported property types.
4195
4196 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
4197 but give a compile time error if the value is greater than 0xff.
4198
4199 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
4200 accidentally not being installed or uninstalled.
4201
4202 25. The pcre.h file was built from pcre.h.in, but the only changes that were
4203 made were to insert the current release number. This seemed silly, because
4204 it made things harder for people building PCRE on systems that don't run
4205 "configure". I have turned pcre.h into a distributed file, no longer built
4206 by "configure", with the version identification directly included. There is
4207 no longer a pcre.h.in file.
4208
4209 However, this change necessitated a change to the pcre-config script as
4210 well. It is built from pcre-config.in, and one of the substitutions was the
4211 release number. I have updated configure.ac so that ./configure now finds
4212 the release number by grepping pcre.h.
4213
4214 26. Added the ability to run the tests under valgrind.
4215
4216
4217 Version 6.4 05-Sep-05
4218 ---------------------
4219
4220 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
4221 "--" to be printed when multiple files were scanned, even when none of the
4222 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
4223 consider it to be a bug, and have restored the previous behaviour.
4224
4225 2. A couple of code tidies to get rid of compiler warnings.
4226
4227 3. The pcretest program used to cheat by referring to symbols in the library
4228 whose names begin with _pcre_. These are internal symbols that are not
4229 really supposed to be visible externally, and in some environments it is
4230 possible to suppress them. The cheating is now confined to including
4231 certain files from the library's source, which is a bit cleaner.
4232
4233 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
4234 file's purpose clearer.
4235
4236 5. Reorganized pcre_ucp_findchar().
4237
4238
4239 Version 6.3 15-Aug-05
4240 ---------------------
4241
4242 1. The file libpcre.pc.in did not have general read permission in the tarball.
4243
4244 2. There were some problems when building without C++ support:
4245
4246 (a) If C++ support was not built, "make install" and "make test" still
4247 tried to test it.
4248
4249 (b) There were problems when the value of CXX was explicitly set. Some
4250 changes have been made to try to fix these, and ...
4251
4252 (c) --disable-cpp can now be used to explicitly disable C++ support.
4253
4254 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
4255 backslash in a target when C++ was disabled. This confuses some
4256 versions of "make", apparently. Using an intermediate variable solves
4257 this. (Same for CPP_LOBJ.)
4258
4259 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
4260 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
4261 necessary on certain architectures.
4262
4263 4. Added a setting of -export-symbols-regex to the link command to remove
4264 those symbols that are exported in the C sense, but actually are local
4265 within the library, and not documented. Their names all begin with
4266 "_pcre_". This is not a perfect job, because (a) we have to except some
4267 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
4268 available (and never for static libraries). I have made a note to try to
4269 find a way round (a) in the future.
4270
4271
4272 Version 6.2 01-Aug-05
4273 ---------------------
4274
4275 1. There was no test for integer overflow of quantifier values. A construction
4276 such as {1111111111111111} would give undefined results. What is worse, if
4277 a minimum quantifier for a parenthesized subpattern overflowed and became
4278 negative, the calculation of the memory size went wrong. This could have
4279 led to memory overwriting.
4280
4281 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
4282
4283 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
4284 operating environments where this matters.
4285
4286 4. Applied Giuseppe Maxia's patch to add additional features for controlling
4287 PCRE options from within the C++ wrapper.
4288
4289 5. Named capturing subpatterns were not being correctly counted when a pattern
4290 was compiled. This caused two problems: (a) If there were more than 100
4291 such subpatterns, the calculation of the memory needed for the whole
4292 compiled pattern went wrong, leading to an overflow error. (b) Numerical
4293 back references of the form \12, where the number was greater than 9, were
4294 not recognized as back references, even though there were sufficient
4295 previous subpatterns.
4296
4297 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
4298 versions of gcc, e.g. 2.95.4.
4299
4300
4301 Version 6.1 21-Jun-05
4302 ---------------------
4303
4304 1. There was one reference to the variable "posix" in pcretest.c that was not
4305 surrounded by "#if !defined NOPOSIX".
4306
4307 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
4308 the cross-check on the old pcre_info() function, for the benefit of the
4309 cut-down version of PCRE that is currently imported into Exim.
4310
4311 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
4312 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
4313 patterns that start (?i) but I don't think that matters. The use of (?i) is
4314 just an example; this all applies to the other options as well.
4315
4316 4. Since libtool seems to echo the compile commands it is issuing, the output
4317 from "make" can be reduced a bit by putting "@" in front of each libtool
4318 compile command.
4319
4320 5. Patch from the folks at Google for configure.in to be a bit more thorough
4321 in checking for a suitable C++ installation before trying to compile the
4322 C++ stuff. This should fix a reported problem when a compiler was present,
4323 but no suitable headers.
4324
4325 6. The man pages all had just "PCRE" as their title. I have changed them to
4326 be the relevant file name. I have also arranged that these names are
4327 retained in the file doc/pcre.txt, which is a concatenation in text format
4328 of all the man pages except the little individual ones for each function.
4329
4330 7. The NON-UNIX-USE file had not been updated for the different set of source
4331 files that come with release 6. I also added a few comments about the C++
4332 wrapper.
4333
4334
4335 Version 6.0 07-Jun-05
4336 ---------------------
4337
4338 1. Some minor internal re-organization to help with my DFA experiments.
4339
4340 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
4341 didn't matter for the library itself when fully configured, but did matter
4342 when compiling without UCP support, or within Exim, where the ucp files are
4343 not imported.
4344
4345 3. Refactoring of the library code to split up the various functions into
4346 different source modules. The addition of the new DFA matching code (see
4347 below) to a single monolithic source would have made it really too
4348 unwieldy, quite apart from causing all the code to be include in a
4349 statically linked application, when only some functions are used. This is
4350 relevant even without the DFA addition now that patterns can be compiled in
4351 one application and matched in another.
4352
4353 The downside of splitting up is that there have to be some external
4354 functions and data tables that are used internally in different modules of
4355 the library but which are not part of the API. These have all had their
4356 names changed to start with "_pcre_" so that they are unlikely to clash
4357 with other external names.
4358
4359 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
4360 a different (DFA) algorithm. Although it is slower than the original
4361 function, it does have some advantages for certain types of matching
4362 problem.
4363
4364 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
4365 including restarting after a partial match.
4366
4367 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
4368 defined when compiling for Windows was sent to me. I have put it into the
4369 code, though I have no means of testing or verifying it.
4370
4371 7. Added the pcre_refcount() auxiliary function.
4372
4373 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
4374 match before or at the first newline in the subject string. In pcretest,
4375 the /f option on a pattern can be used to set this.
4376
4377 9. A repeated \w when used in UTF-8 mode with characters greater than 256
4378 would behave wrongly. This has been present in PCRE since release 4.0.
4379
4380 10. A number of changes to the pcregrep command:
4381
4382 (a) Refactored how -x works; insert ^(...)$ instead of setting
4383 PCRE_ANCHORED and checking the length, in preparation for adding
4384 something similar for -w.
4385
4386 (b) Added the -w (match as a word) option.
4387
4388 (c) Refactored the way lines are read and buffered so as to have more
4389 than one at a time available.
4390
4391 (d) Implemented a pcregrep test script.
4392
4393 (e) Added the -M (multiline match) option. This allows patterns to match
4394 over several lines of the subject. The buffering ensures that at least
4395 8K, or the rest of the document (whichever is the shorter) is available
4396 for matching (and similarly the previous 8K for lookbehind assertions).
4397
4398 (f) Changed the --help output so that it now says
4399
4400 -w, --word-regex(p)
4401
4402 instead of two lines, one with "regex" and the other with "regexp"
4403 because that confused at least one person since the short forms are the
4404 same. (This required a bit of code, as the output is generated
4405 automatically from a table. It wasn't just a text change.)
4406
4407 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4408 option but starts with a hyphen. Could be a pattern or a path name
4409 starting with a hyphen, for instance.
4410
4411 (h) "-" can be given as a file name to represent stdin.
4412
4413 (i) When file names are being printed, "(standard input)" is used for
4414 the standard input, for compatibility with GNU grep. Previously
4415 "<stdin>" was used.
4416
4417 (j) The option --label=xxx can be used to supply a name to be used for
4418 stdin when file names are being printed. There is no short form.
4419
4420 (k) Re-factored the options decoding logic because we are going to add
4421 two more options that take data. Such options can now be given in four
4422 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4423
4424 (l) Added the -A, -B, and -C options for requesting that lines of context
4425 around matches be printed.
4426
4427 (m) Added the -L option to print the names of files that do not contain
4428 any matching lines, that is, the complement of -l.
4429
4430 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4431 continue to scan other files.
4432
4433 (o) The -s option was incorrectly implemented. For compatibility with other
4434 greps, it now suppresses the error message for a non-existent or non-
4435 accessible file (but not the return code). There is a new option called
4436 -q that suppresses the output of matching lines, which was what -s was
4437 previously doing.
4438
4439 (p) Added --include and --exclude options to specify files for inclusion
4440 and exclusion when recursing.
4441
4442 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4443 Hopefully, it now does.
4444
4445 12. Missing cast in pcre_study().
4446
4447 13. Added an "uninstall" target to the makefile.
4448
4449 14. Replaced "extern" in the function prototypes in Makefile.in with
4450 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4451 world, but is set differently for Windows.
4452
4453 15. Added a second compiling function called pcre_compile2(). The only
4454 difference is that it has an extra argument, which is a pointer to an
4455 integer error code. When there is a compile-time failure, this is set
4456 non-zero, in addition to the error test pointer being set to point to an
4457 error message. The new argument may be NULL if no error number is required
4458 (but then you may as well call pcre_compile(), which is now just a
4459 wrapper). This facility is provided because some applications need a
4460 numeric error indication, but it has also enabled me to tidy up the way
4461 compile-time errors are handled in the POSIX wrapper.
4462
4463 16. Added VPATH=.libs to the makefile; this should help when building with one
4464 prefix path and installing with another. (Or so I'm told by someone who
4465 knows more about this stuff than I do.)
4466
4467 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4468 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4469 match everything, including newlines. This is not POSIX-compatible, but
4470 somebody wanted the feature. From pcretest it can be activated by using
4471 both the P and the s flags.
4472
4473 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4474
4475 19. libpcre.pc was being incorrectly installed as executable.
4476
4477 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4478 it now also looks for '\r' so that it will work unmodified on Windows.
4479
4480 21. Added Google's contributed C++ wrapper to the distribution.
4481
4482 22. Added some untidy missing memory free() calls in pcretest, to keep
4483 Electric Fence happy when testing.
4484
4485
4486
4487 Version 5.0 13-Sep-04
4488 ---------------------
4489
4490 1. Internal change: literal characters are no longer packed up into items
4491 containing multiple characters in a single byte-string. Each character
4492 is now matched using a separate opcode. However, there may be more than one
4493 byte in the character in UTF-8 mode.
4494
4495 2. The pcre_callout_block structure has two new fields: pattern_position and
4496 next_item_length. These contain the offset in the pattern to the next match
4497 item, and its length, respectively.
4498
4499 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4500 insertion of callouts before each pattern item. Added the /C option to
4501 pcretest to make use of this.
4502
4503 4. On the advice of a Windows user, the lines
4504
4505 #if defined(_WIN32) || defined(WIN32)
4506 _setmode( _fileno( stdout ), 0x8000 );
4507 #endif /* defined(_WIN32) || defined(WIN32) */
4508
4509 have been added to the source of pcretest. This apparently does useful
4510 magic in relation to line terminators.
4511
4512 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4513 for the benefit of those environments where the "b" makes a difference.
4514
4515 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4516 to know about it. I have put a hack into configure.in that adds in code
4517 to set GCC=yes if CC=icc. This seems to end up at a point in the
4518 generated configure script that is early enough to affect the setting of
4519 compiler options, which is what is needed, but I have no means of testing
4520 whether it really works. (The user who reported this had patched the
4521 generated configure script, which of course I cannot do.)
4522
4523 LATER: After change 22 below (new libtool files), the configure script
4524 seems to know about icc (and also ecc). Therefore, I have commented out
4525 this hack in configure.in.
4526
4527 7. Added support for pkg-config (2 patches were sent in).
4528
4529 8. Negated POSIX character classes that used a combination of internal tables
4530 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4531 [[:^ascii]]. Typically, they would match almost any characters. The other
4532 POSIX classes were not broken in this way.
4533
4534 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4535 to find the match, as PCRE was deluded into thinking that the match had to
4536 start at the start point or following a newline. The same bug applied to
4537 patterns with negative forward assertions or any backward assertions
4538 preceding ".*" at the start, unless the pattern required a fixed first
4539 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4540
4541 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4542 starting at the last subject character, bytes beyond the end of the subject
4543 string were read.
4544
4545 11. Renamed the variable "class" as "classbits" to make life easier for C++
4546 users. (Previously there was a macro definition, but it apparently wasn't
4547 enough.)
4548
4549 12. Added the new field "tables" to the extra data so that tables can be passed
4550 in at exec time, or the internal tables can be re-selected. This allows
4551 a compiled regex to be saved and re-used at a later time by a different
4552 program that might have everything at different addresses.
4553
4554 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4555 -R library as well as a -L library.
4556
4557 14. The debugging options of pcretest (-d on the command line or D on a
4558 pattern) showed incorrect output for anything following an extended class
4559 that contained multibyte characters and which was followed by a quantifier.
4560
4561 15. Added optional support for general category Unicode character properties
4562 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4563 support. It adds about 90K to the size of the library. The meanings of the
4564 inbuilt class escapes such as \d and \s have NOT been changed.
4565
4566 16. Updated pcredemo.c to include calls to free() to release the memory for the
4567 compiled pattern.
4568
4569 17. The generated file chartables.c was being created in the source directory
4570 instead of in the building directory. This caused the build to fail if the
4571 source directory was different from the building directory, and was
4572 read-only.
4573
4574 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4575 file. No doubt somebody will tell me if they don't make sense... Also added
4576 Dan Mooney's comments about building on OpenVMS.
4577
4578 19. Added support for partial matching via the PCRE_PARTIAL option for
4579 pcre_exec() and the \P data escape in pcretest.
4580
4581 20. Extended pcretest with 3 new pattern features:
4582
4583 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4584 write the compiled pattern to the file whose name is "rest-of-line".
4585 This is a straight binary dump of the data, with the saved pointer to
4586 the character tables forced to be NULL. The study data, if any, is
4587 written too. After writing, pcretest reads a new pattern.
4588
4589 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4590 compiled pattern from the given file. There must not be any
4591 occurrences of "<" in the file name (pretty unlikely); if there are,
4592 pcretest will instead treat the initial "<" as a pattern delimiter.
4593 After reading in the pattern, pcretest goes on to read data lines as
4594 usual.
4595
4596 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4597 and 16-bit fields in a compiled pattern, to simulate a pattern that
4598 was compiled on a host of opposite endianness.
4599
4600 21. The pcre-exec() function can now cope with patterns that were compiled on
4601 hosts of opposite endianness, with this restriction:
4602
4603 As for any compiled expression that is saved and used later, the tables
4604 pointer field cannot be preserved; the extra_data field in the arguments
4605 to pcre_exec() should be used to pass in a tables address if a value
4606 other than the default internal tables were used at compile time.
4607
4608 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4609 now diagnosed as an error. Previously, most of the time, a negative number
4610 would have been treated as zero, but if in addition "ovector" was passed as
4611 NULL, a crash could occur.
4612
4613 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4614 new versions from the libtool 1.5 distribution (the last one is a copy of
4615 a file called libtool.m4). This seems to have fixed the need to patch
4616 "configure" to support Darwin 1.3 (which I used to do). However, I still
4617 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4618 workstation).
4619
4620 24. Changed the PCRE licence to be the more standard "BSD" licence.
4621
4622
4623 Version 4.5 01-Dec-03
4624 ---------------------
4625
4626 1. There has been some re-arrangement of the code for the match() function so
4627 that it can be compiled in a version that does not call itself recursively.
4628 Instead, it keeps those local variables that need separate instances for
4629 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4630 needs to "recurse". Keeping track of where control must go is done by means
4631 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4632 hide most of the details from the main code, and operates only if
4633 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4634 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4635 operating.
4636
4637 To make it easier for callers to provide specially tailored get/free
4638 functions for this usage, two new functions, pcre_stack_malloc, and
4639 pcre_stack_free, are used. They are always called in strict stacking order,
4640 and the size of block requested is always the same.
4641
4642 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4643 PCRE has been compiled to use the stack or the heap for recursion. The
4644 -C option of pcretest uses this to show which version is compiled.
4645
4646 A new data escape \S, is added to pcretest; it causes the amounts of store
4647 obtained and freed by both kinds of malloc/free at match time to be added
4648 to the output.
4649
4650 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4651 what's available on my current Linux desktop machine.
4652
4653 3. When matching a UTF-8 string, the test for a valid string at the start has
4654 been extended. If start_offset is not zero, PCRE now checks that it points
4655 to a byte that is the start of a UTF-8 character. If not, it returns
4656 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4657 this is necessary because there may be backward assertions in the pattern.
4658 When matching the same subject several times, it may save resources to use
4659 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4660
4661 4. The code for checking the validity of UTF-8 strings has been tightened so
4662 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4663 containing "overlong sequences".
4664
4665 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4666 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4667 should have been "&", but it just so happened that all the cases this let
4668 through by mistake were picked up later in the function.
4669
4670 6. I had used a variable called "isblank" - this is a C99 function, causing
4671 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4672
4673 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4674 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4675 so the version is shown; (c) stop "make test" echoing "./RunTest".
4676
4677 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4678
4679 9. The prototype for memmove() for systems that don't have it was using
4680 size_t, but the inclusion of the header that defines size_t was later. I've
4681 moved the #includes for the C headers earlier to avoid this.
4682
4683 10. Added some adjustments to the code to make it easier to compiler on certain
4684 special systems:
4685
4686 (a) Some "const" qualifiers were missing.
4687 (b) Added the macro EXPORT before all exported functions; by default this
4688 is defined to be empty.
4689 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4690 that it reads its output file name as an argument instead of writing
4691 to the standard output and assuming this can be redirected.
4692
4693 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4694 class containing characters with values greater than 255, PCRE compilation
4695 went into a loop.
4696
4697 12. A recursive reference to a subpattern that was within another subpattern
4698 that had a minimum quantifier of zero caused PCRE to crash. For example,
4699 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4700 recursion. If the recursively-called subpattern itself had a zero repeat,
4701 that was OK.
4702
4703 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4704 buffer into which it was copied (for escape processing) was still set at
4705 1024, so long lines caused crashes.
4706
4707 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4708 "internal error: code overflow...". This applied to any character class
4709 that was followed by a possessive quantifier.
4710
4711 15. Modified the Makefile to add libpcre.la as a prerequisite for
4712 libpcreposix.la because I was told this is needed for a parallel build to
4713 work.
4714
4715 16. If a pattern that contained .* following optional items at the start was
4716 studied, the wrong optimizing data was generated, leading to matching
4717 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4718 matching string must start with a or b or c. The correct conclusion for
4719 this pattern is that a match can start with any character.
4720
4721
4722 Version 4.4 13-Aug-03
4723 ---------------------
4724
4725 1. In UTF-8 mode, a character class containing characters with values between
4726 127 and 255 was not handled correctly if the compiled pattern was studied.
4727 In fixing this, I have also improved the studying algorithm for such
4728 classes (slightly).
4729
4730 2. Three internal functions had redundant arguments passed to them. Removal
4731 might give a very teeny performance improvement.
4732
4733 3. Documentation bug: the value of the capture_top field in a callout is *one
4734 more than* the number of the hightest numbered captured substring.
4735
4736 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4737 in incorrectly linking with a previously installed version. They now link
4738 explicitly with libpcre.la.
4739
4740 5. configure.in no longer needs to recognize Cygwin specially.
4741
4742 6. A problem in pcre.in for Windows platforms is fixed.
4743
4744 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4745 pcretest, it used to include the size of the study block as part of its
4746 output. Unfortunately, the structure contains a field that has a different
4747 size on different hardware architectures. This meant that the tests that
4748 showed this size failed. As the block is currently always of a fixed size,
4749 this information isn't actually particularly useful in pcretest output, so
4750 I have just removed it.
4751
4752 8. Three pre-processor statements accidentally did not start in column 1.
4753 Sadly, there are *still* compilers around that complain, even though
4754 standard C has not required this for well over a decade. Sigh.
4755
4756 9. In pcretest, the code for checking callouts passed small integers in the
4757 callout_data field, which is a void * field. However, some picky compilers
4758 complained about the casts involved for this on 64-bit systems. Now
4759 pcretest passes the address of the small integer instead, which should get
4760 rid of the warnings.
4761
4762 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4763 both compile and run time, and gives an error if an invalid UTF-8 sequence
4764 is found. There is a option for disabling this check in cases where the
4765 string is known to be correct and/or the maximum performance is wanted.
4766
4767 11. In response to a bug report, I changed one line in Makefile.in from
4768
4769 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4770 to
4771 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4772
4773 to look similar to other lines, but I have no way of telling whether this
4774 is the right thing to do, as I do not use Windows. No doubt I'll get told
4775 if it's wrong...
4776
4777
4778 Version 4.3 21-May-03
4779 ---------------------
4780
4781 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4782 Makefile.
4783
4784 2. Some refactoring to improve the quality of the code:
4785
4786 (i) The utf8_table... variables are now declared "const".
4787
4788 (ii) The code for \cx, which used the "case flipping" table to upper case
4789 lower case letters, now just substracts 32. This is ASCII-specific,
4790 but the whole concept of \cx is ASCII-specific, so it seems
4791 reasonable.
4792
4793 (iii) PCRE was using its character types table to recognize decimal and
4794 hexadecimal digits in the pattern. This is silly, because it handles
4795 only 0-9, a-f, and A-F, but the character types table is locale-
4796 specific, which means strange things might happen. A private
4797 table is now used for this - though it costs 256 bytes, a table is
4798 much faster than multiple explicit tests. Of course, the standard
4799 character types table is still used for matching digits in subject
4800 strings against \d.
4801
4802 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4803 ending in _t are). So I've renamed it as ESC_tee.
4804
4805 3. The first argument for regexec() in the POSIX wrapper should have been
4806 defined as "const".
4807
4808 4. Changed pcretest to use malloc() for its buffers so that they can be
4809 Electric Fenced for debugging.
4810
4811 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4812 to read one or more bytes before the start of the subject string. Often this
4813 had no effect on PCRE's behaviour, but in some circumstances it could
4814 provoke a segmentation fault.
4815
4816 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4817 to try to read one or more bytes before the start of the subject string.
4818
4819 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4820 UTF-8 support could misbehave in various ways if the subject string
4821 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4822 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4823 back over UTF-8 characters.)
4824
4825
4826 Version 4.2 14-Apr-03
4827 ---------------------
4828
4829 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4830
4831 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4832 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4833 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4834 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4835 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4836 and BUILD_EXEEXT
4837 Note: automatic setting of the BUILD variables is not yet working
4838 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4839 compile-time but not at link-time
4840 [LINK]: use for linking executables only
4841 make different versions for Windows and non-Windows
4842 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4843 libraries
4844 [LINK_FOR_BUILD]: new variable
4845 [OBJEXT]: use throughout
4846 [EXEEXT]: use throughout
4847 <winshared>: new target
4848 <wininstall>: new target
4849 <dftables.o>: use native compiler
4850 <dftables>: use native linker
4851 <install>: handle Windows platform correctly
4852 <clean>: ditto
4853 <check>: ditto
4854 copy DLL to top builddir before testing
4855
4856 As part of these changes, -no-undefined was removed again. This was reported
4857 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4858 in any case.
4859
4860 3. Some tidies to get rid of compiler warnings:
4861
4862 . In the match_data structure, match_limit was an unsigned long int, whereas
4863 match_call_count was an int. I've made them both unsigned long ints.
4864
4865 . In pcretest the fact that a const uschar * doesn't automatically cast to
4866 a void * provoked a warning.
4867
4868 . Turning on some more compiler warnings threw up some "shadow" variables
4869 and a few more missing casts.
4870
4871 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4872 option, a class that contained a single character with a value between 128
4873 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4874
4875 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4876 option, a class that contained several characters, but with at least one
4877 whose value was between 128 and 255 caused PCRE to crash.
4878
4879
4880 Version 4.1 12-Mar-03
4881 ---------------------
4882
4883 1. Compiling with gcc -pedantic found a couple of places where casts were
4884 needed, and a string in dftables.c that was longer than standard compilers are
4885 required to support.
4886
4887 2. Compiling with Sun's compiler found a few more places where the code could
4888 be tidied up in order to avoid warnings.
4889
4890 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4891 first of these names is deprecated in the latest Autoconf in favour of the name
4892 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4893 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4894 analogy I have changed it to CFLAGS_FOR_BUILD.
4895
4896 4. Added -no-undefined to the linking command in the Makefile, because this is
4897 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4898 linking step for the pcreposix library.
4899
4900 5. PCRE was failing to diagnose the case of two named groups with the same
4901 name.
4902
4903 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4904 literal character that is needed in the subject for a match, and scans along to
4905 ensure that it is present before embarking on the full matching process. This
4906 saves time in cases of nested unlimited repeats that are never going to match.
4907 Problem: the scan can take a lot of time if the subject is very long (e.g.
4908 megabytes), thus penalizing straightforward matches. It is now done only if the
4909 amount of subject to be scanned is less than 1000 bytes.
4910
4911 7. A lesser problem with the same optimization is that it was recording the
4912 first character of an anchored pattern as "needed", thus provoking a search
4913 right along the subject, even when the first match of the pattern was going to
4914 fail. The "needed" character is now not set for anchored patterns, unless it
4915 follows something in the pattern that is of non-fixed length. Thus, it still
4916 fulfils its original purpose of finding quick non-matches in cases of nested
4917 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4918
4919
4920 Version 4.0 17-Feb-03
4921 ---------------------
4922
4923 1. If a comment in an extended regex that started immediately after a meta-item
4924 extended to the end of string, PCRE compiled incorrect data. This could lead to
4925 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4926
4927 2. Moved to autoconf 2.53 and libtool 1.4.2.
4928
4929 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4930 the special perltest8 script is no longer needed - all the tests can be run
4931 from a single perltest script.
4932
4933 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4934 by \s. It has now been removed in PCRE. This means it isn't recognized as
4935 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4936 class [:space:] *does* include VT, thereby creating a mess.
4937
4938 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4939 space and tab.
4940
4941 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4942 its new features into the main test script, reducing the number of scripts.
4943
4944 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4945 were backward compatible, and made the (?i) apply to the whole pattern, as if
4946 /i were given. Now it behaves more logically, and applies the option setting
4947 only to what follows. PCRE has been changed to follow suit. However, if it
4948 finds options settings right at the start of the pattern, it extracts them into
4949 the global options, as before. Thus, they show up in the info data.
4950
4951 8. Added support for the \Q...\E escape sequence. Characters in between are
4952 treated as literals. This is slightly different from Perl in that $ and @ are
4953 also handled as literals inside the quotes. In Perl, they will cause variable
4954 interpolation. Note the following examples:
4955
4956 Pattern PCRE matches Perl matches
4957
4958 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4959 \Qabc\$xyz\E abc\$xyz abc\$xyz
4960 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4961
4962 For compatibility with Perl, \Q...\E sequences are recognized inside character
4963 classes as well as outside them.
4964
4965 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4966 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4967 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4968 signed/unsigned warnings.
4969
4970 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4971 option for pcretest, so I've replaced it by a simple function that does just
4972 that job.
4973
4974 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4975 "pcregrep -".
4976
4977 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4978 Java package. This provides some syntactic sugar for simple cases of what my
4979 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4980 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4981 item, you can use this simplified notation. Note that only makes sense with
4982 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4983 greediness, whatever the setting of the PCRE_UNGREEDY option.
4984
4985 13. A change of greediness default within a pattern was not taking effect at
4986 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4987 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4988 was abstracted outside.
4989
4990 14. PCRE now supports the \G assertion. It is true when the current matching
4991 position is at the start point of the match. This differs from \A when the
4992 starting offset is non-zero. Used with the /g option of pcretest (or similar
4993 code), it works in the same way as it does for Perl's /g option. If all
4994 alternatives of a regex begin with \G, the expression is anchored to the start
4995 match position, and the "anchored" flag is set in the compiled expression.
4996
4997 15. Some bugs concerning the handling of certain option changes within patterns
4998 have been fixed. These applied to options other than (?ims). For example,
4999 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
5000 been the other way round. Some of this was related to change 7 above.
5001
5002 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
5003 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
5004 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
5005 POSIX classes only within a class (e.g. /[[:alpha:]]/).
5006
5007 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
5008 mode. Unlike ".", it always matches newline, whatever the setting of
5009 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
5010 assertions. Perl allows it, but it doesn't (in general) work because it can't
5011 calculate the length of the lookbehind. At least, that's the case for Perl
5012 5.8.0 - I've been told they are going to document that it doesn't work in
5013 future.
5014
5015 18. Added an error diagnosis for escapes that PCRE does not support: these are
5016 \L, \l, \N, \P, \p, \U, \u, and \X.
5017
5018 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
5019 reading past the end of the pattern in cases such as /[abcd/.
5020
5021 20. PCRE was getting more memory than necessary for patterns with classes that
5022 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
5023
5024 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
5025 compiling PCRE for use with Virtual Pascal.
5026
5027 22. Small fix to the Makefile to make it work properly if the build is done
5028 outside the source tree.
5029
5030 23. Added a new extension: a condition to go with recursion. If a conditional
5031 subpattern starts with (?(R) the "true" branch is used if recursion has
5032 happened, whereas the "false" branch is used only at the top level.
5033
5034 24. When there was a very long string of literal characters (over 255 bytes
5035 without UTF support, over 250 bytes with UTF support), the computation of how
5036 much memory was required could be incorrect, leading to segfaults or other
5037 strange effects.
5038
5039 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
5040 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
5041 there was a subsequent back reference to those brackets. This meant that, for
5042 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
5043 possible to check for precisely this case. All we can do is abandon the
5044 optimization if .* occurs inside capturing brackets when there are any back
5045 references whatsoever. (See below for a better fix that came later.)
5046
5047 26. The handling of the optimization for finding the first character of a
5048 non-anchored pattern, and for finding a character that is required later in the
5049 match were failing in some cases. This didn't break the matching; it just
5050 failed to optimize when it could. The way this is done has been re-implemented.
5051
5052 27. Fixed typo in error message for invalid (?R item (it said "(?p").
5053
5054 28. Added a new feature that provides some of the functionality that Perl
5055 provides with (?{...}). The facility is termed a "callout". The way it is done
5056 in PCRE is for the caller to provide an optional function, by setting
5057 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
5058 global variable. By default it is unset, which disables all calling out. To get
5059 the function called, the regex must include (?C) at appropriate points. This
5060 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
5061 This provides a means of identifying different callout points. When PCRE
5062 reaches such a point in the regex, if pcre_callout has been set, the external
5063 function is called. It is provided with data in a structure called
5064 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
5065 matching continues; if it returns a non-zero value, the match at the current
5066 point fails. However, backtracking will occur if possible. [This was changed
5067 later and other features added - see item 49 below.]
5068
5069 29. pcretest is upgraded to test the callout functionality. It provides a
5070 callout function that displays information. By default, it shows the start of
5071 the match and the current position in the text. There are some new data escapes
5072 to vary what happens:
5073
5074 \C+ in addition, show current contents of captured substrings
5075 \C- do not supply a callout function
5076 \C!n return 1 when callout number n is reached
5077 \C!n!m return 1 when callout number n is reached for the mth time
5078
5079 30. If pcregrep was called with the -l option and just a single file name, it
5080 output "<stdin>" if a match was found, instead of the file name.
5081
5082 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
5083 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
5084 pcre_exec(). This saves a malloc/free per call. The default value of
5085 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
5086 when configuring.
5087
5088 32. The default maximum size of a compiled pattern is 64K. There have been a
5089 few cases of people hitting this limit. The code now uses macros to handle the
5090 storing of links as offsets within the compiled pattern. It defaults to 2-byte
5091 links, but this can be changed to 3 or 4 bytes by --with-link-size when
5092 configuring. Tests 2 and 5 work only with 2-byte links because they output
5093 debugging information about compiled patterns.
5094
5095 33. Internal code re-arrangements:
5096
5097 (a) Moved the debugging function for printing out a compiled regex into
5098 its own source file (printint.c) and used #include to pull it into
5099 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
5100 separate copies.
5101
5102 (b) Defined the list of op-code names for debugging as a macro in
5103 internal.h so that it is next to the definition of the opcodes.
5104
5105 (c) Defined a table of op-code lengths for simpler skipping along compiled
5106 code. This is again a macro in internal.h so that it is next to the
5107 definition of the opcodes.
5108
5109 34. Added support for recursive calls to individual subpatterns, along the
5110 lines of Robin Houston's patch (but implemented somewhat differently).
5111
5112 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
5113 allow it to read and process whole directories in Win32. This code was
5114 contributed by Lionel Fourquaux; it has not been tested by me.
5115
5116 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
5117 used to name a group. Names consist of alphanumerics and underscores, and must
5118 be unique. Back references use the syntax (?P=name) and recursive calls use
5119 (?P>name) which is a PCRE extension to the Python extension. Groups still have
5120 numbers. The function pcre_fullinfo() can be used after compilation to extract
5121 a name/number map. There are three relevant calls:
5122
5123 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
5124 PCRE_INFO_NAMECOUNT yields the number of entries
5125 PCRE_INFO_NAMETABLE yields a pointer to the map.
5126
5127 The map is a vector of fixed-size entries. The size of each entry depends on
5128 the length of the longest name used. The first two bytes of each entry are the
5129 group number, most significant byte first. There follows the corresponding
5130 name, zero terminated. The names are in alphabetical order.
5131
5132 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
5133 case instead of 255. Making it the same both with and without UTF-8 support
5134 means that the same test output works with both.
5135
5136 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
5137 calling malloc() with a zero argument.
5138
5139 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
5140 optimization. I've improved things by keeping a bitmap of backreferences with
5141 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
5142 fact referenced, the optimization can be applied. It is unlikely that a
5143 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
5144 the match to follow \n) will appear inside brackets with a number greater than
5145 31, but if it does, any back reference > 31 suppresses the optimization.
5146
5147 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
5148 of disabling numbered capturing parentheses. Any opening parenthesis that is
5149 not followed by ? behaves as if it were followed by ?: but named parentheses
5150 can still be used for capturing (and they will acquire numbers in the usual
5151 way).
5152
5153 41. Redesigned the return codes from the match() function into yes/no/error so
5154 that errors can be passed back from deep inside the nested calls. A malloc
5155 failure while inside a recursive subpattern call now causes the
5156 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
5157
5158 42. It is now possible to set a limit on the number of times the match()
5159 function is called in a call to pcre_exec(). This facility makes it possible to
5160 limit the amount of recursion and backtracking, though not in a directly
5161 obvious way, because the match() function is used in a number of different
5162 circumstances. The count starts from zero for each position in the subject
5163 string (for non-anchored patterns). The default limit is, for compatibility, a
5164 large number, namely 10 000 000. You can change this in two ways:
5165
5166 (a) When configuring PCRE before making, you can use --with-match-limit=n
5167 to set a default value for the compiled library.
5168
5169 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
5170 a different value is set. See 45 below.
5171
5172 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
5173
5174 43. Added a new function pcre_config(int, void *) to enable run-time extraction
5175 of things that can be changed at compile time. The first argument specifies
5176 what is wanted and the second points to where the information is to be placed.
5177 The current list of available information is:
5178
5179 PCRE_CONFIG_UTF8
5180
5181 The output is an integer that is set to one if UTF-8 support is available;
5182 otherwise it is set to zero.
5183
5184 PCRE_CONFIG_NEWLINE
5185
5186 The output is an integer that it set to the value of the code that is used for
5187 newline. It is either LF (10) or CR (13).
5188
5189 PCRE_CONFIG_LINK_SIZE
5190
5191 The output is an integer that contains the number of bytes used for internal
5192 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
5193
5194 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
5195
5196 The output is an integer that contains the threshold above which the POSIX
5197 interface uses malloc() for output vectors. See item 31 above.
5198
5199 PCRE_CONFIG_MATCH_LIMIT
5200
5201 The output is an unsigned integer that contains the default limit of the number
5202 of match() calls in a pcre_exec() execution. See 42 above.
5203
5204 44. pcretest has been upgraded by the addition of the -C option. This causes it
5205 to extract all the available output from the new pcre_config() function, and to
5206 output it. The program then exits immediately.
5207
5208 45. A need has arisen to pass over additional data with calls to pcre_exec() in
5209 order to support additional features. One way would have been to define
5210 pcre_exec2() (for example) with extra arguments, but this would not have been
5211 extensible, and would also have required all calls to the original function to
5212 be mapped to the new one. Instead, I have chosen to extend the mechanism that
5213 is used for passing in "extra" data from pcre_study().
5214
5215 The pcre_extra structure is now exposed and defined in pcre.h. It currently
5216 contains the following fields:
5217
5218 flags a bitmap indicating which of the following fields are set
5219 study_data opaque data from pcre_study()
5220 match_limit a way of specifying a limit on match() calls for a specific
5221 call to pcre_exec()
5222 callout_data data for callouts (see 49 below)
5223
5224 The flag bits are also defined in pcre.h, and are
5225
5226 PCRE_EXTRA_STUDY_DATA
5227 PCRE_EXTRA_MATCH_LIMIT
5228 PCRE_EXTRA_CALLOUT_DATA
5229
5230 The pcre_study() function now returns one of these new pcre_extra blocks, with
5231 the actual study data pointed to by the study_data field, and the
5232 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
5233 before. That is, this change is entirely upwards-compatible and requires no
5234 change to existing code.
5235
5236 If you want to pass in additional data to pcre_exec(), you can either place it
5237 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
5238 block.
5239
5240 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
5241 data string contains the escape sequence \M, pcretest calls pcre_exec() several
5242 times with different match limits, until it finds the minimum value needed for
5243 pcre_exec() to complete. The value is then output. This can be instructive; for
5244 most simple matches the number is quite small, but for pathological cases it
5245 gets very large very quickly.
5246
5247 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
5248 returns the size of the data block pointed to by the study_data field in a
5249 pcre_extra block, that is, the value that was passed as the argument to
5250 pcre_malloc() when PCRE was getting memory in which to place the information
5251 created by pcre_study(). The fourth argument should point to a size_t variable.
5252 pcretest has been extended so that this information is shown after a successful
5253 pcre_study() call when information about the compiled regex is being displayed.
5254
5255 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
5256 because what follows is always an absolute path. (Later: it turns out that this
5257 is more than cosmetic for MinGW, because it doesn't like empty path
5258 components.)
5259
5260 49. Some changes have been made to the callout feature (see 28 above):
5261
5262 (i) A callout function now has three choices for what it returns:
5263
5264 0 => success, carry on matching
5265 > 0 => failure at this point, but backtrack if possible
5266 < 0 => serious error, return this value from pcre_exec()
5267
5268 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
5269 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
5270 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
5271 use by callout functions. It will never be used by PCRE itself.
5272
5273 (ii) The pcre_extra structure (see 45 above) has a void * field called
5274 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
5275 pcre_callout_block structure has a field of the same name. The contents of
5276 the field passed in the pcre_extra structure are passed to the callout
5277 function in the corresponding field in the callout block. This makes it
5278 easier to use the same callout-containing regex from multiple threads. For
5279 testing, the pcretest program has a new data escape
5280
5281 \C*n pass the number n (may be negative) as callout_data
5282
5283 If the callout function in pcretest receives a non-zero value as
5284 callout_data, it returns that value.
5285
5286 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
5287 there were some redundant $(CFLAGS) in commands that are now specified as
5288 $(LINK), which already includes $(CFLAGS).
5289
5290 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
5291 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
5292 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
5293 one-byte characters throughout. Note that case-insensitive matching applies
5294 only to characters whose values are less than 256. PCRE doesn't support the
5295 notion of cases for higher-valued characters.
5296
5297 (i) A character class whose characters are all within 0-255 is handled as
5298 a bit map, and the map is inverted for negative classes. Previously, a
5299 character > 255 always failed to match such a class; however it should
5300 match if the class was a negative one (e.g. [^ab]). This has been fixed.
5301
5302 (ii) A negated character class with a single character < 255 is coded as
5303 "not this character" (OP_NOT). This wasn't working properly when the test
5304 character was multibyte, either singly or repeated.
5305
5306 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
5307 mode, for example: \x{100}{2,3}.
5308
5309 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
5310 singly or repeated) now correctly test multibyte characters. However,
5311 PCRE doesn't recognize any characters with values greater than 255 as
5312 digits, spaces, or word characters. Such characters always match \D, \S,
5313 and \W, and never match \d, \s, or \w.
5314
5315 (v) Classes may now contain characters and character ranges with values
5316 greater than 255. For example: [ab\x{100}-\x{400}].
5317
5318 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
5319 PCRE in UTF-8 mode.
5320
5321 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
5322 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
5323 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
5324 value.)
5325
5326 53. The single man page has become too large. I have therefore split it up into
5327 a number of separate man pages. These also give rise to individual HTML pages;
5328 these are now put in a separate directory, and there is an index.html page that
5329 lists them all. Some hyperlinking between the pages has been installed.
5330
5331 54. Added convenience functions for handling named capturing parentheses.
5332
5333 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
5334 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
5335 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
5336 are faulted.
5337
5338 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
5339 calling configure. These values are used when compiling the dftables.c program
5340 which is run to generate the source of the default character tables. They
5341 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
5342 you will need to set these values.
5343
5344 57. Updated the building process for Windows DLL, as provided by Fred Cox.
5345
5346
5347 Version 3.9 02-Jan-02
5348 ---------------------
5349
5350 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
5351
5352 2. If --disable-static was given, the building process failed when trying to
5353 build pcretest and pcregrep. (For some reason it was using libtool to compile
5354 them, which is not right, as they aren't part of the library.)
5355
5356
5357 Version 3.8 18-Dec-01
5358 ---------------------
5359
5360 1. The experimental UTF-8 code was completely screwed up. It was packing the
5361 bytes in the wrong order. How dumb can you get?
5362
5363
5364 Version 3.7 29-Oct-01
5365 ---------------------
5366
5367 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
5368 This caused pcretest, when used on the test data, to segfault. Unfortunately,
5369 this didn't happen under Solaris 8, where I normally test things.
5370
5371 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
5372 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
5373 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
5374 this fix an hour or so after the initial 3.7 release.)
5375
5376
5377 Version 3.6 23-Oct-01
5378 ---------------------
5379
5380 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
5381 offsets passed as NULL with zero offset count.
5382
5383 2. The config.guess and config.sub files had not been updated when I moved to
5384 the latest autoconf.
5385
5386
5387 Version 3.5 15-Aug-01
5388 ---------------------
5389
5390 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
5391 had been forgotten.
5392
5393 2. By using declared but undefined structures, we can avoid using "void"
5394 definitions in pcre.h while keeping the internal definitions of the structures
5395 private.
5396
5397 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
5398 user point of view, this means that both static and shared libraries are built
5399 by default, but this can be individually controlled. More of the work of
5400 handling this static/shared cases is now inside libtool instead of PCRE's make
5401 file.
5402
5403 4. The pcretest utility is now installed along with pcregrep because it is
5404 useful for users (to test regexs) and by doing this, it automatically gets
5405 relinked by libtool. The documentation has been turned into a man page, so
5406 there are now .1, .txt, and .html versions in /doc.
5407
5408 5. Upgrades to pcregrep:
5409 (i) Added long-form option names like gnu grep.
5410 (ii) Added --help to list all options with an explanatory phrase.
5411 (iii) Added -r, --recursive to recurse into sub-directories.
5412 (iv) Added -f, --file to read patterns from a file.
5413
5414 6. pcre_exec() was referring to its "code" argument before testing that
5415 argument for NULL (and giving an error if it was NULL).
5416
5417 7. Upgraded Makefile.in to allow for compiling in a different directory from
5418 the source directory.
5419
5420 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5421 options bits, the pointer it was passed was to an int instead of to an unsigned
5422 long int. This mattered only on 64-bit systems.
5423
5424 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5425 generated) instead of pcre.in, which it its source. Also made the same change
5426 in several of the .c files.
5427
5428 10. A new release of gcc defines printf() as a macro, which broke pcretest
5429 because it had an ifdef in the middle of a string argument for printf(). Fixed
5430 by using separate calls to printf().
5431
5432 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5433 script, to force use of CR or LF instead of \n in the source. On non-Unix
5434 systems, the value can be set in config.h.
5435
5436 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5437 absolute limit. Changed the text of the error message to make this clear, and
5438 likewise updated the man page.
5439
5440 13. The limit of 99 on the number of capturing subpatterns has been removed.
5441 The new limit is 65535, which I hope will not be a "real" limit.
5442
5443
5444 Version 3.4 22-Aug-00
5445 ---------------------
5446
5447 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5448
5449 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5450
5451
5452 Version 3.3 01-Aug-00
5453 ---------------------
5454
5455 1. If an octal character was given, but the value was greater than \377, it
5456 was not getting masked to the least significant bits, as documented. This could
5457 lead to crashes in some systems.
5458
5459 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5460 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5461
5462 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5463 These just pass their arguments on to (pcre_free)(), but they are provided
5464 because some uses of PCRE bind it to non-C systems that can call its functions,
5465 but cannot call free() or pcre_free() directly.
5466
5467 4. Add "make test" as a synonym for "make check". Corrected some comments in
5468 the Makefile.
5469
5470 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5471 Makefile.
5472
5473 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5474 command called pgrep for grepping around the active processes.
5475
5476 7. Added the beginnings of support for UTF-8 character strings.
5477
5478 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5479 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5480 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5481 out for the ar command.)
5482
5483
5484 Version 3.2 12-May-00
5485 ---------------------
5486
5487 This is purely a bug fixing release.
5488
5489 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5490 of ZA. This was just one example of several cases that could provoke this bug,
5491 which was introduced by change 9 of version 2.00. The code for breaking
5492 infinite loops after an iteration that matches an empty string was't working
5493 correctly.
5494
5495 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5496 when matched against abbab (for example). After matching an empty string, it
5497 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5498 caused it to match further down the string than it should.
5499
5500 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5501 was there because it doesn't seem to be needed, and it causes trouble on some
5502 systems, as it is not a Standard C header. It has been removed.
5503
5504 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5505 were reported on the Macintosh. The changes were from
5506
5507 while ((c = *(++ptr)) != 0 && c != '\n');
5508 to
5509 while ((c = *(++ptr)) != 0 && c != '\n') ;
5510
5511 Totally extraordinary, but if that's what it takes...
5512
5513 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5514 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5515 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5516 assumes the way PCRE uses memmove() (always moving upwards).
5517
5518 6. PCRE is being used in one environment where strchr() is not available. There
5519 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5520 faster code anyway.
5521
5522
5523 Version 3.1 09-Feb-00
5524 ---------------------
5525
5526 The only change in this release is the fixing of some bugs in Makefile.in for
5527 the "install" target:
5528
5529 (1) It was failing to install pcreposix.h.
5530
5531 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5532
5533
5534 Version 3.0 01-Feb-00
5535 ---------------------
5536
5537 1. Add support for the /+ modifier to perltest (to output $` like it does in
5538 pcretest).
5539
5540 2. Add support for the /g modifier to perltest.
5541
5542 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5543 matches null strings.
5544
5545 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5546 pattern. Perl treats empty patterns specially - it reuses the most recent
5547 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5548 effect.
5549
5550 5. The POSIX interface was broken in that it was just handing over the POSIX
5551 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5552 required a bigger vector, with some working space on the end. This means that
5553 the POSIX wrapper now has to get and free some memory, and copy the results.
5554
5555 6. Added some simple autoconf support, placing the test data and the
5556 documentation in separate directories, re-organizing some of the
5557 information files, and making it build pcre-config (a GNU standard). Also added
5558 libtool support for building PCRE as a shared library, which is now the
5559 default.
5560
5561 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5562 09 are not valid octal constants. Single digits will be used for minor values
5563 less than 10.
5564
5565 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5566 existing programs that set these in the POSIX interface can use PCRE without
5567 modification.
5568
5569 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5570 return all that pcre_info() returns, plus additional data. The pcre_info()
5571 function is retained for compatibility, but is considered to be obsolete.
5572
5573 10. Added experimental recursion feature (?R) to handle one common case that
5574 Perl 5.6 will be able to do with (?p{...}).
5575
5576 11. Added support for POSIX character classes like [:alpha:], which Perl is
5577 adopting.
5578
5579
5580 Version 2.08 31-Aug-99
5581 ----------------------
5582
5583 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5584 trying to match at the startoffset position, but instead was moving forward to
5585 the next newline as if a previous match had failed.
5586
5587 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5588 and could get into a loop if a null string was matched other than at the start
5589 of the subject.
5590
5591 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5592 be distinguished at compile time, and for completeness also added PCRE_DATE.
5593
5594 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5595 in GnuWin32 environments.
5596
5597
5598 Version 2.07 29-Jul-99
5599 ----------------------
5600
5601 1. The documentation is now supplied in plain text form and HTML as well as in
5602 the form of man page sources.
5603
5604 2. C++ compilers don't like assigning (void *) values to other pointer types.
5605 In particular this affects malloc(). Although there is no problem in Standard
5606 C, I've put in casts to keep C++ compilers happy.
5607
5608 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5609 should be (const char *).
5610
5611 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5612 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5613 However, I haven't made this a standard facility. The documentation doesn't
5614 mention it, and the Makefile doesn't support it.
5615
5616 5. The Makefile now contains an "install" target, with editable destinations at
5617 the top of the file. The pcretest program is not installed.
5618
5619 6. pgrep -V now gives the PCRE version number and date.
5620
5621 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5622 causing the entire string to be ignored, instead of just the last character.
5623
5624 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5625 non-matching string, it can take a very, very long time, even for strings of
5626 quite modest length, because of the nested recursion. PCRE now does better in
5627 some of these cases. It does this by remembering the last required literal
5628 character in the pattern, and pre-searching the subject to ensure it is present
5629 before running the real match. In other words, it applies a heuristic to detect
5630 some types of certain failure quickly, and in the above example, if presented
5631 with a string that has no trailing " it gives "no match" very quickly.
5632
5633 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5634 other alternatives are tried instead.
5635
5636
5637 Version 2.06 09-Jun-99
5638 ----------------------
5639
5640 1. Change pcretest's output for amount of store used to show just the code
5641 space, because the remainder (the data block) varies in size between 32-bit and
5642 64-bit systems.
5643
5644 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5645 start matching at. This allows lookbehinds to work when searching for multiple
5646 occurrences in a string.
5647
5648 3. Added additional options to pcretest for testing multiple occurrences:
5649
5650 /+ outputs the rest of the string that follows a match
5651 /g loops for multiple occurrences, using the new startoffset argument
5652 /G loops for multiple occurrences by passing an incremented pointer
5653
5654 4. PCRE wasn't doing the "first character" optimization for patterns starting
5655 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5656 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5657 the letter 'x'. On long subject strings, this gives a significant speed-up.
5658
5659
5660 Version 2.05 21-Apr-99
5661 ----------------------
5662
5663 1. Changed the type of magic_number from int to long int so that it works
5664 properly on 16-bit systems.
5665
5666 2. Fixed a bug which caused patterns starting with .* not to work correctly
5667 when the subject string contained newline characters. PCRE was assuming
5668 anchoring for such patterns in all cases, which is not correct because .* will
5669 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5670 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5671 must be retried after every newline in the subject.
5672
5673
5674 Version 2.04 18-Feb-99
5675 ----------------------
5676
5677 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5678 computation of the store needed to hold the pattern was incorrect (too large).
5679 If such patterns were nested a few deep, this could multiply and become a real
5680 problem.
5681
5682 2. Added /M option to pcretest to show the memory requirement of a specific
5683 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5684
5685 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5686 compiled in such a way that the backtracking after subsequent failure was
5687 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5688 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5689
5690
5691 Version 2.03 02-Feb-99
5692 ----------------------
5693
5694 1. Fixed typo and small mistake in man page.
5695
5696 2. Added 4th condition (GPL supersedes if conflict) and created separate
5697 LICENCE file containing the conditions.
5698
5699 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5700 Perl, that is the internal \ allows the delimiter to be included in the
5701 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5702 the final delimiter, add \ to the end of the pattern (to test the error).
5703
5704 4. Added the convenience functions for extracting substrings after a successful
5705 match. Updated pcretest to make it able to test these functions.
5706
5707
5708 Version 2.02 14-Jan-99
5709 ----------------------
5710
5711 1. Initialized the working variables associated with each extraction so that
5712 their saving and restoring doesn't refer to uninitialized store.
5713
5714 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5715 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5716 fix the problem.
5717
5718 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5719 calls, and wasn't printing the correct value for compiling calls. Increased the
5720 default value of LOOPREPEAT, and the number of significant figures in the
5721 times.
5722
5723 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5724
5725 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5726 a building problem on Windows NT with a FAT file system.
5727
5728
5729 Version 2.01 21-Oct-98
5730 ----------------------
5731
5732 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5733 to character tables built by pcre_maketables() in the current locale. If NULL
5734 is passed, the default tables are used.
5735
5736
5737 Version 2.00 24-Sep-98
5738 ----------------------
5739
5740 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5741 it any more.
5742
5743 2. Allow quantification of (?>) groups, and make it work correctly.
5744
5745 3. The first character computation wasn't working for (?>) groups.
5746
5747 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5748 end of the subject) and add 5.005's \z, which really does match only at the
5749 very end of the subject.
5750
5751 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5752
5753 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5754 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5755 localized options. All options to pcre_study() were also removed.
5756
5757 7. Add other new features from 5.005:
5758
5759 $(?<= positive lookbehind
5760 $(?<! negative lookbehind
5761 (?imsx-imsx) added the unsetting capability
5762 such a setting is global if at outer level; local otherwise
5763 (?imsx-imsx:) non-capturing groups with option setting
5764 (?(cond)re|re) conditional pattern matching
5765
5766 A backreference to itself in a repeated group matches the previous
5767 captured string.
5768
5769 8. General tidying up of studying (both automatic and via "study")
5770 consequential on the addition of new assertions.
5771
5772 9. As in 5.005, unlimited repeated groups that could match an empty substring
5773 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5774 runtime if any iteration does actually match an empty substring.
5775
5776 10. Include the RunTest script in the distribution.
5777
5778 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5779 discrepancies, some of which were old and were also with respect to 5.004. They
5780 have now been fixed.
5781
5782
5783 Version 1.09 28-Apr-98
5784 ----------------------
5785
5786 1. A negated single character class followed by a quantifier with a minimum
5787 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5788 program crashes, or just wrong answers. This did not apply to negated classes
5789 containing more than one character, or to minima other than one.
5790
5791
5792 Version 1.08 27-Mar-98
5793 ----------------------
5794
5795 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5796
5797 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5798 latter must appear before anything that relies on it in the pattern.
5799
5800
5801 Version 1.07 16-Feb-98
5802 ----------------------
5803
5804 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5805 repeat of a potentially empty string).
5806
5807
5808 Version 1.06 23-Jan-98
5809 ----------------------
5810
5811 1. Added Markus Oberhumer's little patches for C++.
5812
5813 2. Literal strings longer than 255 characters were broken.
5814
5815
5816 Version 1.05 23-Dec-97
5817 ----------------------
5818
5819 1. Negated character classes containing more than one character were failing if
5820 PCRE_CASELESS was set at run time.
5821
5822
5823 Version 1.04 19-Dec-97
5824 ----------------------
5825
5826 1. Corrected the man page, where some "const" qualifiers had been omitted.
5827
5828 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5829 input syntax.
5830
5831 3. Fixed memory leak which occurred when a regex with back references was
5832 matched with an offsets vector that wasn't big enough. The temporary memory
5833 that is used in this case wasn't being freed if the match failed.
5834
5835 4. Tidied pcretest to ensure it frees memory that it gets.
5836
5837 5. Temporary memory was being obtained in the case where the passed offsets
5838 vector was exactly big enough.
5839
5840 6. Corrected definition of offsetof() from change 5 below.
5841
5842 7. I had screwed up change 6 below and broken the rules for the use of
5843 setjmp(). Now fixed.
5844
5845
5846 Version 1.03 18-Dec-97
5847 ----------------------
5848
5849 1. A erroneous regex with a missing opening parenthesis was correctly
5850 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5851 on some systems.
5852
5853 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5854 it was reported that one broken compiler failed on the former because "code" is
5855 also an independent variable.
5856
5857 3. The erroneous regex a[]b caused an array overrun reference.
5858
5859 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5860 fail on data ending with that character. (It was going on too far, and checking
5861 the next character, typically a binary zero.) This was specific to the
5862 optimized code for single-character negative classes.
5863
5864 5. Added a contributed patch from the TIN world which does the following:
5865
5866 + Add an undef for memmove, in case the the system defines a macro for it.
5867
5868 + Add a definition of offsetof(), in case there isn't one. (I don't know
5869 the reason behind this - offsetof() is part of the ANSI standard - but
5870 it does no harm).
5871
5872 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5873 most of the places where whitespace preceded '#'. I have given up and
5874 allowed the remaining 2 cases to be at the margin.
5875
5876 + Rename some variables in pcre to eliminate shadowing. This seems very
5877 pedantic, but does no harm, of course.
5878
5879 6. Moved the call to setjmp() into its own function, to get rid of warnings
5880 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5881
5882 7. Constructs such as \d{8,} were compiling into the equivalent of
5883 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5884 outcome, but in this particular case used more store than had been allocated,
5885 which caused the bug to be discovered because it threw up an internal error.
5886
5887 8. The debugging code in both pcre and pcretest for outputting the compiled
5888 form of a regex was going wrong in the case of back references followed by
5889 curly-bracketed repeats.
5890
5891
5892 Version 1.02 12-Dec-97
5893 ----------------------
5894
5895 1. Typos in pcre.3 and comments in the source fixed.
5896
5897 2. Applied a contributed patch to get rid of places where it used to remove
5898 'const' from variables, and fixed some signed/unsigned and uninitialized
5899 variable warnings.
5900
5901 3. Added the "runtest" target to Makefile.
5902
5903 4. Set default compiler flag to -O2 rather than just -O.
5904
5905
5906 Version 1.01 19-Nov-97
5907 ----------------------
5908
5909 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5910 like /([ab]*)*/, that is, for classes with more than one character in them.
5911
5912 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5913 as /((?>a*))*/ (a PCRE_EXTRA facility).
5914
5915
5916 Version 1.00 18-Nov-97
5917 ----------------------
5918
5919 1. Added compile-time macros to support systems such as SunOS4 which don't have
5920 memmove() or strerror() but have other things that can be used instead.
5921
5922 2. Arranged that "make clean" removes the executables.
5923
5924
5925 Version 0.99 27-Oct-97
5926 ----------------------
5927
5928 1. Fixed bug in code for optimizing classes with only one character. It was
5929 initializing a 32-byte map regardless, which could cause it to run off the end
5930 of the memory it had got.
5931
5932 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5933
5934
5935 Version 0.98 22-Oct-97
5936 ----------------------
5937
5938 1. Fixed bug in code for handling temporary memory usage when there are more
5939 back references than supplied space in the ovector. This could cause segfaults.
5940
5941
5942 Version 0.97 21-Oct-97
5943 ----------------------
5944
5945 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5946
5947 2. Optimized negated single characters not to use a bit map.
5948
5949 3. Brought error texts together as macro definitions; clarified some of them;
5950 fixed one that was wrong - it said "range out of order" when it meant "invalid
5951 escape sequence".
5952
5953 4. Changed some char * arguments to const char *.
5954
5955 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5956
5957 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5958 pcretest.
5959
5960
5961 Version 0.96 16-Oct-97
5962 ----------------------
5963
5964 1. Added a simple "pgrep" utility to the distribution.
5965
5966 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5967 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5968 where "ddd" means "one or more decimal digits".
5969
5970 3. Fixed serious bug. If a pattern had a back reference, but the call to
5971 pcre_exec() didn't supply a large enough ovector to record the related
5972 identifying subpattern, the match always failed. PCRE now remembers the number
5973 of the largest back reference, and gets some temporary memory in which to save
5974 the offsets during matching if necessary, in order to ensure that
5975 backreferences always work.
5976
5977 4. Increased the compatibility with Perl in a number of ways:
5978
5979 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5980 to request this handling. The option can be set at compile or exec time.
5981
5982 (b) $ matches before a terminating newline by default; an option
5983 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5984 mode). The option can be set at compile or exec time.
5985
5986 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5987 the same as Perl's. If the decimal number it represents is less than 10
5988 or there aren't that many previous left capturing parentheses, an octal
5989 escape is read. Inside a character class, it's always an octal escape,
5990 even if it is a single digit.
5991
5992 (d) An escaped but undefined alphabetic character is taken as a literal,
5993 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5994 escapes.
5995
5996 (e) {0} is now permitted. (The previous item is removed from the compiled
5997 pattern).
5998
5999 5. Changed all the names of code files so that the basic parts are no longer
6000 than 10 characters, and abolished the teeny "globals.c" file.
6001
6002 6. Changed the handling of character classes; they are now done with a 32-byte
6003 bit map always.
6004
6005 7. Added the -d and /D options to pcretest to make it possible to look at the
6006 internals of compilation without having to recompile pcre.
6007
6008
6009 Version 0.95 23-Sep-97
6010 ----------------------
6011
6012 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
6013 \x20 at the start of a run of normal characters. These were being treated as
6014 real characters, instead of the source characters being re-checked.
6015
6016
6017 Version 0.94 18-Sep-97
6018 ----------------------
6019
6020 1. The functions are now thread-safe, with the caveat that the global variables
6021 containing pointers to malloc() and free() or alternative functions are the
6022 same for all threads.
6023
6024 2. Get pcre_study() to generate a bitmap of initial characters for non-
6025 anchored patterns when this is possible, and use it if passed to pcre_exec().
6026
6027
6028 Version 0.93 15-Sep-97
6029 ----------------------
6030
6031 1. /(b)|(:+)/ was computing an incorrect first character.
6032
6033 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
6034 but not actually doing anything yet.
6035
6036 3. Treat "-" characters in classes that cannot be part of ranges as literals,
6037 as Perl does (e.g. [-az] or [az-]).
6038
6039 4. Set the anchored flag if a branch starts with .* or .*? because that tests
6040 all possible positions.
6041
6042 5. Split up into different modules to avoid including unneeded functions in a
6043 compiled binary. However, compile and exec are still in one module. The "study"
6044 function is split off.
6045
6046 6. The character tables are now in a separate module whose source is generated
6047 by an auxiliary program - but can then be edited by hand if required. There are
6048 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
6049 toupper() in the code.
6050
6051 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
6052 make them global. Abolish the function for setting them, as the caller can now
6053 set them directly.
6054
6055
6056 Version 0.92 11-Sep-97
6057 ----------------------
6058
6059 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
6060