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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1535 - (show annotations)
Wed Mar 25 16:51:51 2015 UTC (4 years, 6 months ago) by ph10
File size: 274563 byte(s)
Error occurred while calculating annotation data.
Add recursion depth limit to auto-possessification.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.37 xx-xxx-2015
5 ------------------------
6
7 1. When an (*ACCEPT) is triggered inside capturing parentheses, it arranges
8 for those parentheses to be closed with whatever has been captured so far.
9 However, it was failing to mark any other groups between the hightest
10 capture so far and the currrent group as "unset". Thus, the ovector for
11 those groups contained whatever was previously there. An example is the
12 pattern /(x)|((*ACCEPT))/ when matched against "abcd".
13
14 2. If an assertion condition was quantified with a minimum of zero (an odd
15 thing to do, but it happened), SIGSEGV or other misbehaviour could occur.
16
17 3. If a pattern in pcretest input had the P (POSIX) modifier followed by an
18 unrecognized modifier, a crash could occur.
19
20 4. An attempt to do global matching in pcretest with a zero-length ovector
21 caused a crash.
22
23 5. Fixed a memory leak during matching that could occur for a subpattern
24 subroutine call (recursive or otherwise) if the number of captured groups
25 that had to be saved was greater than ten.
26
27 6. Catch a bad opcode during auto-possessification after compiling a bad UTF
28 string with NO_UTF_CHECK. This is a tidyup, not a bug fix, as passing bad
29 UTF with NO_UTF_CHECK is documented as having an undefined outcome.
30
31 7. A UTF pattern containing a "not" match of a non-ASCII character and a
32 subroutine reference could loop at compile time. Example: /[^\xff]((?1))/.
33
34 8. When a pattern is compiled, it remembers the highest back reference so that
35 when matching, if the ovector is too small, extra memory can be obtained to
36 use instead. A conditional subpattern whose condition is a check on a
37 capture having happened, such as, for example in the pattern
38 /^(?:(a)|b)(?(1)A|B)/, is another kind of back reference, but it was not
39 setting the highest backreference number. This mattered only if pcre_exec()
40 was called with an ovector that was too small to hold the capture, and there
41 was no other kind of back reference (a situation which is probably quite
42 rare). The effect of the bug was that the condition was always treated as
43 FALSE when the capture could not be consulted, leading to a incorrect
44 behaviour by pcre2_match(). This bug has been fixed.
45
46 9. A reference to a duplicated named group (either a back reference or a test
47 for being set in a conditional) that occurred in a part of the pattern where
48 PCRE_DUPNAMES was not set caused the amount of memory needed for the pattern
49 to be incorrectly calculated, leading to overwriting.
50
51 10. A mutually recursive set of back references such as (\2)(\1) caused a
52 segfault at study time (while trying to find the minimum matching length).
53 The infinite loop is now broken (with the minimum length unset, that is,
54 zero).
55
56 11. If an assertion that was used as a condition was quantified with a minimum
57 of zero, matching went wrong. In particular, if the whole group had
58 unlimited repetition and could match an empty string, a segfault was
59 likely. The pattern (?(?=0)?)+ is an example that caused this. Perl allows
60 assertions to be quantified, but not if they are being used as conditions,
61 so the above pattern is faulted by Perl. PCRE has now been changed so that
62 it also rejects such patterns.
63
64 12. A possessive capturing group such as (a)*+ with a minimum repeat of zero
65 failed to allow the zero-repeat case if pcre2_exec() was called with an
66 ovector too small to capture the group.
67
68 13. Fixed two bugs in pcretest that were discovered by fuzzing and reported by
69 Red Hat Product Security:
70
71 (a) A crash if /K and /F were both set with the option to save the compiled
72 pattern.
73
74 (b) Another crash if the option to print captured substrings in a callout
75 was combined with setting a null ovector, for example \O\C+ as a subject
76 string.
77
78 14. A pattern such as "((?2){0,1999}())?", which has a group containing a
79 forward reference repeated a large (but limited) number of times within a
80 repeated outer group that has a zero minimum quantifier, caused incorrect
81 code to be compiled, leading to the error "internal error:
82 previously-checked referenced subpattern not found" when an incorrect
83 memory address was read. This bug was reported as "heap overflow",
84 discovered by Kai Lu of Fortinet's FortiGuard Labs and given the CVE number
85 CVE-2015-2325.
86
87 23. A pattern such as "((?+1)(\1))/" containing a forward reference subroutine
88 call within a group that also contained a recursive back reference caused
89 incorrect code to be compiled. This bug was reported as "heap overflow",
90 discovered by Kai Lu of Fortinet's FortiGuard Labs, and given the CVE
91 number CVE-2015-2326.
92
93 24. Computing the size of the JIT read-only data in advance has been a source
94 of various issues, and new ones are still appear unfortunately. To fix
95 existing and future issues, size computation is eliminated from the code,
96 and replaced by on-demand memory allocation.
97
98 25. A pattern such as /(?i)[A-`]/, where characters in the other case are
99 adjacent to the end of the range, and the range contained characters with
100 more than one other case, caused incorrect behaviour when compiled in UTF
101 mode. In that example, the range a-j was left out of the class.
102
103 26. Fix JIT compilation of conditional blocks, which assertion
104 is converted to (*FAIL). E.g: /(?(?!))/.
105
106 27. The pattern /(?(?!)^)/ caused references to random memory. This bug was
107 discovered by the LLVM fuzzer.
108
109 28. The assertion (?!) is optimized to (*FAIL). This was not handled correctly
110 when this assertion was used as a condition, for example (?(?!)a|b). In
111 pcre2_match() it worked by luck; in pcre2_dfa_match() it gave an incorrect
112 error about an unsupported item.
113
114 29. For some types of pattern, for example /Z*(|d*){216}/, the auto-
115 possessification code could take exponential time to complete. A recursion
116 depth limit of 1000 has been imposed to limit the resources used by this
117 optimization.
118
119
120 Version 8.36 26-September-2014
121 ------------------------------
122
123 1. Got rid of some compiler warnings in the C++ modules that were shown up by
124 -Wmissing-field-initializers and -Wunused-parameter.
125
126 2. The tests for quantifiers being too big (greater than 65535) were being
127 applied after reading the number, and stupidly assuming that integer
128 overflow would give a negative number. The tests are now applied as the
129 numbers are read.
130
131 3. Tidy code in pcre_exec.c where two branches that used to be different are
132 now the same.
133
134 4. The JIT compiler did not generate match limit checks for certain
135 bracketed expressions with quantifiers. This may lead to exponential
136 backtracking, instead of returning with PCRE_ERROR_MATCHLIMIT. This
137 issue should be resolved now.
138
139 5. Fixed an issue, which occures when nested alternatives are optimized
140 with table jumps.
141
142 6. Inserted two casts and changed some ints to size_t in the light of some
143 reported 64-bit compiler warnings (Bugzilla 1477).
144
145 7. Fixed a bug concerned with zero-minimum possessive groups that could match
146 an empty string, which sometimes were behaving incorrectly in the
147 interpreter (though correctly in the JIT matcher). This pcretest input is
148 an example:
149
150 '\A(?:[^"]++|"(?:[^"]*+|"")*+")++'
151 NON QUOTED "QUOT""ED" AFTER "NOT MATCHED
152
153 the interpreter was reporting a match of 'NON QUOTED ' only, whereas the
154 JIT matcher and Perl both matched 'NON QUOTED "QUOT""ED" AFTER '. The test
155 for an empty string was breaking the inner loop and carrying on at a lower
156 level, when possessive repeated groups should always return to a higher
157 level as they have no backtrack points in them. The empty string test now
158 occurs at the outer level.
159
160 8. Fixed a bug that was incorrectly auto-possessifying \w+ in the pattern
161 ^\w+(?>\s*)(?<=\w) which caused it not to match "test test".
162
163 9. Give a compile-time error for \o{} (as Perl does) and for \x{} (which Perl
164 doesn't).
165
166 10. Change 8.34/15 introduced a bug that caused the amount of memory needed
167 to hold a pattern to be incorrectly computed (too small) when there were
168 named back references to duplicated names. This could cause "internal
169 error: code overflow" or "double free or corruption" or other memory
170 handling errors.
171
172 11. When named subpatterns had the same prefixes, back references could be
173 confused. For example, in this pattern:
174
175 /(?P<Name>a)?(?P<Name2>b)?(?(<Name>)c|d)*l/
176
177 the reference to 'Name' was incorrectly treated as a reference to a
178 duplicate name.
179
180 12. A pattern such as /^s?c/mi8 where the optional character has more than
181 one "other case" was incorrectly compiled such that it would only try to
182 match starting at "c".
183
184 13. When a pattern starting with \s was studied, VT was not included in the
185 list of possible starting characters; this should have been part of the
186 8.34/18 patch.
187
188 14. If a character class started [\Qx]... where x is any character, the class
189 was incorrectly terminated at the ].
190
191 15. If a pattern that started with a caseless match for a character with more
192 than one "other case" was studied, PCRE did not set up the starting code
193 unit bit map for the list of possible characters. Now it does. This is an
194 optimization improvement, not a bug fix.
195
196 16. The Unicode data tables have been updated to Unicode 7.0.0.
197
198 17. Fixed a number of memory leaks in pcregrep.
199
200 18. Avoid a compiler warning (from some compilers) for a function call with
201 a cast that removes "const" from an lvalue by using an intermediate
202 variable (to which the compiler does not object).
203
204 19. Incorrect code was compiled if a group that contained an internal recursive
205 back reference was optional (had quantifier with a minimum of zero). This
206 example compiled incorrect code: /(((a\2)|(a*)\g<-1>))*/ and other examples
207 caused segmentation faults because of stack overflows at compile time.
208
209 20. A pattern such as /((?(R)a|(?1)))+/, which contains a recursion within a
210 group that is quantified with an indefinite repeat, caused a compile-time
211 loop which used up all the system stack and provoked a segmentation fault.
212 This was not the same bug as 19 above.
213
214 21. Add PCRECPP_EXP_DECL declaration to operator<< in pcre_stringpiece.h.
215 Patch by Mike Frysinger.
216
217
218 Version 8.35 04-April-2014
219 --------------------------
220
221 1. A new flag is set, when property checks are present in an XCLASS.
222 When this flag is not set, PCRE can perform certain optimizations
223 such as studying these XCLASS-es.
224
225 2. The auto-possessification of character sets were improved: a normal
226 and an extended character set can be compared now. Furthermore
227 the JIT compiler optimizes more character set checks.
228
229 3. Got rid of some compiler warnings for potentially uninitialized variables
230 that show up only when compiled with -O2.
231
232 4. A pattern such as (?=ab\K) that uses \K in an assertion can set the start
233 of a match later then the end of the match. The pcretest program was not
234 handling the case sensibly - it was outputting from the start to the next
235 binary zero. It now reports this situation in a message, and outputs the
236 text from the end to the start.
237
238 5. Fast forward search is improved in JIT. Instead of the first three
239 characters, any three characters with fixed position can be searched.
240 Search order: first, last, middle.
241
242 6. Improve character range checks in JIT. Characters are read by an inprecise
243 function now, which returns with an unknown value if the character code is
244 above a certain threshold (e.g: 256). The only limitation is that the value
245 must be bigger than the threshold as well. This function is useful when
246 the characters above the threshold are handled in the same way.
247
248 7. The macros whose names start with RAWUCHAR are placeholders for a future
249 mode in which only the bottom 21 bits of 32-bit data items are used. To
250 make this more memorable for those maintaining the code, the names have
251 been changed to start with UCHAR21, and an extensive comment has been added
252 to their definition.
253
254 8. Add missing (new) files sljitNativeTILEGX.c and sljitNativeTILEGX-encoder.c
255 to the export list in Makefile.am (they were accidentally omitted from the
256 8.34 tarball).
257
258 9. The informational output from pcretest used the phrase "starting byte set"
259 which is inappropriate for the 16-bit and 32-bit libraries. As the output
260 for "first char" and "need char" really means "non-UTF-char", I've changed
261 "byte" to "char", and slightly reworded the output. The documentation about
262 these values has also been (I hope) clarified.
263
264 10. Another JIT related optimization: use table jumps for selecting the correct
265 backtracking path, when more than four alternatives are present inside a
266 bracket.
267
268 11. Empty match is not possible, when the minimum length is greater than zero,
269 and there is no \K in the pattern. JIT should avoid empty match checks in
270 such cases.
271
272 12. In a caseless character class with UCP support, when a character with more
273 than one alternative case was not the first character of a range, not all
274 the alternative cases were added to the class. For example, s and \x{17f}
275 are both alternative cases for S: the class [RST] was handled correctly,
276 but [R-T] was not.
277
278 13. The configure.ac file always checked for pthread support when JIT was
279 enabled. This is not used in Windows, so I have put this test inside a
280 check for the presence of windows.h (which was already tested for).
281
282 14. Improve pattern prefix search by a simplified Boyer-Moore algorithm in JIT.
283 The algorithm provides a way to skip certain starting offsets, and usually
284 faster than linear prefix searches.
285
286 15. Change 13 for 8.20 updated RunTest to check for the 'fr' locale as well
287 as for 'fr_FR' and 'french'. For some reason, however, it then used the
288 Windows-specific input and output files, which have 'french' screwed in.
289 So this could never have worked. One of the problems with locales is that
290 they aren't always the same. I have now updated RunTest so that it checks
291 the output of the locale test (test 3) against three different output
292 files, and it allows the test to pass if any one of them matches. With luck
293 this should make the test pass on some versions of Solaris where it was
294 failing. Because of the uncertainty, the script did not used to stop if
295 test 3 failed; it now does. If further versions of a French locale ever
296 come to light, they can now easily be added.
297
298 16. If --with-pcregrep-bufsize was given a non-integer value such as "50K",
299 there was a message during ./configure, but it did not stop. This now
300 provokes an error. The invalid example in README has been corrected.
301 If a value less than the minimum is given, the minimum value has always
302 been used, but now a warning is given.
303
304 17. If --enable-bsr-anycrlf was set, the special 16/32-bit test failed. This
305 was a bug in the test system, which is now fixed. Also, the list of various
306 configurations that are tested for each release did not have one with both
307 16/32 bits and --enable-bar-anycrlf. It now does.
308
309 18. pcretest was missing "-C bsr" for displaying the \R default setting.
310
311 19. Little endian PowerPC systems are supported now by the JIT compiler.
312
313 20. The fast forward newline mechanism could enter to an infinite loop on
314 certain invalid UTF-8 input. Although we don't support these cases
315 this issue can be fixed by a performance optimization.
316
317 21. Change 33 of 8.34 is not sufficient to ensure stack safety because it does
318 not take account if existing stack usage. There is now a new global
319 variable called pcre_stack_guard that can be set to point to an external
320 function to check stack availability. It is called at the start of
321 processing every parenthesized group.
322
323 22. A typo in the code meant that in ungreedy mode the max/min qualifier
324 behaved like a min-possessive qualifier, and, for example, /a{1,3}b/U did
325 not match "ab".
326
327 23. When UTF was disabled, the JIT program reported some incorrect compile
328 errors. These messages are silenced now.
329
330 24. Experimental support for ARM-64 and MIPS-64 has been added to the JIT
331 compiler.
332
333 25. Change all the temporary files used in RunGrepTest to be different to those
334 used by RunTest so that the tests can be run simultaneously, for example by
335 "make -j check".
336
337
338 Version 8.34 15-December-2013
339 -----------------------------
340
341 1. Add pcre[16|32]_jit_free_unused_memory to forcibly free unused JIT
342 executable memory. Patch inspired by Carsten Klein.
343
344 2. ./configure --enable-coverage defined SUPPORT_GCOV in config.h, although
345 this macro is never tested and has no effect, because the work to support
346 coverage involves only compiling and linking options and special targets in
347 the Makefile. The comment in config.h implied that defining the macro would
348 enable coverage support, which is totally false. There was also support for
349 setting this macro in the CMake files (my fault, I just copied it from
350 configure). SUPPORT_GCOV has now been removed.
351
352 3. Make a small performance improvement in strlen16() and strlen32() in
353 pcretest.
354
355 4. Change 36 for 8.33 left some unreachable statements in pcre_exec.c,
356 detected by the Solaris compiler (gcc doesn't seem to be able to diagnose
357 these cases). There was also one in pcretest.c.
358
359 5. Cleaned up a "may be uninitialized" compiler warning in pcre_exec.c.
360
361 6. In UTF mode, the code for checking whether a group could match an empty
362 string (which is used for indefinitely repeated groups to allow for
363 breaking an infinite loop) was broken when the group contained a repeated
364 negated single-character class with a character that occupied more than one
365 data item and had a minimum repetition of zero (for example, [^\x{100}]* in
366 UTF-8 mode). The effect was undefined: the group might or might not be
367 deemed as matching an empty string, or the program might have crashed.
368
369 7. The code for checking whether a group could match an empty string was not
370 recognizing that \h, \H, \v, \V, and \R must match a character.
371
372 8. Implemented PCRE_INFO_MATCH_EMPTY, which yields 1 if the pattern can match
373 an empty string. If it can, pcretest shows this in its information output.
374
375 9. Fixed two related bugs that applied to Unicode extended grapheme clusters
376 that were repeated with a maximizing qualifier (e.g. \X* or \X{2,5}) when
377 matched by pcre_exec() without using JIT:
378
379 (a) If the rest of the pattern did not match after a maximal run of
380 grapheme clusters, the code for backing up to try with fewer of them
381 did not always back up over a full grapheme when characters that do not
382 have the modifier quality were involved, e.g. Hangul syllables.
383
384 (b) If the match point in a subject started with modifier character, and
385 there was no match, the code could incorrectly back up beyond the match
386 point, and potentially beyond the first character in the subject,
387 leading to a segfault or an incorrect match result.
388
389 10. A conditional group with an assertion condition could lead to PCRE
390 recording an incorrect first data item for a match if no other first data
391 item was recorded. For example, the pattern (?(?=ab)ab) recorded "a" as a
392 first data item, and therefore matched "ca" after "c" instead of at the
393 start.
394
395 11. Change 40 for 8.33 (allowing pcregrep to find empty strings) showed up a
396 bug that caused the command "echo a | ./pcregrep -M '|a'" to loop.
397
398 12. The source of pcregrep now includes z/OS-specific code so that it can be
399 compiled for z/OS as part of the special z/OS distribution.
400
401 13. Added the -T and -TM options to pcretest.
402
403 14. The code in pcre_compile.c for creating the table of named capturing groups
404 has been refactored. Instead of creating the table dynamically during the
405 actual compiling pass, the information is remembered during the pre-compile
406 pass (on the stack unless there are more than 20 named groups, in which
407 case malloc() is used) and the whole table is created before the actual
408 compile happens. This has simplified the code (it is now nearly 150 lines
409 shorter) and prepared the way for better handling of references to groups
410 with duplicate names.
411
412 15. A back reference to a named subpattern when there is more than one of the
413 same name now checks them in the order in which they appear in the pattern.
414 The first one that is set is used for the reference. Previously only the
415 first one was inspected. This change makes PCRE more compatible with Perl.
416
417 16. Unicode character properties were updated from Unicode 6.3.0.
418
419 17. The compile-time code for auto-possessification has been refactored, based
420 on a patch by Zoltan Herczeg. It now happens after instead of during
421 compilation. The code is cleaner, and more cases are handled. The option
422 PCRE_NO_AUTO_POSSESS is added for testing purposes, and the -O and /O
423 options in pcretest are provided to set it. It can also be set by
424 (*NO_AUTO_POSSESS) at the start of a pattern.
425
426 18. The character VT has been added to the default ("C" locale) set of
427 characters that match \s and are generally treated as white space,
428 following this same change in Perl 5.18. There is now no difference between
429 "Perl space" and "POSIX space". Whether VT is treated as white space in
430 other locales depends on the locale.
431
432 19. The code for checking named groups as conditions, either for being set or
433 for being recursed, has been refactored (this is related to 14 and 15
434 above). Processing unduplicated named groups should now be as fast at
435 numerical groups, and processing duplicated groups should be faster than
436 before.
437
438 20. Two patches to the CMake build system, by Alexander Barkov:
439
440 (1) Replace the "source" command by "." in CMakeLists.txt because
441 "source" is a bash-ism.
442
443 (2) Add missing HAVE_STDINT_H and HAVE_INTTYPES_H to config-cmake.h.in;
444 without these the CMake build does not work on Solaris.
445
446 21. Perl has changed its handling of \8 and \9. If there is no previously
447 encountered capturing group of those numbers, they are treated as the
448 literal characters 8 and 9 instead of a binary zero followed by the
449 literals. PCRE now does the same.
450
451 22. Following Perl, added \o{} to specify codepoints in octal, making it
452 possible to specify values greater than 0777 and also making them
453 unambiguous.
454
455 23. Perl now gives an error for missing closing braces after \x{... instead of
456 treating the string as literal. PCRE now does the same.
457
458 24. RunTest used to grumble if an inappropriate test was selected explicitly,
459 but just skip it when running all tests. This make it awkward to run ranges
460 of tests when one of them was inappropriate. Now it just skips any
461 inappropriate tests, as it always did when running all tests.
462
463 25. If PCRE_AUTO_CALLOUT and PCRE_UCP were set for a pattern that contained
464 character types such as \d or \w, too many callouts were inserted, and the
465 data that they returned was rubbish.
466
467 26. In UCP mode, \s was not matching two of the characters that Perl matches,
468 namely NEL (U+0085) and MONGOLIAN VOWEL SEPARATOR (U+180E), though they
469 were matched by \h. The code has now been refactored so that the lists of
470 the horizontal and vertical whitespace characters used for \h and \v (which
471 are defined only in one place) are now also used for \s.
472
473 27. Add JIT support for the 64 bit TileGX architecture.
474 Patch by Jiong Wang (Tilera Corporation).
475
476 28. Possessive quantifiers for classes (both explicit and automatically
477 generated) now use special opcodes instead of wrapping in ONCE brackets.
478
479 29. Whereas an item such as A{4}+ ignored the possessivenes of the quantifier
480 (because it's meaningless), this was not happening when PCRE_CASELESS was
481 set. Not wrong, but inefficient.
482
483 30. Updated perltest.pl to add /u (force Unicode mode) when /W (use Unicode
484 properties for \w, \d, etc) is present in a test regex. Otherwise if the
485 test contains no characters greater than 255, Perl doesn't realise it
486 should be using Unicode semantics.
487
488 31. Upgraded the handling of the POSIX classes [:graph:], [:print:], and
489 [:punct:] when PCRE_UCP is set so as to include the same characters as Perl
490 does in Unicode mode.
491
492 32. Added the "forbid" facility to pcretest so that putting tests into the
493 wrong test files can sometimes be quickly detected.
494
495 33. There is now a limit (default 250) on the depth of nesting of parentheses.
496 This limit is imposed to control the amount of system stack used at compile
497 time. It can be changed at build time by --with-parens-nest-limit=xxx or
498 the equivalent in CMake.
499
500 34. Character classes such as [A-\d] or [a-[:digit:]] now cause compile-time
501 errors. Perl warns for these when in warning mode, but PCRE has no facility
502 for giving warnings.
503
504 35. Change 34 for 8.13 allowed quantifiers on assertions, because Perl does.
505 However, this was not working for (?!) because it is optimized to (*FAIL),
506 for which PCRE does not allow quantifiers. The optimization is now disabled
507 when a quantifier follows (?!). I can't see any use for this, but it makes
508 things uniform.
509
510 36. Perl no longer allows group names to start with digits, so I have made this
511 change also in PCRE. It simplifies the code a bit.
512
513 37. In extended mode, Perl ignores spaces before a + that indicates a
514 possessive quantifier. PCRE allowed a space before the quantifier, but not
515 before the possessive +. It now does.
516
517 38. The use of \K (reset reported match start) within a repeated possessive
518 group such as (a\Kb)*+ was not working.
519
520 40. Document that the same character tables must be used at compile time and
521 run time, and that the facility to pass tables to pcre_exec() and
522 pcre_dfa_exec() is for use only with saved/restored patterns.
523
524 41. Applied Jeff Trawick's patch CMakeLists.txt, which "provides two new
525 features for Builds with MSVC:
526
527 1. Support pcre.rc and/or pcreposix.rc (as is already done for MinGW
528 builds). The .rc files can be used to set FileDescription and many other
529 attributes.
530
531 2. Add an option (-DINSTALL_MSVC_PDB) to enable installation of .pdb files.
532 This allows higher-level build scripts which want .pdb files to avoid
533 hard-coding the exact files needed."
534
535 42. Added support for [[:<:]] and [[:>:]] as used in the BSD POSIX library to
536 mean "start of word" and "end of word", respectively, as a transition aid.
537
538 43. A minimizing repeat of a class containing codepoints greater than 255 in
539 non-UTF 16-bit or 32-bit modes caused an internal error when PCRE was
540 compiled to use the heap for recursion.
541
542 44. Got rid of some compiler warnings for unused variables when UTF but not UCP
543 is configured.
544
545
546 Version 8.33 28-May-2013
547 ------------------------
548
549 1. Added 'U' to some constants that are compared to unsigned integers, to
550 avoid compiler signed/unsigned warnings. Added (int) casts to unsigned
551 variables that are added to signed variables, to ensure the result is
552 signed and can be negated.
553
554 2. Applied patch by Daniel Richard G for quashing MSVC warnings to the
555 CMake config files.
556
557 3. Revise the creation of config.h.generic so that all boolean macros are
558 #undefined, whereas non-boolean macros are #ifndef/#endif-ed. This makes
559 overriding via -D on the command line possible.
560
561 4. Changing the definition of the variable "op" in pcre_exec.c from pcre_uchar
562 to unsigned int is reported to make a quite noticeable speed difference in
563 a specific Windows environment. Testing on Linux did also appear to show
564 some benefit (and it is clearly not harmful). Also fixed the definition of
565 Xop which should be unsigned.
566
567 5. Related to (4), changing the definition of the intermediate variable cc
568 in repeated character loops from pcre_uchar to pcre_uint32 also gave speed
569 improvements.
570
571 6. Fix forward search in JIT when link size is 3 or greater. Also removed some
572 unnecessary spaces.
573
574 7. Adjust autogen.sh and configure.ac to lose warnings given by automake 1.12
575 and later.
576
577 8. Fix two buffer over read issues in 16 and 32 bit modes. Affects JIT only.
578
579 9. Optimizing fast_forward_start_bits in JIT.
580
581 10. Adding support for callouts in JIT, and fixing some issues revealed
582 during this work. Namely:
583
584 (a) Unoptimized capturing brackets incorrectly reset on backtrack.
585
586 (b) Minimum length was not checked before the matching is started.
587
588 11. The value of capture_last that is passed to callouts was incorrect in some
589 cases when there was a capture on one path that was subsequently abandoned
590 after a backtrack. Also, the capture_last value is now reset after a
591 recursion, since all captures are also reset in this case.
592
593 12. The interpreter no longer returns the "too many substrings" error in the
594 case when an overflowing capture is in a branch that is subsequently
595 abandoned after a backtrack.
596
597 13. In the pathological case when an offset vector of size 2 is used, pcretest
598 now prints out the matched string after a yield of 0 or 1.
599
600 14. Inlining subpatterns in recursions, when certain conditions are fulfilled.
601 Only supported by the JIT compiler at the moment.
602
603 15. JIT compiler now supports 32 bit Macs thanks to Lawrence Velazquez.
604
605 16. Partial matches now set offsets[2] to the "bumpalong" value, that is, the
606 offset of the starting point of the matching process, provided the offsets
607 vector is large enough.
608
609 17. The \A escape now records a lookbehind value of 1, though its execution
610 does not actually inspect the previous character. This is to ensure that,
611 in partial multi-segment matching, at least one character from the old
612 segment is retained when a new segment is processed. Otherwise, if there
613 are no lookbehinds in the pattern, \A might match incorrectly at the start
614 of a new segment.
615
616 18. Added some #ifdef __VMS code into pcretest.c to help VMS implementations.
617
618 19. Redefined some pcre_uchar variables in pcre_exec.c as pcre_uint32; this
619 gives some modest performance improvement in 8-bit mode.
620
621 20. Added the PCRE-specific property \p{Xuc} for matching characters that can
622 be expressed in certain programming languages using Universal Character
623 Names.
624
625 21. Unicode validation has been updated in the light of Unicode Corrigendum #9,
626 which points out that "non characters" are not "characters that may not
627 appear in Unicode strings" but rather "characters that are reserved for
628 internal use and have only local meaning".
629
630 22. When a pattern was compiled with automatic callouts (PCRE_AUTO_CALLOUT) and
631 there was a conditional group that depended on an assertion, if the
632 assertion was false, the callout that immediately followed the alternation
633 in the condition was skipped when pcre_exec() was used for matching.
634
635 23. Allow an explicit callout to be inserted before an assertion that is the
636 condition for a conditional group, for compatibility with automatic
637 callouts, which always insert a callout at this point.
638
639 24. In 8.31, (*COMMIT) was confined to within a recursive subpattern. Perl also
640 confines (*SKIP) and (*PRUNE) in the same way, and this has now been done.
641
642 25. (*PRUNE) is now supported by the JIT compiler.
643
644 26. Fix infinite loop when /(?<=(*SKIP)ac)a/ is matched against aa.
645
646 27. Fix the case where there are two or more SKIPs with arguments that may be
647 ignored.
648
649 28. (*SKIP) is now supported by the JIT compiler.
650
651 29. (*THEN) is now supported by the JIT compiler.
652
653 30. Update RunTest with additional test selector options.
654
655 31. The way PCRE handles backtracking verbs has been changed in two ways.
656
657 (1) Previously, in something like (*COMMIT)(*SKIP), COMMIT would override
658 SKIP. Now, PCRE acts on whichever backtracking verb is reached first by
659 backtracking. In some cases this makes it more Perl-compatible, but Perl's
660 rather obscure rules do not always do the same thing.
661
662 (2) Previously, backtracking verbs were confined within assertions. This is
663 no longer the case for positive assertions, except for (*ACCEPT). Again,
664 this sometimes improves Perl compatibility, and sometimes does not.
665
666 32. A number of tests that were in test 2 because Perl did things differently
667 have been moved to test 1, because either Perl or PCRE has changed, and
668 these tests are now compatible.
669
670 32. Backtracking control verbs are now handled in the same way in JIT and
671 interpreter.
672
673 33. An opening parenthesis in a MARK/PRUNE/SKIP/THEN name in a pattern that
674 contained a forward subroutine reference caused a compile error.
675
676 34. Auto-detect and optimize limited repetitions in JIT.
677
678 35. Implement PCRE_NEVER_UTF to lock out the use of UTF, in particular,
679 blocking (*UTF) etc.
680
681 36. In the interpreter, maximizing pattern repetitions for characters and
682 character types now use tail recursion, which reduces stack usage.
683
684 37. The value of the max lookbehind was not correctly preserved if a compiled
685 and saved regex was reloaded on a host of different endianness.
686
687 38. Implemented (*LIMIT_MATCH) and (*LIMIT_RECURSION). As part of the extension
688 of the compiled pattern block, expand the flags field from 16 to 32 bits
689 because it was almost full.
690
691 39. Try madvise first before posix_madvise.
692
693 40. Change 7 for PCRE 7.9 made it impossible for pcregrep to find empty lines
694 with a pattern such as ^$. It has taken 4 years for anybody to notice! The
695 original change locked out all matches of empty strings. This has been
696 changed so that one match of an empty string per line is recognized.
697 Subsequent searches on the same line (for colouring or for --only-matching,
698 for example) do not recognize empty strings.
699
700 41. Applied a user patch to fix a number of spelling mistakes in comments.
701
702 42. Data lines longer than 65536 caused pcretest to crash.
703
704 43. Clarified the data type for length and startoffset arguments for pcre_exec
705 and pcre_dfa_exec in the function-specific man pages, where they were
706 explicitly stated to be in bytes, never having been updated. I also added
707 some clarification to the pcreapi man page.
708
709 44. A call to pcre_dfa_exec() with an output vector size less than 2 caused
710 a segmentation fault.
711
712
713 Version 8.32 30-November-2012
714 -----------------------------
715
716 1. Improved JIT compiler optimizations for first character search and single
717 character iterators.
718
719 2. Supporting IBM XL C compilers for PPC architectures in the JIT compiler.
720 Patch by Daniel Richard G.
721
722 3. Single character iterator optimizations in the JIT compiler.
723
724 4. Improved JIT compiler optimizations for character ranges.
725
726 5. Rename the "leave" variable names to "quit" to improve WinCE compatibility.
727 Reported by Giuseppe D'Angelo.
728
729 6. The PCRE_STARTLINE bit, indicating that a match can occur only at the start
730 of a line, was being set incorrectly in cases where .* appeared inside
731 atomic brackets at the start of a pattern, or where there was a subsequent
732 *PRUNE or *SKIP.
733
734 7. Improved instruction cache flush for POWER/PowerPC.
735 Patch by Daniel Richard G.
736
737 8. Fixed a number of issues in pcregrep, making it more compatible with GNU
738 grep:
739
740 (a) There is now no limit to the number of patterns to be matched.
741
742 (b) An error is given if a pattern is too long.
743
744 (c) Multiple uses of --exclude, --exclude-dir, --include, and --include-dir
745 are now supported.
746
747 (d) --exclude-from and --include-from (multiple use) have been added.
748
749 (e) Exclusions and inclusions now apply to all files and directories, not
750 just to those obtained from scanning a directory recursively.
751
752 (f) Multiple uses of -f and --file-list are now supported.
753
754 (g) In a Windows environment, the default for -d has been changed from
755 "read" (the GNU grep default) to "skip", because otherwise the presence
756 of a directory in the file list provokes an error.
757
758 (h) The documentation has been revised and clarified in places.
759
760 9. Improve the matching speed of capturing brackets.
761
762 10. Changed the meaning of \X so that it now matches a Unicode extended
763 grapheme cluster.
764
765 11. Patch by Daniel Richard G to the autoconf files to add a macro for sorting
766 out POSIX threads when JIT support is configured.
767
768 12. Added support for PCRE_STUDY_EXTRA_NEEDED.
769
770 13. In the POSIX wrapper regcomp() function, setting re_nsub field in the preg
771 structure could go wrong in environments where size_t is not the same size
772 as int.
773
774 14. Applied user-supplied patch to pcrecpp.cc to allow PCRE_NO_UTF8_CHECK to be
775 set.
776
777 15. The EBCDIC support had decayed; later updates to the code had included
778 explicit references to (e.g.) \x0a instead of CHAR_LF. There has been a
779 general tidy up of EBCDIC-related issues, and the documentation was also
780 not quite right. There is now a test that can be run on ASCII systems to
781 check some of the EBCDIC-related things (but is it not a full test).
782
783 16. The new PCRE_STUDY_EXTRA_NEEDED option is now used by pcregrep, resulting
784 in a small tidy to the code.
785
786 17. Fix JIT tests when UTF is disabled and both 8 and 16 bit mode are enabled.
787
788 18. If the --only-matching (-o) option in pcregrep is specified multiple
789 times, each one causes appropriate output. For example, -o1 -o2 outputs the
790 substrings matched by the 1st and 2nd capturing parentheses. A separating
791 string can be specified by --om-separator (default empty).
792
793 19. Improving the first n character searches.
794
795 20. Turn case lists for horizontal and vertical white space into macros so that
796 they are defined only once.
797
798 21. This set of changes together give more compatible Unicode case-folding
799 behaviour for characters that have more than one other case when UCP
800 support is available.
801
802 (a) The Unicode property table now has offsets into a new table of sets of
803 three or more characters that are case-equivalent. The MultiStage2.py
804 script that generates these tables (the pcre_ucd.c file) now scans
805 CaseFolding.txt instead of UnicodeData.txt for character case
806 information.
807
808 (b) The code for adding characters or ranges of characters to a character
809 class has been abstracted into a generalized function that also handles
810 case-independence. In UTF-mode with UCP support, this uses the new data
811 to handle characters with more than one other case.
812
813 (c) A bug that is fixed as a result of (b) is that codepoints less than 256
814 whose other case is greater than 256 are now correctly matched
815 caselessly. Previously, the high codepoint matched the low one, but not
816 vice versa.
817
818 (d) The processing of \h, \H, \v, and \ in character classes now makes use
819 of the new class addition function, using character lists defined as
820 macros alongside the case definitions of 20 above.
821
822 (e) Caseless back references now work with characters that have more than
823 one other case.
824
825 (f) General caseless matching of characters with more than one other case
826 is supported.
827
828 22. Unicode character properties were updated from Unicode 6.2.0
829
830 23. Improved CMake support under Windows. Patch by Daniel Richard G.
831
832 24. Add support for 32-bit character strings, and UTF-32
833
834 25. Major JIT compiler update (code refactoring and bugfixing).
835 Experimental Sparc 32 support is added.
836
837 26. Applied a modified version of Daniel Richard G's patch to create
838 pcre.h.generic and config.h.generic by "make" instead of in the
839 PrepareRelease script.
840
841 27. Added a definition for CHAR_NULL (helpful for the z/OS port), and use it in
842 pcre_compile.c when checking for a zero character.
843
844 28. Introducing a native interface for JIT. Through this interface, the compiled
845 machine code can be directly executed. The purpose of this interface is to
846 provide fast pattern matching, so several sanity checks are not performed.
847 However, feature tests are still performed. The new interface provides
848 1.4x speedup compared to the old one.
849
850 29. If pcre_exec() or pcre_dfa_exec() was called with a negative value for
851 the subject string length, the error given was PCRE_ERROR_BADOFFSET, which
852 was confusing. There is now a new error PCRE_ERROR_BADLENGTH for this case.
853
854 30. In 8-bit UTF-8 mode, pcretest failed to give an error for data codepoints
855 greater than 0x7fffffff (which cannot be represented in UTF-8, even under
856 the "old" RFC 2279). Instead, it ended up passing a negative length to
857 pcre_exec().
858
859 31. Add support for GCC's visibility feature to hide internal functions.
860
861 32. Running "pcretest -C pcre8" or "pcretest -C pcre16" gave a spurious error
862 "unknown -C option" after outputting 0 or 1.
863
864 33. There is now support for generating a code coverage report for the test
865 suite in environments where gcc is the compiler and lcov is installed. This
866 is mainly for the benefit of the developers.
867
868 34. If PCRE is built with --enable-valgrind, certain memory regions are marked
869 unaddressable using valgrind annotations, allowing valgrind to detect
870 invalid memory accesses. This is mainly for the benefit of the developers.
871
872 25. (*UTF) can now be used to start a pattern in any of the three libraries.
873
874 26. Give configure error if --enable-cpp but no C++ compiler found.
875
876
877 Version 8.31 06-July-2012
878 -------------------------
879
880 1. Fixing a wrong JIT test case and some compiler warnings.
881
882 2. Removed a bashism from the RunTest script.
883
884 3. Add a cast to pcre_exec.c to fix the warning "unary minus operator applied
885 to unsigned type, result still unsigned" that was given by an MS compiler
886 on encountering the code "-sizeof(xxx)".
887
888 4. Partial matching support is added to the JIT compiler.
889
890 5. Fixed several bugs concerned with partial matching of items that consist
891 of more than one character:
892
893 (a) /^(..)\1/ did not partially match "aba" because checking references was
894 done on an "all or nothing" basis. This also applied to repeated
895 references.
896
897 (b) \R did not give a hard partial match if \r was found at the end of the
898 subject.
899
900 (c) \X did not give a hard partial match after matching one or more
901 characters at the end of the subject.
902
903 (d) When newline was set to CRLF, a pattern such as /a$/ did not recognize
904 a partial match for the string "\r".
905
906 (e) When newline was set to CRLF, the metacharacter "." did not recognize
907 a partial match for a CR character at the end of the subject string.
908
909 6. If JIT is requested using /S++ or -s++ (instead of just /S+ or -s+) when
910 running pcretest, the text "(JIT)" added to the output whenever JIT is
911 actually used to run the match.
912
913 7. Individual JIT compile options can be set in pcretest by following -s+[+]
914 or /S+[+] with a digit between 1 and 7.
915
916 8. OP_NOT now supports any UTF character not just single-byte ones.
917
918 9. (*MARK) control verb is now supported by the JIT compiler.
919
920 10. The command "./RunTest list" lists the available tests without actually
921 running any of them. (Because I keep forgetting what they all are.)
922
923 11. Add PCRE_INFO_MAXLOOKBEHIND.
924
925 12. Applied a (slightly modified) user-supplied patch that improves performance
926 when the heap is used for recursion (compiled with --disable-stack-for-
927 recursion). Instead of malloc and free for each heap frame each time a
928 logical recursion happens, frames are retained on a chain and re-used where
929 possible. This sometimes gives as much as 30% improvement.
930
931 13. As documented, (*COMMIT) is now confined to within a recursive subpattern
932 call.
933
934 14. As documented, (*COMMIT) is now confined to within a positive assertion.
935
936 15. It is now possible to link pcretest with libedit as an alternative to
937 libreadline.
938
939 16. (*COMMIT) control verb is now supported by the JIT compiler.
940
941 17. The Unicode data tables have been updated to Unicode 6.1.0.
942
943 18. Added --file-list option to pcregrep.
944
945 19. Added binary file support to pcregrep, including the -a, --binary-files,
946 -I, and --text options.
947
948 20. The madvise function is renamed for posix_madvise for QNX compatibility
949 reasons. Fixed by Giuseppe D'Angelo.
950
951 21. Fixed a bug for backward assertions with REVERSE 0 in the JIT compiler.
952
953 22. Changed the option for creating symbolic links for 16-bit man pages from
954 -s to -sf so that re-installing does not cause issues.
955
956 23. Support PCRE_NO_START_OPTIMIZE in JIT as (*MARK) support requires it.
957
958 24. Fixed a very old bug in pcretest that caused errors with restarted DFA
959 matches in certain environments (the workspace was not being correctly
960 retained). Also added to pcre_dfa_exec() a simple plausibility check on
961 some of the workspace data at the beginning of a restart.
962
963 25. \s*\R was auto-possessifying the \s* when it should not, whereas \S*\R
964 was not doing so when it should - probably a typo introduced by SVN 528
965 (change 8.10/14).
966
967 26. When PCRE_UCP was not set, \w+\x{c4} was incorrectly auto-possessifying the
968 \w+ when the character tables indicated that \x{c4} was a word character.
969 There were several related cases, all because the tests for doing a table
970 lookup were testing for characters less than 127 instead of 255.
971
972 27. If a pattern contains capturing parentheses that are not used in a match,
973 their slots in the ovector are set to -1. For those that are higher than
974 any matched groups, this happens at the end of processing. In the case when
975 there were back references that the ovector was too small to contain
976 (causing temporary malloc'd memory to be used during matching), and the
977 highest capturing number was not used, memory off the end of the ovector
978 was incorrectly being set to -1. (It was using the size of the temporary
979 memory instead of the true size.)
980
981 28. To catch bugs like 27 using valgrind, when pcretest is asked to specify an
982 ovector size, it uses memory at the end of the block that it has got.
983
984 29. Check for an overlong MARK name and give an error at compile time. The
985 limit is 255 for the 8-bit library and 65535 for the 16-bit library.
986
987 30. JIT compiler update.
988
989 31. JIT is now supported on jailbroken iOS devices. Thanks for Ruiger
990 Rill for the patch.
991
992 32. Put spaces around SLJIT_PRINT_D in the JIT compiler. Required by CXX11.
993
994 33. Variable renamings in the PCRE-JIT compiler. No functionality change.
995
996 34. Fixed typos in pcregrep: in two places there was SUPPORT_LIBZ2 instead of
997 SUPPORT_LIBBZ2. This caused a build problem when bzip2 but not gzip (zlib)
998 was enabled.
999
1000 35. Improve JIT code generation for greedy plus quantifier.
1001
1002 36. When /((?:a?)*)*c/ or /((?>a?)*)*c/ was matched against "aac", it set group
1003 1 to "aa" instead of to an empty string. The bug affected repeated groups
1004 that could potentially match an empty string.
1005
1006 37. Optimizing single character iterators in JIT.
1007
1008 38. Wide characters specified with \uxxxx in JavaScript mode are now subject to
1009 the same checks as \x{...} characters in non-JavaScript mode. Specifically,
1010 codepoints that are too big for the mode are faulted, and in a UTF mode,
1011 disallowed codepoints are also faulted.
1012
1013 39. If PCRE was compiled with UTF support, in three places in the DFA
1014 matcher there was code that should only have been obeyed in UTF mode, but
1015 was being obeyed unconditionally. In 8-bit mode this could cause incorrect
1016 processing when bytes with values greater than 127 were present. In 16-bit
1017 mode the bug would be provoked by values in the range 0xfc00 to 0xdc00. In
1018 both cases the values are those that cannot be the first data item in a UTF
1019 character. The three items that might have provoked this were recursions,
1020 possessively repeated groups, and atomic groups.
1021
1022 40. Ensure that libpcre is explicitly listed in the link commands for pcretest
1023 and pcregrep, because some OS require shared objects to be explicitly
1024 passed to ld, causing the link step to fail if they are not.
1025
1026 41. There were two incorrect #ifdefs in pcre_study.c, meaning that, in 16-bit
1027 mode, patterns that started with \h* or \R* might be incorrectly matched.
1028
1029
1030 Version 8.30 04-February-2012
1031 -----------------------------
1032
1033 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
1034 name is defined in ctype.h.
1035
1036 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
1037 only in quite long subpatterns.
1038
1039 3. Removed the function pcre_info(), which has been obsolete and deprecated
1040 since it was replaced by pcre_fullinfo() in February 2000.
1041
1042 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
1043 match a (*MARK), and the match failed at the start of the subject, a
1044 reference to memory before the start of the subject could occur. This bug
1045 was introduced by fix 17 of release 8.21.
1046
1047 5. A reference to an unset group with zero minimum repetition was giving
1048 totally wrong answers (in non-JavaScript-compatibility mode). For example,
1049 /(another)?(\1?)test/ matched against "hello world test". This bug was
1050 introduced in release 8.13.
1051
1052 6. Add support for 16-bit character strings (a large amount of work involving
1053 many changes and refactorings).
1054
1055 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
1056 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
1057 from a file.
1058
1059 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
1060 rounding is not applied in this particular case).
1061
1062 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
1063 if they appear, or are escaped, in patterns.
1064
1065 10. Get rid of a number of -Wunused-but-set-variable warnings.
1066
1067 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
1068 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
1069 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
1070 also returns the mark "x". This bug applied to capturing parentheses,
1071 non-capturing parentheses, and atomic parentheses. It also applied to some
1072 assertions.
1073
1074 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
1075 information out of configure.ac instead of relying on pcre.h.generic, which
1076 is not stored in the repository.
1077
1078 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
1079 -lreadline.
1080
1081 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
1082
1083 15. Applied Graycode's patch to put the top-level frame on the stack rather
1084 than the heap when not using the stack for recursion. This gives a
1085 performance improvement in many cases when recursion is not deep.
1086
1087 16. Experimental code added to "pcretest -C" to output the stack frame size.
1088
1089
1090 Version 8.21 12-Dec-2011
1091 ------------------------
1092
1093 1. Updating the JIT compiler.
1094
1095 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
1096 are added as well.
1097
1098 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
1099 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
1100 calling _pcre_jit_exec. Some extra comments are added.
1101
1102 4. (*MARK) settings inside atomic groups that do not contain any capturing
1103 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
1104 was introduced by change 18 for 8.20.
1105
1106 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
1107 ECMA-262 standard.
1108
1109 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
1110 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
1111 This bug was probably introduced by change 9 of 8.13.
1112
1113 7. While fixing 6 above, I noticed that a number of other items were being
1114 incorrectly rejected as "not fixed length". This arose partly because newer
1115 opcodes had not been added to the fixed-length checking code. I have (a)
1116 corrected the bug and added tests for these items, and (b) arranged for an
1117 error to occur if an unknown opcode is encountered while checking for fixed
1118 length instead of just assuming "not fixed length". The items that were
1119 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
1120 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
1121 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
1122
1123 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
1124 being incorrectly compiled and would have given unpredicatble results.
1125
1126 9. A possessively repeated subpattern with minimum repeat count greater than
1127 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
1128 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
1129 the first (A) could occur when it should not.
1130
1131 10. Add a cast and remove a redundant test from the code.
1132
1133 11. JIT should use pcre_malloc/pcre_free for allocation.
1134
1135 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
1136 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
1137 is anything other than /usr, -L is still shown).
1138
1139 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
1140
1141 14. Perl does not support \N without a following name in a [] class; PCRE now
1142 also gives an error.
1143
1144 15. If a forward reference was repeated with an upper limit of around 2000,
1145 it caused the error "internal error: overran compiling workspace". The
1146 maximum number of forward references (including repeats) was limited by the
1147 internal workspace, and dependent on the LINK_SIZE. The code has been
1148 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
1149 the default depends on LINK_SIZE. There is a new upper limit (for safety)
1150 of around 200,000 forward references. While doing this, I also speeded up
1151 the filling in of repeated forward references.
1152
1153 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
1154 incorrectly expecting the subject to contain another "a" after the start.
1155
1156 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
1157 in the match, the SKIP should be ignored. This was not happening; instead
1158 the SKIP was being treated as NOMATCH. For patterns such as
1159 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
1160 tested.
1161
1162 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
1163 now much more compatible with Perl, in particular in cases where the result
1164 is a non-match for a non-anchored pattern. For example, if
1165 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
1166 "m", where previously it did not return a name. A side effect of this
1167 change is that for partial matches, the last encountered mark name is
1168 returned, as for non matches. A number of tests that were previously not
1169 Perl-compatible have been moved into the Perl-compatible test files. The
1170 refactoring has had the pleasing side effect of removing one argument from
1171 the match() function, thus reducing its stack requirements.
1172
1173 19. If the /S+ option was used in pcretest to study a pattern using JIT,
1174 subsequent uses of /S (without +) incorrectly behaved like /S+.
1175
1176 21. Retrieve executable code size support for the JIT compiler and fixing
1177 some warnings.
1178
1179 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
1180 not work when the shorter character appeared right at the end of the
1181 subject string.
1182
1183 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
1184 systems.
1185
1186 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
1187 output it when the /M option is used in pcretest.
1188
1189 25. The CheckMan script was not being included in the distribution. Also, added
1190 an explicit "perl" to run Perl scripts from the PrepareRelease script
1191 because this is reportedly needed in Windows.
1192
1193 26. If study data was being save in a file and studying had not found a set of
1194 "starts with" bytes for the pattern, the data written to the file (though
1195 never used) was taken from uninitialized memory and so caused valgrind to
1196 complain.
1197
1198 27. Updated RunTest.bat as provided by Sheri Pierce.
1199
1200 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
1201
1202 29. Computation of memory usage for the table of capturing group names was
1203 giving an unnecessarily large value.
1204
1205
1206 Version 8.20 21-Oct-2011
1207 ------------------------
1208
1209 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
1210 a POSIX class. After further experiments with Perl, which convinced me that
1211 Perl has bugs and confusions, a closing square bracket is no longer allowed
1212 in a POSIX name. This bug also affected patterns with classes that started
1213 with full stops.
1214
1215 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
1216 captured substring, but while checking the failing first alternative,
1217 substring 1 is temporarily captured. If the output vector supplied to
1218 pcre_exec() was not big enough for this capture, the yield of the function
1219 was still zero ("insufficient space for captured substrings"). This cannot
1220 be totally fixed without adding another stack variable, which seems a lot
1221 of expense for a edge case. However, I have improved the situation in cases
1222 such as /(a)(b)x|abc/ matched against "abc", where the return code
1223 indicates that fewer than the maximum number of slots in the ovector have
1224 been set.
1225
1226 3. Related to (2) above: when there are more back references in a pattern than
1227 slots in the output vector, pcre_exec() uses temporary memory during
1228 matching, and copies in the captures as far as possible afterwards. It was
1229 using the entire output vector, but this conflicts with the specification
1230 that only 2/3 is used for passing back captured substrings. Now it uses
1231 only the first 2/3, for compatibility. This is, of course, another edge
1232 case.
1233
1234 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
1235 main code base, and can be used by building with --enable-jit. When this is
1236 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
1237 runtime --no-jit option is given.
1238
1239 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
1240 ovector, the return from the function was zero, implying that there were
1241 other matches that did not fit. The correct "exactly full" value is now
1242 returned.
1243
1244 6. If a subpattern that was called recursively or as a subroutine contained
1245 (*PRUNE) or any other control that caused it to give a non-standard return,
1246 invalid errors such as "Error -26 (nested recursion at the same subject
1247 position)" or even infinite loops could occur.
1248
1249 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
1250 computing the minimum length on reaching *ACCEPT, and so ended up with the
1251 wrong value of 1 rather than 0. Further investigation indicates that
1252 computing a minimum subject length in the presence of *ACCEPT is difficult
1253 (think back references, subroutine calls), and so I have changed the code
1254 so that no minimum is registered for a pattern that contains *ACCEPT.
1255
1256 8. If (*THEN) was present in the first (true) branch of a conditional group,
1257 it was not handled as intended. [But see 16 below.]
1258
1259 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
1260 Sheri Pierce.
1261
1262 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
1263 the first byte in a match must be "a".
1264
1265 11. Change 17 for 8.13 increased the recursion depth for patterns like
1266 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
1267 pattern contains any instances of (*THEN). If it does not, the old
1268 optimizations are restored. It would be nice to do this on a per-group
1269 basis, but at the moment that is not feasible.
1270
1271 12. In some environments, the output of pcretest -C is CRLF terminated. This
1272 broke RunTest's code that checks for the link size. A single white space
1273 character after the value is now allowed for.
1274
1275 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
1276 For "fr", it uses the Windows-specific input and output files.
1277
1278 14. If (*THEN) appeared in a group that was called recursively or as a
1279 subroutine, it did not work as intended. [But see next item.]
1280
1281 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
1282 pattern fragments (but not containing any | characters). If A and B are
1283 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
1284 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
1285 D. In other words, Perl considers parentheses that do not contain any |
1286 characters to be part of a surrounding alternative, whereas PCRE was
1287 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
1288 differently. PCRE now behaves in the same way as Perl, except in the case
1289 of subroutine/recursion calls such as (?1) which have in any case always
1290 been different (but PCRE had them first :-).
1291
1292 16. Related to 15 above: Perl does not treat the | in a conditional group as
1293 creating alternatives. Such a group is treated in the same way as an
1294 ordinary group without any | characters when processing (*THEN). PCRE has
1295 been changed to match Perl's behaviour.
1296
1297 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
1298 RunGrepTest script failed.
1299
1300 18. Change 22 for version 13 caused atomic groups to use more stack. This is
1301 inevitable for groups that contain captures, but it can lead to a lot of
1302 stack use in large patterns. The old behaviour has been restored for atomic
1303 groups that do not contain any capturing parentheses.
1304
1305 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
1306 suppress the check for a minimum subject length at run time. (If it was
1307 given to pcre_exec() or pcre_dfa_exec() it did work.)
1308
1309 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
1310 fail to work when decoding hex characters in data strings in EBCDIC
1311 environments.
1312
1313 21. It appears that in at least one Mac OS environment, the isxdigit() function
1314 is implemented as a macro that evaluates to its argument more than once,
1315 contravening the C 90 Standard (I haven't checked a later standard). There
1316 was an instance in pcretest which caused it to go wrong when processing
1317 \x{...} escapes in subject strings. The has been rewritten to avoid using
1318 things like p++ in the argument of isxdigit().
1319
1320
1321 Version 8.13 16-Aug-2011
1322 ------------------------
1323
1324 1. The Unicode data tables have been updated to Unicode 6.0.0.
1325
1326 2. Two minor typos in pcre_internal.h have been fixed.
1327
1328 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
1329 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
1330 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
1331
1332 4. There were a number of related bugs in the code for matching backrefences
1333 caselessly in UTF-8 mode when codes for the characters concerned were
1334 different numbers of bytes. For example, U+023A and U+2C65 are an upper
1335 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
1336 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
1337 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
1338 2-byte code at the end of the subject (it thought there wasn't enough data
1339 left).
1340
1341 5. Comprehensive information about what went wrong is now returned by
1342 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
1343 as the output vector has at least 2 elements. The offset of the start of
1344 the failing character and a reason code are placed in the vector.
1345
1346 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
1347 now returned is for the first byte of the failing character, instead of the
1348 last byte inspected. This is an incompatible change, but I hope it is small
1349 enough not to be a problem. It makes the returned offset consistent with
1350 pcre_exec() and pcre_dfa_exec().
1351
1352 7. pcretest now gives a text phrase as well as the error number when
1353 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
1354 failure, the offset and reason code are output.
1355
1356 8. When \R was used with a maximizing quantifier it failed to skip backwards
1357 over a \r\n pair if the subsequent match failed. Instead, it just skipped
1358 back over a single character (\n). This seems wrong (because it treated the
1359 two characters as a single entity when going forwards), conflicts with the
1360 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
1361 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
1362 has been changed.
1363
1364 9. Some internal refactoring has changed the processing so that the handling
1365 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
1366 time (the PCRE_DOTALL option was changed this way some time ago: version
1367 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
1368 which was always a bit of a fudge. It also means that there is one less
1369 argument for the match() function, which reduces its stack requirements
1370 slightly. This change also fixes an incompatibility with Perl: the pattern
1371 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
1372
1373 10. More internal refactoring has drastically reduced the number of recursive
1374 calls to match() for possessively repeated groups such as (abc)++ when
1375 using pcre_exec().
1376
1377 11. While implementing 10, a number of bugs in the handling of groups were
1378 discovered and fixed:
1379
1380 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
1381 (a|)*(?1) gave a compile-time internal error.
1382 ((a|)+)+ did not notice that the outer group could match an empty string.
1383 (^a|^)+ was not marked as anchored.
1384 (.*a|.*)+ was not marked as matching at start or after a newline.
1385
1386 12. Yet more internal refactoring has removed another argument from the match()
1387 function. Special calls to this function are now indicated by setting a
1388 value in a variable in the "match data" data block.
1389
1390 13. Be more explicit in pcre_study() instead of relying on "default" for
1391 opcodes that mean there is no starting character; this means that when new
1392 ones are added and accidentally left out of pcre_study(), testing should
1393 pick them up.
1394
1395 14. The -s option of pcretest has been documented for ages as being an old
1396 synonym of -m (show memory usage). I have changed it to mean "force study
1397 for every regex", that is, assume /S for every regex. This is similar to -i
1398 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
1399 using it. It makes it easier to run collections of tests with and without
1400 study enabled, and thereby test pcre_study() more easily. All the standard
1401 tests are now run with and without -s (but some patterns can be marked as
1402 "never study" - see 20 below).
1403
1404 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
1405 restoration of the capturing data to the outer values was not happening
1406 correctly.
1407
1408 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
1409 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
1410 pattern had matched an empty string, and so incorrectly returned a no
1411 match.
1412
1413 17. There was optimizing code for the last branch of non-capturing parentheses,
1414 and also for the obeyed branch of a conditional subexpression, which used
1415 tail recursion to cut down on stack usage. Unfortunately, now that there is
1416 the possibility of (*THEN) occurring in these branches, tail recursion is
1417 no longer possible because the return has to be checked for (*THEN). These
1418 two optimizations have therefore been removed. [But see 8.20/11 above.]
1419
1420 18. If a pattern containing \R was studied, it was assumed that \R always
1421 matched two bytes, thus causing the minimum subject length to be
1422 incorrectly computed because \R can also match just one byte.
1423
1424 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
1425 was incorrectly computed.
1426
1427 20. If /S is present twice on a test pattern in pcretest input, it now
1428 *disables* studying, thereby overriding the use of -s on the command line
1429 (see 14 above). This is necessary for one or two tests to keep the output
1430 identical in both cases.
1431
1432 21. When (*ACCEPT) was used in an assertion that matched an empty string and
1433 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
1434
1435 22. When an atomic group that contained a capturing parenthesis was
1436 successfully matched, but the branch in which it appeared failed, the
1437 capturing was not being forgotten if a higher numbered group was later
1438 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
1439 group 1 to "a", when in fact it should be unset. This applied to multi-
1440 branched capturing and non-capturing groups, repeated or not, and also to
1441 positive assertions (capturing in negative assertions does not happen
1442 in PCRE) and also to nested atomic groups.
1443
1444 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
1445 subject after a captured substring, to make it easier to tell which of a
1446 number of identical substrings has been captured.
1447
1448 24. The way atomic groups are processed by pcre_exec() has been changed so that
1449 if they are repeated, backtracking one repetition now resets captured
1450 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
1451 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
1452 "aaa". Previously, it would have been "a". As part of this code
1453 refactoring, the way recursive calls are handled has also been changed.
1454
1455 25. If an assertion condition captured any substrings, they were not passed
1456 back unless some other capturing happened later. For example, if
1457 (?(?=(a))a) was matched against "a", no capturing was returned.
1458
1459 26. When studying a pattern that contained subroutine calls or assertions,
1460 the code for finding the minimum length of a possible match was handling
1461 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
1462 group 1 called group 2 while simultaneously a separate group 2 called group
1463 1). A stack overflow occurred in this case. I have fixed this by limiting
1464 the recursion depth to 10.
1465
1466 27. Updated RunTest.bat in the distribution to the version supplied by Tom
1467 Fortmann. This supports explicit test numbers on the command line, and has
1468 argument validation and error reporting.
1469
1470 28. An instance of \X with an unlimited repeat could fail if at any point the
1471 first character it looked at was a mark character.
1472
1473 29. Some minor code refactoring concerning Unicode properties and scripts
1474 should reduce the stack requirement of match() slightly.
1475
1476 30. Added the '=' option to pcretest to check the setting of unused capturing
1477 slots at the end of the pattern, which are documented as being -1, but are
1478 not included in the return count.
1479
1480 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
1481 compiled something random. Now it gives a compile-time error (as does
1482 Perl).
1483
1484 32. A *MARK encountered during the processing of a positive assertion is now
1485 recorded and passed back (compatible with Perl).
1486
1487 33. If --only-matching or --colour was set on a pcregrep call whose pattern
1488 had alternative anchored branches, the search for a second match in a line
1489 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
1490 matched the line "0102" twice. The same bug affected patterns that started
1491 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
1492 twice.
1493
1494 34. Previously, PCRE did not allow quantification of assertions. However, Perl
1495 does, and because of capturing effects, quantifying parenthesized
1496 assertions may at times be useful. Quantifiers are now allowed for
1497 parenthesized assertions.
1498
1499 35. A minor code tidy in pcre_compile() when checking options for \R usage.
1500
1501 36. \g was being checked for fancy things in a character class, when it should
1502 just be a literal "g".
1503
1504 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
1505 appearance of a nested POSIX class supersedes an apparent external class.
1506 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
1507 unescaped square brackets may also appear as part of class names. For
1508 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
1509 more like Perl. (But see 8.20/1 above.)
1510
1511 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
1512 was because it thought it was \N{name}, which is not supported).
1513
1514 39. Add minix to OS list not supporting the -S option in pcretest.
1515
1516 40. PCRE tries to detect cases of infinite recursion at compile time, but it
1517 cannot analyze patterns in sufficient detail to catch mutual recursions
1518 such as ((?1))((?2)). There is now a runtime test that gives an error if a
1519 subgroup is called recursively as a subpattern for a second time at the
1520 same position in the subject string. In previous releases this might have
1521 been caught by the recursion limit, or it might have run out of stack.
1522
1523 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
1524 happen only once. PCRE was, however incorrectly giving a compile time error
1525 "recursive call could loop indefinitely" because it cannot analyze the
1526 pattern in sufficient detail. The compile time test no longer happens when
1527 PCRE is compiling a conditional subpattern, but actual runaway loops are
1528 now caught at runtime (see 40 above).
1529
1530 42. It seems that Perl allows any characters other than a closing parenthesis
1531 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
1532 has been changed to be the same.
1533
1534 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
1535 as not to get warnings when autogen.sh is called. Also changed
1536 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
1537
1538 44. To help people who use pcregrep to scan files containing exceedingly long
1539 lines, the following changes have been made:
1540
1541 (a) The default value of the buffer size parameter has been increased from
1542 8K to 20K. (The actual buffer used is three times this size.)
1543
1544 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
1545 PCRE is built.
1546
1547 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
1548 to be set at run time.
1549
1550 (d) Numerical values in pcregrep options can be followed by K or M, for
1551 example --buffer-size=50K.
1552
1553 (e) If a line being scanned overflows pcregrep's buffer, an error is now
1554 given and the return code is set to 2.
1555
1556 45. Add a pointer to the latest mark to the callout data block.
1557
1558 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
1559 partial match of an empty string instead of no match. This was specific to
1560 the use of ".".
1561
1562 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
1563 complete match instead of a partial match. This bug was dependent on both
1564 the PCRE_UTF8 and PCRE_DOTALL options being set.
1565
1566 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
1567 starting byte set, because \b was not being ignored.
1568
1569
1570 Version 8.12 15-Jan-2011
1571 ------------------------
1572
1573 1. Fixed some typos in the markup of the man pages, and wrote a script that
1574 checks for such things as part of the documentation building process.
1575
1576 2. On a big-endian 64-bit system, pcregrep did not correctly process the
1577 --match-limit and --recursion-limit options (added for 8.11). In
1578 particular, this made one of the standard tests fail. (The integer value
1579 went into the wrong half of a long int.)
1580
1581 3. If the --colour option was given to pcregrep with -v (invert match), it
1582 did strange things, either producing crazy output, or crashing. It should,
1583 of course, ignore a request for colour when reporting lines that do not
1584 match.
1585
1586 4. Another pcregrep bug caused similar problems if --colour was specified with
1587 -M (multiline) and the pattern match finished with a line ending.
1588
1589 5. In pcregrep, when a pattern that ended with a literal newline sequence was
1590 matched in multiline mode, the following line was shown as part of the
1591 match. This seems wrong, so I have changed it.
1592
1593 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
1594 the check for further matches in the same line (so they could be coloured)
1595 to overrun the end of the current line. If another match was found, it was
1596 incorrectly shown (and then shown again when found in the next line).
1597
1598 7. If pcregrep was compiled under Windows, there was a reference to the
1599 function pcregrep_exit() before it was defined. I am assuming this was
1600 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
1601 reported by a user. I've moved the definition above the reference.
1602
1603
1604 Version 8.11 10-Dec-2010
1605 ------------------------
1606
1607 1. (*THEN) was not working properly if there were untried alternatives prior
1608 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
1609 backtracked to try for "b" instead of moving to the next alternative branch
1610 at the same level (in this case, to look for "c"). The Perl documentation
1611 is clear that when (*THEN) is backtracked onto, it goes to the "next
1612 alternative in the innermost enclosing group".
1613
1614 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
1615 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
1616 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
1617 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
1618 (*THEN).
1619
1620 3. If \s appeared in a character class, it removed the VT character from
1621 the class, even if it had been included by some previous item, for example
1622 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
1623 of \s, but is part of the POSIX "space" class.)
1624
1625 4. A partial match never returns an empty string (because you can always
1626 match an empty string at the end of the subject); however the checking for
1627 an empty string was starting at the "start of match" point. This has been
1628 changed to the "earliest inspected character" point, because the returned
1629 data for a partial match starts at this character. This means that, for
1630 example, /(?<=abc)def/ gives a partial match for the subject "abc"
1631 (previously it gave "no match").
1632
1633 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
1634 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
1635 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
1636 has an implication that the given string is incomplete (because a partial
1637 match is preferred over a full match). For this reason, these items now
1638 give a partial match in this situation. [Aside: previously, the one case
1639 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
1640 match rather than a full match, which was wrong by the old rules, but is
1641 now correct.]
1642
1643 6. There was a bug in the handling of #-introduced comments, recognized when
1644 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
1645 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
1646 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
1647 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
1648 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
1649 places in pcre_compile().
1650
1651 7. Related to (6) above, when pcre_compile() was skipping #-introduced
1652 comments when looking ahead for named forward references to subpatterns,
1653 the only newline sequence it recognized was NL. It now handles newlines
1654 according to the set newline convention.
1655
1656 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
1657 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
1658 cater for a lack of strerror(). These oversights have been fixed.
1659
1660 9. Added --match-limit and --recursion-limit to pcregrep.
1661
1662 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
1663
1664 11. When the -o option was used, pcregrep was setting a return code of 1, even
1665 when matches were found, and --line-buffered was not being honoured.
1666
1667 12. Added an optional parentheses number to the -o and --only-matching options
1668 of pcregrep.
1669
1670 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
1671 can match an empty string. The code to do it in pcretest and pcredemo
1672 needed fixing:
1673
1674 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
1675 only one byte after an empty string match just before CRLF (this case
1676 just got forgotten; "any" and "anycrlf" were OK).
1677
1678 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
1679 mode when an empty string match preceded an ASCII character followed by
1680 a non-ASCII character. (The code for advancing by one character rather
1681 than one byte was nonsense.)
1682
1683 (c) The pcredemo.c sample program did not have any code at all to handle
1684 the cases when CRLF is a valid newline sequence.
1685
1686 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
1687 as a starting offset was within the subject string. There is now a new
1688 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
1689 negative or greater than the length of the string. In order to test this,
1690 pcretest is extended to allow the setting of negative starting offsets.
1691
1692 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
1693 starting offset points to the beginning of a UTF-8 character was
1694 unnecessarily clumsy. I tidied it up.
1695
1696 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
1697 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
1698
1699 17. Nobody had reported that the --include_dir option, which was added in
1700 release 7.7 should have been called --include-dir (hyphen, not underscore)
1701 for compatibility with GNU grep. I have changed it to --include-dir, but
1702 left --include_dir as an undocumented synonym, and the same for
1703 --exclude-dir, though that is not available in GNU grep, at least as of
1704 release 2.5.4.
1705
1706 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
1707 characters from a string of bytes) have been redefined so as not to use
1708 loops, in order to improve performance in some environments. At the same
1709 time, I abstracted some of the common code into auxiliary macros to save
1710 repetition (this should not affect the compiled code).
1711
1712 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
1713 compile-time error is now given if \c is not followed by an ASCII
1714 character, that is, a byte less than 128. (In EBCDIC mode, the code is
1715 different, and any byte value is allowed.)
1716
1717 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
1718 START_OPTIMIZE option, which is now allowed at compile time - but just
1719 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
1720 to pcregrep and other applications that have no direct access to PCRE
1721 options. The new /Y option in pcretest sets this option when calling
1722 pcre_compile().
1723
1724 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
1725 back references. Groups containing recursive back references were forced to
1726 be atomic by that change, but in the case of named groups, the amount of
1727 memory required was incorrectly computed, leading to "Failed: internal
1728 error: code overflow". This has been fixed.
1729
1730 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
1731 pcretest.c, to avoid build problems in some Borland environments.
1732
1733
1734 Version 8.10 25-Jun-2010
1735 ------------------------
1736
1737 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
1738 THEN.
1739
1740 2. (*ACCEPT) was not working when inside an atomic group.
1741
1742 3. Inside a character class, \B is treated as a literal by default, but
1743 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
1744 causes the error). The code is unchanged, but I tidied the documentation.
1745
1746 4. Inside a character class, PCRE always treated \R and \X as literals,
1747 whereas Perl faults them if its -w option is set. I have changed PCRE so
1748 that it faults them when PCRE_EXTRA is set.
1749
1750 5. Added support for \N, which always matches any character other than
1751 newline. (It is the same as "." when PCRE_DOTALL is not set.)
1752
1753 6. When compiling pcregrep with newer versions of gcc which may have
1754 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
1755 declared with attribute warn_unused_result" were given. Just casting the
1756 result to (void) does not stop the warnings; a more elaborate fudge is
1757 needed. I've used a macro to implement this.
1758
1759 7. Minor change to pcretest.c to avoid a compiler warning.
1760
1761 8. Added four artifical Unicode properties to help with an option to make
1762 \s etc use properties (see next item). The new properties are: Xan
1763 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
1764
1765 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
1766 use Unicode properties. (*UCP) at the start of a pattern can be used to set
1767 this option. Modified pcretest to add /W to test this facility. Added
1768 REG_UCP to make it available via the POSIX interface.
1769
1770 10. Added --line-buffered to pcregrep.
1771
1772 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
1773 studied, and the match started with a letter with a code point greater than
1774 127 whose first byte was different to the first byte of the other case of
1775 the letter, the other case of this starting letter was not recognized
1776 (#976).
1777
1778 12. If a pattern that was studied started with a repeated Unicode property
1779 test, for example, \p{Nd}+, there was the theoretical possibility of
1780 setting up an incorrect bitmap of starting bytes, but fortunately it could
1781 not have actually happened in practice until change 8 above was made (it
1782 added property types that matched character-matching opcodes).
1783
1784 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
1785 possible starting bytes for non-anchored patterns.
1786
1787 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
1788 \R, and also a number of cases that involve Unicode properties, both
1789 explicit and implicit when PCRE_UCP is set.
1790
1791 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
1792 input, it could crash or give wrong results if characters with values
1793 greater than 0xc0 were present in the subject string. (Detail: it assumed
1794 UTF-8 input when processing these items.)
1795
1796 16. Added a lot of (int) casts to avoid compiler warnings in systems where
1797 size_t is 64-bit (#991).
1798
1799 17. Added a check for running out of memory when PCRE is compiled with
1800 --disable-stack-for-recursion (#990).
1801
1802 18. If the last data line in a file for pcretest does not have a newline on
1803 the end, a newline was missing in the output.
1804
1805 19. The default pcre_chartables.c file recognizes only ASCII characters (values
1806 less than 128) in its various bitmaps. However, there is a facility for
1807 generating tables according to the current locale when PCRE is compiled. It
1808 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
1809 characters, are recognized by isspace() and therefore were getting set in
1810 these tables, and indeed these tables seem to approximate to ISO 8859. This
1811 caused a problem in UTF-8 mode when pcre_study() was used to create a list
1812 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
1813 which of course cannot start UTF-8 characters. I have changed the code so
1814 that only real ASCII characters (less than 128) and the correct starting
1815 bytes for UTF-8 encodings are set for characters greater than 127 when in
1816 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
1817 altogether.)
1818
1819 20. Added the /T option to pcretest so as to be able to run tests with non-
1820 standard character tables, thus making it possible to include the tests
1821 used for 19 above in the standard set of tests.
1822
1823 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
1824 reference to a subpattern the other side of a comment that contains an
1825 opening parenthesis caused either an internal compiling error, or a
1826 reference to the wrong subpattern.
1827
1828
1829 Version 8.02 19-Mar-2010
1830 ------------------------
1831
1832 1. The Unicode data tables have been updated to Unicode 5.2.0.
1833
1834 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
1835 configured.
1836
1837 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
1838 original author of that file, following a query about its status.
1839
1840 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
1841 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
1842
1843 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
1844 quantifier applied to a forward-referencing subroutine call, could compile
1845 incorrect code or give the error "internal error: previously-checked
1846 referenced subpattern not found".
1847
1848 6. Both MS Visual Studio and Symbian OS have problems with initializing
1849 variables to point to external functions. For these systems, therefore,
1850 pcre_malloc etc. are now initialized to local functions that call the
1851 relevant global functions.
1852
1853 7. There were two entries missing in the vectors called coptable and poptable
1854 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
1855 I've fixed the data, and added a kludgy way of testing at compile time that
1856 the lengths are correct (equal to the number of opcodes).
1857
1858 8. Following on from 7, I added a similar kludge to check the length of the
1859 eint vector in pcreposix.c.
1860
1861 9. Error texts for pcre_compile() are held as one long string to avoid too
1862 much relocation at load time. To find a text, the string is searched,
1863 counting zeros. There was no check for running off the end of the string,
1864 which could happen if a new error number was added without updating the
1865 string.
1866
1867 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
1868
1869 11. \K was not working if it appeared in an atomic group or in a group that
1870 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
1871 \K is "not well defined" if used in an assertion. PCRE now accepts it if
1872 the assertion is positive, but not if it is negative.
1873
1874 12. Change 11 fortuitously reduced the size of the stack frame used in the
1875 "match()" function of pcre_exec.c by one pointer. Forthcoming
1876 implementation of support for (*MARK) will need an extra pointer on the
1877 stack; I have reserved it now, so that the stack frame size does not
1878 decrease.
1879
1880 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
1881 item in branch that calls a recursion is a subroutine call - as in the
1882 second branch in the above example - was incorrectly given the compile-
1883 time error "recursive call could loop indefinitely" because pcre_compile()
1884 was not correctly checking the subroutine for matching a non-empty string.
1885
1886 14. The checks for overrunning compiling workspace could trigger after an
1887 overrun had occurred. This is a "should never occur" error, but it can be
1888 triggered by pathological patterns such as hundreds of nested parentheses.
1889 The checks now trigger 100 bytes before the end of the workspace.
1890
1891 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
1892
1893
1894 Version 8.01 19-Jan-2010
1895 ------------------------
1896
1897 1. If a pattern contained a conditional subpattern with only one branch (in
1898 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
1899 computed the wrong minimum data length (which is of course zero for such
1900 subpatterns). This could cause incorrect "no match" results.
1901
1902 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
1903 the pattern is reset in the first branch, pcre_compile() failed with
1904 "internal error: code overflow at offset...". This happened only when
1905 the reset was to the original external option setting. (An optimization
1906 abstracts leading options settings into an external setting, which was the
1907 cause of this.)
1908
1909 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
1910 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
1911 assertion pattern did not match (meaning that the assertion was true), it
1912 was incorrectly treated as false if the SKIP had been reached during the
1913 matching. This also applied to assertions used as conditions.
1914
1915 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
1916 assertion subpattern, including such a pattern used as a condition,
1917 unpredictable results occurred, instead of the error return
1918 PCRE_ERROR_DFA_UITEM.
1919
1920 5. The C++ GlobalReplace function was not working like Perl for the special
1921 situation when an empty string is matched. It now does the fancy magic
1922 stuff that is necessary.
1923
1924 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
1925 removed. (These were left over from very, very early versions of PCRE.)
1926
1927 7. Some cosmetic changes to the code to make life easier when compiling it
1928 as part of something else:
1929
1930 (a) Change DEBUG to PCRE_DEBUG.
1931
1932 (b) In pcre_compile(), rename the member of the "branch_chain" structure
1933 called "current" as "current_branch", to prevent a collision with the
1934 Linux macro when compiled as a kernel module.
1935
1936 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
1937 prevent a collision with the Linux macro when compiled as a kernel
1938 module.
1939
1940 8. In pcre_compile() there are some checks for integer overflows that used to
1941 cast potentially large values to (double). This has been changed to that
1942 when building, a check for int64_t is made, and if it is found, it is used
1943 instead, thus avoiding the use of floating point arithmetic. (There is no
1944 other use of FP in PCRE.) If int64_t is not found, the fallback is to
1945 double.
1946
1947 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
1948 2005 (difference between two addresses compared to an unsigned value).
1949
1950 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
1951 custom one, because of the following reported problem in Windows:
1952
1953 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
1954 under Win32.
1955 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
1956 therefore missing the function definition.
1957 - The compiler thus generates a "C" signature for the test function.
1958 - The linker fails to find the "C" function.
1959 - PCRE fails to configure if asked to do so against libbz2.
1960
1961 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
1962 messages were output:
1963
1964 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
1965 rerunning libtoolize, to keep the correct libtool macros in-tree.
1966 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
1967
1968 I have done both of these things.
1969
1970 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
1971 most of the time, it *can* run out if it is given a pattern that contains a
1972 runaway infinite recursion. I updated the discussion in the pcrestack man
1973 page.
1974
1975 13. Now that we have gone to the x.xx style of version numbers, the minor
1976 version may start with zero. Using 08 or 09 is a bad idea because users
1977 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
1978 interpreted as invalid octal numbers. I've updated the previous comment in
1979 configure.ac, and also added a check that gives an error if 08 or 09 are
1980 used.
1981
1982 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
1983 causing partial matching to fail when the end of the subject matched \W
1984 in a UTF-8 pattern where \W was quantified with a minimum of 3.
1985
1986 15. There were some discrepancies between the declarations in pcre_internal.h
1987 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
1988 their definitions. The declarations used "const uschar *" and the
1989 definitions used USPTR. Even though USPTR is normally defined as "const
1990 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
1991 reported that: "This difference in casting confuses some C++ compilers, for
1992 example, SunCC recognizes above declarations as different functions and
1993 generates broken code for hbpcre." I have changed the declarations to use
1994 USPTR.
1995
1996 16. GNU libtool is named differently on some systems. The autogen.sh script now
1997 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
1998 (FreeBSD).
1999
2000 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
2001 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
2002 comment: "Figure out how to create a longlong from a string: strtoll and
2003 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
2004 instance, but it only takes 2 args instead of 3!"
2005
2006 18. A subtle bug concerned with back references has been fixed by a change of
2007 specification, with a corresponding code fix. A pattern such as
2008 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
2009 refers, was giving matches when it shouldn't. For example, xa=xaaa would
2010 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
2011 same bug. Such groups have to be quantified to be useful, or contained
2012 inside another quantified group. (If there's no repetition, the reference
2013 can never match.) The problem arises because, having left the group and
2014 moved on to the rest of the pattern, a later failure that backtracks into
2015 the group uses the captured value from the final iteration of the group
2016 rather than the correct earlier one. I have fixed this in PCRE by forcing
2017 any group that contains a reference to itself to be an atomic group; that
2018 is, there cannot be any backtracking into it once it has completed. This is
2019 similar to recursive and subroutine calls.
2020
2021
2022 Version 8.00 19-Oct-09
2023 ----------------------
2024
2025 1. The table for translating pcre_compile() error codes into POSIX error codes
2026 was out-of-date, and there was no check on the pcre_compile() error code
2027 being within the table. This could lead to an OK return being given in
2028 error.
2029
2030 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
2031 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
2032 in a Windows environment.
2033
2034 3. The pcregrep --count option prints the count for each file even when it is
2035 zero, as does GNU grep. However, pcregrep was also printing all files when
2036 --files-with-matches was added. Now, when both options are given, it prints
2037 counts only for those files that have at least one match. (GNU grep just
2038 prints the file name in this circumstance, but including the count seems
2039 more useful - otherwise, why use --count?) Also ensured that the
2040 combination -clh just lists non-zero counts, with no names.
2041
2042 4. The long form of the pcregrep -F option was incorrectly implemented as
2043 --fixed_strings instead of --fixed-strings. This is an incompatible change,
2044 but it seems right to fix it, and I didn't think it was worth preserving
2045 the old behaviour.
2046
2047 5. The command line items --regex=pattern and --regexp=pattern were not
2048 recognized by pcregrep, which required --regex pattern or --regexp pattern
2049 (with a space rather than an '='). The man page documented the '=' forms,
2050 which are compatible with GNU grep; these now work.
2051
2052 6. No libpcreposix.pc file was created for pkg-config; there was just
2053 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
2054
2055 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
2056 when UCP support is not needed, by modifying the Python script that
2057 generates it from Unicode data files. This should not matter if the module
2058 is correctly used as a library, but I received one complaint about 50K of
2059 unwanted data. My guess is that the person linked everything into his
2060 program rather than using a library. Anyway, it does no harm.
2061
2062 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
2063 was a minimum greater than 1 for a wide character in a possessive
2064 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
2065 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
2066 character. Chaos in the form of incorrect output or a compiling loop could
2067 result.
2068
2069 9. The restrictions on what a pattern can contain when partial matching is
2070 requested for pcre_exec() have been removed. All patterns can now be
2071 partially matched by this function. In addition, if there are at least two
2072 slots in the offset vector, the offset of the earliest inspected character
2073 for the match and the offset of the end of the subject are set in them when
2074 PCRE_ERROR_PARTIAL is returned.
2075
2076 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
2077 synonymous with PCRE_PARTIAL, for backwards compatibility, and
2078 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
2079 and may be more useful for multi-segment matching.
2080
2081 11. Partial matching with pcre_exec() is now more intuitive. A partial match
2082 used to be given if ever the end of the subject was reached; now it is
2083 given only if matching could not proceed because another character was
2084 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
2085 string "Z", which now yields "no match" instead of "partial match". In the
2086 case of pcre_dfa_exec(), "no match" is given if every matching path for the
2087 final character ended with (*FAIL).
2088
2089 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
2090 if the pattern had a "must contain" character that was already found in the
2091 earlier partial match, unless partial matching was again requested. For
2092 example, with the pattern /dog.(body)?/, the "must contain" character is
2093 "g". If the first part-match was for the string "dog", restarting with
2094 "sbody" failed. This bug has been fixed.
2095
2096 13. The string returned by pcre_dfa_exec() after a partial match has been
2097 changed so that it starts at the first inspected character rather than the
2098 first character of the match. This makes a difference only if the pattern
2099 starts with a lookbehind assertion or \b or \B (\K is not supported by
2100 pcre_dfa_exec()). It's an incompatible change, but it makes the two
2101 matching functions compatible, and I think it's the right thing to do.
2102
2103 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
2104 so that the demonstration program is easily available in environments where
2105 PCRE has not been installed from source.
2106
2107 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
2108 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
2109 library.
2110
2111 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
2112 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
2113 is not the first non-POSIX option to be added. Clearly some people find
2114 these options useful.
2115
2116 17. If a caller to the POSIX matching function regexec() passes a non-zero
2117 value for nmatch with a NULL value for pmatch, the value of
2118 nmatch is forced to zero.
2119
2120 18. RunGrepTest did not have a test for the availability of the -u option of
2121 the diff command, as RunTest does. It now checks in the same way as
2122 RunTest, and also checks for the -b option.
2123
2124 19. If an odd number of negated classes containing just a single character
2125 interposed, within parentheses, between a forward reference to a named
2126 subpattern and the definition of the subpattern, compilation crashed with
2127 an internal error, complaining that it could not find the referenced
2128 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
2129 [The bug was that it was starting one character too far in when skipping
2130 over the character class, thus treating the ] as data rather than
2131 terminating the class. This meant it could skip too much.]
2132
2133 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
2134 /g option in pcretest when the pattern contains \K, which makes it possible
2135 to have an empty string match not at the start, even when the pattern is
2136 anchored. Updated pcretest and pcredemo to use this option.
2137
2138 21. If the maximum number of capturing subpatterns in a recursion was greater
2139 than the maximum at the outer level, the higher number was returned, but
2140 with unset values at the outer level. The correct (outer level) value is
2141 now given.
2142
2143 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
2144 PCRE did not set those parentheses (unlike Perl). I have now found a way to
2145 make it do so. The string so far is captured, making this feature
2146 compatible with Perl.
2147
2148 23. The tests have been re-organized, adding tests 11 and 12, to make it
2149 possible to check the Perl 5.10 features against Perl 5.10.
2150
2151 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
2152 pattern matches a fixed length string. PCRE did not allow this; now it
2153 does. Neither allows recursion.
2154
2155 25. I finally figured out how to implement a request to provide the minimum
2156 length of subject string that was needed in order to match a given pattern.
2157 (It was back references and recursion that I had previously got hung up
2158 on.) This code has now been added to pcre_study(); it finds a lower bound
2159 to the length of subject needed. It is not necessarily the greatest lower
2160 bound, but using it to avoid searching strings that are too short does give
2161 some useful speed-ups. The value is available to calling programs via
2162 pcre_fullinfo().
2163
2164 26. While implementing 25, I discovered to my embarrassment that pcretest had
2165 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
2166 study optimizations had never been tested with that matching function.
2167 Oops. What is worse, even when it was passed study data, there was a bug in
2168 pcre_dfa_exec() that meant it never actually used it. Double oops. There
2169 were also very few tests of studied patterns with pcre_dfa_exec().
2170
2171 27. If (?| is used to create subpatterns with duplicate numbers, they are now
2172 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
2173 on the other side of the coin, they are no longer allowed to have different
2174 names, because these cannot be distinguished in PCRE, and this has caused
2175 confusion. (This is a difference from Perl.)
2176
2177 28. When duplicate subpattern names are present (necessarily with different
2178 numbers, as required by 27 above), and a test is made by name in a
2179 conditional pattern, either for a subpattern having been matched, or for
2180 recursion in such a pattern, all the associated numbered subpatterns are
2181 tested, and the overall condition is true if the condition is true for any
2182 one of them. This is the way Perl works, and is also more like the way
2183 testing by number works.
2184
2185
2186 Version 7.9 11-Apr-09
2187 ---------------------
2188
2189 1. When building with support for bzlib/zlib (pcregrep) and/or readline
2190 (pcretest), all targets were linked against these libraries. This included
2191 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
2192 libraries. This caused unwanted dependencies to be created. This problem
2193 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
2194 pcretest is linked with readline.
2195
2196 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
2197 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
2198 moved outside it again, because FALSE and TRUE are already defined in AIX,
2199 but BOOL is not.
2200
2201 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
2202 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
2203
2204 4. The pcregrep documentation said spaces were inserted as well as colons (or
2205 hyphens) following file names and line numbers when outputting matching
2206 lines. This is not true; no spaces are inserted. I have also clarified the
2207 wording for the --colour (or --color) option.
2208
2209 5. In pcregrep, when --colour was used with -o, the list of matching strings
2210 was not coloured; this is different to GNU grep, so I have changed it to be
2211 the same.
2212
2213 6. When --colo(u)r was used in pcregrep, only the first matching substring in
2214 each matching line was coloured. Now it goes on to look for further matches
2215 of any of the test patterns, which is the same behaviour as GNU grep.
2216
2217 7. A pattern that could match an empty string could cause pcregrep to loop; it
2218 doesn't make sense to accept an empty string match in pcregrep, so I have
2219 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
2220 seems to be how GNU grep behaves. [But see later change 40 for release
2221 8.33.]
2222
2223 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
2224 start or after a newline", because the conditional assertion was not being
2225 correctly handled. The rule now is that both the assertion and what follows
2226 in the first alternative must satisfy the test.
2227
2228 9. If auto-callout was enabled in a pattern with a conditional group whose
2229 condition was an assertion, PCRE could crash during matching, both with
2230 pcre_exec() and pcre_dfa_exec().
2231
2232 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
2233 used for matching.
2234
2235 11. Unicode property support in character classes was not working for
2236 characters (bytes) greater than 127 when not in UTF-8 mode.
2237
2238 12. Added the -M command line option to pcretest.
2239
2240 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
2241
2242 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
2243
2244 16. Added comments and documentation about mis-use of no_arg in the C++
2245 wrapper.
2246
2247 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
2248 from Martin Jerabek that uses macro names for all relevant character and
2249 string constants.
2250
2251 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
2252 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
2253 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
2254 these, but not everybody uses configure.
2255
2256 19. A conditional group that had only one branch was not being correctly
2257 recognized as an item that could match an empty string. This meant that an
2258 enclosing group might also not be so recognized, causing infinite looping
2259 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
2260 with the subject "ab", where knowledge that the repeated group can match
2261 nothing is needed in order to break the loop.
2262
2263 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
2264 exec(), but without supplying a callout function, matching went wrong.
2265
2266 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
2267 leak if the size of the offset vector was greater than 30. When the vector
2268 is smaller, the saved offsets during recursion go onto a local stack
2269 vector, but for larger vectors malloc() is used. It was failing to free
2270 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
2271 error, in fact).
2272
2273 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
2274 heapframe that is used only when UTF-8 support is enabled. This caused no
2275 problem, but was untidy.
2276
2277 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
2278 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
2279 included within another project.
2280
2281 24. Steven Van Ingelgem's patches to add more options to the CMake support,
2282 slightly modified by me:
2283
2284 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
2285 not building pcregrep.
2286
2287 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
2288 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
2289
2290 25. Forward references, both numeric and by name, in patterns that made use of
2291 duplicate group numbers, could behave incorrectly or give incorrect errors,
2292 because when scanning forward to find the reference group, PCRE was not
2293 taking into account the duplicate group numbers. A pattern such as
2294 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
2295
2296 26. Changed a few more instances of "const unsigned char *" to USPTR, making
2297 the feature of a custom pointer more persuasive (as requested by a user).
2298
2299 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
2300 pcretest.c, inside #ifndefs, because it seems they are sometimes already
2301 pre-defined.
2302
2303 28. Added support for (*UTF8) at the start of a pattern.
2304
2305 29. Arrange for flags added by the "release type" setting in CMake to be shown
2306 in the configuration summary.
2307
2308
2309 Version 7.8 05-Sep-08
2310 ---------------------
2311
2312 1. Replaced UCP searching code with optimized version as implemented for Ad
2313 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
2314 stage table and inline lookup instead of a function, giving speed ups of 2
2315 to 5 times on some simple patterns that I tested. Permission was given to
2316 distribute the MultiStage2.py script that generates the tables (it's not in
2317 the tarball, but is in the Subversion repository).
2318
2319 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
2320 scripts.
2321
2322 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
2323 a group with a zero qualifier. The result of the study could be incorrect,
2324 or the function might crash, depending on the pattern.
2325
2326 4. Caseless matching was not working for non-ASCII characters in back
2327 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
2328 It now works when Unicode Property Support is available.
2329
2330 5. In pcretest, an escape such as \x{de} in the data was always generating
2331 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
2332 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
2333 truncation.
2334
2335 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
2336
2337 7. Added two (int) casts to pcregrep when printing the difference of two
2338 pointers, in case they are 64-bit values.
2339
2340 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
2341 test 2 if it fails.
2342
2343 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
2344 and a #define of that name to empty if it is not externally set. This is to
2345 allow users of MSVC to set it if necessary.
2346
2347 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
2348 the convenience functions in the pcre_get.c source file.
2349
2350 11. An option change at the start of a pattern that had top-level alternatives
2351 could cause overwriting and/or a crash. This command provoked a crash in
2352 some environments:
2353
2354 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
2355
2356 This potential security problem was recorded as CVE-2008-2371.
2357
2358 12. For a pattern where the match had to start at the beginning or immediately
2359 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
2360 pcre_dfa_exec() could read past the end of the passed subject if there was
2361 no match. To help with detecting such bugs (e.g. with valgrind), I modified
2362 pcretest so that it places the subject at the end of its malloc-ed buffer.
2363
2364 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
2365 exec() might read past the end of the data buffer in UTF-8 mode.
2366
2367 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
2368 the data contained the byte 0x85 as part of a UTF-8 character within its
2369 first line. This applied both to normal and DFA matching.
2370
2371 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
2372 /^[^d]*?$/8 failed to match "abc".
2373
2374 16. Added a missing copyright notice to pcrecpp_internal.h.
2375
2376 17. Make it more clear in the documentation that values returned from
2377 pcre_exec() in ovector are byte offsets, not character counts.
2378
2379 18. Tidied a few places to stop certain compilers from issuing warnings.
2380
2381 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
2382 supplied by Stefan Weber. I made a further small update for 7.8 because
2383 there is a change of source arrangements: the pcre_searchfuncs.c module is
2384 replaced by pcre_ucd.c.
2385
2386
2387 Version 7.7 07-May-08
2388 ---------------------
2389
2390 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
2391 a string to a long long, pretend we don't even have a long long." This is
2392 done by checking for the strtoq, strtoll, and _strtoi64 functions.
2393
2394 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
2395 pre-7.6 versions, which defined a global no_arg variable instead of putting
2396 it in the RE class. (See also #8 below.)
2397
2398 3. Remove a line of dead code, identified by coverity and reported by Nuno
2399 Lopes.
2400
2401 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
2402
2403 (1) The include/exclude patterns were being applied to the whole pathnames
2404 of files, instead of just to the final components.
2405
2406 (2) If there was more than one level of directory, the subdirectories were
2407 skipped unless they satisfied the include/exclude conditions. This is
2408 inconsistent with GNU grep (and could even be seen as contrary to the
2409 pcregrep specification - which I improved to make it absolutely clear).
2410 The action now is always to scan all levels of directory, and just
2411 apply the include/exclude patterns to regular files.
2412
2413 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
2414 --exclude_dir in the tests to avoid scanning .svn directories.
2415
2416 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
2417 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
2418 doesn't support NULs in patterns.
2419
2420 7. Added some missing "const"s to declarations of static tables in
2421 pcre_compile.c and pcre_dfa_exec.c.
2422
2423 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
2424 caused by fix #2 above. (Subsequently also a second patch to fix the
2425 first patch. And a third patch - this was a messy problem.)
2426
2427 9. Applied Craig's patch to remove the use of push_back().
2428
2429 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
2430 matching function regexec().
2431
2432 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
2433 which, however, unlike Perl's \g{...}, are subroutine calls, not back
2434 references. PCRE supports relative numbers with this syntax (I don't think
2435 Oniguruma does).
2436
2437 12. Previously, a group with a zero repeat such as (...){0} was completely
2438 omitted from the compiled regex. However, this means that if the group
2439 was called as a subroutine from elsewhere in the pattern, things went wrong
2440 (an internal error was given). Such groups are now left in the compiled
2441 pattern, with a new opcode that causes them to be skipped at execution
2442 time.
2443
2444 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
2445 to the way PCRE behaves:
2446
2447 (a) A lone ] character is dis-allowed (Perl treats it as data).
2448
2449 (b) A back reference to an unmatched subpattern matches an empty string
2450 (Perl fails the current match path).
2451
2452 (c) A data ] in a character class must be notated as \] because if the
2453 first data character in a class is ], it defines an empty class. (In
2454 Perl it is not possible to have an empty class.) The empty class []
2455 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
2456 The negative empty class [^] matches any one character, independently
2457 of the DOTALL setting.
2458
2459 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
2460 non-existent subpattern following a character class starting with ']' and
2461 containing () gave an internal compiling error instead of "reference to
2462 non-existent subpattern". Fortunately, when the pattern did exist, the
2463 compiled code was correct. (When scanning forwards to check for the
2464 existence of the subpattern, it was treating the data ']' as terminating
2465 the class, so got the count wrong. When actually compiling, the reference
2466 was subsequently set up correctly.)
2467
2468 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
2469 it was being rejected as not supported by pcre_dfa_exec(), even though
2470 other assertions are supported. I have made pcre_dfa_exec() support
2471 (*FAIL).
2472
2473 16. The implementation of 13c above involved the invention of a new opcode,
2474 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
2475 cannot be changed at match time, I realized I could make a small
2476 improvement to matching performance by compiling OP_ALLANY instead of
2477 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
2478 on the OP_ANY path.
2479
2480 17. Compiling pcretest on Windows with readline support failed without the
2481 following two fixes: (1) Make the unistd.h include conditional on
2482 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
2483
2484 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
2485 ncurses library to be included for pcretest when ReadLine support is
2486 requested, but also to allow for it to be overridden. This patch came from
2487 Daniel Bergström.
2488
2489 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
2490 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
2491 any errors with the current Unicode tables. Thanks to Peter Kankowski for
2492 spotting this.
2493
2494
2495 Version 7.6 28-Jan-08
2496 ---------------------
2497
2498 1. A character class containing a very large number of characters with
2499 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
2500 overflow.
2501
2502 2. Patch to cut out the "long long" test in pcrecpp_unittest when
2503 HAVE_LONG_LONG is not defined.
2504
2505 3. Applied Christian Ehrlicher's patch to update the CMake build files to
2506 bring them up to date and include new features. This patch includes:
2507
2508 - Fixed PH's badly added libz and libbz2 support.
2509 - Fixed a problem with static linking.
2510 - Added pcredemo. [But later removed - see 7 below.]
2511 - Fixed dftables problem and added an option.
2512 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
2513 HAVE_LONG_LONG.
2514 - Added readline support for pcretest.
2515 - Added an listing of the option settings after cmake has run.
2516
2517 4. A user submitted a patch to Makefile that makes it easy to create
2518 "pcre.dll" under mingw when using Configure/Make. I added stuff to
2519 Makefile.am that cause it to include this special target, without
2520 affecting anything else. Note that the same mingw target plus all
2521 the other distribution libraries and programs are now supported
2522 when configuring with CMake (see 6 below) instead of with
2523 Configure/Make.
2524
2525 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
2526 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
2527 exported in the Windows port". It has not yet been confirmed that the patch
2528 solves the problem, but it does no harm.
2529
2530 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
2531 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
2532 with CMake, and also correct the comment about stack recursion.
2533
2534 7. Remove the automatic building of pcredemo from the ./configure system and
2535 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
2536 of a program that users should build themselves after PCRE is installed, so
2537 building it automatically is not really right. What is more, it gave
2538 trouble in some build environments.
2539
2540 8. Further tidies to CMakeLists.txt from Sheri and Christian.
2541
2542
2543 Version 7.5 10-Jan-08
2544 ---------------------
2545
2546 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
2547 values in parens when parsing an RE using the C++ wrapper."
2548
2549 2. Negative specials like \S did not work in character classes in UTF-8 mode.
2550 Characters greater than 255 were excluded from the class instead of being
2551 included.
2552
2553 3. The same bug as (2) above applied to negated POSIX classes such as
2554 [:^space:].
2555
2556 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
2557 defined or documented. It seems to have been a typo for PCRE_STATIC, so
2558 I have changed it.
2559
2560 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
2561 first named subpattern) and a construct such as (?&a) would reference the
2562 first named subpattern whose name started with "a" (in other words, the
2563 length check was missing). Both these problems are fixed. "Subpattern name
2564 expected" is now given for (?&) (a zero-length name), and this patch also
2565 makes it give the same error for \k'' (previously it complained that that
2566 was a reference to a non-existent subpattern).
2567
2568 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
2569 this is right because (?- can be followed by option settings as well as by
2570 digits. I have, however, made the messages clearer.
2571
2572 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
2573 than the number used in the conditional) now cause a compile-time error.
2574 This is actually not compatible with Perl, which accepts such patterns, but
2575 treats the conditional as always being FALSE (as PCRE used to), but it
2576 seems to me that giving a diagnostic is better.
2577
2578 8. Change "alphameric" to the more common word "alphanumeric" in comments
2579 and messages.
2580
2581 9. Fix two occurrences of "backslash" in comments that should have been
2582 "backspace".
2583
2584 10. Remove two redundant lines of code that can never be obeyed (their function
2585 was moved elsewhere).
2586
2587 11. The program that makes PCRE's Unicode character property table had a bug
2588 which caused it to generate incorrect table entries for sequences of
2589 characters that have the same character type, but are in different scripts.
2590 It amalgamated them into a single range, with the script of the first of
2591 them. In other words, some characters were in the wrong script. There were
2592 thirteen such cases, affecting characters in the following ranges:
2593
2594 U+002b0 - U+002c1
2595 U+0060c - U+0060d
2596 U+0061e - U+00612
2597 U+0064b - U+0065e
2598 U+0074d - U+0076d
2599 U+01800 - U+01805
2600 U+01d00 - U+01d77
2601 U+01d9b - U+01dbf
2602 U+0200b - U+0200f
2603 U+030fc - U+030fe
2604 U+03260 - U+0327f
2605 U+0fb46 - U+0fbb1
2606 U+10450 - U+1049d
2607
2608 12. The -o option (show only the matching part of a line) for pcregrep was not
2609 compatible with GNU grep in that, if there was more than one match in a
2610 line, it showed only the first of them. It now behaves in the same way as
2611 GNU grep.
2612
2613 13. If the -o and -v options were combined for pcregrep, it printed a blank
2614 line for every non-matching line. GNU grep prints nothing, and pcregrep now
2615 does the same. The return code can be used to tell if there were any
2616 non-matching lines.
2617
2618 14. Added --file-offsets and --line-offsets to pcregrep.
2619
2620 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
2621 infinitely looping recursion. The bug was that positive lookaheads were not
2622 being skipped when checking for a possible empty match (negative lookaheads
2623 and both kinds of lookbehind were skipped).
2624
2625 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
2626 inclusion of <windows.h> to before rather than after the definition of
2627 INVALID_FILE_ATTRIBUTES (patch from David Byron).
2628
2629 17. Specifying a possessive quantifier with a specific limit for a Unicode
2630 character property caused pcre_compile() to compile bad code, which led at
2631 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
2632 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
2633 caused the error; without that there was no problem.
2634
2635 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
2636
2637 19. Added --enable-pcretest-libreadline.
2638
2639 20. In pcrecpp.cc, the variable 'count' was incremented twice in
2640 RE::GlobalReplace(). As a result, the number of replacements returned was
2641 double what it should be. I removed one of the increments, but Craig sent a
2642 later patch that removed the other one (the right fix) and added unit tests
2643 that check the return values (which was not done before).
2644
2645 21. Several CMake things:
2646
2647 (1) Arranged that, when cmake is used on Unix, the libraries end up with
2648 the names libpcre and libpcreposix, not just pcre and pcreposix.
2649
2650 (2) The above change means that pcretest and pcregrep are now correctly
2651 linked with the newly-built libraries, not previously installed ones.
2652
2653 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
2654
2655 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
2656 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
2657 UTF-8 newline character). The key issue is that the pattern starts .*;
2658 this means that the match must be either at the beginning, or after a
2659 newline. The bug was in the code for advancing after a failed match and
2660 checking that the new position followed a newline. It was not taking
2661 account of UTF-8 characters correctly.
2662
2663 23. PCRE was behaving differently from Perl in the way it recognized POSIX
2664 character classes. PCRE was not treating the sequence [:...:] as a
2665 character class unless the ... were all letters. Perl, however, seems to
2666 allow any characters between [: and :], though of course it rejects as
2667 unknown any "names" that contain non-letters, because all the known class
2668 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
2669 for example, whereas PCRE did not - it did not recognize a POSIX character
2670 class. This seemed a bit dangerous, so the code has been changed to be
2671 closer to Perl. The behaviour is not identical to Perl, because PCRE will
2672 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
2673 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
2674 Perl does, and where it didn't before.
2675
2676 24. Rewrite so as to remove the single use of %n from pcregrep because in some
2677 Windows environments %n is disabled by default.
2678
2679
2680 Version 7.4 21-Sep-07
2681 ---------------------
2682
2683 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
2684 means that a class such as [\s] counted as "explicit reference to CR or
2685 LF". That isn't really right - the whole point of the change was to try to
2686 help when there was an actual mention of one of the two characters. So now
2687 the change happens only if \r or \n (or a literal CR or LF) character is
2688 encountered.
2689
2690 2. The 32-bit options word was also used for 6 internal flags, but the numbers
2691 of both had grown to the point where there were only 3 bits left.
2692 Fortunately, there was spare space in the data structure, and so I have
2693 moved the internal flags into a new 16-bit field to free up more option
2694 bits.
2695
2696 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
2697 but did not set the internal JCHANGED flag - either of these is enough to
2698 control the way the "get" function works - but the PCRE_INFO_JCHANGED
2699 facility is supposed to tell if (?J) was ever used, so now (?J) at the
2700 start sets both bits.
2701
2702 4. Added options (at build time, compile time, exec time) to change \R from
2703 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
2704
2705 5. doc/pcresyntax.html was missing from the distribution.
2706
2707 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
2708 compatibility, even though it is no longer used.
2709
2710 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
2711 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
2712 windows.h file is present (where different names are used). [This was
2713 reversed later after testing - see 16 below.]
2714
2715 8. Changed all #include <config.h> to #include "config.h". There were also
2716 some further <pcre.h> cases that I changed to "pcre.h".
2717
2718 9. When pcregrep was used with the --colour option, it missed the line ending
2719 sequence off the lines that it output.
2720
2721 10. It was pointed out to me that arrays of string pointers cause lots of
2722 relocations when a shared library is dynamically loaded. A technique of
2723 using a single long string with a table of offsets can drastically reduce
2724 these. I have refactored PCRE in four places to do this. The result is
2725 dramatic:
2726
2727 Originally: 290
2728 After changing UCP table: 187
2729 After changing error message table: 43
2730 After changing table of "verbs" 36
2731 After changing table of Posix names 22
2732
2733 Thanks to the folks working on Gregex for glib for this insight.
2734
2735 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
2736 unicode-properties was also set.
2737
2738 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
2739
2740 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
2741 checked only for CRLF.
2742
2743 14. Added casts to pcretest.c to avoid compiler warnings.
2744
2745 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
2746
2747 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
2748 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
2749 entirely. This removes changes made in 7 above.
2750
2751 17. The CMake files have been updated, and there is now more information about
2752 building with CMake in the NON-UNIX-USE document.
2753
2754
2755 Version 7.3 28-Aug-07
2756 ---------------------
2757
2758 1. In the rejigging of the build system that eventually resulted in 7.1, the
2759 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
2760 brackets there is not right, since it causes compilers to look for an
2761 installed pcre.h, not the version that is in the source that is being
2762 compiled (which of course may be different). I have changed it back to:
2763
2764 #include "pcre.h"
2765
2766 I have a vague recollection that the change was concerned with compiling in
2767 different directories, but in the new build system, that is taken care of
2768 by the VPATH setting the Makefile.
2769
2770 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
2771 when the subject happened to end in the byte 0x85 (e.g. if the last
2772 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
2773 characters but of course it shouldn't be taken as a newline when it is part
2774 of another character. The bug was that, for an unlimited repeat of . in
2775 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
2776 characters when looking for a newline.
2777
2778 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
2779
2780 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
2781 in debug output.
2782
2783 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
2784 long printing in the pcrecpp unittest when running under MinGW.
2785
2786 6. ESC_K was left out of the EBCDIC table.
2787
2788 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
2789 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
2790 limit also applies to "virtual nesting" when a pattern is recursive, and in
2791 this case 1000 isn't so big. I have been able to remove this limit at the
2792 expense of backing off one optimization in certain circumstances. Normally,
2793 when pcre_exec() would call its internal match() function recursively and
2794 immediately return the result unconditionally, it uses a "tail recursion"
2795 feature to save stack. However, when a subpattern that can match an empty
2796 string has an unlimited repetition quantifier, it no longer makes this
2797 optimization. That gives it a stack frame in which to save the data for
2798 checking that an empty string has been matched. Previously this was taken
2799 from the 1000-entry workspace that had been reserved. So now there is no
2800 explicit limit, but more stack is used.
2801
2802 8. Applied Daniel's patches to solve problems with the import/export magic
2803 syntax that is required for Windows, and which was going wrong for the
2804 pcreposix and pcrecpp parts of the library. These were overlooked when this
2805 problem was solved for the main library.
2806
2807 9. There were some crude static tests to avoid integer overflow when computing
2808 the size of patterns that contain repeated groups with explicit upper
2809 limits. As the maximum quantifier is 65535, the maximum group length was
2810 set at 30,000 so that the product of these two numbers did not overflow a
2811 32-bit integer. However, it turns out that people want to use groups that
2812 are longer than 30,000 bytes (though not repeat them that many times).
2813 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
2814 made it possible to implement the integer overflow checks in a much more
2815 dynamic way, which I have now done. The artificial limitation on group
2816 length has been removed - we now have only the limit on the total length of
2817 the compiled pattern, which depends on the LINK_SIZE setting.
2818
2819 10. Fixed a bug in the documentation for get/copy named substring when
2820 duplicate names are permitted. If none of the named substrings are set, the
2821 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
2822 empty string.
2823
2824 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
2825 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
2826 because the ] is interpreted as the first data character and the
2827 terminating ] is not found. PCRE has been made compatible with Perl in this
2828 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
2829 cause memory overwriting.
2830
2831 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
2832 string has been matched (to stop an infinite loop). It was not recognizing
2833 a conditional subpattern that could match an empty string if that
2834 subpattern was within another subpattern. For example, it looped when
2835 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
2836 condition was not nested. This bug has been fixed.
2837
2838 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
2839 past the start of the subject in the presence of bytes with the top bit
2840 set, for example "\x8aBCD".
2841
2842 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
2843 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
2844
2845 14. Optimized (?!) to (*FAIL).
2846
2847 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
2848 This restricts code points to be within the range 0 to 0x10FFFF, excluding
2849 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
2850 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
2851 does: it's just the validity check that is more restrictive.
2852
2853 16. Inserted checks for integer overflows during escape sequence (backslash)
2854 processing, and also fixed erroneous offset values for syntax errors during
2855 backslash processing.
2856
2857 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
2858 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
2859
2860 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
2861 caused an overrun.
2862
2863 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
2864 something other than just ASCII characters) inside a group that had an
2865 unlimited repeat caused a loop at compile time (while checking to see
2866 whether the group could match an empty string).
2867
2868 20. Debugging a pattern containing \p or \P could cause a crash. For example,
2869 [\P{Any}] did so. (Error in the code for printing property names.)
2870
2871 21. An orphan \E inside a character class could cause a crash.
2872
2873 22. A repeated capturing bracket such as (A)? could cause a wild memory
2874 reference during compilation.
2875
2876 23. There are several functions in pcre_compile() that scan along a compiled
2877 expression for various reasons (e.g. to see if it's fixed length for look
2878 behind). There were bugs in these functions when a repeated \p or \P was
2879 present in the pattern. These operators have additional parameters compared
2880 with \d, etc, and these were not being taken into account when moving along
2881 the compiled data. Specifically:
2882
2883 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
2884 length.
2885
2886 (b) An item such as \pL+ within a repeated group could cause crashes or
2887 loops.
2888
2889 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
2890 "reference to non-existent subpattern" error.
2891
2892 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
2893
2894 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
2895 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
2896
2897 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
2898
2899 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
2900 character were causing crashes (broken optimization).
2901
2902 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
2903 \p or \P) caused a compile-time loop.
2904
2905 28. More problems have arisen in unanchored patterns when CRLF is a valid line
2906 break. For example, the unstudied pattern [\r\n]A does not match the string
2907 "\r\nA" because change 7.0/46 below moves the current point on by two
2908 characters after failing to match at the start. However, the pattern \nA
2909 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
2910 the same is true. There doesn't seem any very clean way out of this, but
2911 what I have chosen to do makes the common cases work: PCRE now takes note
2912 of whether there can be an explicit match for \r or \n anywhere in the
2913 pattern, and if so, 7.0/46 no longer applies. As part of this change,
2914 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
2915 pattern has explicit CR or LF references.
2916
2917 29. Added (*CR) etc for changing newline setting at start of pattern.
2918
2919
2920 Version 7.2 19-Jun-07
2921 ---------------------
2922
2923 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
2924 which is apparently normally available under Windows.
2925
2926 2. Re-jig the pcregrep tests with different newline settings in an attempt
2927 to make them independent of the local environment's newline setting.
2928
2929 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
2930
2931 4. Some of the "internals" tests were previously cut out when the link size
2932 was not 2, because the output contained actual offsets. The recent new
2933 "Z" feature of pcretest means that these can be cut out, making the tests
2934 usable with all link sizes.
2935
2936 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
2937 stack recursion. This gives a massive performance boost under BSD, but just
2938 a small improvement under Linux. However, it saves one field in the frame
2939 in all cases.
2940
2941 6. Added more features from the forthcoming Perl 5.10:
2942
2943 (a) (?-n) (where n is a string of digits) is a relative subroutine or
2944 recursion call. It refers to the nth most recently opened parentheses.
2945
2946 (b) (?+n) is also a relative subroutine call; it refers to the nth next
2947 to be opened parentheses.
2948
2949 (c) Conditions that refer to capturing parentheses can be specified
2950 relatively, for example, (?(-2)... or (?(+3)...
2951
2952 (d) \K resets the start of the current match so that everything before
2953 is not part of it.
2954
2955 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
2956
2957 (f) \g{name} is another synonym - part of Perl 5.10's unification of
2958 reference syntax.
2959
2960 (g) (?| introduces a group in which the numbering of parentheses in each
2961 alternative starts with the same number.
2962
2963 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
2964
2965 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
2966 PCRE_INFO_JCHANGED.
2967
2968 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
2969 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
2970 for detecting groups that can match an empty string.
2971
2972 9. A pattern with a very large number of alternatives (more than several
2973 hundred) was running out of internal workspace during the pre-compile
2974 phase, where pcre_compile() figures out how much memory will be needed. A
2975 bit of new cunning has reduced the workspace needed for groups with
2976 alternatives. The 1000-alternative test pattern now uses 12 bytes of
2977 workspace instead of running out of the 4096 that are available.
2978
2979 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
2980
2981 11. Applied patch from Google to remove an optimization that didn't quite work.
2982 The report of the bug said:
2983
2984 pcrecpp::RE("a*").FullMatch("aaa") matches, while
2985 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
2986 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
2987
2988 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
2989 it matched the wrong number of bytes.
2990
2991
2992 Version 7.1 24-Apr-07
2993 ---------------------
2994
2995 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
2996 that is more "standard", making use of automake and other Autotools. There
2997 is some re-arrangement of the files and adjustment of comments consequent
2998 on this.
2999
3000 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
3001 for recursive directory scanning broke on some systems because the files
3002 are not scanned in any specific order and on different systems the order
3003 was different. A call to "sort" has been inserted into RunGrepTest for the
3004 approprate test as a short-term fix. In the longer term there may be an
3005 alternative.
3006
3007 3. I had an email from Eric Raymond about problems translating some of PCRE's
3008 man pages to HTML (despite the fact that I distribute HTML pages, some
3009 people do their own conversions for various reasons). The problems
3010 concerned the use of low-level troff macros .br and .in. I have therefore
3011 removed all such uses from the man pages (some were redundant, some could
3012 be replaced by .nf/.fi pairs). The 132html script that I use to generate
3013 HTML has been updated to handle .nf/.fi and to complain if it encounters
3014 .br or .in.
3015
3016 4. Updated comments in configure.ac that get placed in config.h.in and also
3017 arranged for config.h to be included in the distribution, with the name
3018 config.h.generic, for the benefit of those who have to compile without
3019 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
3020
3021 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
3022 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
3023 makevp.bat for the current PCRE, using the additional files
3024 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
3025
3026 6. A Windows user reported a minor discrepancy with test 2, which turned out
3027 to be caused by a trailing space on an input line that had got lost in his
3028 copy. The trailing space was an accident, so I've just removed it.
3029
3030 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
3031 that is needed.
3032
3033 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
3034 as "const" (a) because they are and (b) because it helps the PHP
3035 maintainers who have recently made a script to detect big data structures
3036 in the php code that should be moved to the .rodata section. I remembered
3037 to update Builducptable as well, so it won't revert if ucptable.h is ever
3038 re-created.
3039
3040 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
3041 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
3042 order to be able to cut out the UTF-8 tables in the latter when UTF-8
3043 support is not required. This saves 1.5-2K of code, which is important in
3044 some applications.
3045
3046 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
3047 so as not to refer to the tables, even though these functions will never be
3048 called when UTF-8 support is disabled. Otherwise there are problems with a
3049 shared library.
3050
3051 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
3052
3053 (a) It was defining its arguments as char * instead of void *.
3054
3055 (b) It was assuming that all moves were upwards in memory; this was true
3056 a long time ago when I wrote it, but is no longer the case.
3057
3058 The emulated memove() is provided for those environments that have neither
3059 memmove() nor bcopy(). I didn't think anyone used it these days, but that
3060 is clearly not the case, as these two bugs were recently reported.
3061
3062 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
3063 and Detrail to create the HTML documentation, the .txt form of the man
3064 pages, and it removes trailing spaces from listed files. It also creates
3065 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
3066 case, it wraps all the #defines with #ifndefs. This script should be run
3067 before "make dist".
3068
3069 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
3070 with Unicode property support.
3071
3072 (a) For a maximizing quantifier, if the two different cases of the
3073 character were of different lengths in their UTF-8 codings (there are
3074 some cases like this - I found 11), and the matching function had to
3075 back up over a mixture of the two cases, it incorrectly assumed they
3076 were both the same length.
3077
3078 (b) When PCRE was configured to use the heap rather than the stack for
3079 recursion during matching, it was not correctly preserving the data for
3080 the other case of a UTF-8 character when checking ahead for a match
3081 while processing a minimizing repeat. If the check also involved
3082 matching a wide character, but failed, corruption could cause an
3083 erroneous result when trying to check for a repeat of the original
3084 character.
3085
3086 13. Some tidying changes to the testing mechanism:
3087
3088 (a) The RunTest script now detects the internal link size and whether there
3089 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
3090 values substituted by "configure". (The RunGrepTest script already did
3091 this for UTF-8.) The configure.ac script no longer substitutes the
3092 relevant variables.
3093
3094 (b) The debugging options /B and /D in pcretest show the compiled bytecode
3095 with length and offset values. This means that the output is different
3096 for different internal link sizes. Test 2 is skipped for link sizes
3097 other than 2 because of this, bypassing the problem. Unfortunately,
3098 there was also a test in test 3 (the locale tests) that used /B and
3099 failed for link sizes other than 2. Rather than cut the whole test out,
3100 I have added a new /Z option to pcretest that replaces the length and
3101 offset values with spaces. This is now used to make test 3 independent
3102 of link size. (Test 2 will be tidied up later.)
3103
3104 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
3105 segmentation fault instead of returning the appropriate error message.
3106
3107 15. In multiline mode when the newline sequence was set to "any", the pattern
3108 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
3109 This doesn't seem right; it now treats the CRLF combination as the line
3110 ending, and so does not match in that case. It's only a pattern such as ^$
3111 that would hit this one: something like ^ABC$ would have failed after \r
3112 and then tried again after \r\n.
3113
3114 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
3115 in an attempt to make files that differ only in their line terminators
3116 compare equal. This works on Linux.
3117
3118 17. Under certain error circumstances pcregrep might try to free random memory
3119 as it exited. This is now fixed, thanks to valgrind.
3120
3121 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
3122 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
3123 was because its rules for how to advance for /g after matching an empty
3124 string at the end of a line did not allow for this case. They now check for
3125 it specially.
3126
3127 20. pcretest is supposed to handle patterns and data of any length, by
3128 extending its buffers when necessary. It was getting this wrong when the
3129 buffer for a data line had to be extended.
3130
3131 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
3132 CRLF as a newline sequence.
3133
3134 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
3135 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
3136 I have nevertheless tidied it up.
3137
3138 23. Added some casts to kill warnings from HP-UX ia64 compiler.
3139
3140 24. Added a man page for pcre-config.
3141
3142
3143 Version 7.0 19-Dec-06
3144 ---------------------
3145
3146 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
3147 moving to gcc 4.1.1.
3148
3149 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
3150 sys/time.h, which is documented as needed for this function. It doesn't
3151 seem to matter on Linux, but it showed up on some releases of OS X.
3152
3153 3. It seems that there are systems where bytes whose values are greater than
3154 127 match isprint() in the "C" locale. The "C" locale should be the
3155 default when a C program starts up. In most systems, only ASCII printing
3156 characters match isprint(). This difference caused the output from pcretest
3157 to vary, making some of the tests fail. I have changed pcretest so that:
3158
3159 (a) When it is outputting text in the compiled version of a pattern, bytes
3160 other than 32-126 are always shown as hex escapes.
3161
3162 (b) When it is outputting text that is a matched part of a subject string,
3163 it does the same, unless a different locale has been set for the match
3164 (using the /L modifier). In this case, it uses isprint() to decide.
3165
3166 4. Fixed a major bug that caused incorrect computation of the amount of memory
3167 required for a compiled pattern when options that changed within the
3168 pattern affected the logic of the preliminary scan that determines the
3169 length. The relevant options are -x, and -i in UTF-8 mode. The result was
3170 that the computed length was too small. The symptoms of this bug were
3171 either the PCRE error "internal error: code overflow" from pcre_compile(),
3172 or a glibc crash with a message such as "pcretest: free(): invalid next
3173 size (fast)". Examples of patterns that provoked this bug (shown in
3174 pcretest format) are:
3175
3176 /(?-x: )/x
3177 /(?x)(?-x: \s*#\s*)/
3178 /((?i)[\x{c0}])/8
3179 /(?i:[\x{c0}])/8
3180
3181 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
3182 is now done differently.
3183
3184 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
3185 wrapper classes; (b) implement a new function in the C++ scanner that is
3186 more efficient than the old way of doing things because it avoids levels of
3187 recursion in the regex matching; (c) add a paragraph to the documentation
3188 for the FullMatch() function.
3189
3190 6. The escape sequence \n was being treated as whatever was defined as
3191 "newline". Not only was this contrary to the documentation, which states
3192 that \n is character 10 (hex 0A), but it also went horribly wrong when
3193 "newline" was defined as CRLF. This has been fixed.
3194
3195 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
3196 was being set to -1 for the "end of line" case (supposedly a value that no
3197 character can have). Though this value is never used (the check for end of
3198 line is "zero bytes in current character"), it caused compiler complaints.
3199 I've changed it to 0xffffffff.
3200
3201 8. In pcre_version.c, the version string was being built by a sequence of
3202 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
3203 string (as it is for production releases) called a macro with an empty
3204 argument. The C standard says the result of this is undefined. The gcc
3205 compiler treats it as an empty string (which was what was wanted) but it is
3206 reported that Visual C gives an error. The source has been hacked around to
3207 avoid this problem.
3208
3209 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
3210 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
3211 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
3212 of them did).
3213
3214 10. Originally, pcretest opened its input and output without "b"; then I was
3215 told that "b" was needed in some environments, so it was added for release
3216 5.0 to both the input and output. (It makes no difference on Unix-like
3217 systems.) Later I was told that it is wrong for the input on Windows. I've
3218 now abstracted the modes into two macros, to make it easier to fiddle with
3219 them, and removed "b" from the input mode under Windows.
3220
3221 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
3222
3223 12. Added -help and --help to pcretest as an official way of being reminded
3224 of the options.
3225
3226 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
3227 and pcrecpp.cc because they annoy compilers at high warning levels.
3228
3229 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
3230
3231 15. Fixed an occurrence of == in configure.ac that should have been = (shell
3232 scripts are not C programs :-) and which was not noticed because it works
3233 on Linux.
3234
3235 16. pcretest is supposed to handle any length of pattern and data line (as one
3236 line or as a continued sequence of lines) by extending its input buffer if
3237 necessary. This feature was broken for very long pattern lines, leading to
3238 a string of junk being passed to pcre_compile() if the pattern was longer
3239 than about 50K.
3240
3241 17. I have done a major re-factoring of the way pcre_compile() computes the
3242 amount of memory needed for a compiled pattern. Previously, there was code
3243 that made a preliminary scan of the pattern in order to do this. That was
3244 OK when PCRE was new, but as the facilities have expanded, it has become
3245 harder and harder to keep it in step with the real compile phase, and there
3246 have been a number of bugs (see for example, 4 above). I have now found a
3247 cunning way of running the real compile function in a "fake" mode that
3248 enables it to compute how much memory it would need, while actually only
3249 ever using a few hundred bytes of working memory and without too many
3250 tests of the mode. This should make future maintenance and development
3251 easier. A side effect of this work is that the limit of 200 on the nesting
3252 depth of parentheses has been removed (though this was never a serious
3253 limitation, I suspect). However, there is a downside: pcre_compile() now
3254 runs more slowly than before (30% or more, depending on the pattern). I
3255 hope this isn't a big issue. There is no effect on runtime performance.
3256
3257 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
3258 newline (only possible for the last line of a file) and it was a
3259 pattern that set a locale (followed by /Lsomething), pcretest crashed.
3260
3261 19. Added additional timing features to pcretest. (1) The -tm option now times
3262 matching only, not compiling. (2) Both -t and -tm can be followed, as a
3263 separate command line item, by a number that specifies the number of
3264 repeats to use when timing. The default is 50000; this gives better
3265 precision, but takes uncomfortably long for very large patterns.
3266
3267 20. Extended pcre_study() to be more clever in cases where a branch of a
3268 subpattern has no definite first character. For example, (a*|b*)[cd] would
3269 previously give no result from pcre_study(). Now it recognizes that the
3270 first character must be a, b, c, or d.
3271
3272 21. There was an incorrect error "recursive call could loop indefinitely" if
3273 a subpattern (or the entire pattern) that was being tested for matching an
3274 empty string contained only one non-empty item after a nested subpattern.
3275 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
3276 incorrectly, because the \d was being skipped in the check.
3277
3278 22. The pcretest program now has a new pattern option /B and a command line
3279 option -b, which is equivalent to adding /B to every pattern. This causes
3280 it to show the compiled bytecode, without the additional information that
3281 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
3282 is the same as /B/I).
3283
3284 23. A new optimization is now able automatically to treat some sequences such
3285 as a*b as a*+b. More specifically, if something simple (such as a character
3286 or a simple class like \d) has an unlimited quantifier, and is followed by
3287 something that cannot possibly match the quantified thing, the quantifier
3288 is automatically "possessified".
3289
3290 24. A recursive reference to a subpattern whose number was greater than 39
3291 went wrong under certain circumstances in UTF-8 mode. This bug could also
3292 have affected the operation of pcre_study().
3293
3294 25. Realized that a little bit of performance could be had by replacing
3295 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
3296
3297 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
3298
3299 27. Possessive quantifiers such as a++ were previously implemented by turning
3300 them into atomic groups such as ($>a+). Now they have their own opcodes,
3301 which improves performance. This includes the automatically created ones
3302 from 23 above.
3303
3304 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
3305 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
3306 the first matched character to be a colon. This applied both to named and
3307 numbered groups.
3308
3309 29. The ucpinternal.h header file was missing its idempotency #ifdef.
3310
3311 30. I was sent a "project" file called libpcre.a.dev which I understand makes
3312 building PCRE on Windows easier, so I have included it in the distribution.
3313
3314 31. There is now a check in pcretest against a ridiculously large number being
3315 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
3316 loop, the loop is abandoned.
3317
3318 32. Forward references to subpatterns in conditions such as (?(2)...) where
3319 subpattern 2 is defined later cause pcre_compile() to search forwards in
3320 the pattern for the relevant set of parentheses. This search went wrong
3321 when there were unescaped parentheses in a character class, parentheses
3322 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
3323
3324 33. "Subroutine" calls and backreferences were previously restricted to
3325 referencing subpatterns earlier in the regex. This restriction has now
3326 been removed.
3327
3328 34. Added a number of extra features that are going to be in Perl 5.10. On the
3329 whole, these are just syntactic alternatives for features that PCRE had
3330 previously implemented using the Python syntax or my own invention. The
3331 other formats are all retained for compatibility.
3332
3333 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
3334 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
3335 also .NET compatible.
3336
3337 (b) A recursion or subroutine call to a named group can now be defined as
3338 (?&name) as well as (?P>name).
3339
3340 (c) A backreference to a named group can now be defined as \k<name> or
3341 \k'name' as well as (?P=name). The new forms, as well as being in Perl
3342 5.10, are also .NET compatible.
3343
3344 (d) A conditional reference to a named group can now use the syntax
3345 (?(<name>) or (?('name') as well as (?(name).
3346
3347 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
3348 groups (named and numbered) that are never evaluated inline, but can be
3349 called as "subroutines" from elsewhere. In effect, the DEFINE condition
3350 is always false. There may be only one alternative in such a group.
3351
3352 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
3353 as the simple (?(R). The condition is true only if the most recent
3354 recursion is that of the given number or name. It does not search out
3355 through the entire recursion stack.
3356
3357 (g) The escape \gN or \g{N} has been added, where N is a positive or
3358 negative number, specifying an absolute or relative reference.
3359
3360 35. Tidied to get rid of some further signed/unsigned compiler warnings and
3361 some "unreachable code" warnings.
3362
3363 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
3364 things, this adds five new scripts.
3365
3366 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
3367 There were also incompatibilities regarding the handling of \Q..\E inside
3368 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
3369 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
3370
3371 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
3372 matches an empty string, and forcibly breaks the loop. There were bugs in
3373 this code in non-simple cases. For a pattern such as ^(a()*)* matched
3374 against aaaa the result was just "a" rather than "aaaa", for example. Two
3375 separate and independent bugs (that affected different cases) have been
3376 fixed.
3377
3378 39. Refactored the code to abolish the use of different opcodes for small
3379 capturing bracket numbers. This is a tidy that I avoided doing when I
3380 removed the limit on the number of capturing brackets for 3.5 back in 2001.
3381 The new approach is not only tidier, it makes it possible to reduce the
3382 memory needed to fix the previous bug (38).
3383
3384 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
3385 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
3386 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
3387 mode.
3388
3389 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
3390 report.
3391
3392 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
3393 copy construction and assignment in the C++ wrapper.
3394
3395 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
3396 couple of bugs that could have given wrong results in the "--newline=crlf"
3397 case.
3398
3399 44. Added a number of casts and did some reorganization of signed/unsigned int
3400 variables following suggestions from Dair Grant. Also renamed the variable
3401 "this" as "item" because it is a C++ keyword.
3402
3403 45. Arranged for dftables to add
3404
3405 #include "pcre_internal.h"
3406
3407 to pcre_chartables.c because without it, gcc 4.x may remove the array
3408 definition from the final binary if PCRE is built into a static library and
3409 dead code stripping is activated.
3410
3411 46. For an unanchored pattern, if a match attempt fails at the start of a
3412 newline sequence, and the newline setting is CRLF or ANY, and the next two
3413 characters are CRLF, advance by two characters instead of one.
3414
3415
3416 Version 6.7 04-Jul-06
3417 ---------------------
3418
3419 1. In order to handle tests when input lines are enormously long, pcretest has
3420 been re-factored so that it automatically extends its buffers when
3421 necessary. The code is crude, but this _is_ just a test program. The
3422 default size has been increased from 32K to 50K.
3423
3424 2. The code in pcre_study() was using the value of the re argument before
3425 testing it for NULL. (Of course, in any sensible call of the function, it
3426 won't be NULL.)
3427
3428 3. The memmove() emulation function in pcre_internal.h, which is used on
3429 systems that lack both memmove() and bcopy() - that is, hardly ever -
3430 was missing a "static" storage class specifier.
3431
3432 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
3433 containing an extended class (one that cannot be represented by a bitmap
3434 because it contains high-valued characters or Unicode property items, e.g.
3435 [\pZ]). Almost always one would set UTF-8 mode when processing such a
3436 pattern, but PCRE should not loop if you do not (it no longer does).
3437 [Detail: two cases were found: (a) a repeated subpattern containing an
3438 extended class; (b) a recursive reference to a subpattern that followed a
3439 previous extended class. It wasn't skipping over the extended class
3440 correctly when UTF-8 mode was not set.]
3441
3442 5. A negated single-character class was not being recognized as fixed-length
3443 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
3444 compile error "lookbehind assertion is not fixed length".
3445
3446 6. The RunPerlTest auxiliary script was showing an unexpected difference
3447 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
3448 write a Perl script that can interpret lines of an input file either as
3449 byte characters or as UTF-8, which is what "perltest" was being required to
3450 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
3451 can't do is switch easily at run time between having the "use utf8;" pragma
3452 or not. In the end, I fudged it by using the RunPerlTest script to insert
3453 "use utf8;" explicitly for the UTF-8 tests.
3454
3455 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
3456 the end of the subject string, contrary to the documentation and to what
3457 Perl does. This was true of both matching functions. Now it matches only at
3458 the start of the subject and immediately after *internal* newlines.
3459
3460 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
3461 a pointer to an int instead of a pointer to an unsigned long int. This
3462 caused problems on 64-bit systems.
3463
3464 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
3465 instance of the 'standard' template library not being so standard".
3466
3467 10. There was no check on the number of named subpatterns nor the maximum
3468 length of a subpattern name. The product of these values is used to compute
3469 the size of the memory block for a compiled pattern. By supplying a very
3470 long subpattern name and a large number of named subpatterns, the size
3471 computation could be caused to overflow. This is now prevented by limiting
3472 the length of names to 32 characters, and the number of named subpatterns
3473 to 10,000.
3474
3475 11. Subpatterns that are repeated with specific counts have to be replicated in
3476 the compiled pattern. The size of memory for this was computed from the
3477 length of the subpattern and the repeat count. The latter is limited to
3478 65535, but there was no limit on the former, meaning that integer overflow
3479 could in principle occur. The compiled length of a repeated subpattern is
3480 now limited to 30,000 bytes in order to prevent this.
3481
3482 12. Added the optional facility to have named substrings with the same name.
3483
3484 13. Added the ability to use a named substring as a condition, using the
3485 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
3486 are numbers (not recommended). Forward references are permitted.
3487
3488 14. Added forward references in named backreferences (if you see what I mean).
3489
3490 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
3491 pattern could run off the end of the subject. For example, the pattern
3492 "(?s)(.{1,5})"8 did this with the subject "ab".
3493
3494 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
3495 PCRE_CASELESS was set when matching characters that were quantified with ?
3496 or *.
3497
3498 17. A character class other than a single negated character that had a minimum
3499 but no maximum quantifier - for example [ab]{6,} - was not handled
3500 correctly by pce_dfa_exec(). It would match only one character.
3501
3502 18. A valid (though odd) pattern that looked like a POSIX character
3503 class but used an invalid character after [ (for example [[,abc,]]) caused
3504 pcre_compile() to give the error "Failed: internal error: code overflow" or
3505 in some cases to crash with a glibc free() error. This could even happen if
3506 the pattern terminated after [[ but there just happened to be a sequence of
3507 letters, a binary zero, and a closing ] in the memory that followed.
3508
3509 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
3510 over the years. Originally (before any Unicode support), just the bottom 8
3511 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
3512 output from "man perlunicode" includes this:
3513
3514 The regular expression compiler produces polymorphic opcodes. That
3515 is, the pattern adapts to the data and automatically switches to
3516 the Unicode character scheme when presented with Unicode data--or
3517 instead uses a traditional byte scheme when presented with byte
3518 data.
3519
3520 Sadly, a wide octal escape does not cause a switch, and in a string with
3521 no other multibyte characters, these octal escapes are treated as before.
3522 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
3523 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
3524 Unicode string.
3525
3526 I have not perpetrated such confusion in PCRE. Up till now, it took just
3527 the bottom 8 bits, as in old Perl. I have now made octal escapes with
3528 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
3529 translate to the appropriate multibyte character.
3530
3531 29. Applied some refactoring to reduce the number of warnings from Microsoft
3532 and Borland compilers. This has included removing the fudge introduced
3533 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
3534 a warning about an unused variable.
3535
3536 21. PCRE has not included VT (character 0x0b) in the set of whitespace
3537 characters since release 4.0, because Perl (from release 5.004) does not.
3538 [Or at least, is documented not to: some releases seem to be in conflict
3539 with the documentation.] However, when a pattern was studied with
3540 pcre_study() and all its branches started with \s, PCRE still included VT
3541 as a possible starting character. Of course, this did no harm; it just
3542 caused an unnecessary match attempt.
3543
3544 22. Removed a now-redundant internal flag bit that recorded the fact that case
3545 dependency changed within the pattern. This was once needed for "required
3546 byte" processing, but is no longer used. This recovers a now-scarce options
3547 bit. Also moved the least significant internal flag bit to the most-
3548 significant bit of the word, which was not previously used (hangover from
3549 the days when it was an int rather than a uint) to free up another bit for
3550 the future.
3551
3552 23. Added support for CRLF line endings as well as CR and LF. As well as the
3553 default being selectable at build time, it can now be changed at runtime
3554 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
3555 specify that it is scanning data with non-default line endings.
3556
3557 24. Changed the definition of CXXLINK to make it agree with the definition of
3558 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
3559
3560 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
3561 recursions. This makes a big different to stack usage for some patterns.
3562
3563 26. If a subpattern containing a named recursion or subroutine reference such
3564 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
3565 the space required for the compiled pattern went wrong and gave too small a
3566 value. Depending on the environment, this could lead to "Failed: internal
3567 error: code overflow at offset 49" or "glibc detected double free or
3568 corruption" errors.
3569
3570 27. Applied patches from Google (a) to support the new newline modes and (b) to
3571 advance over multibyte UTF-8 characters in GlobalReplace.
3572
3573 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
3574 difference for some implementation of PCRE in some Windows version.
3575
3576 29. Added some extra testing facilities to pcretest:
3577
3578 \q<number> in a data line sets the "match limit" value
3579 \Q<number> in a data line sets the "match recursion limt" value
3580 -S <number> sets the stack size, where <number> is in megabytes
3581
3582 The -S option isn't available for Windows.
3583
3584
3585 Version 6.6 06-Feb-06
3586 ---------------------
3587
3588 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
3589 in pcreposix.h. I have copied the definition from pcre.h.
3590
3591 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
3592 because pcre.h is no longer a built file.
3593
3594 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
3595 not normally included in the compiled code.
3596
3597
3598 Version 6.5 01-Feb-06
3599 ---------------------
3600
3601 1. When using the partial match feature with pcre_dfa_exec(), it was not
3602 anchoring the second and subsequent partial matches at the new starting
3603 point. This could lead to incorrect results. For example, with the pattern
3604 /1234/, partially matching against "123" and then "a4" gave a match.
3605
3606 2. Changes to pcregrep:
3607
3608 (a) All non-match returns from pcre_exec() were being treated as failures
3609 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
3610 error message is output. Some extra information is given for the
3611 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
3612 probably the only errors that are likely to be caused by users (by
3613 specifying a regex that has nested indefinite repeats, for instance).
3614 If there are more than 20 of these errors, pcregrep is abandoned.
3615
3616 (b) A binary zero was treated as data while matching, but terminated the
3617 output line if it was written out. This has been fixed: binary zeroes
3618 are now no different to any other data bytes.
3619
3620 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
3621 used to set a locale for matching. The --locale=xxxx long option has
3622 been added (no short equivalent) to specify a locale explicitly on the
3623 pcregrep command, overriding the environment variables.
3624
3625 (d) When -B was used with -n, some line numbers in the output were one less
3626 than they should have been.
3627
3628 (e) Added the -o (--only-matching) option.
3629
3630 (f) If -A or -C was used with -c (count only), some lines of context were
3631 accidentally printed for the final match.
3632
3633 (g) Added the -H (--with-filename) option.
3634
3635 (h) The combination of options -rh failed to suppress file names for files
3636 that were found from directory arguments.
3637
3638 (i) Added the -D (--devices) and -d (--directories) options.
3639
3640 (j) Added the -F (--fixed-strings) option.
3641
3642 (k) Allow "-" to be used as a file name for -f as well as for a data file.
3643
3644 (l) Added the --colo(u)r option.
3645
3646 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
3647 is not present by default.
3648
3649 3. A nasty bug was discovered in the handling of recursive patterns, that is,
3650 items such as (?R) or (?1), when the recursion could match a number of
3651 alternatives. If it matched one of the alternatives, but subsequently,
3652 outside the recursion, there was a failure, the code tried to back up into
3653 the recursion. However, because of the way PCRE is implemented, this is not
3654 possible, and the result was an incorrect result from the match.
3655
3656 In order to prevent this happening, the specification of recursion has
3657 been changed so that all such subpatterns are automatically treated as
3658 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
3659
3660 4. I had overlooked the fact that, in some locales, there are characters for
3661 which isalpha() is true but neither isupper() nor islower() are true. In
3662 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
3663 and ordfeminine) are like this. This affected the treatment of \w and \W
3664 when they appeared in character classes, but not when they appeared outside
3665 a character class. The bit map for "word" characters is now created
3666 separately from the results of isalnum() instead of just taking it from the
3667 upper, lower, and digit maps. (Plus the underscore character, of course.)
3668
3669 5. The above bug also affected the handling of POSIX character classes such as
3670 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
3671 permanent tables. Instead, the bit maps for such a class were previously
3672 created as the appropriate unions of the upper, lower, and digit bitmaps.
3673 Now they are created by subtraction from the [[:word:]] class, which has
3674 its own bitmap.
3675
3676 6. The [[:blank:]] character class matches horizontal, but not vertical space.
3677 It is created by subtracting the vertical space characters (\x09, \x0a,
3678 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
3679 subtraction was done in the overall bitmap for a character class, meaning
3680 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
3681 be recognized. This bug has been fixed.
3682
3683 7. Patches from the folks at Google:
3684
3685 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
3686 real life, but is still worth protecting against".
3687
3688 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
3689 regular expressions".
3690
3691 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
3692 have it.
3693
3694 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
3695 "configure" and the latter not, in order to fix a problem somebody had
3696 with compiling the Arg class on HP-UX.
3697
3698 (e) Improve the error-handling of the C++ wrapper a little bit.
3699
3700 (f) New tests for checking recursion limiting.
3701
3702 8. The pcre_memmove() function, which is used only if the environment does not
3703 have a standard memmove() function (and is therefore rarely compiled),
3704 contained two bugs: (a) use of int instead of size_t, and (b) it was not
3705 returning a result (though PCRE never actually uses the result).
3706
3707 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
3708 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
3709 returned instead of calling malloc() with an overflowing number that would
3710 most likely cause subsequent chaos.
3711
3712 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
3713
3714 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
3715 with this option is matched, the nmatch and pmatch options of regexec() are
3716 ignored.
3717
3718 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
3719 provided in case anyone wants to the the POSIX interface with UTF-8
3720 strings.
3721
3722 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
3723 C++ linking (needed for some HP-UX environments).
3724
3725 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
3726 (unused parameter) and in the pcre_printint() function (omitted "default"
3727 switch label when the default is to do nothing).
3728
3729 15. Added some code to make it possible, when PCRE is compiled as a C++
3730 library, to replace subject pointers for pcre_exec() with a smart pointer
3731 class, thus making it possible to process discontinuous strings.
3732
3733 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
3734 much the same function. They were added by different people who were trying
3735 to make PCRE easy to compile on non-Unix systems. It has been suggested
3736 that PCRE_EXPORT be abolished now that there is more automatic apparatus
3737 for compiling on Windows systems. I have therefore replaced it with
3738 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
3739 defaults to "extern" for C or "extern C" for C++, which works fine on
3740 Unix-like systems. It is now possible to override the value of PCRE_DATA_
3741 SCOPE with something explicit in config.h. In addition:
3742
3743 (a) pcreposix.h still had just "extern" instead of either of these macros;
3744 I have replaced it with PCRE_DATA_SCOPE.
3745
3746 (b) Functions such as _pcre_xclass(), which are internal to the library,
3747 but external in the C sense, all had PCRE_EXPORT in their definitions.
3748 This is apparently wrong for the Windows case, so I have removed it.
3749 (It makes no difference on Unix-like systems.)
3750
3751 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
3752 of recursive calls to match(). This is different to MATCH_LIMIT because
3753 that limits the total number of calls to match(), not all of which increase
3754 the depth of recursion. Limiting the recursion depth limits the amount of
3755 stack (or heap if NO_RECURSE is set) that is used. The default can be set
3756 when PCRE is compiled, and changed at run time. A patch from Google adds
3757 this functionality to the C++ interface.
3758
3759 18. Changes to the handling of Unicode character properties:
3760
3761 (a) Updated the table to Unicode 4.1.0.
3762
3763 (b) Recognize characters that are not in the table as "Cn" (undefined).
3764
3765 (c) I revised the way the table is implemented to a much improved format
3766 which includes recognition of ranges. It now supports the ranges that
3767 are defined in UnicodeData.txt, and it also amalgamates other
3768 characters into ranges. This has reduced the number of entries in the
3769 table from around 16,000 to around 3,000, thus reducing its size
3770 considerably. I realized I did not need to use a tree structure after
3771 all - a binary chop search is just as efficient. Having reduced the
3772 number of entries, I extended their size from 6 bytes to 8 bytes to
3773 allow for more data.
3774
3775 (d) Added support for Unicode script names via properties such as \p{Han}.
3776
3777 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
3778 matching that character.
3779
3780 20. When matching a repeated Unicode property with a minimum greater than zero,
3781 (for example \pL{2,}), PCRE could look past the end of the subject if it
3782 reached it while seeking the minimum number of characters. This could
3783 happen only if some of the characters were more than one byte long, because
3784 there is a check for at least the minimum number of bytes.
3785
3786 21. Refactored the implementation of \p and \P so as to be more general, to
3787 allow for more different types of property in future. This has changed the
3788 compiled form incompatibly. Anybody with saved compiled patterns that use
3789 \p or \P will have to recompile them.
3790
3791 22. Added "Any" and "L&" to the supported property types.
3792
3793 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
3794 but give a compile time error if the value is greater than 0xff.
3795
3796 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
3797 accidentally not being installed or uninstalled.
3798
3799 25. The pcre.h file was built from pcre.h.in, but the only changes that were
3800 made were to insert the current release number. This seemed silly, because
3801 it made things harder for people building PCRE on systems that don't run
3802 "configure". I have turned pcre.h into a distributed file, no longer built
3803 by "configure", with the version identification directly included. There is
3804 no longer a pcre.h.in file.
3805
3806 However, this change necessitated a change to the pcre-config script as
3807 well. It is built from pcre-config.in, and one of the substitutions was the
3808 release number. I have updated configure.ac so that ./configure now finds
3809 the release number by grepping pcre.h.
3810
3811 26. Added the ability to run the tests under valgrind.
3812
3813
3814 Version 6.4 05-Sep-05
3815 ---------------------
3816
3817 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
3818 "--" to be printed when multiple files were scanned, even when none of the
3819 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
3820 consider it to be a bug, and have restored the previous behaviour.
3821
3822 2. A couple of code tidies to get rid of compiler warnings.
3823
3824 3. The pcretest program used to cheat by referring to symbols in the library
3825 whose names begin with _pcre_. These are internal symbols that are not
3826 really supposed to be visible externally, and in some environments it is
3827 possible to suppress them. The cheating is now confined to including
3828 certain files from the library's source, which is a bit cleaner.
3829
3830 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
3831 file's purpose clearer.
3832
3833 5. Reorganized pcre_ucp_findchar().
3834
3835
3836 Version 6.3 15-Aug-05
3837 ---------------------
3838
3839 1. The file libpcre.pc.in did not have general read permission in the tarball.
3840
3841 2. There were some problems when building without C++ support:
3842
3843 (a) If C++ support was not built, "make install" and "make test" still
3844 tried to test it.
3845
3846 (b) There were problems when the value of CXX was explicitly set. Some
3847 changes have been made to try to fix these, and ...
3848
3849 (c) --disable-cpp can now be used to explicitly disable C++ support.
3850
3851 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
3852 backslash in a target when C++ was disabled. This confuses some
3853 versions of "make", apparently. Using an intermediate variable solves
3854 this. (Same for CPP_LOBJ.)
3855
3856 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
3857 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
3858 necessary on certain architectures.
3859
3860 4. Added a setting of -export-symbols-regex to the link command to remove
3861 those symbols that are exported in the C sense, but actually are local
3862 within the library, and not documented. Their names all begin with
3863 "_pcre_". This is not a perfect job, because (a) we have to except some
3864 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
3865 available (and never for static libraries). I have made a note to try to
3866 find a way round (a) in the future.
3867
3868
3869 Version 6.2 01-Aug-05
3870 ---------------------
3871
3872 1. There was no test for integer overflow of quantifier values. A construction
3873 such as {1111111111111111} would give undefined results. What is worse, if
3874 a minimum quantifier for a parenthesized subpattern overflowed and became
3875 negative, the calculation of the memory size went wrong. This could have
3876 led to memory overwriting.
3877
3878 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
3879
3880 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
3881 operating environments where this matters.
3882
3883 4. Applied Giuseppe Maxia's patch to add additional features for controlling
3884 PCRE options from within the C++ wrapper.
3885
3886 5. Named capturing subpatterns were not being correctly counted when a pattern
3887 was compiled. This caused two problems: (a) If there were more than 100
3888 such subpatterns, the calculation of the memory needed for the whole
3889 compiled pattern went wrong, leading to an overflow error. (b) Numerical
3890 back references of the form \12, where the number was greater than 9, were
3891 not recognized as back references, even though there were sufficient
3892 previous subpatterns.
3893
3894 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
3895 versions of gcc, e.g. 2.95.4.
3896
3897
3898 Version 6.1 21-Jun-05
3899 ---------------------
3900
3901 1. There was one reference to the variable "posix" in pcretest.c that was not
3902 surrounded by "#if !defined NOPOSIX".
3903
3904 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
3905 the cross-check on the old pcre_info() function, for the benefit of the
3906 cut-down version of PCRE that is currently imported into Exim.
3907
3908 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
3909 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
3910 patterns that start (?i) but I don't think that matters. The use of (?i) is
3911 just an example; this all applies to the other options as well.
3912
3913 4. Since libtool seems to echo the compile commands it is issuing, the output
3914 from "make" can be reduced a bit by putting "@" in front of each libtool
3915 compile command.
3916
3917 5. Patch from the folks at Google for configure.in to be a bit more thorough
3918 in checking for a suitable C++ installation before trying to compile the
3919 C++ stuff. This should fix a reported problem when a compiler was present,
3920 but no suitable headers.
3921
3922 6. The man pages all had just "PCRE" as their title. I have changed them to
3923 be the relevant file name. I have also arranged that these names are
3924 retained in the file doc/pcre.txt, which is a concatenation in text format
3925 of all the man pages except the little individual ones for each function.
3926
3927 7. The NON-UNIX-USE file had not been updated for the different set of source
3928 files that come with release 6. I also added a few comments about the C++
3929 wrapper.
3930
3931
3932 Version 6.0 07-Jun-05
3933 ---------------------
3934
3935 1. Some minor internal re-organization to help with my DFA experiments.
3936
3937 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
3938 didn't matter for the library itself when fully configured, but did matter
3939 when compiling without UCP support, or within Exim, where the ucp files are
3940 not imported.
3941
3942 3. Refactoring of the library code to split up the various functions into
3943 different source modules. The addition of the new DFA matching code (see
3944 below) to a single monolithic source would have made it really too
3945 unwieldy, quite apart from causing all the code to be include in a
3946 statically linked application, when only some functions are used. This is
3947 relevant even without the DFA addition now that patterns can be compiled in
3948 one application and matched in another.
3949
3950 The downside of splitting up is that there have to be some external
3951 functions and data tables that are used internally in different modules of
3952 the library but which are not part of the API. These have all had their
3953 names changed to start with "_pcre_" so that they are unlikely to clash
3954 with other external names.
3955
3956 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
3957 a different (DFA) algorithm. Although it is slower than the original
3958 function, it does have some advantages for certain types of matching
3959 problem.
3960
3961 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
3962 including restarting after a partial match.
3963
3964 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
3965 defined when compiling for Windows was sent to me. I have put it into the
3966 code, though I have no means of testing or verifying it.
3967
3968 7. Added the pcre_refcount() auxiliary function.
3969
3970 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
3971 match before or at the first newline in the subject string. In pcretest,
3972 the /f option on a pattern can be used to set this.
3973
3974 9. A repeated \w when used in UTF-8 mode with characters greater than 256
3975 would behave wrongly. This has been present in PCRE since release 4.0.
3976
3977 10. A number of changes to the pcregrep command:
3978
3979 (a) Refactored how -x works; insert ^(...)$ instead of setting
3980 PCRE_ANCHORED and checking the length, in preparation for adding
3981 something similar for -w.
3982
3983 (b) Added the -w (match as a word) option.
3984
3985 (c) Refactored the way lines are read and buffered so as to have more
3986 than one at a time available.
3987
3988 (d) Implemented a pcregrep test script.
3989
3990 (e) Added the -M (multiline match) option. This allows patterns to match
3991 over several lines of the subject. The buffering ensures that at least
3992 8K, or the rest of the document (whichever is the shorter) is available
3993 for matching (and similarly the previous 8K for lookbehind assertions).
3994
3995 (f) Changed the --help output so that it now says
3996
3997 -w, --word-regex(p)
3998
3999 instead of two lines, one with "regex" and the other with "regexp"
4000 because that confused at least one person since the short forms are the
4001 same. (This required a bit of code, as the output is generated
4002 automatically from a table. It wasn't just a text change.)
4003
4004 (g) -- can be used to terminate pcregrep options if the next thing isn't an
4005 option but starts with a hyphen. Could be a pattern or a path name
4006 starting with a hyphen, for instance.
4007
4008 (h) "-" can be given as a file name to represent stdin.
4009
4010 (i) When file names are being printed, "(standard input)" is used for
4011 the standard input, for compatibility with GNU grep. Previously
4012 "<stdin>" was used.
4013
4014 (j) The option --label=xxx can be used to supply a name to be used for
4015 stdin when file names are being printed. There is no short form.
4016
4017 (k) Re-factored the options decoding logic because we are going to add
4018 two more options that take data. Such options can now be given in four
4019 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
4020
4021 (l) Added the -A, -B, and -C options for requesting that lines of context
4022 around matches be printed.
4023
4024 (m) Added the -L option to print the names of files that do not contain
4025 any matching lines, that is, the complement of -l.
4026
4027 (n) The return code is 2 if any file cannot be opened, but pcregrep does
4028 continue to scan other files.
4029
4030 (o) The -s option was incorrectly implemented. For compatibility with other
4031 greps, it now suppresses the error message for a non-existent or non-
4032 accessible file (but not the return code). There is a new option called
4033 -q that suppresses the output of matching lines, which was what -s was
4034 previously doing.
4035
4036 (p) Added --include and --exclude options to specify files for inclusion
4037 and exclusion when recursing.
4038
4039 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
4040 Hopefully, it now does.
4041
4042 12. Missing cast in pcre_study().
4043
4044 13. Added an "uninstall" target to the makefile.
4045
4046 14. Replaced "extern" in the function prototypes in Makefile.in with
4047 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
4048 world, but is set differently for Windows.
4049
4050 15. Added a second compiling function called pcre_compile2(). The only
4051 difference is that it has an extra argument, which is a pointer to an
4052 integer error code. When there is a compile-time failure, this is set
4053 non-zero, in addition to the error test pointer being set to point to an
4054 error message. The new argument may be NULL if no error number is required
4055 (but then you may as well call pcre_compile(), which is now just a
4056 wrapper). This facility is provided because some applications need a
4057 numeric error indication, but it has also enabled me to tidy up the way
4058 compile-time errors are handled in the POSIX wrapper.
4059
4060 16. Added VPATH=.libs to the makefile; this should help when building with one
4061 prefix path and installing with another. (Or so I'm told by someone who
4062 knows more about this stuff than I do.)
4063
4064 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
4065 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
4066 match everything, including newlines. This is not POSIX-compatible, but
4067 somebody wanted the feature. From pcretest it can be activated by using
4068 both the P and the s flags.
4069
4070 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
4071
4072 19. libpcre.pc was being incorrectly installed as executable.
4073
4074 20. A couple of places in pcretest check for end-of-line by looking for '\n';
4075 it now also looks for '\r' so that it will work unmodified on Windows.
4076
4077 21. Added Google's contributed C++ wrapper to the distribution.
4078
4079 22. Added some untidy missing memory free() calls in pcretest, to keep
4080 Electric Fence happy when testing.
4081
4082
4083
4084 Version 5.0 13-Sep-04
4085 ---------------------
4086
4087 1. Internal change: literal characters are no longer packed up into items
4088 containing multiple characters in a single byte-string. Each character
4089 is now matched using a separate opcode. However, there may be more than one
4090 byte in the character in UTF-8 mode.
4091
4092 2. The pcre_callout_block structure has two new fields: pattern_position and
4093 next_item_length. These contain the offset in the pattern to the next match
4094 item, and its length, respectively.
4095
4096 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
4097 insertion of callouts before each pattern item. Added the /C option to
4098 pcretest to make use of this.
4099
4100 4. On the advice of a Windows user, the lines
4101
4102 #if defined(_WIN32) || defined(WIN32)
4103 _setmode( _fileno( stdout ), 0x8000 );
4104 #endif /* defined(_WIN32) || defined(WIN32) */
4105
4106 have been added to the source of pcretest. This apparently does useful
4107 magic in relation to line terminators.
4108
4109 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
4110 for the benefit of those environments where the "b" makes a difference.
4111
4112 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
4113 to know about it. I have put a hack into configure.in that adds in code
4114 to set GCC=yes if CC=icc. This seems to end up at a point in the
4115 generated configure script that is early enough to affect the setting of
4116 compiler options, which is what is needed, but I have no means of testing
4117 whether it really works. (The user who reported this had patched the
4118 generated configure script, which of course I cannot do.)
4119
4120 LATER: After change 22 below (new libtool files), the configure script
4121 seems to know about icc (and also ecc). Therefore, I have commented out
4122 this hack in configure.in.
4123
4124 7. Added support for pkg-config (2 patches were sent in).
4125
4126 8. Negated POSIX character classes that used a combination of internal tables
4127 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
4128 [[:^ascii]]. Typically, they would match almost any characters. The other
4129 POSIX classes were not broken in this way.
4130
4131 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
4132 to find the match, as PCRE was deluded into thinking that the match had to
4133 start at the start point or following a newline. The same bug applied to
4134 patterns with negative forward assertions or any backward assertions
4135 preceding ".*" at the start, unless the pattern required a fixed first
4136 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
4137
4138 10. In UTF-8 mode, when moving forwards in the subject after a failed match
4139 starting at the last subject character, bytes beyond the end of the subject
4140 string were read.
4141
4142 11. Renamed the variable "class" as "classbits" to make life easier for C++
4143 users. (Previously there was a macro definition, but it apparently wasn't
4144 enough.)
4145
4146 12. Added the new field "tables" to the extra data so that tables can be passed
4147 in at exec time, or the internal tables can be re-selected. This allows
4148 a compiled regex to be saved and re-used at a later time by a different
4149 program that might have everything at different addresses.
4150
4151 13. Modified the pcre-config script so that, when run on Solaris, it shows a
4152 -R library as well as a -L library.
4153
4154 14. The debugging options of pcretest (-d on the command line or D on a
4155 pattern) showed incorrect output for anything following an extended class
4156 that contained multibyte characters and which was followed by a quantifier.
4157
4158 15. Added optional support for general category Unicode character properties
4159 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
4160 support. It adds about 90K to the size of the library. The meanings of the
4161 inbuilt class escapes such as \d and \s have NOT been changed.
4162
4163 16. Updated pcredemo.c to include calls to free() to release the memory for the
4164 compiled pattern.
4165
4166 17. The generated file chartables.c was being created in the source directory
4167 instead of in the building directory. This caused the build to fail if the
4168 source directory was different from the building directory, and was
4169 read-only.
4170
4171 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
4172 file. No doubt somebody will tell me if they don't make sense... Also added
4173 Dan Mooney's comments about building on OpenVMS.
4174
4175 19. Added support for partial matching via the PCRE_PARTIAL option for
4176 pcre_exec() and the \P data escape in pcretest.
4177
4178 20. Extended pcretest with 3 new pattern features:
4179
4180 (i) A pattern option of the form ">rest-of-line" causes pcretest to
4181 write the compiled pattern to the file whose name is "rest-of-line".
4182 This is a straight binary dump of the data, with the saved pointer to
4183 the character tables forced to be NULL. The study data, if any, is
4184 written too. After writing, pcretest reads a new pattern.
4185
4186 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
4187 compiled pattern from the given file. There must not be any
4188 occurrences of "<" in the file name (pretty unlikely); if there are,
4189 pcretest will instead treat the initial "<" as a pattern delimiter.
4190 After reading in the pattern, pcretest goes on to read data lines as
4191 usual.
4192
4193 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
4194 and 16-bit fields in a compiled pattern, to simulate a pattern that
4195 was compiled on a host of opposite endianness.
4196
4197 21. The pcre-exec() function can now cope with patterns that were compiled on
4198 hosts of opposite endianness, with this restriction:
4199
4200 As for any compiled expression that is saved and used later, the tables
4201 pointer field cannot be preserved; the extra_data field in the arguments
4202 to pcre_exec() should be used to pass in a tables address if a value
4203 other than the default internal tables were used at compile time.
4204
4205 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
4206 now diagnosed as an error. Previously, most of the time, a negative number
4207 would have been treated as zero, but if in addition "ovector" was passed as
4208 NULL, a crash could occur.
4209
4210 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
4211 new versions from the libtool 1.5 distribution (the last one is a copy of
4212 a file called libtool.m4). This seems to have fixed the need to patch
4213 "configure" to support Darwin 1.3 (which I used to do). However, I still
4214 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
4215 workstation).
4216
4217 24. Changed the PCRE licence to be the more standard "BSD" licence.
4218
4219
4220 Version 4.5 01-Dec-03
4221 ---------------------
4222
4223 1. There has been some re-arrangement of the code for the match() function so
4224 that it can be compiled in a version that does not call itself recursively.
4225 Instead, it keeps those local variables that need separate instances for
4226 each "recursion" in a frame on the heap, and gets/frees frames whenever it
4227 needs to "recurse". Keeping track of where control must go is done by means
4228 of setjmp/longjmp. The whole thing is implemented by a set of macros that
4229 hide most of the details from the main code, and operates only if
4230 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
4231 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
4232 operating.
4233
4234 To make it easier for callers to provide specially tailored get/free
4235 functions for this usage, two new functions, pcre_stack_malloc, and
4236 pcre_stack_free, are used. They are always called in strict stacking order,
4237 and the size of block requested is always the same.
4238
4239 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
4240 PCRE has been compiled to use the stack or the heap for recursion. The
4241 -C option of pcretest uses this to show which version is compiled.
4242
4243 A new data escape \S, is added to pcretest; it causes the amounts of store
4244 obtained and freed by both kinds of malloc/free at match time to be added
4245 to the output.
4246
4247 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
4248 what's available on my current Linux desktop machine.
4249
4250 3. When matching a UTF-8 string, the test for a valid string at the start has
4251 been extended. If start_offset is not zero, PCRE now checks that it points
4252 to a byte that is the start of a UTF-8 character. If not, it returns
4253 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
4254 this is necessary because there may be backward assertions in the pattern.
4255 When matching the same subject several times, it may save resources to use
4256 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
4257
4258 4. The code for checking the validity of UTF-8 strings has been tightened so
4259 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
4260 containing "overlong sequences".
4261
4262 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
4263 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
4264 should have been "&", but it just so happened that all the cases this let
4265 through by mistake were picked up later in the function.
4266
4267 6. I had used a variable called "isblank" - this is a C99 function, causing
4268 some compilers to warn. To avoid this, I renamed it (as "blankclass").
4269
4270 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
4271 prompting; (b) run "./pcretest /dev/null" at the start of the test script
4272 so the version is shown; (c) stop "make test" echoing "./RunTest".
4273
4274 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
4275
4276 9. The prototype for memmove() for systems that don't have it was using
4277 size_t, but the inclusion of the header that defines size_t was later. I've
4278 moved the #includes for the C headers earlier to avoid this.
4279
4280 10. Added some adjustments to the code to make it easier to compiler on certain
4281 special systems:
4282
4283 (a) Some "const" qualifiers were missing.
4284 (b) Added the macro EXPORT before all exported functions; by default this
4285 is defined to be empty.
4286 (c) Changed the dftables auxiliary program (that builds chartables.c) so
4287 that it reads its output file name as an argument instead of writing
4288 to the standard output and assuming this can be redirected.
4289
4290 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
4291 class containing characters with values greater than 255, PCRE compilation
4292 went into a loop.
4293
4294 12. A recursive reference to a subpattern that was within another subpattern
4295 that had a minimum quantifier of zero caused PCRE to crash. For example,
4296 (x(y(?2))z)? provoked this bug with a subject that got as far as the
4297 recursion. If the recursively-called subpattern itself had a zero repeat,
4298 that was OK.
4299
4300 13. In pcretest, the buffer for reading a data line was set at 30K, but the
4301 buffer into which it was copied (for escape processing) was still set at
4302 1024, so long lines caused crashes.
4303
4304 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
4305 "internal error: code overflow...". This applied to any character class
4306 that was followed by a possessive quantifier.
4307
4308 15. Modified the Makefile to add libpcre.la as a prerequisite for
4309 libpcreposix.la because I was told this is needed for a parallel build to
4310 work.
4311
4312 16. If a pattern that contained .* following optional items at the start was
4313 studied, the wrong optimizing data was generated, leading to matching
4314 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
4315 matching string must start with a or b or c. The correct conclusion for
4316 this pattern is that a match can start with any character.
4317
4318
4319 Version 4.4 13-Aug-03
4320 ---------------------
4321
4322 1. In UTF-8 mode, a character class containing characters with values between
4323 127 and 255 was not handled correctly if the compiled pattern was studied.
4324 In fixing this, I have also improved the studying algorithm for such
4325 classes (slightly).
4326
4327 2. Three internal functions had redundant arguments passed to them. Removal
4328 might give a very teeny performance improvement.
4329
4330 3. Documentation bug: the value of the capture_top field in a callout is *one
4331 more than* the number of the hightest numbered captured substring.
4332
4333 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
4334 in incorrectly linking with a previously installed version. They now link
4335 explicitly with libpcre.la.
4336
4337 5. configure.in no longer needs to recognize Cygwin specially.
4338
4339 6. A problem in pcre.in for Windows platforms is fixed.
4340
4341 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
4342 pcretest, it used to include the size of the study block as part of its
4343 output. Unfortunately, the structure contains a field that has a different
4344 size on different hardware architectures. This meant that the tests that
4345 showed this size failed. As the block is currently always of a fixed size,
4346 this information isn't actually particularly useful in pcretest output, so
4347 I have just removed it.
4348
4349 8. Three pre-processor statements accidentally did not start in column 1.
4350 Sadly, there are *still* compilers around that complain, even though
4351 standard C has not required this for well over a decade. Sigh.
4352
4353 9. In pcretest, the code for checking callouts passed small integers in the
4354 callout_data field, which is a void * field. However, some picky compilers
4355 complained about the casts involved for this on 64-bit systems. Now
4356 pcretest passes the address of the small integer instead, which should get
4357 rid of the warnings.
4358
4359 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
4360 both compile and run time, and gives an error if an invalid UTF-8 sequence
4361 is found. There is a option for disabling this check in cases where the
4362 string is known to be correct and/or the maximum performance is wanted.
4363
4364 11. In response to a bug report, I changed one line in Makefile.in from
4365
4366 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
4367 to
4368 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
4369
4370 to look similar to other lines, but I have no way of telling whether this
4371 is the right thing to do, as I do not use Windows. No doubt I'll get told
4372 if it's wrong...
4373
4374
4375 Version 4.3 21-May-03
4376 ---------------------
4377
4378 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
4379 Makefile.
4380
4381 2. Some refactoring to improve the quality of the code:
4382
4383 (i) The utf8_table... variables are now declared "const".
4384
4385 (ii) The code for \cx, which used the "case flipping" table to upper case
4386 lower case letters, now just substracts 32. This is ASCII-specific,
4387 but the whole concept of \cx is ASCII-specific, so it seems
4388 reasonable.
4389
4390 (iii) PCRE was using its character types table to recognize decimal and
4391 hexadecimal digits in the pattern. This is silly, because it handles
4392 only 0-9, a-f, and A-F, but the character types table is locale-
4393 specific, which means strange things might happen. A private
4394 table is now used for this - though it costs 256 bytes, a table is
4395 much faster than multiple explicit tests. Of course, the standard
4396 character types table is still used for matching digits in subject
4397 strings against \d.
4398
4399 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
4400 ending in _t are). So I've renamed it as ESC_tee.
4401
4402 3. The first argument for regexec() in the POSIX wrapper should have been
4403 defined as "const".
4404
4405 4. Changed pcretest to use malloc() for its buffers so that they can be
4406 Electric Fenced for debugging.
4407
4408 5. There were several places in the code where, in UTF-8 mode, PCRE would try
4409 to read one or more bytes before the start of the subject string. Often this
4410 had no effect on PCRE's behaviour, but in some circumstances it could
4411 provoke a segmentation fault.
4412
4413 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
4414 to try to read one or more bytes before the start of the subject string.
4415
4416 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
4417 UTF-8 support could misbehave in various ways if the subject string
4418 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
4419 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
4420 back over UTF-8 characters.)
4421
4422
4423 Version 4.2 14-Apr-03
4424 ---------------------
4425
4426 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
4427
4428 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
4429 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
4430 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
4431 [WIN_PREFIX]: new variable, "cyg" for Cygwin
4432 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
4433 and BUILD_EXEEXT
4434 Note: automatic setting of the BUILD variables is not yet working
4435 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
4436 compile-time but not at link-time
4437 [LINK]: use for linking executables only
4438 make different versions for Windows and non-Windows
4439 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
4440 libraries
4441 [LINK_FOR_BUILD]: new variable
4442 [OBJEXT]: use throughout
4443 [EXEEXT]: use throughout
4444 <winshared>: new target
4445 <wininstall>: new target
4446 <dftables.o>: use native compiler
4447 <dftables>: use native linker
4448 <install>: handle Windows platform correctly
4449 <clean>: ditto
4450 <check>: ditto
4451 copy DLL to top builddir before testing
4452
4453 As part of these changes, -no-undefined was removed again. This was reported
4454 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
4455 in any case.
4456
4457 3. Some tidies to get rid of compiler warnings:
4458
4459 . In the match_data structure, match_limit was an unsigned long int, whereas
4460 match_call_count was an int. I've made them both unsigned long ints.
4461
4462 . In pcretest the fact that a const uschar * doesn't automatically cast to
4463 a void * provoked a warning.
4464
4465 . Turning on some more compiler warnings threw up some "shadow" variables
4466 and a few more missing casts.
4467
4468 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
4469 option, a class that contained a single character with a value between 128
4470 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
4471
4472 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
4473 option, a class that contained several characters, but with at least one
4474 whose value was between 128 and 255 caused PCRE to crash.
4475
4476
4477 Version 4.1 12-Mar-03
4478 ---------------------
4479
4480 1. Compiling with gcc -pedantic found a couple of places where casts were
4481 needed, and a string in dftables.c that was longer than standard compilers are
4482 required to support.
4483
4484 2. Compiling with Sun's compiler found a few more places where the code could
4485 be tidied up in order to avoid warnings.
4486
4487 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
4488 first of these names is deprecated in the latest Autoconf in favour of the name
4489 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
4490 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
4491 analogy I have changed it to CFLAGS_FOR_BUILD.
4492
4493 4. Added -no-undefined to the linking command in the Makefile, because this is
4494 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
4495 linking step for the pcreposix library.
4496
4497 5. PCRE was failing to diagnose the case of two named groups with the same
4498 name.
4499
4500 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
4501 literal character that is needed in the subject for a match, and scans along to
4502 ensure that it is present before embarking on the full matching process. This
4503 saves time in cases of nested unlimited repeats that are never going to match.
4504 Problem: the scan can take a lot of time if the subject is very long (e.g.
4505 megabytes), thus penalizing straightforward matches. It is now done only if the
4506 amount of subject to be scanned is less than 1000 bytes.
4507
4508 7. A lesser problem with the same optimization is that it was recording the
4509 first character of an anchored pattern as "needed", thus provoking a search
4510 right along the subject, even when the first match of the pattern was going to
4511 fail. The "needed" character is now not set for anchored patterns, unless it
4512 follows something in the pattern that is of non-fixed length. Thus, it still
4513 fulfils its original purpose of finding quick non-matches in cases of nested
4514 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
4515
4516
4517 Version 4.0 17-Feb-03
4518 ---------------------
4519
4520 1. If a comment in an extended regex that started immediately after a meta-item
4521 extended to the end of string, PCRE compiled incorrect data. This could lead to
4522 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
4523
4524 2. Moved to autoconf 2.53 and libtool 1.4.2.
4525
4526 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
4527 the special perltest8 script is no longer needed - all the tests can be run
4528 from a single perltest script.
4529
4530 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
4531 by \s. It has now been removed in PCRE. This means it isn't recognized as
4532 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
4533 class [:space:] *does* include VT, thereby creating a mess.
4534
4535 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
4536 space and tab.
4537
4538 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
4539 its new features into the main test script, reducing the number of scripts.
4540
4541 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
4542 were backward compatible, and made the (?i) apply to the whole pattern, as if
4543 /i were given. Now it behaves more logically, and applies the option setting
4544 only to what follows. PCRE has been changed to follow suit. However, if it
4545 finds options settings right at the start of the pattern, it extracts them into
4546 the global options, as before. Thus, they show up in the info data.
4547
4548 8. Added support for the \Q...\E escape sequence. Characters in between are
4549 treated as literals. This is slightly different from Perl in that $ and @ are
4550 also handled as literals inside the quotes. In Perl, they will cause variable
4551 interpolation. Note the following examples:
4552
4553 Pattern PCRE matches Perl matches
4554
4555 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
4556 \Qabc\$xyz\E abc\$xyz abc\$xyz
4557 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
4558
4559 For compatibility with Perl, \Q...\E sequences are recognized inside character
4560 classes as well as outside them.
4561
4562 9. Re-organized 3 code statements in pcretest to avoid "overflow in
4563 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
4564 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
4565 signed/unsigned warnings.
4566
4567 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
4568 option for pcretest, so I've replaced it by a simple function that does just
4569 that job.
4570
4571 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
4572 "pcregrep -".
4573
4574 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
4575 Java package. This provides some syntactic sugar for simple cases of what my
4576 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
4577 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
4578 item, you can use this simplified notation. Note that only makes sense with
4579 greedy quantifiers. Consequently, the use of the possessive quantifier forces
4580 greediness, whatever the setting of the PCRE_UNGREEDY option.
4581
4582 13. A change of greediness default within a pattern was not taking effect at
4583 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
4584 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
4585 was abstracted outside.
4586
4587 14. PCRE now supports the \G assertion. It is true when the current matching
4588 position is at the start point of the match. This differs from \A when the
4589 starting offset is non-zero. Used with the /g option of pcretest (or similar
4590 code), it works in the same way as it does for Perl's /g option. If all
4591 alternatives of a regex begin with \G, the expression is anchored to the start
4592 match position, and the "anchored" flag is set in the compiled expression.
4593
4594 15. Some bugs concerning the handling of certain option changes within patterns
4595 have been fixed. These applied to options other than (?ims). For example,
4596 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
4597 been the other way round. Some of this was related to change 7 above.
4598
4599 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
4600 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
4601 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
4602 POSIX classes only within a class (e.g. /[[:alpha:]]/).
4603
4604 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
4605 mode. Unlike ".", it always matches newline, whatever the setting of
4606 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
4607 assertions. Perl allows it, but it doesn't (in general) work because it can't
4608 calculate the length of the lookbehind. At least, that's the case for Perl
4609 5.8.0 - I've been told they are going to document that it doesn't work in
4610 future.
4611
4612 18. Added an error diagnosis for escapes that PCRE does not support: these are
4613 \L, \l, \N, \P, \p, \U, \u, and \X.
4614
4615 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
4616 reading past the end of the pattern in cases such as /[abcd/.
4617
4618 20. PCRE was getting more memory than necessary for patterns with classes that
4619 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
4620
4621 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
4622 compiling PCRE for use with Virtual Pascal.
4623
4624 22. Small fix to the Makefile to make it work properly if the build is done
4625 outside the source tree.
4626
4627 23. Added a new extension: a condition to go with recursion. If a conditional
4628 subpattern starts with (?(R) the "true" branch is used if recursion has
4629 happened, whereas the "false" branch is used only at the top level.
4630
4631 24. When there was a very long string of literal characters (over 255 bytes
4632 without UTF support, over 250 bytes with UTF support), the computation of how
4633 much memory was required could be incorrect, leading to segfaults or other
4634 strange effects.
4635
4636 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
4637 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
4638 there was a subsequent back reference to those brackets. This meant that, for
4639 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
4640 possible to check for precisely this case. All we can do is abandon the
4641 optimization if .* occurs inside capturing brackets when there are any back
4642 references whatsoever. (See below for a better fix that came later.)
4643
4644 26. The handling of the optimization for finding the first character of a
4645 non-anchored pattern, and for finding a character that is required later in the
4646 match were failing in some cases. This didn't break the matching; it just
4647 failed to optimize when it could. The way this is done has been re-implemented.
4648
4649 27. Fixed typo in error message for invalid (?R item (it said "(?p").
4650
4651 28. Added a new feature that provides some of the functionality that Perl
4652 provides with (?{...}). The facility is termed a "callout". The way it is done
4653 in PCRE is for the caller to provide an optional function, by setting
4654 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
4655 global variable. By default it is unset, which disables all calling out. To get
4656 the function called, the regex must include (?C) at appropriate points. This
4657 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
4658 This provides a means of identifying different callout points. When PCRE
4659 reaches such a point in the regex, if pcre_callout has been set, the external
4660 function is called. It is provided with data in a structure called
4661 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
4662 matching continues; if it returns a non-zero value, the match at the current
4663 point fails. However, backtracking will occur if possible. [This was changed
4664 later and other features added - see item 49 below.]
4665
4666 29. pcretest is upgraded to test the callout functionality. It provides a
4667 callout function that displays information. By default, it shows the start of
4668 the match and the current position in the text. There are some new data escapes
4669 to vary what happens:
4670
4671 \C+ in addition, show current contents of captured substrings
4672 \C- do not supply a callout function
4673 \C!n return 1 when callout number n is reached
4674 \C!n!m return 1 when callout number n is reached for the mth time
4675
4676 30. If pcregrep was called with the -l option and just a single file name, it
4677 output "<stdin>" if a match was found, instead of the file name.
4678
4679 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
4680 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
4681 pcre_exec(). This saves a malloc/free per call. The default value of
4682 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
4683 when configuring.
4684
4685 32. The default maximum size of a compiled pattern is 64K. There have been a
4686 few cases of people hitting this limit. The code now uses macros to handle the
4687 storing of links as offsets within the compiled pattern. It defaults to 2-byte
4688 links, but this can be changed to 3 or 4 bytes by --with-link-size when
4689 configuring. Tests 2 and 5 work only with 2-byte links because they output
4690 debugging information about compiled patterns.
4691
4692 33. Internal code re-arrangements:
4693
4694 (a) Moved the debugging function for printing out a compiled regex into
4695 its own source file (printint.c) and used #include to pull it into
4696 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
4697 separate copies.
4698
4699 (b) Defined the list of op-code names for debugging as a macro in
4700 internal.h so that it is next to the definition of the opcodes.
4701
4702 (c) Defined a table of op-code lengths for simpler skipping along compiled
4703 code. This is again a macro in internal.h so that it is next to the
4704 definition of the opcodes.
4705
4706 34. Added support for recursive calls to individual subpatterns, along the
4707 lines of Robin Houston's patch (but implemented somewhat differently).
4708
4709 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
4710 allow it to read and process whole directories in Win32. This code was
4711 contributed by Lionel Fourquaux; it has not been tested by me.
4712
4713 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
4714 used to name a group. Names consist of alphanumerics and underscores, and must
4715 be unique. Back references use the syntax (?P=name) and recursive calls use
4716 (?P>name) which is a PCRE extension to the Python extension. Groups still have
4717 numbers. The function pcre_fullinfo() can be used after compilation to extract
4718 a name/number map. There are three relevant calls:
4719
4720 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
4721 PCRE_INFO_NAMECOUNT yields the number of entries
4722 PCRE_INFO_NAMETABLE yields a pointer to the map.
4723
4724 The map is a vector of fixed-size entries. The size of each entry depends on
4725 the length of the longest name used. The first two bytes of each entry are the
4726 group number, most significant byte first. There follows the corresponding
4727 name, zero terminated. The names are in alphabetical order.
4728
4729 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
4730 case instead of 255. Making it the same both with and without UTF-8 support
4731 means that the same test output works with both.
4732
4733 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
4734 calling malloc() with a zero argument.
4735
4736 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
4737 optimization. I've improved things by keeping a bitmap of backreferences with
4738 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
4739 fact referenced, the optimization can be applied. It is unlikely that a
4740 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
4741 the match to follow \n) will appear inside brackets with a number greater than
4742 31, but if it does, any back reference > 31 suppresses the optimization.
4743
4744 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
4745 of disabling numbered capturing parentheses. Any opening parenthesis that is
4746 not followed by ? behaves as if it were followed by ?: but named parentheses
4747 can still be used for capturing (and they will acquire numbers in the usual
4748 way).
4749
4750 41. Redesigned the return codes from the match() function into yes/no/error so
4751 that errors can be passed back from deep inside the nested calls. A malloc
4752 failure while inside a recursive subpattern call now causes the
4753 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
4754
4755 42. It is now possible to set a limit on the number of times the match()
4756 function is called in a call to pcre_exec(). This facility makes it possible to
4757 limit the amount of recursion and backtracking, though not in a directly
4758 obvious way, because the match() function is used in a number of different
4759 circumstances. The count starts from zero for each position in the subject
4760 string (for non-anchored patterns). The default limit is, for compatibility, a
4761 large number, namely 10 000 000. You can change this in two ways:
4762
4763 (a) When configuring PCRE before making, you can use --with-match-limit=n
4764 to set a default value for the compiled library.
4765
4766 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
4767 a different value is set. See 45 below.
4768
4769 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
4770
4771 43. Added a new function pcre_config(int, void *) to enable run-time extraction
4772 of things that can be changed at compile time. The first argument specifies
4773 what is wanted and the second points to where the information is to be placed.
4774 The current list of available information is:
4775
4776 PCRE_CONFIG_UTF8
4777
4778 The output is an integer that is set to one if UTF-8 support is available;
4779 otherwise it is set to zero.
4780
4781 PCRE_CONFIG_NEWLINE
4782
4783 The output is an integer that it set to the value of the code that is used for
4784 newline. It is either LF (10) or CR (13).
4785
4786 PCRE_CONFIG_LINK_SIZE
4787
4788 The output is an integer that contains the number of bytes used for internal
4789 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
4790
4791 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
4792
4793 The output is an integer that contains the threshold above which the POSIX
4794 interface uses malloc() for output vectors. See item 31 above.
4795
4796 PCRE_CONFIG_MATCH_LIMIT
4797
4798 The output is an unsigned integer that contains the default limit of the number
4799 of match() calls in a pcre_exec() execution. See 42 above.
4800
4801 44. pcretest has been upgraded by the addition of the -C option. This causes it
4802 to extract all the available output from the new pcre_config() function, and to
4803 output it. The program then exits immediately.
4804
4805 45. A need has arisen to pass over additional data with calls to pcre_exec() in
4806 order to support additional features. One way would have been to define
4807 pcre_exec2() (for example) with extra arguments, but this would not have been
4808 extensible, and would also have required all calls to the original function to
4809 be mapped to the new one. Instead, I have chosen to extend the mechanism that
4810 is used for passing in "extra" data from pcre_study().
4811
4812 The pcre_extra structure is now exposed and defined in pcre.h. It currently
4813 contains the following fields:
4814
4815 flags a bitmap indicating which of the following fields are set
4816 study_data opaque data from pcre_study()
4817 match_limit a way of specifying a limit on match() calls for a specific
4818 call to pcre_exec()
4819 callout_data data for callouts (see 49 below)
4820
4821 The flag bits are also defined in pcre.h, and are
4822
4823 PCRE_EXTRA_STUDY_DATA
4824 PCRE_EXTRA_MATCH_LIMIT
4825 PCRE_EXTRA_CALLOUT_DATA
4826
4827 The pcre_study() function now returns one of these new pcre_extra blocks, with
4828 the actual study data pointed to by the study_data field, and the
4829 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
4830 before. That is, this change is entirely upwards-compatible and requires no
4831 change to existing code.
4832
4833 If you want to pass in additional data to pcre_exec(), you can either place it
4834 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
4835 block.
4836
4837 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
4838 data string contains the escape sequence \M, pcretest calls pcre_exec() several
4839 times with different match limits, until it finds the minimum value needed for
4840 pcre_exec() to complete. The value is then output. This can be instructive; for
4841 most simple matches the number is quite small, but for pathological cases it
4842 gets very large very quickly.
4843
4844 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
4845 returns the size of the data block pointed to by the study_data field in a
4846 pcre_extra block, that is, the value that was passed as the argument to
4847 pcre_malloc() when PCRE was getting memory in which to place the information
4848 created by pcre_study(). The fourth argument should point to a size_t variable.
4849 pcretest has been extended so that this information is shown after a successful
4850 pcre_study() call when information about the compiled regex is being displayed.
4851
4852 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
4853 because what follows is always an absolute path. (Later: it turns out that this
4854 is more than cosmetic for MinGW, because it doesn't like empty path
4855 components.)
4856
4857 49. Some changes have been made to the callout feature (see 28 above):
4858
4859 (i) A callout function now has three choices for what it returns:
4860
4861 0 => success, carry on matching
4862 > 0 => failure at this point, but backtrack if possible
4863 < 0 => serious error, return this value from pcre_exec()
4864
4865 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
4866 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
4867 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
4868 use by callout functions. It will never be used by PCRE itself.
4869
4870 (ii) The pcre_extra structure (see 45 above) has a void * field called
4871 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
4872 pcre_callout_block structure has a field of the same name. The contents of
4873 the field passed in the pcre_extra structure are passed to the callout
4874 function in the corresponding field in the callout block. This makes it
4875 easier to use the same callout-containing regex from multiple threads. For
4876 testing, the pcretest program has a new data escape
4877
4878 \C*n pass the number n (may be negative) as callout_data
4879
4880 If the callout function in pcretest receives a non-zero value as
4881 callout_data, it returns that value.
4882
4883 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
4884 there were some redundant $(CFLAGS) in commands that are now specified as
4885 $(LINK), which already includes $(CFLAGS).
4886
4887 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
4888 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
4889 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
4890 one-byte characters throughout. Note that case-insensitive matching applies
4891 only to characters whose values are less than 256. PCRE doesn't support the
4892 notion of cases for higher-valued characters.
4893
4894 (i) A character class whose characters are all within 0-255 is handled as
4895 a bit map, and the map is inverted for negative classes. Previously, a
4896 character > 255 always failed to match such a class; however it should
4897 match if the class was a negative one (e.g. [^ab]). This has been fixed.
4898
4899 (ii) A negated character class with a single character < 255 is coded as
4900 "not this character" (OP_NOT). This wasn't working properly when the test
4901 character was multibyte, either singly or repeated.
4902
4903 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
4904 mode, for example: \x{100}{2,3}.
4905
4906 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
4907 singly or repeated) now correctly test multibyte characters. However,
4908 PCRE doesn't recognize any characters with values greater than 255 as
4909 digits, spaces, or word characters. Such characters always match \D, \S,
4910 and \W, and never match \d, \s, or \w.
4911
4912 (v) Classes may now contain characters and character ranges with values
4913 greater than 255. For example: [ab\x{100}-\x{400}].
4914
4915 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
4916 PCRE in UTF-8 mode.
4917
4918 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
4919 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
4920 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
4921 value.)
4922
4923 53. The single man page has become too large. I have therefore split it up into
4924 a number of separate man pages. These also give rise to individual HTML pages;
4925 these are now put in a separate directory, and there is an index.html page that
4926 lists them all. Some hyperlinking between the pages has been installed.
4927
4928 54. Added convenience functions for handling named capturing parentheses.
4929
4930 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
4931 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
4932 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
4933 are faulted.
4934
4935 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
4936 calling configure. These values are used when compiling the dftables.c program
4937 which is run to generate the source of the default character tables. They
4938 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
4939 you will need to set these values.
4940
4941 57. Updated the building process for Windows DLL, as provided by Fred Cox.
4942
4943
4944 Version 3.9 02-Jan-02
4945 ---------------------
4946
4947 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
4948
4949 2. If --disable-static was given, the building process failed when trying to
4950 build pcretest and pcregrep. (For some reason it was using libtool to compile
4951 them, which is not right, as they aren't part of the library.)
4952
4953
4954 Version 3.8 18-Dec-01
4955 ---------------------
4956
4957 1. The experimental UTF-8 code was completely screwed up. It was packing the
4958 bytes in the wrong order. How dumb can you get?
4959
4960
4961 Version 3.7 29-Oct-01
4962 ---------------------
4963
4964 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
4965 This caused pcretest, when used on the test data, to segfault. Unfortunately,
4966 this didn't happen under Solaris 8, where I normally test things.
4967
4968 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
4969 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
4970 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
4971 this fix an hour or so after the initial 3.7 release.)
4972
4973
4974 Version 3.6 23-Oct-01
4975 ---------------------
4976
4977 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
4978 offsets passed as NULL with zero offset count.
4979
4980 2. The config.guess and config.sub files had not been updated when I moved to
4981 the latest autoconf.
4982
4983
4984 Version 3.5 15-Aug-01
4985 ---------------------
4986
4987 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
4988 had been forgotten.
4989
4990 2. By using declared but undefined structures, we can avoid using "void"
4991 definitions in pcre.h while keeping the internal definitions of the structures
4992 private.
4993
4994 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
4995 user point of view, this means that both static and shared libraries are built
4996 by default, but this can be individually controlled. More of the work of
4997 handling this static/shared cases is now inside libtool instead of PCRE's make
4998 file.
4999
5000 4. The pcretest utility is now installed along with pcregrep because it is
5001 useful for users (to test regexs) and by doing this, it automatically gets
5002 relinked by libtool. The documentation has been turned into a man page, so
5003 there are now .1, .txt, and .html versions in /doc.
5004
5005 5. Upgrades to pcregrep:
5006 (i) Added long-form option names like gnu grep.
5007 (ii) Added --help to list all options with an explanatory phrase.
5008 (iii) Added -r, --recursive to recurse into sub-directories.
5009 (iv) Added -f, --file to read patterns from a file.
5010
5011 6. pcre_exec() was referring to its "code" argument before testing that
5012 argument for NULL (and giving an error if it was NULL).
5013
5014 7. Upgraded Makefile.in to allow for compiling in a different directory from
5015 the source directory.
5016
5017 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
5018 options bits, the pointer it was passed was to an int instead of to an unsigned
5019 long int. This mattered only on 64-bit systems.
5020
5021 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
5022 generated) instead of pcre.in, which it its source. Also made the same change
5023 in several of the .c files.
5024
5025 10. A new release of gcc defines printf() as a macro, which broke pcretest
5026 because it had an ifdef in the middle of a string argument for printf(). Fixed
5027 by using separate calls to printf().
5028
5029 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
5030 script, to force use of CR or LF instead of \n in the source. On non-Unix
5031 systems, the value can be set in config.h.
5032
5033 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
5034 absolute limit. Changed the text of the error message to make this clear, and
5035 likewise updated the man page.
5036
5037 13. The limit of 99 on the number of capturing subpatterns has been removed.
5038 The new limit is 65535, which I hope will not be a "real" limit.
5039
5040
5041 Version 3.4 22-Aug-00
5042 ---------------------
5043
5044 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
5045
5046 2. Diagnose condition (?(0) as an error instead of crashing on matching.
5047
5048
5049 Version 3.3 01-Aug-00
5050 ---------------------
5051
5052 1. If an octal character was given, but the value was greater than \377, it
5053 was not getting masked to the least significant bits, as documented. This could
5054 lead to crashes in some systems.
5055
5056 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
5057 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
5058
5059 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
5060 These just pass their arguments on to (pcre_free)(), but they are provided
5061 because some uses of PCRE bind it to non-C systems that can call its functions,
5062 but cannot call free() or pcre_free() directly.
5063
5064 4. Add "make test" as a synonym for "make check". Corrected some comments in
5065 the Makefile.
5066
5067 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
5068 Makefile.
5069
5070 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
5071 command called pgrep for grepping around the active processes.
5072
5073 7. Added the beginnings of support for UTF-8 character strings.
5074
5075 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
5076 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
5077 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
5078 out for the ar command.)
5079
5080
5081 Version 3.2 12-May-00
5082 ---------------------
5083
5084 This is purely a bug fixing release.
5085
5086 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
5087 of ZA. This was just one example of several cases that could provoke this bug,
5088 which was introduced by change 9 of version 2.00. The code for breaking
5089 infinite loops after an iteration that matches an empty string was't working
5090 correctly.
5091
5092 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
5093 when matched against abbab (for example). After matching an empty string, it
5094 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
5095 caused it to match further down the string than it should.
5096
5097 3. The code contained an inclusion of sys/types.h. It isn't clear why this
5098 was there because it doesn't seem to be needed, and it causes trouble on some
5099 systems, as it is not a Standard C header. It has been removed.
5100
5101 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
5102 were reported on the Macintosh. The changes were from
5103
5104 while ((c = *(++ptr)) != 0 && c != '\n');
5105 to
5106 while ((c = *(++ptr)) != 0 && c != '\n') ;
5107
5108 Totally extraordinary, but if that's what it takes...
5109
5110 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
5111 available. Added HAVE_BCOPY and an autoconf test for it; if neither
5112 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
5113 assumes the way PCRE uses memmove() (always moving upwards).
5114
5115 6. PCRE is being used in one environment where strchr() is not available. There
5116 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
5117 faster code anyway.
5118
5119
5120 Version 3.1 09-Feb-00
5121 ---------------------
5122
5123 The only change in this release is the fixing of some bugs in Makefile.in for
5124 the "install" target:
5125
5126 (1) It was failing to install pcreposix.h.
5127
5128 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
5129
5130
5131 Version 3.0 01-Feb-00
5132 ---------------------
5133
5134 1. Add support for the /+ modifier to perltest (to output $` like it does in
5135 pcretest).
5136
5137 2. Add support for the /g modifier to perltest.
5138
5139 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
5140 matches null strings.
5141
5142 4. Fix perltest so that it doesn't do unwanted things when fed an empty
5143 pattern. Perl treats empty patterns specially - it reuses the most recent
5144 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
5145 effect.
5146
5147 5. The POSIX interface was broken in that it was just handing over the POSIX
5148 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
5149 required a bigger vector, with some working space on the end. This means that
5150 the POSIX wrapper now has to get and free some memory, and copy the results.
5151
5152 6. Added some simple autoconf support, placing the test data and the
5153 documentation in separate directories, re-organizing some of the
5154 information files, and making it build pcre-config (a GNU standard). Also added
5155 libtool support for building PCRE as a shared library, which is now the
5156 default.
5157
5158 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
5159 09 are not valid octal constants. Single digits will be used for minor values
5160 less than 10.
5161
5162 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
5163 existing programs that set these in the POSIX interface can use PCRE without
5164 modification.
5165
5166 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
5167 return all that pcre_info() returns, plus additional data. The pcre_info()
5168 function is retained for compatibility, but is considered to be obsolete.
5169
5170 10. Added experimental recursion feature (?R) to handle one common case that
5171 Perl 5.6 will be able to do with (?p{...}).
5172
5173 11. Added support for POSIX character classes like [:alpha:], which Perl is
5174 adopting.
5175
5176
5177 Version 2.08 31-Aug-99
5178 ----------------------
5179
5180 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
5181 trying to match at the startoffset position, but instead was moving forward to
5182 the next newline as if a previous match had failed.
5183
5184 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
5185 and could get into a loop if a null string was matched other than at the start
5186 of the subject.
5187
5188 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
5189 be distinguished at compile time, and for completeness also added PCRE_DATE.
5190
5191 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
5192 in GnuWin32 environments.
5193
5194
5195 Version 2.07 29-Jul-99
5196 ----------------------
5197
5198 1. The documentation is now supplied in plain text form and HTML as well as in
5199 the form of man page sources.
5200
5201 2. C++ compilers don't like assigning (void *) values to other pointer types.
5202 In particular this affects malloc(). Although there is no problem in Standard
5203 C, I've put in casts to keep C++ compilers happy.
5204
5205 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
5206 should be (const char *).
5207
5208 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
5209 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
5210 However, I haven't made this a standard facility. The documentation doesn't
5211 mention it, and the Makefile doesn't support it.
5212
5213 5. The Makefile now contains an "install" target, with editable destinations at
5214 the top of the file. The pcretest program is not installed.
5215
5216 6. pgrep -V now gives the PCRE version number and date.
5217
5218 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
5219 causing the entire string to be ignored, instead of just the last character.
5220
5221 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
5222 non-matching string, it can take a very, very long time, even for strings of
5223 quite modest length, because of the nested recursion. PCRE now does better in
5224 some of these cases. It does this by remembering the last required literal
5225 character in the pattern, and pre-searching the subject to ensure it is present
5226 before running the real match. In other words, it applies a heuristic to detect
5227 some types of certain failure quickly, and in the above example, if presented
5228 with a string that has no trailing " it gives "no match" very quickly.
5229
5230 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
5231 other alternatives are tried instead.
5232
5233
5234 Version 2.06 09-Jun-99
5235 ----------------------
5236
5237 1. Change pcretest's output for amount of store used to show just the code
5238 space, because the remainder (the data block) varies in size between 32-bit and
5239 64-bit systems.
5240
5241 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
5242 start matching at. This allows lookbehinds to work when searching for multiple
5243 occurrences in a string.
5244
5245 3. Added additional options to pcretest for testing multiple occurrences:
5246
5247 /+ outputs the rest of the string that follows a match
5248 /g loops for multiple occurrences, using the new startoffset argument
5249 /G loops for multiple occurrences by passing an incremented pointer
5250
5251 4. PCRE wasn't doing the "first character" optimization for patterns starting
5252 with \b or \B, though it was doing it for other lookbehind assertions. That is,
5253 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
5254 the letter 'x'. On long subject strings, this gives a significant speed-up.
5255
5256
5257 Version 2.05 21-Apr-99
5258 ----------------------
5259
5260 1. Changed the type of magic_number from int to long int so that it works
5261 properly on 16-bit systems.
5262
5263 2. Fixed a bug which caused patterns starting with .* not to work correctly
5264 when the subject string contained newline characters. PCRE was assuming
5265 anchoring for such patterns in all cases, which is not correct because .* will
5266 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
5267 DOTALL is set at top level; otherwise it knows that patterns starting with .*
5268 must be retried after every newline in the subject.
5269
5270
5271 Version 2.04 18-Feb-99
5272 ----------------------
5273
5274 1. For parenthesized subpatterns with repeats whose minimum was zero, the
5275 computation of the store needed to hold the pattern was incorrect (too large).
5276 If such patterns were nested a few deep, this could multiply and become a real
5277 problem.
5278
5279 2. Added /M option to pcretest to show the memory requirement of a specific
5280 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
5281
5282 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
5283 compiled in such a way that the backtracking after subsequent failure was
5284 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
5285 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
5286
5287
5288 Version 2.03 02-Feb-99
5289 ----------------------
5290
5291 1. Fixed typo and small mistake in man page.
5292
5293 2. Added 4th condition (GPL supersedes if conflict) and created separate
5294 LICENCE file containing the conditions.
5295
5296 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
5297 Perl, that is the internal \ allows the delimiter to be included in the
5298 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
5299 the final delimiter, add \ to the end of the pattern (to test the error).
5300
5301 4. Added the convenience functions for extracting substrings after a successful
5302 match. Updated pcretest to make it able to test these functions.
5303
5304
5305 Version 2.02 14-Jan-99
5306 ----------------------
5307
5308 1. Initialized the working variables associated with each extraction so that
5309 their saving and restoring doesn't refer to uninitialized store.
5310
5311 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
5312 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
5313 fix the problem.
5314
5315 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
5316 calls, and wasn't printing the correct value for compiling calls. Increased the
5317 default value of LOOPREPEAT, and the number of significant figures in the
5318 times.
5319
5320 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
5321
5322 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
5323 a building problem on Windows NT with a FAT file system.
5324
5325
5326 Version 2.01 21-Oct-98
5327 ----------------------
5328
5329 1. Changed the API for pcre_compile() to allow for the provision of a pointer
5330 to character tables built by pcre_maketables() in the current locale. If NULL
5331 is passed, the default tables are used.
5332
5333
5334 Version 2.00 24-Sep-98
5335 ----------------------
5336
5337 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
5338 it any more.
5339
5340 2. Allow quantification of (?>) groups, and make it work correctly.
5341
5342 3. The first character computation wasn't working for (?>) groups.
5343
5344 4. Correct the implementation of \Z (it is permitted to match on the \n at the
5345 end of the subject) and add 5.005's \z, which really does match only at the
5346 very end of the subject.
5347
5348 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
5349
5350 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
5351 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
5352 localized options. All options to pcre_study() were also removed.
5353
5354 7. Add other new features from 5.005:
5355
5356 $(?<= positive lookbehind
5357 $(?<! negative lookbehind
5358 (?imsx-imsx) added the unsetting capability
5359 such a setting is global if at outer level; local otherwise
5360 (?imsx-imsx:) non-capturing groups with option setting
5361 (?(cond)re|re) conditional pattern matching
5362
5363 A backreference to itself in a repeated group matches the previous
5364 captured string.
5365
5366 8. General tidying up of studying (both automatic and via "study")
5367 consequential on the addition of new assertions.
5368
5369 9. As in 5.005, unlimited repeated groups that could match an empty substring
5370 are no longer faulted at compile time. Instead, the loop is forcibly broken at
5371 runtime if any iteration does actually match an empty substring.
5372
5373 10. Include the RunTest script in the distribution.
5374
5375 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
5376 discrepancies, some of which were old and were also with respect to 5.004. They
5377 have now been fixed.
5378
5379
5380 Version 1.09 28-Apr-98
5381 ----------------------
5382
5383 1. A negated single character class followed by a quantifier with a minimum
5384 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
5385 program crashes, or just wrong answers. This did not apply to negated classes
5386 containing more than one character, or to minima other than one.
5387
5388
5389 Version 1.08 27-Mar-98
5390 ----------------------
5391
5392 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
5393
5394 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
5395 latter must appear before anything that relies on it in the pattern.
5396
5397
5398 Version 1.07 16-Feb-98
5399 ----------------------
5400
5401 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
5402 repeat of a potentially empty string).
5403
5404
5405 Version 1.06 23-Jan-98
5406 ----------------------
5407
5408 1. Added Markus Oberhumer's little patches for C++.
5409
5410 2. Literal strings longer than 255 characters were broken.
5411
5412
5413 Version 1.05 23-Dec-97
5414 ----------------------
5415
5416 1. Negated character classes containing more than one character were failing if
5417 PCRE_CASELESS was set at run time.
5418
5419
5420 Version 1.04 19-Dec-97
5421 ----------------------
5422
5423 1. Corrected the man page, where some "const" qualifiers had been omitted.
5424
5425 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
5426 input syntax.
5427
5428 3. Fixed memory leak which occurred when a regex with back references was
5429 matched with an offsets vector that wasn't big enough. The temporary memory
5430 that is used in this case wasn't being freed if the match failed.
5431
5432 4. Tidied pcretest to ensure it frees memory that it gets.
5433
5434 5. Temporary memory was being obtained in the case where the passed offsets
5435 vector was exactly big enough.
5436
5437 6. Corrected definition of offsetof() from change 5 below.
5438
5439 7. I had screwed up change 6 below and broken the rules for the use of
5440 setjmp(). Now fixed.
5441
5442
5443 Version 1.03 18-Dec-97
5444 ----------------------
5445
5446 1. A erroneous regex with a missing opening parenthesis was correctly
5447 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
5448 on some systems.
5449
5450 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
5451 it was reported that one broken compiler failed on the former because "code" is
5452 also an independent variable.
5453
5454 3. The erroneous regex a[]b caused an array overrun reference.
5455
5456 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
5457 fail on data ending with that character. (It was going on too far, and checking
5458 the next character, typically a binary zero.) This was specific to the
5459 optimized code for single-character negative classes.
5460
5461 5. Added a contributed patch from the TIN world which does the following:
5462
5463 + Add an undef for memmove, in case the the system defines a macro for it.
5464
5465 + Add a definition of offsetof(), in case there isn't one. (I don't know
5466 the reason behind this - offsetof() is part of the ANSI standard - but
5467 it does no harm).
5468
5469 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
5470 most of the places where whitespace preceded '#'. I have given up and
5471 allowed the remaining 2 cases to be at the margin.
5472
5473 + Rename some variables in pcre to eliminate shadowing. This seems very
5474 pedantic, but does no harm, of course.
5475
5476 6. Moved the call to setjmp() into its own function, to get rid of warnings
5477 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
5478
5479 7. Constructs such as \d{8,} were compiling into the equivalent of
5480 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
5481 outcome, but in this particular case used more store than had been allocated,
5482 which caused the bug to be discovered because it threw up an internal error.
5483
5484 8. The debugging code in both pcre and pcretest for outputting the compiled
5485 form of a regex was going wrong in the case of back references followed by
5486 curly-bracketed repeats.
5487
5488
5489 Version 1.02 12-Dec-97
5490 ----------------------
5491
5492 1. Typos in pcre.3 and comments in the source fixed.
5493
5494 2. Applied a contributed patch to get rid of places where it used to remove
5495 'const' from variables, and fixed some signed/unsigned and uninitialized
5496 variable warnings.
5497
5498 3. Added the "runtest" target to Makefile.
5499
5500 4. Set default compiler flag to -O2 rather than just -O.
5501
5502
5503 Version 1.01 19-Nov-97
5504 ----------------------
5505
5506 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
5507 like /([ab]*)*/, that is, for classes with more than one character in them.
5508
5509 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
5510 as /((?>a*))*/ (a PCRE_EXTRA facility).
5511
5512
5513 Version 1.00 18-Nov-97
5514 ----------------------
5515
5516 1. Added compile-time macros to support systems such as SunOS4 which don't have
5517 memmove() or strerror() but have other things that can be used instead.
5518
5519 2. Arranged that "make clean" removes the executables.
5520
5521
5522 Version 0.99 27-Oct-97
5523 ----------------------
5524
5525 1. Fixed bug in code for optimizing classes with only one character. It was
5526 initializing a 32-byte map regardless, which could cause it to run off the end
5527 of the memory it had got.
5528
5529 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
5530
5531
5532 Version 0.98 22-Oct-97
5533 ----------------------
5534
5535 1. Fixed bug in code for handling temporary memory usage when there are more
5536 back references than supplied space in the ovector. This could cause segfaults.
5537
5538
5539 Version 0.97 21-Oct-97
5540 ----------------------
5541
5542 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
5543
5544 2. Optimized negated single characters not to use a bit map.
5545
5546 3. Brought error texts together as macro definitions; clarified some of them;
5547 fixed one that was wrong - it said "range out of order" when it meant "invalid
5548 escape sequence".
5549
5550 4. Changed some char * arguments to const char *.
5551
5552 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
5553
5554 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
5555 pcretest.
5556
5557
5558 Version 0.96 16-Oct-97
5559 ----------------------
5560
5561 1. Added a simple "pgrep" utility to the distribution.
5562
5563 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
5564 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
5565 where "ddd" means "one or more decimal digits".
5566
5567 3. Fixed serious bug. If a pattern had a back reference, but the call to
5568 pcre_exec() didn't supply a large enough ovector to record the related
5569 identifying subpattern, the match always failed. PCRE now remembers the number
5570 of the largest back reference, and gets some temporary memory in which to save
5571 the offsets during matching if necessary, in order to ensure that
5572 backreferences always work.
5573
5574 4. Increased the compatibility with Perl in a number of ways:
5575
5576 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
5577 to request this handling. The option can be set at compile or exec time.
5578
5579 (b) $ matches before a terminating newline by default; an option
5580 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
5581 mode). The option can be set at compile or exec time.
5582
5583 (c) The handling of \ followed by a digit other than 0 is now supposed to be
5584 the same as Perl's. If the decimal number it represents is less than 10
5585 or there aren't that many previous left capturing parentheses, an octal
5586 escape is read. Inside a character class, it's always an octal escape,
5587 even if it is a single digit.
5588
5589 (d) An escaped but undefined alphabetic character is taken as a literal,
5590 unless PCRE_EXTRA is set. Currently this just reserves the remaining
5591 escapes.
5592
5593 (e) {0} is now permitted. (The previous item is removed from the compiled
5594 pattern).
5595
5596 5. Changed all the names of code files so that the basic parts are no longer
5597 than 10 characters, and abolished the teeny "globals.c" file.
5598
5599 6. Changed the handling of character classes; they are now done with a 32-byte
5600 bit map always.
5601
5602 7. Added the -d and /D options to pcretest to make it possible to look at the
5603 internals of compilation without having to recompile pcre.
5604
5605
5606 Version 0.95 23-Sep-97
5607 ----------------------
5608
5609 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
5610 \x20 at the start of a run of normal characters. These were being treated as
5611 real characters, instead of the source characters being re-checked.
5612
5613
5614 Version 0.94 18-Sep-97
5615 ----------------------
5616
5617 1. The functions are now thread-safe, with the caveat that the global variables
5618 containing pointers to malloc() and free() or alternative functions are the
5619 same for all threads.
5620
5621 2. Get pcre_study() to generate a bitmap of initial characters for non-
5622 anchored patterns when this is possible, and use it if passed to pcre_exec().
5623
5624
5625 Version 0.93 15-Sep-97
5626 ----------------------
5627
5628 1. /(b)|(:+)/ was computing an incorrect first character.
5629
5630 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
5631 but not actually doing anything yet.
5632
5633 3. Treat "-" characters in classes that cannot be part of ranges as literals,
5634 as Perl does (e.g. [-az] or [az-]).
5635
5636 4. Set the anchored flag if a branch starts with .* or .*? because that tests
5637 all possible positions.
5638
5639 5. Split up into different modules to avoid including unneeded functions in a
5640 compiled binary. However, compile and exec are still in one module. The "study"
5641 function is split off.
5642
5643 6. The character tables are now in a separate module whose source is generated
5644 by an auxiliary program - but can then be edited by hand if required. There are
5645 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
5646 toupper() in the code.
5647
5648 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
5649 make them global. Abolish the function for setting them, as the caller can now
5650 set them directly.
5651
5652
5653 Version 0.92 11-Sep-97
5654 ----------------------
5655
5656 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
5657 (e.g. /a{1,3}/) was broken (I mis-optimized it).
5658
5659 2. Caseless matching was not working in character classes if the characters in
5660 the pattern were in upper case.
5661
5662 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
5663
5664 4. Make PCRE_ANCHORED public and accept as a compile option.
5665
5666 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
5667 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
5668 pass them.
5669
5670 6. Give an error if bad option bits passed at compile or run time.
5671
5672 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
5673 pcretest to cause it to pass that flag.
5674
5675 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
5676 options, and the first character, if set.
5677
5678 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
5679
5680
5681 Version 0.91 10-Sep-97
5682 ----------------------
5683
5684 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
5685 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
5686
5687 2. PCRE was looping on encountering an indefinitely repeated back reference to
5688 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
5689 Perl does - treats the match as successful.
5690
5691 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5