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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 792 - (show annotations) (download)
Wed Dec 7 16:44:48 2011 UTC (2 years, 9 months ago) by ph10
File size: 223167 byte(s)
Fix a possible uninitialized memory bug.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12