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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 775 - (show annotations) (download)
Thu Dec 1 10:35:30 2011 UTC (3 years ago) by ph10
File size: 222132 byte(s)
Fix pcretest not forgetting JIT after it has been used once.

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

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