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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 777 - (show annotations) (download)
Thu Dec 1 15:15:31 2011 UTC (2 years, 9 months ago) by zherczeg
File size: 222226 byte(s)
Retrieve executable code size support for the JIT compiler and fixing some warnings.
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 21. Retrieve executable code size support for the JIT compiler and fixing
91 some warnings.
92
93
94 Version 8.20 21-Oct-2011
95 ------------------------
96
97 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
98 a POSIX class. After further experiments with Perl, which convinced me that
99 Perl has bugs and confusions, a closing square bracket is no longer allowed
100 in a POSIX name. This bug also affected patterns with classes that started
101 with full stops.
102
103 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
104 captured substring, but while checking the failing first alternative,
105 substring 1 is temporarily captured. If the output vector supplied to
106 pcre_exec() was not big enough for this capture, the yield of the function
107 was still zero ("insufficient space for captured substrings"). This cannot
108 be totally fixed without adding another stack variable, which seems a lot
109 of expense for a edge case. However, I have improved the situation in cases
110 such as /(a)(b)x|abc/ matched against "abc", where the return code
111 indicates that fewer than the maximum number of slots in the ovector have
112 been set.
113
114 3. Related to (2) above: when there are more back references in a pattern than
115 slots in the output vector, pcre_exec() uses temporary memory during
116 matching, and copies in the captures as far as possible afterwards. It was
117 using the entire output vector, but this conflicts with the specification
118 that only 2/3 is used for passing back captured substrings. Now it uses
119 only the first 2/3, for compatibility. This is, of course, another edge
120 case.
121
122 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
123 main code base, and can be used by building with --enable-jit. When this is
124 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
125 runtime --no-jit option is given.
126
127 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
128 ovector, the return from the function was zero, implying that there were
129 other matches that did not fit. The correct "exactly full" value is now
130 returned.
131
132 6. If a subpattern that was called recursively or as a subroutine contained
133 (*PRUNE) or any other control that caused it to give a non-standard return,
134 invalid errors such as "Error -26 (nested recursion at the same subject
135 position)" or even infinite loops could occur.
136
137 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
138 computing the minimum length on reaching *ACCEPT, and so ended up with the
139 wrong value of 1 rather than 0. Further investigation indicates that
140 computing a minimum subject length in the presence of *ACCEPT is difficult
141 (think back references, subroutine calls), and so I have changed the code
142 so that no minimum is registered for a pattern that contains *ACCEPT.
143
144 8. If (*THEN) was present in the first (true) branch of a conditional group,
145 it was not handled as intended. [But see 16 below.]
146
147 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
148 Sheri Pierce.
149
150 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
151 the first byte in a match must be "a".
152
153 11. Change 17 for 8.13 increased the recursion depth for patterns like
154 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
155 pattern contains any instances of (*THEN). If it does not, the old
156 optimizations are restored. It would be nice to do this on a per-group
157 basis, but at the moment that is not feasible.
158
159 12. In some environments, the output of pcretest -C is CRLF terminated. This
160 broke RunTest's code that checks for the link size. A single white space
161 character after the value is now allowed for.
162
163 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
164 For "fr", it uses the Windows-specific input and output files.
165
166 14. If (*THEN) appeared in a group that was called recursively or as a
167 subroutine, it did not work as intended. [But see next item.]
168
169 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
170 pattern fragments (but not containing any | characters). If A and B are
171 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
172 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
173 D. In other words, Perl considers parentheses that do not contain any |
174 characters to be part of a surrounding alternative, whereas PCRE was
175 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
176 differently. PCRE now behaves in the same way as Perl, except in the case
177 of subroutine/recursion calls such as (?1) which have in any case always
178 been different (but PCRE had them first :-).
179
180 16. Related to 15 above: Perl does not treat the | in a conditional group as
181 creating alternatives. Such a group is treated in the same way as an
182 ordinary group without any | characters when processing (*THEN). PCRE has
183 been changed to match Perl's behaviour.
184
185 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
186 RunGrepTest script failed.
187
188 18. Change 22 for version 13 caused atomic groups to use more stack. This is
189 inevitable for groups that contain captures, but it can lead to a lot of
190 stack use in large patterns. The old behaviour has been restored for atomic
191 groups that do not contain any capturing parentheses.
192
193 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
194 suppress the check for a minimum subject length at run time. (If it was
195 given to pcre_exec() or pcre_dfa_exec() it did work.)
196
197 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
198 fail to work when decoding hex characters in data strings in EBCDIC
199 environments.
200
201 21. It appears that in at least one Mac OS environment, the isxdigit() function
202 is implemented as a macro that evaluates to its argument more than once,
203 contravening the C 90 Standard (I haven't checked a later standard). There
204 was an instance in pcretest which caused it to go wrong when processing
205 \x{...} escapes in subject strings. The has been rewritten to avoid using
206 things like p++ in the argument of isxdigit().
207
208
209 Version 8.13 16-Aug-2011
210 ------------------------
211
212 1. The Unicode data tables have been updated to Unicode 6.0.0.
213
214 2. Two minor typos in pcre_internal.h have been fixed.
215
216 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
217 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
218 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
219
220 4. There were a number of related bugs in the code for matching backrefences
221 caselessly in UTF-8 mode when codes for the characters concerned were
222 different numbers of bytes. For example, U+023A and U+2C65 are an upper
223 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
224 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
225 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
226 2-byte code at the end of the subject (it thought there wasn't enough data
227 left).
228
229 5. Comprehensive information about what went wrong is now returned by
230 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
231 as the output vector has at least 2 elements. The offset of the start of
232 the failing character and a reason code are placed in the vector.
233
234 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
235 now returned is for the first byte of the failing character, instead of the
236 last byte inspected. This is an incompatible change, but I hope it is small
237 enough not to be a problem. It makes the returned offset consistent with
238 pcre_exec() and pcre_dfa_exec().
239
240 7. pcretest now gives a text phrase as well as the error number when
241 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
242 failure, the offset and reason code are output.
243
244 8. When \R was used with a maximizing quantifier it failed to skip backwards
245 over a \r\n pair if the subsequent match failed. Instead, it just skipped
246 back over a single character (\n). This seems wrong (because it treated the
247 two characters as a single entity when going forwards), conflicts with the
248 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
249 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
250 has been changed.
251
252 9. Some internal refactoring has changed the processing so that the handling
253 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
254 time (the PCRE_DOTALL option was changed this way some time ago: version
255 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
256 which was always a bit of a fudge. It also means that there is one less
257 argument for the match() function, which reduces its stack requirements
258 slightly. This change also fixes an incompatibility with Perl: the pattern
259 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
260
261 10. More internal refactoring has drastically reduced the number of recursive
262 calls to match() for possessively repeated groups such as (abc)++ when
263 using pcre_exec().
264
265 11. While implementing 10, a number of bugs in the handling of groups were
266 discovered and fixed:
267
268 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
269 (a|)*(?1) gave a compile-time internal error.
270 ((a|)+)+ did not notice that the outer group could match an empty string.
271 (^a|^)+ was not marked as anchored.
272 (.*a|.*)+ was not marked as matching at start or after a newline.
273
274 12. Yet more internal refactoring has removed another argument from the match()
275 function. Special calls to this function are now indicated by setting a
276 value in a variable in the "match data" data block.
277
278 13. Be more explicit in pcre_study() instead of relying on "default" for
279 opcodes that mean there is no starting character; this means that when new
280 ones are added and accidentally left out of pcre_study(), testing should
281 pick them up.
282
283 14. The -s option of pcretest has been documented for ages as being an old
284 synonym of -m (show memory usage). I have changed it to mean "force study
285 for every regex", that is, assume /S for every regex. This is similar to -i
286 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
287 using it. It makes it easier to run collections of tests with and without
288 study enabled, and thereby test pcre_study() more easily. All the standard
289 tests are now run with and without -s (but some patterns can be marked as
290 "never study" - see 20 below).
291
292 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
293 restoration of the capturing data to the outer values was not happening
294 correctly.
295
296 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
297 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
298 pattern had matched an empty string, and so incorrectly returned a no
299 match.
300
301 17. There was optimizing code for the last branch of non-capturing parentheses,
302 and also for the obeyed branch of a conditional subexpression, which used
303 tail recursion to cut down on stack usage. Unfortunately, now that there is
304 the possibility of (*THEN) occurring in these branches, tail recursion is
305 no longer possible because the return has to be checked for (*THEN). These
306 two optimizations have therefore been removed. [But see 8.20/11 above.]
307
308 18. If a pattern containing \R was studied, it was assumed that \R always
309 matched two bytes, thus causing the minimum subject length to be
310 incorrectly computed because \R can also match just one byte.
311
312 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
313 was incorrectly computed.
314
315 20. If /S is present twice on a test pattern in pcretest input, it now
316 *disables* studying, thereby overriding the use of -s on the command line
317 (see 14 above). This is necessary for one or two tests to keep the output
318 identical in both cases.
319
320 21. When (*ACCEPT) was used in an assertion that matched an empty string and
321 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
322
323 22. When an atomic group that contained a capturing parenthesis was
324 successfully matched, but the branch in which it appeared failed, the
325 capturing was not being forgotten if a higher numbered group was later
326 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
327 group 1 to "a", when in fact it should be unset. This applied to multi-
328 branched capturing and non-capturing groups, repeated or not, and also to
329 positive assertions (capturing in negative assertions does not happen
330 in PCRE) and also to nested atomic groups.
331
332 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
333 subject after a captured substring, to make it easier to tell which of a
334 number of identical substrings has been captured.
335
336 24. The way atomic groups are processed by pcre_exec() has been changed so that
337 if they are repeated, backtracking one repetition now resets captured
338 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
339 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
340 "aaa". Previously, it would have been "a". As part of this code
341 refactoring, the way recursive calls are handled has also been changed.
342
343 25. If an assertion condition captured any substrings, they were not passed
344 back unless some other capturing happened later. For example, if
345 (?(?=(a))a) was matched against "a", no capturing was returned.
346
347 26. When studying a pattern that contained subroutine calls or assertions,
348 the code for finding the minimum length of a possible match was handling
349 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
350 group 1 called group 2 while simultaneously a separate group 2 called group
351 1). A stack overflow occurred in this case. I have fixed this by limiting
352 the recursion depth to 10.
353
354 27. Updated RunTest.bat in the distribution to the version supplied by Tom
355 Fortmann. This supports explicit test numbers on the command line, and has
356 argument validation and error reporting.
357
358 28. An instance of \X with an unlimited repeat could fail if at any point the
359 first character it looked at was a mark character.
360
361 29. Some minor code refactoring concerning Unicode properties and scripts
362 should reduce the stack requirement of match() slightly.
363
364 30. Added the '=' option to pcretest to check the setting of unused capturing
365 slots at the end of the pattern, which are documented as being -1, but are
366 not included in the return count.
367
368 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
369 compiled something random. Now it gives a compile-time error (as does
370 Perl).
371
372 32. A *MARK encountered during the processing of a positive assertion is now
373 recorded and passed back (compatible with Perl).
374
375 33. If --only-matching or --colour was set on a pcregrep call whose pattern
376 had alternative anchored branches, the search for a second match in a line
377 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
378 matched the line "0102" twice. The same bug affected patterns that started
379 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
380 twice.
381
382 34. Previously, PCRE did not allow quantification of assertions. However, Perl
383 does, and because of capturing effects, quantifying parenthesized
384 assertions may at times be useful. Quantifiers are now allowed for
385 parenthesized assertions.
386
387 35. A minor code tidy in pcre_compile() when checking options for \R usage.
388
389 36. \g was being checked for fancy things in a character class, when it should
390 just be a literal "g".
391
392 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
393 appearance of a nested POSIX class supersedes an apparent external class.
394 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
395 unescaped square brackets may also appear as part of class names. For
396 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
397 more like Perl. (But see 8.20/1 above.)
398
399 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
400 was because it thought it was \N{name}, which is not supported).
401
402 39. Add minix to OS list not supporting the -S option in pcretest.
403
404 40. PCRE tries to detect cases of infinite recursion at compile time, but it
405 cannot analyze patterns in sufficient detail to catch mutual recursions
406 such as ((?1))((?2)). There is now a runtime test that gives an error if a
407 subgroup is called recursively as a subpattern for a second time at the
408 same position in the subject string. In previous releases this might have
409 been caught by the recursion limit, or it might have run out of stack.
410
411 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
412 happen only once. PCRE was, however incorrectly giving a compile time error
413 "recursive call could loop indefinitely" because it cannot analyze the
414 pattern in sufficient detail. The compile time test no longer happens when
415 PCRE is compiling a conditional subpattern, but actual runaway loops are
416 now caught at runtime (see 40 above).
417
418 42. It seems that Perl allows any characters other than a closing parenthesis
419 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
420 has been changed to be the same.
421
422 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
423 as not to get warnings when autogen.sh is called. Also changed
424 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
425
426 44. To help people who use pcregrep to scan files containing exceedingly long
427 lines, the following changes have been made:
428
429 (a) The default value of the buffer size parameter has been increased from
430 8K to 20K. (The actual buffer used is three times this size.)
431
432 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
433 PCRE is built.
434
435 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
436 to be set at run time.
437
438 (d) Numerical values in pcregrep options can be followed by K or M, for
439 example --buffer-size=50K.
440
441 (e) If a line being scanned overflows pcregrep's buffer, an error is now
442 given and the return code is set to 2.
443
444 45. Add a pointer to the latest mark to the callout data block.
445
446 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
447 partial match of an empty string instead of no match. This was specific to
448 the use of ".".
449
450 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
451 complete match instead of a partial match. This bug was dependent on both
452 the PCRE_UTF8 and PCRE_DOTALL options being set.
453
454 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
455 starting byte set, because \b was not being ignored.
456
457
458 Version 8.12 15-Jan-2011
459 ------------------------
460
461 1. Fixed some typos in the markup of the man pages, and wrote a script that
462 checks for such things as part of the documentation building process.
463
464 2. On a big-endian 64-bit system, pcregrep did not correctly process the
465 --match-limit and --recursion-limit options (added for 8.11). In
466 particular, this made one of the standard tests fail. (The integer value
467 went into the wrong half of a long int.)
468
469 3. If the --colour option was given to pcregrep with -v (invert match), it
470 did strange things, either producing crazy output, or crashing. It should,
471 of course, ignore a request for colour when reporting lines that do not
472 match.
473
474 4. Another pcregrep bug caused similar problems if --colour was specified with
475 -M (multiline) and the pattern match finished with a line ending.
476
477 5. In pcregrep, when a pattern that ended with a literal newline sequence was
478 matched in multiline mode, the following line was shown as part of the
479 match. This seems wrong, so I have changed it.
480
481 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
482 the check for further matches in the same line (so they could be coloured)
483 to overrun the end of the current line. If another match was found, it was
484 incorrectly shown (and then shown again when found in the next line).
485
486 7. If pcregrep was compiled under Windows, there was a reference to the
487 function pcregrep_exit() before it was defined. I am assuming this was
488 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
489 reported by a user. I've moved the definition above the reference.
490
491
492 Version 8.11 10-Dec-2010
493 ------------------------
494
495 1. (*THEN) was not working properly if there were untried alternatives prior
496 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
497 backtracked to try for "b" instead of moving to the next alternative branch
498 at the same level (in this case, to look for "c"). The Perl documentation
499 is clear that when (*THEN) is backtracked onto, it goes to the "next
500 alternative in the innermost enclosing group".
501
502 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
503 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
504 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
505 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
506 (*THEN).
507
508 3. If \s appeared in a character class, it removed the VT character from
509 the class, even if it had been included by some previous item, for example
510 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
511 of \s, but is part of the POSIX "space" class.)
512
513 4. A partial match never returns an empty string (because you can always
514 match an empty string at the end of the subject); however the checking for
515 an empty string was starting at the "start of match" point. This has been
516 changed to the "earliest inspected character" point, because the returned
517 data for a partial match starts at this character. This means that, for
518 example, /(?<=abc)def/ gives a partial match for the subject "abc"
519 (previously it gave "no match").
520
521 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
522 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
523 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
524 has an implication that the given string is incomplete (because a partial
525 match is preferred over a full match). For this reason, these items now
526 give a partial match in this situation. [Aside: previously, the one case
527 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
528 match rather than a full match, which was wrong by the old rules, but is
529 now correct.]
530
531 6. There was a bug in the handling of #-introduced comments, recognized when
532 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
533 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
534 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
535 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
536 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
537 places in pcre_compile().
538
539 7. Related to (6) above, when pcre_compile() was skipping #-introduced
540 comments when looking ahead for named forward references to subpatterns,
541 the only newline sequence it recognized was NL. It now handles newlines
542 according to the set newline convention.
543
544 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
545 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
546 cater for a lack of strerror(). These oversights have been fixed.
547
548 9. Added --match-limit and --recursion-limit to pcregrep.
549
550 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
551
552 11. When the -o option was used, pcregrep was setting a return code of 1, even
553 when matches were found, and --line-buffered was not being honoured.
554
555 12. Added an optional parentheses number to the -o and --only-matching options
556 of pcregrep.
557
558 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
559 can match an empty string. The code to do it in pcretest and pcredemo
560 needed fixing:
561
562 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
563 only one byte after an empty string match just before CRLF (this case
564 just got forgotten; "any" and "anycrlf" were OK).
565
566 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
567 mode when an empty string match preceded an ASCII character followed by
568 a non-ASCII character. (The code for advancing by one character rather
569 than one byte was nonsense.)
570
571 (c) The pcredemo.c sample program did not have any code at all to handle
572 the cases when CRLF is a valid newline sequence.
573
574 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
575 as a starting offset was within the subject string. There is now a new
576 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
577 negative or greater than the length of the string. In order to test this,
578 pcretest is extended to allow the setting of negative starting offsets.
579
580 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
581 starting offset points to the beginning of a UTF-8 character was
582 unnecessarily clumsy. I tidied it up.
583
584 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
585 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
586
587 17. Nobody had reported that the --include_dir option, which was added in
588 release 7.7 should have been called --include-dir (hyphen, not underscore)
589 for compatibility with GNU grep. I have changed it to --include-dir, but
590 left --include_dir as an undocumented synonym, and the same for
591 --exclude-dir, though that is not available in GNU grep, at least as of
592 release 2.5.4.
593
594 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
595 characters from a string of bytes) have been redefined so as not to use
596 loops, in order to improve performance in some environments. At the same
597 time, I abstracted some of the common code into auxiliary macros to save
598 repetition (this should not affect the compiled code).
599
600 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
601 compile-time error is now given if \c is not followed by an ASCII
602 character, that is, a byte less than 128. (In EBCDIC mode, the code is
603 different, and any byte value is allowed.)
604
605 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
606 START_OPTIMIZE option, which is now allowed at compile time - but just
607 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
608 to pcregrep and other applications that have no direct access to PCRE
609 options. The new /Y option in pcretest sets this option when calling
610 pcre_compile().
611
612 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
613 back references. Groups containing recursive back references were forced to
614 be atomic by that change, but in the case of named groups, the amount of
615 memory required was incorrectly computed, leading to "Failed: internal
616 error: code overflow". This has been fixed.
617
618 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
619 pcretest.c, to avoid build problems in some Borland environments.
620
621
622 Version 8.10 25-Jun-2010
623 ------------------------
624
625 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
626 THEN.
627
628 2. (*ACCEPT) was not working when inside an atomic group.
629
630 3. Inside a character class, \B is treated as a literal by default, but
631 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
632 causes the error). The code is unchanged, but I tidied the documentation.
633
634 4. Inside a character class, PCRE always treated \R and \X as literals,
635 whereas Perl faults them if its -w option is set. I have changed PCRE so
636 that it faults them when PCRE_EXTRA is set.
637
638 5. Added support for \N, which always matches any character other than
639 newline. (It is the same as "." when PCRE_DOTALL is not set.)
640
641 6. When compiling pcregrep with newer versions of gcc which may have
642 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
643 declared with attribute warn_unused_result" were given. Just casting the
644 result to (void) does not stop the warnings; a more elaborate fudge is
645 needed. I've used a macro to implement this.
646
647 7. Minor change to pcretest.c to avoid a compiler warning.
648
649 8. Added four artifical Unicode properties to help with an option to make
650 \s etc use properties (see next item). The new properties are: Xan
651 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
652
653 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
654 use Unicode properties. (*UCP) at the start of a pattern can be used to set
655 this option. Modified pcretest to add /W to test this facility. Added
656 REG_UCP to make it available via the POSIX interface.
657
658 10. Added --line-buffered to pcregrep.
659
660 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
661 studied, and the match started with a letter with a code point greater than
662 127 whose first byte was different to the first byte of the other case of
663 the letter, the other case of this starting letter was not recognized
664 (#976).
665
666 12. If a pattern that was studied started with a repeated Unicode property
667 test, for example, \p{Nd}+, there was the theoretical possibility of
668 setting up an incorrect bitmap of starting bytes, but fortunately it could
669 not have actually happened in practice until change 8 above was made (it
670 added property types that matched character-matching opcodes).
671
672 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
673 possible starting bytes for non-anchored patterns.
674
675 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
676 \R, and also a number of cases that involve Unicode properties, both
677 explicit and implicit when PCRE_UCP is set.
678
679 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
680 input, it could crash or give wrong results if characters with values
681 greater than 0xc0 were present in the subject string. (Detail: it assumed
682 UTF-8 input when processing these items.)
683
684 16. Added a lot of (int) casts to avoid compiler warnings in systems where
685 size_t is 64-bit (#991).
686
687 17. Added a check for running out of memory when PCRE is compiled with
688 --disable-stack-for-recursion (#990).
689
690 18. If the last data line in a file for pcretest does not have a newline on
691 the end, a newline was missing in the output.
692
693 19. The default pcre_chartables.c file recognizes only ASCII characters (values
694 less than 128) in its various bitmaps. However, there is a facility for
695 generating tables according to the current locale when PCRE is compiled. It
696 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
697 characters, are recognized by isspace() and therefore were getting set in
698 these tables, and indeed these tables seem to approximate to ISO 8859. This
699 caused a problem in UTF-8 mode when pcre_study() was used to create a list
700 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
701 which of course cannot start UTF-8 characters. I have changed the code so
702 that only real ASCII characters (less than 128) and the correct starting
703 bytes for UTF-8 encodings are set for characters greater than 127 when in
704 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
705 altogether.)
706
707 20. Added the /T option to pcretest so as to be able to run tests with non-
708 standard character tables, thus making it possible to include the tests
709 used for 19 above in the standard set of tests.
710
711 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
712 reference to a subpattern the other side of a comment that contains an
713 opening parenthesis caused either an internal compiling error, or a
714 reference to the wrong subpattern.
715
716
717 Version 8.02 19-Mar-2010
718 ------------------------
719
720 1. The Unicode data tables have been updated to Unicode 5.2.0.
721
722 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
723 configured.
724
725 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
726 original author of that file, following a query about its status.
727
728 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
729 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
730
731 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
732 quantifier applied to a forward-referencing subroutine call, could compile
733 incorrect code or give the error "internal error: previously-checked
734 referenced subpattern not found".
735
736 6. Both MS Visual Studio and Symbian OS have problems with initializing
737 variables to point to external functions. For these systems, therefore,
738 pcre_malloc etc. are now initialized to local functions that call the
739 relevant global functions.
740
741 7. There were two entries missing in the vectors called coptable and poptable
742 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
743 I've fixed the data, and added a kludgy way of testing at compile time that
744 the lengths are correct (equal to the number of opcodes).
745
746 8. Following on from 7, I added a similar kludge to check the length of the
747 eint vector in pcreposix.c.
748
749 9. Error texts for pcre_compile() are held as one long string to avoid too
750 much relocation at load time. To find a text, the string is searched,
751 counting zeros. There was no check for running off the end of the string,
752 which could happen if a new error number was added without updating the
753 string.
754
755 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
756
757 11. \K was not working if it appeared in an atomic group or in a group that
758 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
759 \K is "not well defined" if used in an assertion. PCRE now accepts it if
760 the assertion is positive, but not if it is negative.
761
762 12. Change 11 fortuitously reduced the size of the stack frame used in the
763 "match()" function of pcre_exec.c by one pointer. Forthcoming
764 implementation of support for (*MARK) will need an extra pointer on the
765 stack; I have reserved it now, so that the stack frame size does not
766 decrease.
767
768 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
769 item in branch that calls a recursion is a subroutine call - as in the
770 second branch in the above example - was incorrectly given the compile-
771 time error "recursive call could loop indefinitely" because pcre_compile()
772 was not correctly checking the subroutine for matching a non-empty string.
773
774 14. The checks for overrunning compiling workspace could trigger after an
775 overrun had occurred. This is a "should never occur" error, but it can be
776 triggered by pathological patterns such as hundreds of nested parentheses.
777 The checks now trigger 100 bytes before the end of the workspace.
778
779 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
780
781
782 Version 8.01 19-Jan-2010
783 ------------------------
784
785 1. If a pattern contained a conditional subpattern with only one branch (in
786 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
787 computed the wrong minimum data length (which is of course zero for such
788 subpatterns). This could cause incorrect "no match" results.
789
790 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
791 the pattern is reset in the first branch, pcre_compile() failed with
792 "internal error: code overflow at offset...". This happened only when
793 the reset was to the original external option setting. (An optimization
794 abstracts leading options settings into an external setting, which was the
795 cause of this.)
796
797 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
798 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
799 assertion pattern did not match (meaning that the assertion was true), it
800 was incorrectly treated as false if the SKIP had been reached during the
801 matching. This also applied to assertions used as conditions.
802
803 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
804 assertion subpattern, including such a pattern used as a condition,
805 unpredictable results occurred, instead of the error return
806 PCRE_ERROR_DFA_UITEM.
807
808 5. The C++ GlobalReplace function was not working like Perl for the special
809 situation when an empty string is matched. It now does the fancy magic
810 stuff that is necessary.
811
812 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
813 removed. (These were left over from very, very early versions of PCRE.)
814
815 7. Some cosmetic changes to the code to make life easier when compiling it
816 as part of something else:
817
818 (a) Change DEBUG to PCRE_DEBUG.
819
820 (b) In pcre_compile(), rename the member of the "branch_chain" structure
821 called "current" as "current_branch", to prevent a collision with the
822 Linux macro when compiled as a kernel module.
823
824 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
825 prevent a collision with the Linux macro when compiled as a kernel
826 module.
827
828 8. In pcre_compile() there are some checks for integer overflows that used to
829 cast potentially large values to (double). This has been changed to that
830 when building, a check for int64_t is made, and if it is found, it is used
831 instead, thus avoiding the use of floating point arithmetic. (There is no
832 other use of FP in PCRE.) If int64_t is not found, the fallback is to
833 double.
834
835 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
836 2005 (difference between two addresses compared to an unsigned value).
837
838 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
839 custom one, because of the following reported problem in Windows:
840
841 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
842 under Win32.
843 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
844 therefore missing the function definition.
845 - The compiler thus generates a "C" signature for the test function.
846 - The linker fails to find the "C" function.
847 - PCRE fails to configure if asked to do so against libbz2.
848
849 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
850 messages were output:
851
852 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
853 rerunning libtoolize, to keep the correct libtool macros in-tree.
854 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
855
856 I have done both of these things.
857
858 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
859 most of the time, it *can* run out if it is given a pattern that contains a
860 runaway infinite recursion. I updated the discussion in the pcrestack man
861 page.
862
863 13. Now that we have gone to the x.xx style of version numbers, the minor
864 version may start with zero. Using 08 or 09 is a bad idea because users
865 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
866 interpreted as invalid octal numbers. I've updated the previous comment in
867 configure.ac, and also added a check that gives an error if 08 or 09 are
868 used.
869
870 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
871 causing partial matching to fail when the end of the subject matched \W
872 in a UTF-8 pattern where \W was quantified with a minimum of 3.
873
874 15. There were some discrepancies between the declarations in pcre_internal.h
875 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
876 their definitions. The declarations used "const uschar *" and the
877 definitions used USPTR. Even though USPTR is normally defined as "const
878 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
879 reported that: "This difference in casting confuses some C++ compilers, for
880 example, SunCC recognizes above declarations as different functions and
881 generates broken code for hbpcre." I have changed the declarations to use
882 USPTR.
883
884 16. GNU libtool is named differently on some systems. The autogen.sh script now
885 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
886 (FreeBSD).
887
888 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
889 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
890 comment: "Figure out how to create a longlong from a string: strtoll and
891 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
892 instance, but it only takes 2 args instead of 3!"
893
894 18. A subtle bug concerned with back references has been fixed by a change of
895 specification, with a corresponding code fix. A pattern such as
896 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
897 refers, was giving matches when it shouldn't. For example, xa=xaaa would
898 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
899 same bug. Such groups have to be quantified to be useful, or contained
900 inside another quantified group. (If there's no repetition, the reference
901 can never match.) The problem arises because, having left the group and
902 moved on to the rest of the pattern, a later failure that backtracks into
903 the group uses the captured value from the final iteration of the group
904 rather than the correct earlier one. I have fixed this in PCRE by forcing
905 any group that contains a reference to itself to be an atomic group; that
906 is, there cannot be any backtracking into it once it has completed. This is
907 similar to recursive and subroutine calls.
908
909
910 Version 8.00 19-Oct-09
911 ----------------------
912
913 1. The table for translating pcre_compile() error codes into POSIX error codes
914 was out-of-date, and there was no check on the pcre_compile() error code
915 being within the table. This could lead to an OK return being given in
916 error.
917
918 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
919 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
920 in a Windows environment.
921
922 3. The pcregrep --count option prints the count for each file even when it is
923 zero, as does GNU grep. However, pcregrep was also printing all files when
924 --files-with-matches was added. Now, when both options are given, it prints
925 counts only for those files that have at least one match. (GNU grep just
926 prints the file name in this circumstance, but including the count seems
927 more useful - otherwise, why use --count?) Also ensured that the
928 combination -clh just lists non-zero counts, with no names.
929
930 4. The long form of the pcregrep -F option was incorrectly implemented as
931 --fixed_strings instead of --fixed-strings. This is an incompatible change,
932 but it seems right to fix it, and I didn't think it was worth preserving
933 the old behaviour.
934
935 5. The command line items --regex=pattern and --regexp=pattern were not
936 recognized by pcregrep, which required --regex pattern or --regexp pattern
937 (with a space rather than an '='). The man page documented the '=' forms,
938 which are compatible with GNU grep; these now work.
939
940 6. No libpcreposix.pc file was created for pkg-config; there was just
941 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
942
943 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
944 when UCP support is not needed, by modifying the Python script that
945 generates it from Unicode data files. This should not matter if the module
946 is correctly used as a library, but I received one complaint about 50K of
947 unwanted data. My guess is that the person linked everything into his
948 program rather than using a library. Anyway, it does no harm.
949
950 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
951 was a minimum greater than 1 for a wide character in a possessive
952 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
953 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
954 character. Chaos in the form of incorrect output or a compiling loop could
955 result.
956
957 9. The restrictions on what a pattern can contain when partial matching is
958 requested for pcre_exec() have been removed. All patterns can now be
959 partially matched by this function. In addition, if there are at least two
960 slots in the offset vector, the offset of the earliest inspected character
961 for the match and the offset of the end of the subject are set in them when
962 PCRE_ERROR_PARTIAL is returned.
963
964 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
965 synonymous with PCRE_PARTIAL, for backwards compatibility, and
966 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
967 and may be more useful for multi-segment matching.
968
969 11. Partial matching with pcre_exec() is now more intuitive. A partial match
970 used to be given if ever the end of the subject was reached; now it is
971 given only if matching could not proceed because another character was
972 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
973 string "Z", which now yields "no match" instead of "partial match". In the
974 case of pcre_dfa_exec(), "no match" is given if every matching path for the
975 final character ended with (*FAIL).
976
977 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
978 if the pattern had a "must contain" character that was already found in the
979 earlier partial match, unless partial matching was again requested. For
980 example, with the pattern /dog.(body)?/, the "must contain" character is
981 "g". If the first part-match was for the string "dog", restarting with
982 "sbody" failed. This bug has been fixed.
983
984 13. The string returned by pcre_dfa_exec() after a partial match has been
985 changed so that it starts at the first inspected character rather than the
986 first character of the match. This makes a difference only if the pattern
987 starts with a lookbehind assertion or \b or \B (\K is not supported by
988 pcre_dfa_exec()). It's an incompatible change, but it makes the two
989 matching functions compatible, and I think it's the right thing to do.
990
991 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
992 so that the demonstration program is easily available in environments where
993 PCRE has not been installed from source.
994
995 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
996 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
997 library.
998
999 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1000 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1001 is not the first non-POSIX option to be added. Clearly some people find
1002 these options useful.
1003
1004 17. If a caller to the POSIX matching function regexec() passes a non-zero
1005 value for nmatch with a NULL value for pmatch, the value of
1006 nmatch is forced to zero.
1007
1008 18. RunGrepTest did not have a test for the availability of the -u option of
1009 the diff command, as RunTest does. It now checks in the same way as
1010 RunTest, and also checks for the -b option.
1011
1012 19. If an odd number of negated classes containing just a single character
1013 interposed, within parentheses, between a forward reference to a named
1014 subpattern and the definition of the subpattern, compilation crashed with
1015 an internal error, complaining that it could not find the referenced
1016 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1017 [The bug was that it was starting one character too far in when skipping
1018 over the character class, thus treating the ] as data rather than
1019 terminating the class. This meant it could skip too much.]
1020
1021 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1022 /g option in pcretest when the pattern contains \K, which makes it possible
1023 to have an empty string match not at the start, even when the pattern is
1024 anchored. Updated pcretest and pcredemo to use this option.
1025
1026 21. If the maximum number of capturing subpatterns in a recursion was greater
1027 than the maximum at the outer level, the higher number was returned, but
1028 with unset values at the outer level. The correct (outer level) value is
1029 now given.
1030
1031 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1032 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1033 make it do so. The string so far is captured, making this feature
1034 compatible with Perl.
1035
1036 23. The tests have been re-organized, adding tests 11 and 12, to make it
1037 possible to check the Perl 5.10 features against Perl 5.10.
1038
1039 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1040 pattern matches a fixed length string. PCRE did not allow this; now it
1041 does. Neither allows recursion.
1042
1043 25. I finally figured out how to implement a request to provide the minimum
1044 length of subject string that was needed in order to match a given pattern.
1045 (It was back references and recursion that I had previously got hung up
1046 on.) This code has now been added to pcre_study(); it finds a lower bound
1047 to the length of subject needed. It is not necessarily the greatest lower
1048 bound, but using it to avoid searching strings that are too short does give
1049 some useful speed-ups. The value is available to calling programs via
1050 pcre_fullinfo().
1051
1052 26. While implementing 25, I discovered to my embarrassment that pcretest had
1053 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1054 study optimizations had never been tested with that matching function.
1055 Oops. What is worse, even when it was passed study data, there was a bug in
1056 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1057 were also very few tests of studied patterns with pcre_dfa_exec().
1058
1059 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1060 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1061 on the other side of the coin, they are no longer allowed to have different
1062 names, because these cannot be distinguished in PCRE, and this has caused
1063 confusion. (This is a difference from Perl.)
1064
1065 28. When duplicate subpattern names are present (necessarily with different
1066 numbers, as required by 27 above), and a test is made by name in a
1067 conditional pattern, either for a subpattern having been matched, or for
1068 recursion in such a pattern, all the associated numbered subpatterns are
1069 tested, and the overall condition is true if the condition is true for any
1070 one of them. This is the way Perl works, and is also more like the way
1071 testing by number works.
1072
1073
1074 Version 7.9 11-Apr-09
1075 ---------------------
1076
1077 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1078 (pcretest), all targets were linked against these libraries. This included
1079 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1080 libraries. This caused unwanted dependencies to be created. This problem
1081 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1082 pcretest is linked with readline.
1083
1084 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1085 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1086 moved outside it again, because FALSE and TRUE are already defined in AIX,
1087 but BOOL is not.
1088
1089 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1090 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1091
1092 4. The pcregrep documentation said spaces were inserted as well as colons (or
1093 hyphens) following file names and line numbers when outputting matching
1094 lines. This is not true; no spaces are inserted. I have also clarified the
1095 wording for the --colour (or --color) option.
1096
1097 5. In pcregrep, when --colour was used with -o, the list of matching strings
1098 was not coloured; this is different to GNU grep, so I have changed it to be
1099 the same.
1100
1101 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1102 each matching line was coloured. Now it goes on to look for further matches
1103 of any of the test patterns, which is the same behaviour as GNU grep.
1104
1105 7. A pattern that could match an empty string could cause pcregrep to loop; it
1106 doesn't make sense to accept an empty string match in pcregrep, so I have
1107 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1108 seems to be how GNU grep behaves.
1109
1110 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1111 start or after a newline", because the conditional assertion was not being
1112 correctly handled. The rule now is that both the assertion and what follows
1113 in the first alternative must satisfy the test.
1114
1115 9. If auto-callout was enabled in a pattern with a conditional group whose
1116 condition was an assertion, PCRE could crash during matching, both with
1117 pcre_exec() and pcre_dfa_exec().
1118
1119 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1120 used for matching.
1121
1122 11. Unicode property support in character classes was not working for
1123 characters (bytes) greater than 127 when not in UTF-8 mode.
1124
1125 12. Added the -M command line option to pcretest.
1126
1127 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1128
1129 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1130
1131 16. Added comments and documentation about mis-use of no_arg in the C++
1132 wrapper.
1133
1134 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1135 from Martin Jerabek that uses macro names for all relevant character and
1136 string constants.
1137
1138 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1139 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1140 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1141 these, but not everybody uses configure.
1142
1143 19. A conditional group that had only one branch was not being correctly
1144 recognized as an item that could match an empty string. This meant that an
1145 enclosing group might also not be so recognized, causing infinite looping
1146 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1147 with the subject "ab", where knowledge that the repeated group can match
1148 nothing is needed in order to break the loop.
1149
1150 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1151 exec(), but without supplying a callout function, matching went wrong.
1152
1153 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1154 leak if the size of the offset vector was greater than 30. When the vector
1155 is smaller, the saved offsets during recursion go onto a local stack
1156 vector, but for larger vectors malloc() is used. It was failing to free
1157 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1158 error, in fact).
1159
1160 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1161 heapframe that is used only when UTF-8 support is enabled. This caused no
1162 problem, but was untidy.
1163
1164 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1165 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1166 included within another project.
1167
1168 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1169 slightly modified by me:
1170
1171 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1172 not building pcregrep.
1173
1174 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1175 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1176
1177 25. Forward references, both numeric and by name, in patterns that made use of
1178 duplicate group numbers, could behave incorrectly or give incorrect errors,
1179 because when scanning forward to find the reference group, PCRE was not
1180 taking into account the duplicate group numbers. A pattern such as
1181 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1182
1183 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1184 the feature of a custom pointer more persuasive (as requested by a user).
1185
1186 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1187 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1188 pre-defined.
1189
1190 28. Added support for (*UTF8) at the start of a pattern.
1191
1192 29. Arrange for flags added by the "release type" setting in CMake to be shown
1193 in the configuration summary.
1194
1195
1196 Version 7.8 05-Sep-08
1197 ---------------------
1198
1199 1. Replaced UCP searching code with optimized version as implemented for Ad
1200 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1201 stage table and inline lookup instead of a function, giving speed ups of 2
1202 to 5 times on some simple patterns that I tested. Permission was given to
1203 distribute the MultiStage2.py script that generates the tables (it's not in
1204 the tarball, but is in the Subversion repository).
1205
1206 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1207 scripts.
1208
1209 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1210 a group with a zero qualifier. The result of the study could be incorrect,
1211 or the function might crash, depending on the pattern.
1212
1213 4. Caseless matching was not working for non-ASCII characters in back
1214 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1215 It now works when Unicode Property Support is available.
1216
1217 5. In pcretest, an escape such as \x{de} in the data was always generating
1218 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1219 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1220 truncation.
1221
1222 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1223
1224 7. Added two (int) casts to pcregrep when printing the difference of two
1225 pointers, in case they are 64-bit values.
1226
1227 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1228 test 2 if it fails.
1229
1230 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1231 and a #define of that name to empty if it is not externally set. This is to
1232 allow users of MSVC to set it if necessary.
1233
1234 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1235 the convenience functions in the pcre_get.c source file.
1236
1237 11. An option change at the start of a pattern that had top-level alternatives
1238 could cause overwriting and/or a crash. This command provoked a crash in
1239 some environments:
1240
1241 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1242
1243 This potential security problem was recorded as CVE-2008-2371.
1244
1245 12. For a pattern where the match had to start at the beginning or immediately
1246 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1247 pcre_dfa_exec() could read past the end of the passed subject if there was
1248 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1249 pcretest so that it places the subject at the end of its malloc-ed buffer.
1250
1251 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1252 exec() might read past the end of the data buffer in UTF-8 mode.
1253
1254 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1255 the data contained the byte 0x85 as part of a UTF-8 character within its
1256 first line. This applied both to normal and DFA matching.
1257
1258 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1259 /^[^d]*?$/8 failed to match "abc".
1260
1261 16. Added a missing copyright notice to pcrecpp_internal.h.
1262
1263 17. Make it more clear in the documentation that values returned from
1264 pcre_exec() in ovector are byte offsets, not character counts.
1265
1266 18. Tidied a few places to stop certain compilers from issuing warnings.
1267
1268 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1269 supplied by Stefan Weber. I made a further small update for 7.8 because
1270 there is a change of source arrangements: the pcre_searchfuncs.c module is
1271 replaced by pcre_ucd.c.
1272
1273
1274 Version 7.7 07-May-08
1275 ---------------------
1276
1277 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1278 a string to a long long, pretend we don't even have a long long." This is
1279 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1280
1281 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1282 pre-7.6 versions, which defined a global no_arg variable instead of putting
1283 it in the RE class. (See also #8 below.)
1284
1285 3. Remove a line of dead code, identified by coverity and reported by Nuno
1286 Lopes.
1287
1288 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1289
1290 (1) The include/exclude patterns were being applied to the whole pathnames
1291 of files, instead of just to the final components.
1292
1293 (2) If there was more than one level of directory, the subdirectories were
1294 skipped unless they satisfied the include/exclude conditions. This is
1295 inconsistent with GNU grep (and could even be seen as contrary to the
1296 pcregrep specification - which I improved to make it absolutely clear).
1297 The action now is always to scan all levels of directory, and just
1298 apply the include/exclude patterns to regular files.
1299
1300 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1301 --exclude_dir in the tests to avoid scanning .svn directories.
1302
1303 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1304 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1305 doesn't support NULs in patterns.
1306
1307 7. Added some missing "const"s to declarations of static tables in
1308 pcre_compile.c and pcre_dfa_exec.c.
1309
1310 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1311 caused by fix #2 above. (Subsequently also a second patch to fix the
1312 first patch. And a third patch - this was a messy problem.)
1313
1314 9. Applied Craig's patch to remove the use of push_back().
1315
1316 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1317 matching function regexec().
1318
1319 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1320 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1321 references. PCRE supports relative numbers with this syntax (I don't think
1322 Oniguruma does).
1323
1324 12. Previously, a group with a zero repeat such as (...){0} was completely
1325 omitted from the compiled regex. However, this means that if the group
1326 was called as a subroutine from elsewhere in the pattern, things went wrong
1327 (an internal error was given). Such groups are now left in the compiled
1328 pattern, with a new opcode that causes them to be skipped at execution
1329 time.
1330
1331 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1332 to the way PCRE behaves:
1333
1334 (a) A lone ] character is dis-allowed (Perl treats it as data).
1335
1336 (b) A back reference to an unmatched subpattern matches an empty string
1337 (Perl fails the current match path).
1338
1339 (c) A data ] in a character class must be notated as \] because if the
1340 first data character in a class is ], it defines an empty class. (In
1341 Perl it is not possible to have an empty class.) The empty class []
1342 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1343 The negative empty class [^] matches any one character, independently
1344 of the DOTALL setting.
1345
1346 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1347 non-existent subpattern following a character class starting with ']' and
1348 containing () gave an internal compiling error instead of "reference to
1349 non-existent subpattern". Fortunately, when the pattern did exist, the
1350 compiled code was correct. (When scanning forwards to check for the
1351 existencd of the subpattern, it was treating the data ']' as terminating
1352 the class, so got the count wrong. When actually compiling, the reference
1353 was subsequently set up correctly.)
1354
1355 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1356 it was being rejected as not supported by pcre_dfa_exec(), even though
1357 other assertions are supported. I have made pcre_dfa_exec() support
1358 (*FAIL).
1359
1360 16. The implementation of 13c above involved the invention of a new opcode,
1361 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1362 cannot be changed at match time, I realized I could make a small
1363 improvement to matching performance by compiling OP_ALLANY instead of
1364 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1365 on the OP_ANY path.
1366
1367 17. Compiling pcretest on Windows with readline support failed without the
1368 following two fixes: (1) Make the unistd.h include conditional on
1369 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1370
1371 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1372 ncurses library to be included for pcretest when ReadLine support is
1373 requested, but also to allow for it to be overridden. This patch came from
1374 Daniel Bergström.
1375
1376 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1377 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1378 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1379 spotting this.
1380
1381
1382 Version 7.6 28-Jan-08
1383 ---------------------
1384
1385 1. A character class containing a very large number of characters with
1386 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1387 overflow.
1388
1389 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1390 HAVE_LONG_LONG is not defined.
1391
1392 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1393 bring them up to date and include new features. This patch includes:
1394
1395 - Fixed PH's badly added libz and libbz2 support.
1396 - Fixed a problem with static linking.
1397 - Added pcredemo. [But later removed - see 7 below.]
1398 - Fixed dftables problem and added an option.
1399 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1400 HAVE_LONG_LONG.
1401 - Added readline support for pcretest.
1402 - Added an listing of the option settings after cmake has run.
1403
1404 4. A user submitted a patch to Makefile that makes it easy to create
1405 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1406 Makefile.am that cause it to include this special target, without
1407 affecting anything else. Note that the same mingw target plus all
1408 the other distribution libraries and programs are now supported
1409 when configuring with CMake (see 6 below) instead of with
1410 Configure/Make.
1411
1412 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1413 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1414 exported in the Windows port". It has not yet been confirmed that the patch
1415 solves the problem, but it does no harm.
1416
1417 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1418 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1419 with CMake, and also correct the comment about stack recursion.
1420
1421 7. Remove the automatic building of pcredemo from the ./configure system and
1422 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1423 of a program that users should build themselves after PCRE is installed, so
1424 building it automatically is not really right. What is more, it gave
1425 trouble in some build environments.
1426
1427 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1428
1429
1430 Version 7.5 10-Jan-08
1431 ---------------------
1432
1433 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1434 values in parens when parsing an RE using the C++ wrapper."
1435
1436 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1437 Characters greater than 255 were excluded from the class instead of being
1438 included.
1439
1440 3. The same bug as (2) above applied to negated POSIX classes such as
1441 [:^space:].
1442
1443 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1444 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1445 I have changed it.
1446
1447 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1448 first named subpattern) and a construct such as (?&a) would reference the
1449 first named subpattern whose name started with "a" (in other words, the
1450 length check was missing). Both these problems are fixed. "Subpattern name
1451 expected" is now given for (?&) (a zero-length name), and this patch also
1452 makes it give the same error for \k'' (previously it complained that that
1453 was a reference to a non-existent subpattern).
1454
1455 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1456 this is right because (?- can be followed by option settings as well as by
1457 digits. I have, however, made the messages clearer.
1458
1459 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1460 than the number used in the conditional) now cause a compile-time error.
1461 This is actually not compatible with Perl, which accepts such patterns, but
1462 treats the conditional as always being FALSE (as PCRE used to), but it
1463 seems to me that giving a diagnostic is better.
1464
1465 8. Change "alphameric" to the more common word "alphanumeric" in comments
1466 and messages.
1467
1468 9. Fix two occurrences of "backslash" in comments that should have been
1469 "backspace".
1470
1471 10. Remove two redundant lines of code that can never be obeyed (their function
1472 was moved elsewhere).
1473
1474 11. The program that makes PCRE's Unicode character property table had a bug
1475 which caused it to generate incorrect table entries for sequences of
1476 characters that have the same character type, but are in different scripts.
1477 It amalgamated them into a single range, with the script of the first of
1478 them. In other words, some characters were in the wrong script. There were
1479 thirteen such cases, affecting characters in the following ranges:
1480
1481 U+002b0 - U+002c1
1482 U+0060c - U+0060d
1483 U+0061e - U+00612
1484 U+0064b - U+0065e
1485 U+0074d - U+0076d
1486 U+01800 - U+01805
1487 U+01d00 - U+01d77
1488 U+01d9b - U+01dbf
1489 U+0200b - U+0200f
1490 U+030fc - U+030fe
1491 U+03260 - U+0327f
1492 U+0fb46 - U+0fbb1
1493 U+10450 - U+1049d
1494
1495 12. The -o option (show only the matching part of a line) for pcregrep was not
1496 compatible with GNU grep in that, if there was more than one match in a
1497 line, it showed only the first of them. It now behaves in the same way as
1498 GNU grep.
1499
1500 13. If the -o and -v options were combined for pcregrep, it printed a blank
1501 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1502 does the same. The return code can be used to tell if there were any
1503 non-matching lines.
1504
1505 14. Added --file-offsets and --line-offsets to pcregrep.
1506
1507 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1508 infinitely looping recursion. The bug was that positive lookaheads were not
1509 being skipped when checking for a possible empty match (negative lookaheads
1510 and both kinds of lookbehind were skipped).
1511
1512 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1513 inclusion of <windows.h> to before rather than after the definition of
1514 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1515
1516 17. Specifying a possessive quantifier with a specific limit for a Unicode
1517 character property caused pcre_compile() to compile bad code, which led at
1518 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1519 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1520 caused the error; without that there was no problem.
1521
1522 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1523
1524 19. Added --enable-pcretest-libreadline.
1525
1526 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1527 RE::GlobalReplace(). As a result, the number of replacements returned was
1528 double what it should be. I removed one of the increments, but Craig sent a
1529 later patch that removed the other one (the right fix) and added unit tests
1530 that check the return values (which was not done before).
1531
1532 21. Several CMake things:
1533
1534 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1535 the names libpcre and libpcreposix, not just pcre and pcreposix.
1536
1537 (2) The above change means that pcretest and pcregrep are now correctly
1538 linked with the newly-built libraries, not previously installed ones.
1539
1540 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1541
1542 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1543 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1544 UTF-8 newline character). The key issue is that the pattern starts .*;
1545 this means that the match must be either at the beginning, or after a
1546 newline. The bug was in the code for advancing after a failed match and
1547 checking that the new position followed a newline. It was not taking
1548 account of UTF-8 characters correctly.
1549
1550 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1551 character classes. PCRE was not treating the sequence [:...:] as a
1552 character class unless the ... were all letters. Perl, however, seems to
1553 allow any characters between [: and :], though of course it rejects as
1554 unknown any "names" that contain non-letters, because all the known class
1555 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1556 for example, whereas PCRE did not - it did not recognize a POSIX character
1557 class. This seemed a bit dangerous, so the code has been changed to be
1558 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1559 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1560 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1561 Perl does, and where it didn't before.
1562
1563 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1564 Windows environments %n is disabled by default.
1565
1566
1567 Version 7.4 21-Sep-07
1568 ---------------------
1569
1570 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1571 means that a class such as [\s] counted as "explicit reference to CR or
1572 LF". That isn't really right - the whole point of the change was to try to
1573 help when there was an actual mention of one of the two characters. So now
1574 the change happens only if \r or \n (or a literal CR or LF) character is
1575 encountered.
1576
1577 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1578 of both had grown to the point where there were only 3 bits left.
1579 Fortunately, there was spare space in the data structure, and so I have
1580 moved the internal flags into a new 16-bit field to free up more option
1581 bits.
1582
1583 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1584 but did not set the internal JCHANGED flag - either of these is enough to
1585 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1586 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1587 start sets both bits.
1588
1589 4. Added options (at build time, compile time, exec time) to change \R from
1590 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1591
1592 5. doc/pcresyntax.html was missing from the distribution.
1593
1594 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1595 compatibility, even though it is no longer used.
1596
1597 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1598 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1599 windows.h file is present (where different names are used). [This was
1600 reversed later after testing - see 16 below.]
1601
1602 8. Changed all #include <config.h> to #include "config.h". There were also
1603 some further <pcre.h> cases that I changed to "pcre.h".
1604
1605 9. When pcregrep was used with the --colour option, it missed the line ending
1606 sequence off the lines that it output.
1607
1608 10. It was pointed out to me that arrays of string pointers cause lots of
1609 relocations when a shared library is dynamically loaded. A technique of
1610 using a single long string with a table of offsets can drastically reduce
1611 these. I have refactored PCRE in four places to do this. The result is
1612 dramatic:
1613
1614 Originally: 290
1615 After changing UCP table: 187
1616 After changing error message table: 43
1617 After changing table of "verbs" 36
1618 After changing table of Posix names 22
1619
1620 Thanks to the folks working on Gregex for glib for this insight.
1621
1622 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1623 unicode-properties was also set.
1624
1625 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1626
1627 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1628 checked only for CRLF.
1629
1630 14. Added casts to pcretest.c to avoid compiler warnings.
1631
1632 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1633
1634 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1635 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1636 entirely. This removes changes made in 7 above.
1637
1638 17. The CMake files have been updated, and there is now more information about
1639 building with CMake in the NON-UNIX-USE document.
1640
1641
1642 Version 7.3 28-Aug-07
1643 ---------------------
1644
1645 1. In the rejigging of the build system that eventually resulted in 7.1, the
1646 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1647 brackets there is not right, since it causes compilers to look for an
1648 installed pcre.h, not the version that is in the source that is being
1649 compiled (which of course may be different). I have changed it back to:
1650
1651 #include "pcre.h"
1652
1653 I have a vague recollection that the change was concerned with compiling in
1654 different directories, but in the new build system, that is taken care of
1655 by the VPATH setting the Makefile.
1656
1657 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1658 when the subject happened to end in the byte 0x85 (e.g. if the last
1659 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1660 characters but of course it shouldn't be taken as a newline when it is part
1661 of another character. The bug was that, for an unlimited repeat of . in
1662 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1663 characters when looking for a newline.
1664
1665 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1666
1667 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1668 in debug output.
1669
1670 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1671 long printing in the pcrecpp unittest when running under MinGW.
1672
1673 6. ESC_K was left out of the EBCDIC table.
1674
1675 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1676 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1677 limit also applies to "virtual nesting" when a pattern is recursive, and in
1678 this case 1000 isn't so big. I have been able to remove this limit at the
1679 expense of backing off one optimization in certain circumstances. Normally,
1680 when pcre_exec() would call its internal match() function recursively and
1681 immediately return the result unconditionally, it uses a "tail recursion"
1682 feature to save stack. However, when a subpattern that can match an empty
1683 string has an unlimited repetition quantifier, it no longer makes this
1684 optimization. That gives it a stack frame in which to save the data for
1685 checking that an empty string has been matched. Previously this was taken
1686 from the 1000-entry workspace that had been reserved. So now there is no
1687 explicit limit, but more stack is used.
1688
1689 8. Applied Daniel's patches to solve problems with the import/export magic
1690 syntax that is required for Windows, and which was going wrong for the
1691 pcreposix and pcrecpp parts of the library. These were overlooked when this
1692 problem was solved for the main library.
1693
1694 9. There were some crude static tests to avoid integer overflow when computing
1695 the size of patterns that contain repeated groups with explicit upper
1696 limits. As the maximum quantifier is 65535, the maximum group length was
1697 set at 30,000 so that the product of these two numbers did not overflow a
1698 32-bit integer. However, it turns out that people want to use groups that
1699 are longer than 30,000 bytes (though not repeat them that many times).
1700 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1701 made it possible to implement the integer overflow checks in a much more
1702 dynamic way, which I have now done. The artificial limitation on group
1703 length has been removed - we now have only the limit on the total length of
1704 the compiled pattern, which depends on the LINK_SIZE setting.
1705
1706 10. Fixed a bug in the documentation for get/copy named substring when
1707 duplicate names are permitted. If none of the named substrings are set, the
1708 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1709 empty string.
1710
1711 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1712 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1713 because the ] is interpreted as the first data character and the
1714 terminating ] is not found. PCRE has been made compatible with Perl in this
1715 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1716 cause memory overwriting.
1717
1718 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1719 string has been matched (to stop an infinite loop). It was not recognizing
1720 a conditional subpattern that could match an empty string if that
1721 subpattern was within another subpattern. For example, it looped when
1722 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1723 condition was not nested. This bug has been fixed.
1724
1725 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1726 past the start of the subject in the presence of bytes with the top bit
1727 set, for example "\x8aBCD".
1728
1729 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1730 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1731
1732 14. Optimized (?!) to (*FAIL).
1733
1734 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1735 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1736 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1737 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1738 does: it's just the validity check that is more restrictive.
1739
1740 16. Inserted checks for integer overflows during escape sequence (backslash)
1741 processing, and also fixed erroneous offset values for syntax errors during
1742 backslash processing.
1743
1744 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1745 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1746
1747 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1748 caused an overrun.
1749
1750 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1751 something other than just ASCII characters) inside a group that had an
1752 unlimited repeat caused a loop at compile time (while checking to see
1753 whether the group could match an empty string).
1754
1755 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1756 [\P{Any}] did so. (Error in the code for printing property names.)
1757
1758 21. An orphan \E inside a character class could cause a crash.
1759
1760 22. A repeated capturing bracket such as (A)? could cause a wild memory
1761 reference during compilation.
1762
1763 23. There are several functions in pcre_compile() that scan along a compiled
1764 expression for various reasons (e.g. to see if it's fixed length for look
1765 behind). There were bugs in these functions when a repeated \p or \P was
1766 present in the pattern. These operators have additional parameters compared
1767 with \d, etc, and these were not being taken into account when moving along
1768 the compiled data. Specifically:
1769
1770 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1771 length.
1772
1773 (b) An item such as \pL+ within a repeated group could cause crashes or
1774 loops.
1775
1776 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1777 "reference to non-existent subpattern" error.
1778
1779 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1780
1781 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1782 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1783
1784 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1785
1786 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1787 character were causing crashes (broken optimization).
1788
1789 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1790 \p or \P) caused a compile-time loop.
1791
1792 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1793 break. For example, the unstudied pattern [\r\n]A does not match the string
1794 "\r\nA" because change 7.0/46 below moves the current point on by two
1795 characters after failing to match at the start. However, the pattern \nA
1796 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1797 the same is true. There doesn't seem any very clean way out of this, but
1798 what I have chosen to do makes the common cases work: PCRE now takes note
1799 of whether there can be an explicit match for \r or \n anywhere in the
1800 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1801 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1802 pattern has explicit CR or LF references.
1803
1804 29. Added (*CR) etc for changing newline setting at start of pattern.
1805
1806
1807 Version 7.2 19-Jun-07
1808 ---------------------
1809
1810 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1811 which is apparently normally available under Windows.
1812
1813 2. Re-jig the pcregrep tests with different newline settings in an attempt
1814 to make them independent of the local environment's newline setting.
1815
1816 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1817
1818 4. Some of the "internals" tests were previously cut out when the link size
1819 was not 2, because the output contained actual offsets. The recent new
1820 "Z" feature of pcretest means that these can be cut out, making the tests
1821 usable with all link sizes.
1822
1823 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1824 stack recursion. This gives a massive performance boost under BSD, but just
1825 a small improvement under Linux. However, it saves one field in the frame
1826 in all cases.
1827
1828 6. Added more features from the forthcoming Perl 5.10:
1829
1830 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1831 recursion call. It refers to the nth most recently opened parentheses.
1832
1833 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1834 to be opened parentheses.
1835
1836 (c) Conditions that refer to capturing parentheses can be specified
1837 relatively, for example, (?(-2)... or (?(+3)...
1838
1839 (d) \K resets the start of the current match so that everything before
1840 is not part of it.
1841
1842 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1843
1844 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1845 reference syntax.
1846
1847 (g) (?| introduces a group in which the numbering of parentheses in each
1848 alternative starts with the same number.
1849
1850 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1851
1852 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1853 PCRE_INFO_JCHANGED.
1854
1855 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1856 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1857 for detecting groups that can match an empty string.
1858
1859 9. A pattern with a very large number of alternatives (more than several
1860 hundred) was running out of internal workspace during the pre-compile
1861 phase, where pcre_compile() figures out how much memory will be needed. A
1862 bit of new cunning has reduced the workspace needed for groups with
1863 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1864 workspace instead of running out of the 4096 that are available.
1865
1866 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1867
1868 11. Applied patch from Google to remove an optimization that didn't quite work.
1869 The report of the bug said:
1870
1871 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1872 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1873 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1874
1875 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1876 it matched the wrong number of bytes.
1877
1878
1879 Version 7.1 24-Apr-07
1880 ---------------------
1881
1882 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1883 that is more "standard", making use of automake and other Autotools. There
1884 is some re-arrangement of the files and adjustment of comments consequent
1885 on this.
1886
1887 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1888 for recursive directory scanning broke on some systems because the files
1889 are not scanned in any specific order and on different systems the order
1890 was different. A call to "sort" has been inserted into RunGrepTest for the
1891 approprate test as a short-term fix. In the longer term there may be an
1892 alternative.
1893
1894 3. I had an email from Eric Raymond about problems translating some of PCRE's
1895 man pages to HTML (despite the fact that I distribute HTML pages, some
1896 people do their own conversions for various reasons). The problems
1897 concerned the use of low-level troff macros .br and .in. I have therefore
1898 removed all such uses from the man pages (some were redundant, some could
1899 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1900 HTML has been updated to handle .nf/.fi and to complain if it encounters
1901 .br or .in.
1902
1903 4. Updated comments in configure.ac that get placed in config.h.in and also
1904 arranged for config.h to be included in the distribution, with the name
1905 config.h.generic, for the benefit of those who have to compile without
1906 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1907
1908 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1909 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1910 makevp.bat for the current PCRE, using the additional files
1911 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1912
1913 6. A Windows user reported a minor discrepancy with test 2, which turned out
1914 to be caused by a trailing space on an input line that had got lost in his
1915 copy. The trailing space was an accident, so I've just removed it.
1916
1917 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1918 that is needed.
1919
1920 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1921 as "const" (a) because they are and (b) because it helps the PHP
1922 maintainers who have recently made a script to detect big data structures
1923 in the php code that should be moved to the .rodata section. I remembered
1924 to update Builducptable as well, so it won't revert if ucptable.h is ever
1925 re-created.
1926
1927 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1928 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1929 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1930 support is not required. This saves 1.5-2K of code, which is important in
1931 some applications.
1932
1933 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1934 so as not to refer to the tables, even though these functions will never be
1935 called when UTF-8 support is disabled. Otherwise there are problems with a
1936 shared library.
1937
1938 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1939
1940 (a) It was defining its arguments as char * instead of void *.
1941
1942 (b) It was assuming that all moves were upwards in memory; this was true
1943 a long time ago when I wrote it, but is no longer the case.
1944
1945 The emulated memove() is provided for those environments that have neither
1946 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1947 is clearly not the case, as these two bugs were recently reported.
1948
1949 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1950 and Detrail to create the HTML documentation, the .txt form of the man
1951 pages, and it removes trailing spaces from listed files. It also creates
1952 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1953 case, it wraps all the #defines with #ifndefs. This script should be run
1954 before "make dist".
1955
1956 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1957 with Unicode property support.
1958
1959 (a) For a maximizing quantifier, if the two different cases of the
1960 character were of different lengths in their UTF-8 codings (there are
1961 some cases like this - I found 11), and the matching function had to
1962 back up over a mixture of the two cases, it incorrectly assumed they
1963 were both the same length.
1964
1965 (b) When PCRE was configured to use the heap rather than the stack for
1966 recursion during matching, it was not correctly preserving the data for
1967 the other case of a UTF-8 character when checking ahead for a match
1968 while processing a minimizing repeat. If the check also involved
1969 matching a wide character, but failed, corruption could cause an
1970 erroneous result when trying to check for a repeat of the original
1971 character.
1972
1973 13. Some tidying changes to the testing mechanism:
1974
1975 (a) The RunTest script now detects the internal link size and whether there
1976 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1977 values substituted by "configure". (The RunGrepTest script already did
1978 this for UTF-8.) The configure.ac script no longer substitutes the
1979 relevant variables.
1980
1981 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1982 with length and offset values. This means that the output is different
1983 for different internal link sizes. Test 2 is skipped for link sizes
1984 other than 2 because of this, bypassing the problem. Unfortunately,
1985 there was also a test in test 3 (the locale tests) that used /B and
1986 failed for link sizes other than 2. Rather than cut the whole test out,
1987 I have added a new /Z option to pcretest that replaces the length and
1988 offset values with spaces. This is now used to make test 3 independent
1989 of link size. (Test 2 will be tidied up later.)
1990
1991 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1992 segmentation fault instead of returning the appropriate error message.
1993
1994 15. In multiline mode when the newline sequence was set to "any", the pattern
1995 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1996 This doesn't seem right; it now treats the CRLF combination as the line
1997 ending, and so does not match in that case. It's only a pattern such as ^$
1998 that would hit this one: something like ^ABC$ would have failed after \r
1999 and then tried again after \r\n.
2000
2001 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2002 in an attempt to make files that differ only in their line terminators
2003 compare equal. This works on Linux.
2004
2005 17. Under certain error circumstances pcregrep might try to free random memory
2006 as it exited. This is now fixed, thanks to valgrind.
2007
2008 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2009 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2010 was because its rules for how to advance for /g after matching an empty
2011 string at the end of a line did not allow for this case. They now check for
2012 it specially.
2013
2014 20. pcretest is supposed to handle patterns and data of any length, by
2015 extending its buffers when necessary. It was getting this wrong when the
2016 buffer for a data line had to be extended.
2017
2018 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2019 CRLF as a newline sequence.
2020
2021 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2022 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2023 I have nevertheless tidied it up.
2024
2025 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2026
2027 24. Added a man page for pcre-config.
2028
2029
2030 Version 7.0 19-Dec-06
2031 ---------------------
2032
2033 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2034 moving to gcc 4.1.1.
2035
2036 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2037 sys/time.h, which is documented as needed for this function. It doesn't
2038 seem to matter on Linux, but it showed up on some releases of OS X.
2039
2040 3. It seems that there are systems where bytes whose values are greater than
2041 127 match isprint() in the "C" locale. The "C" locale should be the
2042 default when a C program starts up. In most systems, only ASCII printing
2043 characters match isprint(). This difference caused the output from pcretest
2044 to vary, making some of the tests fail. I have changed pcretest so that:
2045
2046 (a) When it is outputting text in the compiled version of a pattern, bytes
2047 other than 32-126 are always shown as hex escapes.
2048
2049 (b) When it is outputting text that is a matched part of a subject string,
2050 it does the same, unless a different locale has been set for the match
2051 (using the /L modifier). In this case, it uses isprint() to decide.
2052
2053 4. Fixed a major bug that caused incorrect computation of the amount of memory
2054 required for a compiled pattern when options that changed within the
2055 pattern affected the logic of the preliminary scan that determines the
2056 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2057 that the computed length was too small. The symptoms of this bug were
2058 either the PCRE error "internal error: code overflow" from pcre_compile(),
2059 or a glibc crash with a message such as "pcretest: free(): invalid next
2060 size (fast)". Examples of patterns that provoked this bug (shown in
2061 pcretest format) are:
2062
2063 /(?-x: )/x
2064 /(?x)(?-x: \s*#\s*)/
2065 /((?i)[\x{c0}])/8
2066 /(?i:[\x{c0}])/8
2067
2068 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2069 is now done differently.
2070
2071 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2072 wrapper classes; (b) implement a new function in the C++ scanner that is
2073 more efficient than the old way of doing things because it avoids levels of
2074 recursion in the regex matching; (c) add a paragraph to the documentation
2075 for the FullMatch() function.
2076
2077 6. The escape sequence \n was being treated as whatever was defined as
2078 "newline". Not only was this contrary to the documentation, which states
2079 that \n is character 10 (hex 0A), but it also went horribly wrong when
2080 "newline" was defined as CRLF. This has been fixed.
2081
2082 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2083 was being set to -1 for the "end of line" case (supposedly a value that no
2084 character can have). Though this value is never used (the check for end of
2085 line is "zero bytes in current character"), it caused compiler complaints.
2086 I've changed it to 0xffffffff.
2087
2088 8. In pcre_version.c, the version string was being built by a sequence of
2089 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2090 string (as it is for production releases) called a macro with an empty
2091 argument. The C standard says the result of this is undefined. The gcc
2092 compiler treats it as an empty string (which was what was wanted) but it is
2093 reported that Visual C gives an error. The source has been hacked around to
2094 avoid this problem.
2095
2096 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2097 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2098 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2099 of them did).
2100
2101 10. Originally, pcretest opened its input and output without "b"; then I was
2102 told that "b" was needed in some environments, so it was added for release
2103 5.0 to both the input and output. (It makes no difference on Unix-like
2104 systems.) Later I was told that it is wrong for the input on Windows. I've
2105 now abstracted the modes into two macros, to make it easier to fiddle with
2106 them, and removed "b" from the input mode under Windows.
2107
2108 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2109
2110 12. Added -help and --help to pcretest as an official way of being reminded
2111 of the options.
2112
2113 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2114 and pcrecpp.cc because they annoy compilers at high warning levels.
2115
2116 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2117
2118 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2119 scripts are not C programs :-) and which was not noticed because it works
2120 on Linux.
2121
2122 16. pcretest is supposed to handle any length of pattern and data line (as one
2123 line or as a continued sequence of lines) by extending its input buffer if
2124 necessary. This feature was broken for very long pattern lines, leading to
2125 a string of junk being passed to pcre_compile() if the pattern was longer
2126 than about 50K.
2127
2128 17. I have done a major re-factoring of the way pcre_compile() computes the
2129 amount of memory needed for a compiled pattern. Previously, there was code
2130 that made a preliminary scan of the pattern in order to do this. That was
2131 OK when PCRE was new, but as the facilities have expanded, it has become
2132 harder and harder to keep it in step with the real compile phase, and there
2133 have been a number of bugs (see for example, 4 above). I have now found a
2134 cunning way of running the real compile function in a "fake" mode that
2135 enables it to compute how much memory it would need, while actually only
2136 ever using a few hundred bytes of working memory and without too many
2137 tests of the mode. This should make future maintenance and development
2138 easier. A side effect of this work is that the limit of 200 on the nesting
2139 depth of parentheses has been removed (though this was never a serious
2140 limitation, I suspect). However, there is a downside: pcre_compile() now
2141 runs more slowly than before (30% or more, depending on the pattern). I
2142 hope this isn't a big issue. There is no effect on runtime performance.
2143
2144 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2145 newline (only possible for the last line of a file) and it was a
2146 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2147
2148 19. Added additional timing features to pcretest. (1) The -tm option now times
2149 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2150 separate command line item, by a number that specifies the number of
2151 repeats to use when timing. The default is 50000; this gives better
2152 precision, but takes uncomfortably long for very large patterns.
2153
2154 20. Extended pcre_study() to be more clever in cases where a branch of a
2155 subpattern has no definite first character. For example, (a*|b*)[cd] would
2156 previously give no result from pcre_study(). Now it recognizes that the
2157 first character must be a, b, c, or d.
2158
2159 21. There was an incorrect error "recursive call could loop indefinitely" if
2160 a subpattern (or the entire pattern) that was being tested for matching an
2161 empty string contained only one non-empty item after a nested subpattern.
2162 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2163 incorrectly, because the \d was being skipped in the check.
2164
2165 22. The pcretest program now has a new pattern option /B and a command line
2166 option -b, which is equivalent to adding /B to every pattern. This causes
2167 it to show the compiled bytecode, without the additional information that
2168 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2169 is the same as /B/I).
2170
2171 23. A new optimization is now able automatically to treat some sequences such
2172 as a*b as a*+b. More specifically, if something simple (such as a character
2173 or a simple class like \d) has an unlimited quantifier, and is followed by
2174 something that cannot possibly match the quantified thing, the quantifier
2175 is automatically "possessified".
2176
2177 24. A recursive reference to a subpattern whose number was greater than 39
2178 went wrong under certain circumstances in UTF-8 mode. This bug could also
2179 have affected the operation of pcre_study().
2180
2181 25. Realized that a little bit of performance could be had by replacing
2182 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2183
2184 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2185
2186 27. Possessive quantifiers such as a++ were previously implemented by turning
2187 them into atomic groups such as ($>a+). Now they have their own opcodes,
2188 which improves performance. This includes the automatically created ones
2189 from 23 above.
2190
2191 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2192 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2193 the first matched character to be a colon. This applied both to named and
2194 numbered groups.
2195
2196 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2197
2198 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2199 building PCRE on Windows easier, so I have included it in the distribution.
2200
2201 31. There is now a check in pcretest against a ridiculously large number being
2202 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2203 loop, the loop is abandoned.
2204
2205 32. Forward references to subpatterns in conditions such as (?(2)...) where
2206 subpattern 2 is defined later cause pcre_compile() to search forwards in
2207 the pattern for the relevant set of parentheses. This search went wrong
2208 when there were unescaped parentheses in a character class, parentheses
2209 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2210
2211 33. "Subroutine" calls and backreferences were previously restricted to
2212 referencing subpatterns earlier in the regex. This restriction has now
2213 been removed.
2214
2215 34. Added a number of extra features that are going to be in Perl 5.10. On the
2216 whole, these are just syntactic alternatives for features that PCRE had
2217 previously implemented using the Python syntax or my own invention. The
2218 other formats are all retained for compatibility.
2219
2220 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2221 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2222 also .NET compatible.
2223
2224 (b) A recursion or subroutine call to a named group can now be defined as
2225 (?&name) as well as (?P>name).
2226
2227 (c) A backreference to a named group can now be defined as \k<name> or
2228 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2229 5.10, are also .NET compatible.
2230
2231 (d) A conditional reference to a named group can now use the syntax
2232 (?(<name>) or (?('name') as well as (?(name).
2233
2234 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2235 groups (named and numbered) that are never evaluated inline, but can be
2236 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2237 is always false. There may be only one alternative in such a group.
2238
2239 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2240 as the simple (?(R). The condition is true only if the most recent
2241 recursion is that of the given number or name. It does not search out
2242 through the entire recursion stack.
2243
2244 (g) The escape \gN or \g{N} has been added, where N is a positive or
2245 negative number, specifying an absolute or relative reference.
2246
2247 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2248 some "unreachable code" warnings.
2249
2250 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2251 things, this adds five new scripts.
2252
2253 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2254 There were also incompatibilities regarding the handling of \Q..\E inside
2255 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2256 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2257
2258 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2259 matches an empty string, and forcibly breaks the loop. There were bugs in
2260 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2261 against aaaa the result was just "a" rather than "aaaa", for example. Two
2262 separate and independent bugs (that affected different cases) have been
2263 fixed.
2264
2265 39. Refactored the code to abolish the use of different opcodes for small
2266 capturing bracket numbers. This is a tidy that I avoided doing when I
2267 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2268 The new approach is not only tidier, it makes it possible to reduce the
2269 memory needed to fix the previous bug (38).
2270
2271 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2272 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2273 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2274 mode.
2275
2276 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2277 report.
2278
2279 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2280 copy construction and assignment in the C++ wrapper.
2281
2282 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2283 couple of bugs that could have given wrong results in the "--newline=crlf"
2284 case.
2285
2286 44. Added a number of casts and did some reorganization of signed/unsigned int
2287 variables following suggestions from Dair Grant. Also renamed the variable
2288 "this" as "item" because it is a C++ keyword.
2289
2290 45. Arranged for dftables to add
2291
2292 #include "pcre_internal.h"
2293
2294 to pcre_chartables.c because without it, gcc 4.x may remove the array
2295 definition from the final binary if PCRE is built into a static library and
2296 dead code stripping is activated.
2297
2298 46. For an unanchored pattern, if a match attempt fails at the start of a
2299 newline sequence, and the newline setting is CRLF or ANY, and the next two
2300 characters are CRLF, advance by two characters instead of one.
2301
2302
2303 Version 6.7 04-Jul-06
2304 ---------------------
2305
2306 1. In order to handle tests when input lines are enormously long, pcretest has
2307 been re-factored so that it automatically extends its buffers when
2308 necessary. The code is crude, but this _is_ just a test program. The
2309 default size has been increased from 32K to 50K.
2310
2311 2. The code in pcre_study() was using the value of the re argument before
2312 testing it for NULL. (Of course, in any sensible call of the function, it
2313 won't be NULL.)
2314
2315 3. The memmove() emulation function in pcre_internal.h, which is used on
2316 systems that lack both memmove() and bcopy() - that is, hardly ever -
2317 was missing a "static" storage class specifier.
2318
2319 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2320 containing an extended class (one that cannot be represented by a bitmap
2321 because it contains high-valued characters or Unicode property items, e.g.
2322 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2323 pattern, but PCRE should not loop if you do not (it no longer does).
2324 [Detail: two cases were found: (a) a repeated subpattern containing an
2325 extended class; (b) a recursive reference to a subpattern that followed a
2326 previous extended class. It wasn't skipping over the extended class
2327 correctly when UTF-8 mode was not set.]
2328
2329 5. A negated single-character class was not being recognized as fixed-length
2330 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2331 compile error "lookbehind assertion is not fixed length".
2332
2333 6. The RunPerlTest auxiliary script was showing an unexpected difference
2334 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2335 write a Perl script that can interpret lines of an input file either as
2336 byte characters or as UTF-8, which is what "perltest" was being required to
2337 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2338 can't do is switch easily at run time between having the "use utf8;" pragma
2339 or not. In the end, I fudged it by using the RunPerlTest script to insert
2340 "use utf8;" explicitly for the UTF-8 tests.
2341
2342 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2343 the end of the subject string, contrary to the documentation and to what
2344 Perl does. This was true of both matching functions. Now it matches only at
2345 the start of the subject and immediately after *internal* newlines.
2346
2347 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2348 a pointer to an int instead of a pointer to an unsigned long int. This
2349 caused problems on 64-bit systems.
2350
2351 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2352 instance of the 'standard' template library not being so standard".
2353
2354 10. There was no check on the number of named subpatterns nor the maximum
2355 length of a subpattern name. The product of these values is used to compute
2356 the size of the memory block for a compiled pattern. By supplying a very
2357 long subpattern name and a large number of named subpatterns, the size
2358 computation could be caused to overflow. This is now prevented by limiting
2359 the length of names to 32 characters, and the number of named subpatterns
2360 to 10,000.
2361
2362 11. Subpatterns that are repeated with specific counts have to be replicated in
2363 the compiled pattern. The size of memory for this was computed from the
2364 length of the subpattern and the repeat count. The latter is limited to
2365 65535, but there was no limit on the former, meaning that integer overflow
2366 could in principle occur. The compiled length of a repeated subpattern is
2367 now limited to 30,000 bytes in order to prevent this.
2368
2369 12. Added the optional facility to have named substrings with the same name.
2370
2371 13. Added the ability to use a named substring as a condition, using the
2372 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2373 are numbers (not recommended). Forward references are permitted.
2374
2375 14. Added forward references in named backreferences (if you see what I mean).
2376
2377 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2378 pattern could run off the end of the subject. For example, the pattern
2379 "(?s)(.{1,5})"8 did this with the subject "ab".
2380
2381 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2382 PCRE_CASELESS was set when matching characters that were quantified with ?
2383 or *.
2384
2385 17. A character class other than a single negated character that had a minimum
2386 but no maximum quantifier - for example [ab]{6,} - was not handled
2387 correctly by pce_dfa_exec(). It would match only one character.
2388
2389 18. A valid (though odd) pattern that looked like a POSIX character
2390 class but used an invalid character after [ (for example [[,abc,]]) caused
2391 pcre_compile() to give the error "Failed: internal error: code overflow" or
2392 in some cases to crash with a glibc free() error. This could even happen if
2393 the pattern terminated after [[ but there just happened to be a sequence of
2394 letters, a binary zero, and a closing ] in the memory that followed.
2395
2396 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2397 over the years. Originally (before any Unicode support), just the bottom 8
2398 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2399 output from "man perlunicode" includes this:
2400
2401 The regular expression compiler produces polymorphic opcodes. That
2402 is, the pattern adapts to the data and automatically switches to
2403 the Unicode character scheme when presented with Unicode data--or
2404 instead uses a traditional byte scheme when presented with byte
2405 data.
2406
2407 Sadly, a wide octal escape does not cause a switch, and in a string with
2408 no other multibyte characters, these octal escapes are treated as before.
2409 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2410 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2411 Unicode string.
2412
2413 I have not perpetrated such confusion in PCRE. Up till now, it took just
2414 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2415 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2416 translate to the appropriate multibyte character.
2417
2418 29. Applied some refactoring to reduce the number of warnings from Microsoft
2419 and Borland compilers. This has included removing the fudge introduced
2420 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2421 a warning about an unused variable.
2422
2423 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2424 characters since release 4.0, because Perl (from release 5.004) does not.
2425 [Or at least, is documented not to: some releases seem to be in conflict
2426 with the documentation.] However, when a pattern was studied with
2427 pcre_study() and all its branches started with \s, PCRE still included VT
2428 as a possible starting character. Of course, this did no harm; it just
2429 caused an unnecessary match attempt.
2430
2431 22. Removed a now-redundant internal flag bit that recorded the fact that case
2432 dependency changed within the pattern. This was once needed for "required
2433 byte" processing, but is no longer used. This recovers a now-scarce options
2434 bit. Also moved the least significant internal flag bit to the most-
2435 significant bit of the word, which was not previously used (hangover from
2436 the days when it was an int rather than a uint) to free up another bit for
2437 the future.
2438
2439 23. Added support for CRLF line endings as well as CR and LF. As well as the
2440 default being selectable at build time, it can now be changed at runtime
2441 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2442 specify that it is scanning data with non-default line endings.
2443
2444 24. Changed the definition of CXXLINK to make it agree with the definition of
2445 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2446
2447 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2448 recursions. This makes a big different to stack usage for some patterns.
2449
2450 26. If a subpattern containing a named recursion or subroutine reference such
2451 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2452 the space required for the compiled pattern went wrong and gave too small a
2453 value. Depending on the environment, this could lead to "Failed: internal
2454 error: code overflow at offset 49" or "glibc detected double free or
2455 corruption" errors.
2456
2457 27. Applied patches from Google (a) to support the new newline modes and (b) to
2458 advance over multibyte UTF-8 characters in GlobalReplace.
2459
2460 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2461 difference for some implementation of PCRE in some Windows version.
2462
2463 29. Added some extra testing facilities to pcretest:
2464
2465 \q<number> in a data line sets the "match limit" value
2466 \Q<number> in a data line sets the "match recursion limt" value
2467 -S <number> sets the stack size, where <number> is in megabytes
2468
2469 The -S option isn't available for Windows.
2470
2471
2472 Version 6.6 06-Feb-06
2473 ---------------------
2474
2475 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2476 in pcreposix.h. I have copied the definition from pcre.h.
2477
2478 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2479 because pcre.h is no longer a built file.
2480
2481 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2482 not normally included in the compiled code.
2483
2484
2485 Version 6.5 01-Feb-06
2486 ---------------------
2487
2488 1. When using the partial match feature with pcre_dfa_exec(), it was not
2489 anchoring the second and subsequent partial matches at the new starting
2490 point. This could lead to incorrect results. For example, with the pattern
2491 /1234/, partially matching against "123" and then "a4" gave a match.
2492
2493 2. Changes to pcregrep:
2494
2495 (a) All non-match returns from pcre_exec() were being treated as failures
2496 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2497 error message is output. Some extra information is given for the
2498 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2499 probably the only errors that are likely to be caused by users (by
2500 specifying a regex that has nested indefinite repeats, for instance).
2501 If there are more than 20 of these errors, pcregrep is abandoned.
2502
2503 (b) A binary zero was treated as data while matching, but terminated the
2504 output line if it was written out. This has been fixed: binary zeroes
2505 are now no different to any other data bytes.
2506
2507 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2508 used to set a locale for matching. The --locale=xxxx long option has
2509 been added (no short equivalent) to specify a locale explicitly on the
2510 pcregrep command, overriding the environment variables.
2511
2512 (d) When -B was used with -n, some line numbers in the output were one less
2513 than they should have been.
2514
2515 (e) Added the -o (--only-matching) option.
2516
2517 (f) If -A or -C was used with -c (count only), some lines of context were
2518 accidentally printed for the final match.
2519
2520 (g) Added the -H (--with-filename) option.
2521
2522 (h) The combination of options -rh failed to suppress file names for files
2523 that were found from directory arguments.
2524
2525 (i) Added the -D (--devices) and -d (--directories) options.
2526
2527 (j) Added the -F (--fixed-strings) option.
2528
2529 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2530
2531 (l) Added the --colo(u)r option.
2532
2533 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2534 is not present by default.
2535
2536 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2537 items such as (?R) or (?1), when the recursion could match a number of
2538 alternatives. If it matched one of the alternatives, but subsequently,
2539 outside the recursion, there was a failure, the code tried to back up into
2540 the recursion. However, because of the way PCRE is implemented, this is not
2541 possible, and the result was an incorrect result from the match.
2542
2543 In order to prevent this happening, the specification of recursion has
2544 been changed so that all such subpatterns are automatically treated as
2545 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2546
2547 4. I had overlooked the fact that, in some locales, there are characters for
2548 which isalpha() is true but neither isupper() nor islower() are true. In
2549 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2550 and ordfeminine) are like this. This affected the treatment of \w and \W
2551 when they appeared in character classes, but not when they appeared outside
2552 a character class. The bit map for "word" characters is now created
2553 separately from the results of isalnum() instead of just taking it from the
2554 upper, lower, and digit maps. (Plus the underscore character, of course.)
2555
2556 5. The above bug also affected the handling of POSIX character classes such as
2557 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2558 permanent tables. Instead, the bit maps for such a class were previously
2559 created as the appropriate unions of the upper, lower, and digit bitmaps.
2560 Now they are created by subtraction from the [[:word:]] class, which has
2561 its own bitmap.
2562
2563 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2564 It is created by subtracting the vertical space characters (\x09, \x0a,
2565 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2566 subtraction was done in the overall bitmap for a character class, meaning
2567 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2568 be recognized. This bug has been fixed.
2569
2570 7. Patches from the folks at Google:
2571
2572 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2573 real life, but is still worth protecting against".
2574
2575 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2576 regular expressions".
2577
2578 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2579 have it.
2580
2581 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2582 "configure" and the latter not, in order to fix a problem somebody had
2583 with compiling the Arg class on HP-UX.
2584
2585 (e) Improve the error-handling of the C++ wrapper a little bit.
2586
2587 (f) New tests for checking recursion limiting.
2588
2589 8. The pcre_memmove() function, which is used only if the environment does not
2590 have a standard memmove() function (and is therefore rarely compiled),
2591 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2592 returning a result (though PCRE never actually uses the result).
2593
2594 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2595 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2596 returned instead of calling malloc() with an overflowing number that would
2597 most likely cause subsequent chaos.
2598
2599 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2600
2601 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2602 with this option is matched, the nmatch and pmatch options of regexec() are
2603 ignored.
2604
2605 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2606 provided in case anyone wants to the the POSIX interface with UTF-8
2607 strings.
2608
2609 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2610 C++ linking (needed for some HP-UX environments).
2611
2612 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2613 (unused parameter) and in the pcre_printint() function (omitted "default"
2614 switch label when the default is to do nothing).
2615
2616 15. Added some code to make it possible, when PCRE is compiled as a C++
2617 library, to replace subject pointers for pcre_exec() with a smart pointer
2618 class, thus making it possible to process discontinuous strings.
2619
2620 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2621 much the same function. They were added by different people who were trying
2622 to make PCRE easy to compile on non-Unix systems. It has been suggested
2623 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2624 for compiling on Windows systems. I have therefore replaced it with
2625 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2626 defaults to "extern" for C or "extern C" for C++, which works fine on
2627 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2628 SCOPE with something explicit in config.h. In addition:
2629
2630 (a) pcreposix.h still had just "extern" instead of either of these macros;
2631 I have replaced it with PCRE_DATA_SCOPE.
2632
2633 (b) Functions such as _pcre_xclass(), which are internal to the library,
2634 but external in the C sense, all had PCRE_EXPORT in their definitions.
2635 This is apparently wrong for the Windows case, so I have removed it.
2636 (It makes no difference on Unix-like systems.)
2637
2638 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2639 of recursive calls to match(). This is different to MATCH_LIMIT because
2640 that limits the total number of calls to match(), not all of which increase
2641 the depth of recursion. Limiting the recursion depth limits the amount of
2642 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2643 when PCRE is compiled, and changed at run time. A patch from Google adds
2644 this functionality to the C++ interface.
2645
2646 18. Changes to the handling of Unicode character properties:
2647
2648 (a) Updated the table to Unicode 4.1.0.
2649
2650 (b) Recognize characters that are not in the table as "Cn" (undefined).
2651
2652 (c) I revised the way the table is implemented to a much improved format
2653 which includes recognition of ranges. It now supports the ranges that
2654 are defined in UnicodeData.txt, and it also amalgamates other
2655 characters into ranges. This has reduced the number of entries in the
2656 table from around 16,000 to around 3,000, thus reducing its size
2657 considerably. I realized I did not need to use a tree structure after
2658 all - a binary chop search is just as efficient. Having reduced the
2659 number of entries, I extended their size from 6 bytes to 8 bytes to
2660 allow for more data.
2661
2662 (d) Added support for Unicode script names via properties such as \p{Han}.
2663
2664 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2665 matching that character.
2666
2667 20. When matching a repeated Unicode property with a minimum greater than zero,
2668 (for example \pL{2,}), PCRE could look past the end of the subject if it
2669 reached it while seeking the minimum number of characters. This could
2670 happen only if some of the characters were more than one byte long, because
2671 there is a check for at least the minimum number of bytes.
2672
2673 21. Refactored the implementation of \p and \P so as to be more general, to
2674 allow for more different types of property in future. This has changed the
2675 compiled form incompatibly. Anybody with saved compiled patterns that use
2676 \p or \P will have to recompile them.
2677
2678 22. Added "Any" and "L&" to the supported property types.
2679
2680 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2681 but give a compile time error if the value is greater than 0xff.
2682
2683 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2684 accidentally not being installed or uninstalled.
2685
2686 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2687 made were to insert the current release number. This seemed silly, because
2688 it made things harder for people building PCRE on systems that don't run
2689 "configure". I have turned pcre.h into a distributed file, no longer built
2690 by "configure", with the version identification directly included. There is
2691 no longer a pcre.h.in file.
2692
2693 However, this change necessitated a change to the pcre-config script as
2694 well. It is built from pcre-config.in, and one of the substitutions was the
2695 release number. I have updated configure.ac so that ./configure now finds
2696 the release number by grepping pcre.h.
2697
2698 26. Added the ability to run the tests under valgrind.
2699
2700
2701 Version 6.4 05-Sep-05
2702 ---------------------
2703
2704 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2705 "--" to be printed when multiple files were scanned, even when none of the
2706 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2707 consider it to be a bug, and have restored the previous behaviour.
2708
2709 2. A couple of code tidies to get rid of compiler warnings.
2710
2711 3. The pcretest program used to cheat by referring to symbols in the library
2712 whose names begin with _pcre_. These are internal symbols that are not
2713 really supposed to be visible externally, and in some environments it is
2714 possible to suppress them. The cheating is now confined to including
2715 certain files from the library's source, which is a bit cleaner.
2716
2717 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2718 file's purpose clearer.
2719
2720 5. Reorganized pcre_ucp_findchar().
2721
2722
2723 Version 6.3 15-Aug-05
2724 ---------------------
2725
2726 1. The file libpcre.pc.in did not have general read permission in the tarball.
2727
2728 2. There were some problems when building without C++ support:
2729
2730 (a) If C++ support was not built, "make install" and "make test" still
2731 tried to test it.
2732
2733 (b) There were problems when the value of CXX was explicitly set. Some
2734 changes have been made to try to fix these, and ...
2735
2736 (c) --disable-cpp can now be used to explicitly disable C++ support.
2737
2738 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2739 backslash in a target when C++ was disabled. This confuses some
2740 versions of "make", apparently. Using an intermediate variable solves
2741 this. (Same for CPP_LOBJ.)
2742
2743 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2744 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2745 necessary on certain architectures.
2746
2747 4. Added a setting of -export-symbols-regex to the link command to remove
2748 those symbols that are exported in the C sense, but actually are local
2749 within the library, and not documented. Their names all begin with
2750 "_pcre_". This is not a perfect job, because (a) we have to except some
2751 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2752 available (and never for static libraries). I have made a note to try to
2753 find a way round (a) in the future.
2754
2755
2756 Version 6.2 01-Aug-05
2757 ---------------------
2758
2759 1. There was no test for integer overflow of quantifier values. A construction
2760 such as {1111111111111111} would give undefined results. What is worse, if
2761 a minimum quantifier for a parenthesized subpattern overflowed and became
2762 negative, the calculation of the memory size went wrong. This could have
2763 led to memory overwriting.
2764
2765 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2766
2767 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2768 operating environments where this matters.
2769
2770 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2771 PCRE options from within the C++ wrapper.
2772
2773 5. Named capturing subpatterns were not being correctly counted when a pattern
2774 was compiled. This caused two problems: (a) If there were more than 100
2775 such subpatterns, the calculation of the memory needed for the whole
2776 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2777 back references of the form \12, where the number was greater than 9, were
2778 not recognized as back references, even though there were sufficient
2779 previous subpatterns.
2780
2781 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2782 versions of gcc, e.g. 2.95.4.
2783
2784
2785 Version 6.1 21-Jun-05
2786 ---------------------
2787
2788 1. There was one reference to the variable "posix" in pcretest.c that was not
2789 surrounded by "#if !defined NOPOSIX".
2790
2791 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2792 the cross-check on the old pcre_info() function, for the benefit of the
2793 cut-down version of PCRE that is currently imported into Exim.
2794
2795 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2796 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2797 patterns that start (?i) but I don't think that matters. The use of (?i) is
2798 just an example; this all applies to the other options as well.
2799
2800 4. Since libtool seems to echo the compile commands it is issuing, the output
2801 from "make" can be reduced a bit by putting "@" in front of each libtool
2802 compile command.
2803
2804 5. Patch from the folks at Google for configure.in to be a bit more thorough
2805 in checking for a suitable C++ installation before trying to compile the
2806 C++ stuff. This should fix a reported problem when a compiler was present,
2807 but no suitable headers.
2808
2809 6. The man pages all had just "PCRE" as their title. I have changed them to
2810 be the relevant file name. I have also arranged that these names are
2811 retained in the file doc/pcre.txt, which is a concatenation in text format
2812 of all the man pages except the little individual ones for each function.
2813
2814 7. The NON-UNIX-USE file had not been updated for the different set of source
2815 files that come with release 6. I also added a few comments about the C++
2816 wrapper.
2817
2818
2819 Version 6.0 07-Jun-05
2820 ---------------------
2821
2822 1. Some minor internal re-organization to help with my DFA experiments.
2823
2824 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2825 didn't matter for the library itself when fully configured, but did matter
2826 when compiling without UCP support, or within Exim, where the ucp files are
2827 not imported.
2828
2829 3. Refactoring of the library code to split up the various functions into
2830 different source modules. The addition of the new DFA matching code (see
2831 below) to a single monolithic source would have made it really too
2832 unwieldy, quite apart from causing all the code to be include in a
2833 statically linked application, when only some functions are used. This is
2834 relevant even without the DFA addition now that patterns can be compiled in
2835 one application and matched in another.
2836
2837 The downside of splitting up is that there have to be some external
2838 functions and data tables that are used internally in different modules of
2839 the library but which are not part of the API. These have all had their
2840 names changed to start with "_pcre_" so that they are unlikely to clash
2841 with other external names.
2842
2843 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2844 a different (DFA) algorithm. Although it is slower than the original
2845 function, it does have some advantages for certain types of matching
2846 problem.
2847
2848 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2849 including restarting after a partial match.
2850
2851 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2852 defined when compiling for Windows was sent to me. I have put it into the
2853 code, though I have no means of testing or verifying it.
2854
2855 7. Added the pcre_refcount() auxiliary function.
2856
2857 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2858 match before or at the first newline in the subject string. In pcretest,
2859 the /f option on a pattern can be used to set this.
2860
2861 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2862 would behave wrongly. This has been present in PCRE since release 4.0.
2863
2864 10. A number of changes to the pcregrep command:
2865
2866 (a) Refactored how -x works; insert ^(...)$ instead of setting
2867 PCRE_ANCHORED and checking the length, in preparation for adding
2868 something similar for -w.
2869
2870 (b) Added the -w (match as a word) option.
2871
2872 (c) Refactored the way lines are read and buffered so as to have more
2873 than one at a time available.
2874
2875 (d) Implemented a pcregrep test script.
2876
2877 (e) Added the -M (multiline match) option. This allows patterns to match
2878 over several lines of the subject. The buffering ensures that at least
2879 8K, or the rest of the document (whichever is the shorter) is available
2880 for matching (and similarly the previous 8K for lookbehind assertions).
2881
2882 (f) Changed the --help output so that it now says
2883
2884 -w, --word-regex(p)
2885
2886 instead of two lines, one with "regex" and the other with "regexp"
2887 because that confused at least one person since the short forms are the
2888 same. (This required a bit of code, as the output is generated
2889 automatically from a table. It wasn't just a text change.)
2890
2891 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2892 option but starts with a hyphen. Could be a pattern or a path name
2893 starting with a hyphen, for instance.
2894
2895 (h) "-" can be given as a file name to represent stdin.
2896
2897 (i) When file names are being printed, "(standard input)" is used for
2898 the standard input, for compatibility with GNU grep. Previously
2899 "<stdin>" was used.
2900
2901 (j) The option --label=xxx can be used to supply a name to be used for
2902 stdin when file names are being printed. There is no short form.
2903
2904 (k) Re-factored the options decoding logic because we are going to add
2905 two more options that take data. Such options can now be given in four
2906 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2907
2908 (l) Added the -A, -B, and -C options for requesting that lines of context
2909 around matches be printed.
2910
2911 (m) Added the -L option to print the names of files that do not contain
2912 any matching lines, that is, the complement of -l.
2913
2914 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2915 continue to scan other files.
2916
2917 (o) The -s option was incorrectly implemented. For compatibility with other
2918 greps, it now suppresses the error message for a non-existent or non-
2919 accessible file (but not the return code). There is a new option called
2920 -q that suppresses the output of matching lines, which was what -s was
2921 previously doing.
2922
2923 (p) Added --include and --exclude options to specify files for inclusion
2924 and exclusion when recursing.
2925
2926 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2927 Hopefully, it now does.
2928
2929 12. Missing cast in pcre_study().
2930
2931 13. Added an "uninstall" target to the makefile.
2932
2933 14. Replaced "extern" in the function prototypes in Makefile.in with
2934 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2935 world, but is set differently for Windows.
2936
2937 15. Added a second compiling function called pcre_compile2(). The only
2938 difference is that it has an extra argument, which is a pointer to an
2939 integer error code. When there is a compile-time failure, this is set
2940 non-zero, in addition to the error test pointer being set to point to an
2941 error message. The new argument may be NULL if no error number is required
2942 (but then you may as well call pcre_compile(), which is now just a
2943 wrapper). This facility is provided because some applications need a
2944 numeric error indication, but it has also enabled me to tidy up the way
2945 compile-time errors are handled in the POSIX wrapper.
2946
2947 16. Added VPATH=.libs to the makefile; this should help when building with one
2948 prefix path and installing with another. (Or so I'm told by someone who
2949 knows more about this stuff than I do.)
2950
2951 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2952 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2953 match everything, including newlines. This is not POSIX-compatible, but
2954 somebody wanted the feature. From pcretest it can be activated by using
2955 both the P and the s flags.
2956
2957 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2958
2959 19. libpcre.pc was being incorrectly installed as executable.
2960
2961 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2962 it now also looks for '\r' so that it will work unmodified on Windows.
2963
2964 21. Added Google's contributed C++ wrapper to the distribution.
2965
2966 22. Added some untidy missing memory free() calls in pcretest, to keep
2967 Electric Fence happy when testing.
2968
2969
2970
2971 Version 5.0 13-Sep-04
2972 ---------------------
2973
2974 1. Internal change: literal characters are no longer packed up into items
2975 containing multiple characters in a single byte-string. Each character
2976 is now matched using a separate opcode. However, there may be more than one
2977 byte in the character in UTF-8 mode.
2978
2979 2. The pcre_callout_block structure has two new fields: pattern_position and
2980 next_item_length. These contain the offset in the pattern to the next match
2981 item, and its length, respectively.
2982
2983 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2984 insertion of callouts before each pattern item. Added the /C option to
2985 pcretest to make use of this.
2986
2987 4. On the advice of a Windows user, the lines
2988
2989 #if defined(_WIN32) || defined(WIN32)
2990 _setmode( _fileno( stdout ), 0x8000 );
2991 #endif /* defined(_WIN32) || defined(WIN32) */
2992
2993 have been added to the source of pcretest. This apparently does useful
2994 magic in relation to line terminators.
2995
2996 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2997 for the benefit of those environments where the "b" makes a difference.
2998
2999 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3000 to know about it. I have put a hack into configure.in that adds in code
3001 to set GCC=yes if CC=icc. This seems to end up at a point in the
3002 generated configure script that is early enough to affect the setting of
3003 compiler options, which is what is needed, but I have no means of testing
3004 whether it really works. (The user who reported this had patched the
3005 generated configure script, which of course I cannot do.)
3006
3007 LATER: After change 22 below (new libtool files), the configure script
3008 seems to know about icc (and also ecc). Therefore, I have commented out
3009 this hack in configure.in.
3010
3011 7. Added support for pkg-config (2 patches were sent in).
3012
3013 8. Negated POSIX character classes that used a combination of internal tables
3014 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3015 [[:^ascii]]. Typically, they would match almost any characters. The other
3016 POSIX classes were not broken in this way.
3017
3018 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3019 to find the match, as PCRE was deluded into thinking that the match had to
3020 start at the start point or following a newline. The same bug applied to
3021 patterns with negative forward assertions or any backward assertions
3022 preceding ".*" at the start, unless the pattern required a fixed first
3023 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3024
3025 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3026 starting at the last subject character, bytes beyond the end of the subject
3027 string were read.
3028
3029 11. Renamed the variable "class" as "classbits" to make life easier for C++
3030 users. (Previously there was a macro definition, but it apparently wasn't
3031 enough.)
3032
3033 12. Added the new field "tables" to the extra data so that tables can be passed
3034 in at exec time, or the internal tables can be re-selected. This allows
3035 a compiled regex to be saved and re-used at a later time by a different
3036 program that might have everything at different addresses.
3037
3038 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3039 -R library as well as a -L library.
3040
3041 14. The debugging options of pcretest (-d on the command line or D on a
3042 pattern) showed incorrect output for anything following an extended class
3043 that contained multibyte characters and which was followed by a quantifier.
3044
3045 15. Added optional support for general category Unicode character properties
3046 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3047 support. It adds about 90K to the size of the library. The meanings of the
3048 inbuilt class escapes such as \d and \s have NOT been changed.
3049
3050 16. Updated pcredemo.c to include calls to free() to release the memory for the
3051 compiled pattern.
3052
3053 17. The generated file chartables.c was being created in the source directory
3054 instead of in the building directory. This caused the build to fail if the
3055 source directory was different from the building directory, and was
3056 read-only.
3057
3058 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3059 file. No doubt somebody will tell me if they don't make sense... Also added
3060 Dan Mooney's comments about building on OpenVMS.
3061
3062 19. Added support for partial matching via the PCRE_PARTIAL option for
3063 pcre_exec() and the \P data escape in pcretest.
3064
3065 20. Extended pcretest with 3 new pattern features:
3066
3067 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3068 write the compiled pattern to the file whose name is "rest-of-line".
3069 This is a straight binary dump of the data, with the saved pointer to
3070 the character tables forced to be NULL. The study data, if any, is
3071 written too. After writing, pcretest reads a new pattern.
3072
3073 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3074 compiled pattern from the given file. There must not be any
3075 occurrences of "<" in the file name (pretty unlikely); if there are,
3076 pcretest will instead treat the initial "<" as a pattern delimiter.
3077 After reading in the pattern, pcretest goes on to read data lines as
3078 usual.
3079
3080 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3081 and 16-bit fields in a compiled pattern, to simulate a pattern that
3082 was compiled on a host of opposite endianness.
3083
3084 21. The pcre-exec() function can now cope with patterns that were compiled on
3085 hosts of opposite endianness, with this restriction:
3086
3087 As for any compiled expression that is saved and used later, the tables
3088 pointer field cannot be preserved; the extra_data field in the arguments
3089 to pcre_exec() should be used to pass in a tables address if a value
3090 other than the default internal tables were used at compile time.
3091
3092 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3093 now diagnosed as an error. Previously, most of the time, a negative number
3094 would have been treated as zero, but if in addition "ovector" was passed as
3095 NULL, a crash could occur.
3096
3097 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3098 new versions from the libtool 1.5 distribution (the last one is a copy of
3099 a file called libtool.m4). This seems to have fixed the need to patch
3100 "configure" to support Darwin 1.3 (which I used to do). However, I still
3101 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3102 workstation).
3103
3104 24. Changed the PCRE licence to be the more standard "BSD" licence.
3105
3106
3107 Version 4.5 01-Dec-03
3108 ---------------------
3109
3110 1. There has been some re-arrangement of the code for the match() function so
3111 that it can be compiled in a version that does not call itself recursively.
3112 Instead, it keeps those local variables that need separate instances for
3113 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3114 needs to "recurse". Keeping track of where control must go is done by means
3115 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3116 hide most of the details from the main code, and operates only if
3117 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3118 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3119 operating.
3120
3121 To make it easier for callers to provide specially tailored get/free
3122 functions for this usage, two new functions, pcre_stack_malloc, and
3123 pcre_stack_free, are used. They are always called in strict stacking order,
3124 and the size of block requested is always the same.
3125
3126 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3127 PCRE has been compiled to use the stack or the heap for recursion. The
3128 -C option of pcretest uses this to show which version is compiled.
3129
3130 A new data escape \S, is added to pcretest; it causes the amounts of store
3131 obtained and freed by both kinds of malloc/free at match time to be added
3132 to the output.
3133
3134 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3135 what's available on my current Linux desktop machine.
3136
3137 3. When matching a UTF-8 string, the test for a valid string at the start has
3138 been extended. If start_offset is not zero, PCRE now checks that it points
3139 to a byte that is the start of a UTF-8 character. If not, it returns
3140 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3141 this is necessary because there may be backward assertions in the pattern.
3142 When matching the same subject several times, it may save resources to use
3143 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3144
3145 4. The code for checking the validity of UTF-8 strings has been tightened so
3146 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3147 containing "overlong sequences".
3148
3149 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3150 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3151 should have been "&", but it just so happened that all the cases this let
3152 through by mistake were picked up later in the function.
3153
3154 6. I had used a variable called "isblank" - this is a C99 function, causing
3155 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3156
3157 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3158 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3159 so the version is shown; (c) stop "make test" echoing "./RunTest".
3160
3161 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3162
3163 9. The prototype for memmove() for systems that don't have it was using
3164 size_t, but the inclusion of the header that defines size_t was later. I've
3165 moved the #includes for the C headers earlier to avoid this.
3166
3167 10. Added some adjustments to the code to make it easier to compiler on certain
3168 special systems:
3169
3170 (a) Some "const" qualifiers were missing.
3171 (b) Added the macro EXPORT before all exported functions; by default this
3172 is defined to be empty.
3173 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3174 that it reads its output file name as an argument instead of writing
3175 to the standard output and assuming this can be redirected.
3176
3177 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3178 class containing characters with values greater than 255, PCRE compilation
3179 went into a loop.
3180
3181 12. A recursive reference to a subpattern that was within another subpattern
3182 that had a minimum quantifier of zero caused PCRE to crash. For example,
3183 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3184 recursion. If the recursively-called subpattern itself had a zero repeat,
3185 that was OK.
3186
3187 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3188 buffer into which it was copied (for escape processing) was still set at
3189 1024, so long lines caused crashes.
3190
3191 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3192 "internal error: code overflow...". This applied to any character class
3193 that was followed by a possessive quantifier.
3194
3195 15. Modified the Makefile to add libpcre.la as a prerequisite for
3196 libpcreposix.la because I was told this is needed for a parallel build to
3197 work.
3198
3199 16. If a pattern that contained .* following optional items at the start was
3200 studied, the wrong optimizing data was generated, leading to matching
3201 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3202 matching string must start with a or b or c. The correct conclusion for
3203 this pattern is that a match can start with any character.
3204
3205
3206 Version 4.4 13-Aug-03
3207 ---------------------
3208
3209 1. In UTF-8 mode, a character class containing characters with values between
3210 127 and 255 was not handled correctly if the compiled pattern was studied.
3211 In fixing this, I have also improved the studying algorithm for such
3212 classes (slightly).
3213
3214 2. Three internal functions had redundant arguments passed to them. Removal
3215 might give a very teeny performance improvement.
3216
3217 3. Documentation bug: the value of the capture_top field in a callout is *one
3218 more than* the number of the hightest numbered captured substring.
3219
3220 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3221 in incorrectly linking with a previously installed version. They now link
3222 explicitly with libpcre.la.
3223
3224 5. configure.in no longer needs to recognize Cygwin specially.
3225
3226 6. A problem in pcre.in for Windows platforms is fixed.
3227
3228 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3229 pcretest, it used to include the size of the study block as part of its
3230 output. Unfortunately, the structure contains a field that has a different
3231 size on different hardware architectures. This meant that the tests that
3232 showed this size failed. As the block is currently always of a fixed size,
3233 this information isn't actually particularly useful in pcretest output, so
3234 I have just removed it.
3235
3236 8. Three pre-processor statements accidentally did not start in column 1.
3237 Sadly, there are *still* compilers around that complain, even though
3238 standard C has not required this for well over a decade. Sigh.
3239
3240 9. In pcretest, the code for checking callouts passed small integers in the
3241 callout_data field, which is a void * field. However, some picky compilers
3242 complained about the casts involved for this on 64-bit systems. Now
3243 pcretest passes the address of the small integer instead, which should get
3244 rid of the warnings.
3245
3246 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3247 both compile and run time, and gives an error if an invalid UTF-8 sequence
3248 is found. There is a option for disabling this check in cases where the
3249 string is known to be correct and/or the maximum performance is wanted.
3250
3251 11. In response to a bug report, I changed one line in Makefile.in from
3252
3253 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3254 to
3255 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3256
3257 to look similar to other lines, but I have no way of telling whether this
3258 is the right thing to do, as I do not use Windows. No doubt I'll get told
3259 if it's wrong...
3260
3261
3262 Version 4.3 21-May-03
3263 ---------------------
3264
3265 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3266 Makefile.
3267
3268 2. Some refactoring to improve the quality of the code:
3269
3270 (i) The utf8_table... variables are now declared "const".
3271
3272 (ii) The code for \cx, which used the "case flipping" table to upper case
3273 lower case letters, now just substracts 32. This is ASCII-specific,
3274 but the whole concept of \cx is ASCII-specific, so it seems
3275 reasonable.
3276
3277 (iii) PCRE was using its character types table to recognize decimal and
3278 hexadecimal digits in the pattern. This is silly, because it handles
3279 only 0-9, a-f, and A-F, but the character types table is locale-
3280 specific, which means strange things might happen. A private
3281 table is now used for this - though it costs 256 bytes, a table is
3282 much faster than multiple explicit tests. Of course, the standard
3283 character types table is still used for matching digits in subject
3284 strings against \d.
3285
3286 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3287 ending in _t are). So I've renamed it as ESC_tee.
3288
3289 3. The first argument for regexec() in the POSIX wrapper should have been
3290 defined as "const".
3291
3292 4. Changed pcretest to use malloc() for its buffers so that they can be
3293 Electric Fenced for debugging.
3294
3295 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3296 to read one or more bytes before the start of the subject string. Often this
3297 had no effect on PCRE's behaviour, but in some circumstances it could
3298 provoke a segmentation fault.
3299
3300 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3301 to try to read one or more bytes before the start of the subject string.
3302
3303 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3304 UTF-8 support could misbehave in various ways if the subject string
3305 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3306 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3307 back over UTF-8 characters.)
3308
3309
3310 Version 4.2 14-Apr-03
3311 ---------------------
3312
3313 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3314
3315 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3316 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3317 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3318 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3319 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3320 and BUILD_EXEEXT
3321 Note: automatic setting of the BUILD variables is not yet working
3322 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3323 compile-time but not at link-time
3324 [LINK]: use for linking executables only
3325 make different versions for Windows and non-Windows
3326 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3327 libraries
3328 [LINK_FOR_BUILD]: new variable
3329 [OBJEXT]: use throughout
3330 [EXEEXT]: use throughout
3331 <winshared>: new target
3332 <wininstall>: new target
3333 <dftables.o>: use native compiler
3334 <dftables>: use native linker
3335 <install>: handle Windows platform correctly
3336 <clean>: ditto
3337 <check>: ditto
3338 copy DLL to top builddir before testing
3339
3340 As part of these changes, -no-undefined was removed again. This was reported
3341 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3342 in any case.
3343
3344 3. Some tidies to get rid of compiler warnings:
3345
3346 . In the match_data structure, match_limit was an unsigned long int, whereas
3347 match_call_count was an int. I've made them both unsigned long ints.
3348
3349 . In pcretest the fact that a const uschar * doesn't automatically cast to
3350 a void * provoked a warning.
3351
3352 . Turning on some more compiler warnings threw up some "shadow" variables
3353 and a few more missing casts.
3354
3355 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3356 option, a class that contained a single character with a value between 128
3357 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3358
3359 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3360 option, a class that contained several characters, but with at least one
3361 whose value was between 128 and 255 caused PCRE to crash.
3362
3363
3364 Version 4.1 12-Mar-03
3365 ---------------------
3366
3367 1. Compiling with gcc -pedantic found a couple of places where casts were
3368 needed, and a string in dftables.c that was longer than standard compilers are
3369 required to support.
3370
3371 2. Compiling with Sun's compiler found a few more places where the code could
3372 be tidied up in order to avoid warnings.
3373
3374 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3375 first of these names is deprecated in the latest Autoconf in favour of the name
3376 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3377 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3378 analogy I have changed it to CFLAGS_FOR_BUILD.
3379
3380 4. Added -no-undefined to the linking command in the Makefile, because this is
3381 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3382 linking step for the pcreposix library.
3383
3384 5. PCRE was failing to diagnose the case of two named groups with the same
3385 name.
3386
3387 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3388 literal character that is needed in the subject for a match, and scans along to
3389 ensure that it is present before embarking on the full matching process. This
3390 saves time in cases of nested unlimited repeats that are never going to match.
3391 Problem: the scan can take a lot of time if the subject is very long (e.g.
3392 megabytes), thus penalizing straightforward matches. It is now done only if the
3393 amount of subject to be scanned is less than 1000 bytes.
3394
3395 7. A lesser problem with the same optimization is that it was recording the
3396 first character of an anchored pattern as "needed", thus provoking a search
3397 right along the subject, even when the first match of the pattern was going to
3398 fail. The "needed" character is now not set for anchored patterns, unless it
3399 follows something in the pattern that is of non-fixed length. Thus, it still
3400 fulfils its original purpose of finding quick non-matches in cases of nested
3401 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3402
3403
3404 Version 4.0 17-Feb-03
3405 ---------------------
3406
3407 1. If a comment in an extended regex that started immediately after a meta-item
3408 extended to the end of string, PCRE compiled incorrect data. This could lead to
3409 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3410
3411 2. Moved to autoconf 2.53 and libtool 1.4.2.
3412
3413 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3414 the special perltest8 script is no longer needed - all the tests can be run
3415 from a single perltest script.
3416
3417 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3418 by \s. It has now been removed in PCRE. This means it isn't recognized as
3419 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3420 class [:space:] *does* include VT, thereby creating a mess.
3421
3422 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3423 space and tab.
3424
3425 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3426 its new features into the main test script, reducing the number of scripts.
3427
3428 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3429 were backward compatible, and made the (?i) apply to the whole pattern, as if
3430 /i were given. Now it behaves more logically, and applies the option setting
3431 only to what follows. PCRE has been changed to follow suit. However, if it
3432 finds options settings right at the start of the pattern, it extracts them into
3433 the global options, as before. Thus, they show up in the info data.
3434
3435 8. Added support for the \Q...\E escape sequence. Characters in between are
3436 treated as literals. This is slightly different from Perl in that $ and @ are
3437 also handled as literals inside the quotes. In Perl, they will cause variable
3438 interpolation. Note the following examples:
3439
3440 Pattern PCRE matches Perl matches
3441
3442 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3443 \Qabc\$xyz\E abc\$xyz abc\$xyz
3444 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3445
3446 For compatibility with Perl, \Q...\E sequences are recognized inside character
3447 classes as well as outside them.
3448
3449 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3450 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3451 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3452 signed/unsigned warnings.
3453
3454 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3455 option for pcretest, so I've replaced it by a simple function that does just
3456 that job.
3457
3458 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3459 "pcregrep -".
3460
3461 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3462 Java package. This provides some syntactic sugar for simple cases of what my
3463 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3464 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3465 item, you can use this simplified notation. Note that only makes sense with
3466 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3467 greediness, whatever the setting of the PCRE_UNGREEDY option.
3468
3469 13. A change of greediness default within a pattern was not taking effect at
3470 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3471 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3472 was abstracted outside.
3473
3474 14. PCRE now supports the \G assertion. It is true when the current matching
3475 position is at the start point of the match. This differs from \A when the
3476 starting offset is non-zero. Used with the /g option of pcretest (or similar
3477 code), it works in the same way as it does for Perl's /g option. If all
3478 alternatives of a regex begin with \G, the expression is anchored to the start
3479 match position, and the "anchored" flag is set in the compiled expression.
3480
3481 15. Some bugs concerning the handling of certain option changes within patterns
3482 have been fixed. These applied to options other than (?ims). For example,
3483 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3484 been the other way round. Some of this was related to change 7 above.
3485
3486 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3487 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3488 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3489 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3490
3491 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3492 mode. Unlike ".", it always matches newline, whatever the setting of
3493 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3494 assertions. Perl allows it, but it doesn't (in general) work because it can't
3495 calculate the length of the lookbehind. At least, that's the case for Perl
3496 5.8.0 - I've been told they are going to document that it doesn't work in
3497 future.
3498
3499 18. Added an error diagnosis for escapes that PCRE does not support: these are
3500 \L, \l, \N, \P, \p, \U, \u, and \X.
3501
3502 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3503 reading past the end of the pattern in cases such as /[abcd/.
3504
3505 20. PCRE was getting more memory than necessary for patterns with classes that
3506 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3507
3508 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3509 compiling PCRE for use with Virtual Pascal.
3510
3511 22. Small fix to the Makefile to make it work properly if the build is done
3512 outside the source tree.
3513
3514 23. Added a new extension: a condition to go with recursion. If a conditional
3515 subpattern starts with (?(R) the "true" branch is used if recursion has
3516 happened, whereas the "false" branch is used only at the top level.
3517
3518 24. When there was a very long string of literal characters (over 255 bytes
3519 without UTF support, over 250 bytes with UTF support), the computation of how
3520 much memory was required could be incorrect, leading to segfaults or other
3521 strange effects.
3522
3523 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3524 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3525 there was a subsequent back reference to those brackets. This meant that, for
3526 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3527 possible to check for precisely this case. All we can do is abandon the
3528 optimization if .* occurs inside capturing brackets when there are any back
3529 references whatsoever. (See below for a better fix that came later.)
3530
3531 26. The handling of the optimization for finding the first character of a
3532 non-anchored pattern, and for finding a character that is required later in the
3533 match were failing in some cases. This didn't break the matching; it just
3534 failed to optimize when it could. The way this is done has been re-implemented.
3535
3536 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3537
3538 28. Added a new feature that provides some of the functionality that Perl
3539 provides with (?{...}). The facility is termed a "callout". The way it is done
3540 in PCRE is for the caller to provide an optional function, by setting
3541 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3542 global variable. By default it is unset, which disables all calling out. To get
3543 the function called, the regex must include (?C) at appropriate points. This
3544 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3545 This provides a means of identifying different callout points. When PCRE
3546 reaches such a point in the regex, if pcre_callout has been set, the external
3547 function is called. It is provided with data in a structure called
3548 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3549 matching continues; if it returns a non-zero value, the match at the current
3550 point fails. However, backtracking will occur if possible. [This was changed
3551 later and other features added - see item 49 below.]
3552
3553 29. pcretest is upgraded to test the callout functionality. It provides a
3554 callout function that displays information. By default, it shows the start of
3555 the match and the current position in the text. There are some new data escapes
3556 to vary what happens:
3557
3558 \C+ in addition, show current contents of captured substrings
3559 \C- do not supply a callout function
3560 \C!n return 1 when callout number n is reached
3561 \C!n!m return 1 when callout number n is reached for the mth time
3562
3563 30. If pcregrep was called with the -l option and just a single file name, it
3564 output "<stdin>" if a match was found, instead of the file name.
3565
3566 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3567 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3568 pcre_exec(). This saves a malloc/free per call. The default value of
3569 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3570 when configuring.
3571
3572 32. The default maximum size of a compiled pattern is 64K. There have been a
3573 few cases of people hitting this limit. The code now uses macros to handle the
3574 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3575 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3576 configuring. Tests 2 and 5 work only with 2-byte links because they output
3577 debugging information about compiled patterns.
3578
3579 33. Internal code re-arrangements:
3580
3581 (a) Moved the debugging function for printing out a compiled regex into
3582 its own source file (printint.c) and used #include to pull it into
3583 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3584 separate copies.
3585
3586 (b) Defined the list of op-code names for debugging as a macro in
3587 internal.h so that it is next to the definition of the opcodes.
3588
3589 (c) Defined a table of op-code lengths for simpler skipping along compiled
3590 code. This is again a macro in internal.h so that it is next to the
3591 definition of the opcodes.
3592
3593 34. Added support for recursive calls to individual subpatterns, along the
3594 lines of Robin Houston's patch (but implemented somewhat differently).
3595
3596 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3597 allow it to read and process whole directories in Win32. This code was
3598 contributed by Lionel Fourquaux; it has not been tested by me.
3599
3600 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3601 used to name a group. Names consist of alphanumerics and underscores, and must
3602 be unique. Back references use the syntax (?P=name) and recursive calls use
3603 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3604 numbers. The function pcre_fullinfo() can be used after compilation to extract
3605 a name/number map. There are three relevant calls:
3606
3607 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3608 PCRE_INFO_NAMECOUNT yields the number of entries
3609 PCRE_INFO_NAMETABLE yields a pointer to the map.
3610
3611 The map is a vector of fixed-size entries. The size of each entry depends on
3612 the length of the longest name used. The first two bytes of each entry are the
3613 group number, most significant byte first. There follows the corresponding
3614 name, zero terminated. The names are in alphabetical order.
3615
3616 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3617 case instead of 255. Making it the same both with and without UTF-8 support
3618 means that the same test output works with both.
3619
3620 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3621 calling malloc() with a zero argument.
3622
3623 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3624 optimization. I've improved things by keeping a bitmap of backreferences with
3625 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3626 fact referenced, the optimization can be applied. It is unlikely that a
3627 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3628 the match to follow \n) will appear inside brackets with a number greater than
3629 31, but if it does, any back reference > 31 suppresses the optimization.
3630
3631 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3632 of disabling numbered capturing parentheses. Any opening parenthesis that is
3633 not followed by ? behaves as if it were followed by ?: but named parentheses
3634 can still be used for capturing (and they will acquire numbers in the usual
3635 way).
3636
3637 41. Redesigned the return codes from the match() function into yes/no/error so
3638 that errors can be passed back from deep inside the nested calls. A malloc
3639 failure while inside a recursive subpattern call now causes the
3640 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3641
3642 42. It is now possible to set a limit on the number of times the match()
3643 function is called in a call to pcre_exec(). This facility makes it possible to
3644 limit the amount of recursion and backtracking, though not in a directly
3645 obvious way, because the match() function is used in a number of different
3646 circumstances. The count starts from zero for each position in the subject
3647 string (for non-anchored patterns). The default limit is, for compatibility, a
3648 large number, namely 10 000 000. You can change this in two ways:
3649
3650 (a) When configuring PCRE before making, you can use --with-match-limit=n
3651 to set a default value for the compiled library.
3652
3653 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3654 a different value is set. See 45 below.
3655
3656 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3657
3658 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3659 of things that can be changed at compile time. The first argument specifies
3660 what is wanted and the second points to where the information is to be placed.
3661 The current list of available information is:
3662
3663 PCRE_CONFIG_UTF8
3664
3665 The output is an integer that is set to one if UTF-8 support is available;
3666 otherwise it is set to zero.
3667
3668 PCRE_CONFIG_NEWLINE
3669
3670 The output is an integer that it set to the value of the code that is used for
3671 newline. It is either LF (10) or CR (13).
3672
3673 PCRE_CONFIG_LINK_SIZE
3674
3675 The output is an integer that contains the number of bytes used for internal
3676 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3677
3678 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3679
3680 The output is an integer that contains the threshold above which the POSIX
3681 interface uses malloc() for output vectors. See item 31 above.
3682
3683 PCRE_CONFIG_MATCH_LIMIT
3684
3685 The output is an unsigned integer that contains the default limit of the number
3686 of match() calls in a pcre_exec() execution. See 42 above.
3687
3688 44. pcretest has been upgraded by the addition of the -C option. This causes it
3689 to extract all the available output from the new pcre_config() function, and to
3690 output it. The program then exits immediately.
3691
3692 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3693 order to support additional features. One way would have been to define
3694 pcre_exec2() (for example) with extra arguments, but this would not have been
3695 extensible, and would also have required all calls to the original function to
3696 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3697 is used for passing in "extra" data from pcre_study().
3698
3699 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3700 contains the following fields:
3701
3702 flags a bitmap indicating which of the following fields are set
3703 study_data opaque data from pcre_study()
3704 match_limit a way of specifying a limit on match() calls for a specific
3705 call to pcre_exec()
3706 callout_data data for callouts (see 49 below)
3707
3708 The flag bits are also defined in pcre.h, and are
3709
3710 PCRE_EXTRA_STUDY_DATA
3711 PCRE_EXTRA_MATCH_LIMIT
3712 PCRE_EXTRA_CALLOUT_DATA
3713
3714 The pcre_study() function now returns one of these new pcre_extra blocks, with
3715 the actual study data pointed to by the study_data field, and the
3716 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3717 before. That is, this change is entirely upwards-compatible and requires no
3718 change to existing code.
3719
3720 If you want to pass in additional data to pcre_exec(), you can either place it
3721 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3722 block.
3723
3724 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3725 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3726 times with different match limits, until it finds the minimum value needed for
3727 pcre_exec() to complete. The value is then output. This can be instructive; for
3728 most simple matches the number is quite small, but for pathological cases it
3729 gets very large very quickly.
3730
3731 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3732 returns the size of the data block pointed to by the study_data field in a
3733 pcre_extra block, that is, the value that was passed as the argument to
3734 pcre_malloc() when PCRE was getting memory in which to place the information
3735 created by pcre_study(). The fourth argument should point to a size_t variable.
3736 pcretest has been extended so that this information is shown after a successful
3737 pcre_study() call when information about the compiled regex is being displayed.
3738
3739 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3740 because what follows is always an absolute path. (Later: it turns out that this
3741 is more than cosmetic for MinGW, because it doesn't like empty path
3742 components.)
3743
3744 49. Some changes have been made to the callout feature (see 28 above):
3745
3746 (i) A callout function now has three choices for what it returns:
3747
3748 0 => success, carry on matching
3749 > 0 => failure at this point, but backtrack if possible
3750 < 0 => serious error, return this value from pcre_exec()
3751
3752 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3753 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3754 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3755 use by callout functions. It will never be used by PCRE itself.
3756
3757 (ii) The pcre_extra structure (see 45 above) has a void * field called
3758 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3759 pcre_callout_block structure has a field of the same name. The contents of
3760 the field passed in the pcre_extra structure are passed to the callout
3761 function in the corresponding field in the callout block. This makes it
3762 easier to use the same callout-containing regex from multiple threads. For
3763 testing, the pcretest program has a new data escape
3764
3765 \C*n pass the number n (may be negative) as callout_data
3766
3767 If the callout function in pcretest receives a non-zero value as
3768 callout_data, it returns that value.
3769
3770 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3771 there were some redundant $(CFLAGS) in commands that are now specified as
3772 $(LINK), which already includes $(CFLAGS).
3773
3774 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3775 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3776 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3777 one-byte characters throughout. Note that case-insensitive matching applies
3778 only to characters whose values are less than 256. PCRE doesn't support the
3779 notion of cases for higher-valued characters.
3780
3781 (i) A character class whose characters are all within 0-255 is handled as
3782 a bit map, and the map is inverted for negative classes. Previously, a
3783 character > 255 always failed to match such a class; however it should
3784 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3785
3786 (ii) A negated character class with a single character < 255 is coded as
3787 "not this character" (OP_NOT). This wasn't working properly when the test
3788 character was multibyte, either singly or repeated.
3789
3790 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3791 mode, for example: \x{100}{2,3}.
3792
3793 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3794 singly or repeated) now correctly test multibyte characters. However,
3795 PCRE doesn't recognize any characters with values greater than 255 as
3796 digits, spaces, or word characters. Such characters always match \D, \S,
3797 and \W, and never match \d, \s, or \w.
3798
3799 (v) Classes may now contain characters and character ranges with values
3800 greater than 255. For example: [ab\x{100}-\x{400}].
3801
3802 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3803 PCRE in UTF-8 mode.
3804
3805 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3806 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3807 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3808 value.)
3809
3810 53. The single man page has become too large. I have therefore split it up into
3811 a number of separate man pages. These also give rise to individual HTML pages;
3812 these are now put in a separate directory, and there is an index.html page that
3813 lists them all. Some hyperlinking between the pages has been installed.
3814
3815 54. Added convenience functions for handling named capturing parentheses.
3816
3817 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3818 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3819 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3820 are faulted.
3821
3822 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3823 calling configure. These values are used when compiling the dftables.c program
3824 which is run to generate the source of the default character tables. They
3825 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3826 you will need to set these values.
3827
3828 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3829
3830
3831 Version 3.9 02-Jan-02
3832 ---------------------
3833
3834 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3835
3836 2. If --disable-static was given, the building process failed when trying to
3837 build pcretest and pcregrep. (For some reason it was using libtool to compile
3838 them, which is not right, as they aren't part of the library.)
3839
3840
3841 Version 3.8 18-Dec-01
3842 ---------------------
3843
3844 1. The experimental UTF-8 code was completely screwed up. It was packing the
3845 bytes in the wrong order. How dumb can you get?
3846
3847
3848 Version 3.7 29-Oct-01
3849 ---------------------
3850
3851 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3852 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3853 this didn't happen under Solaris 8, where I normally test things.
3854
3855 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3856 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3857 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3858 this fix an hour or so after the initial 3.7 release.)
3859
3860
3861 Version 3.6 23-Oct-01
3862 ---------------------
3863
3864 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3865 offsets passed as NULL with zero offset count.
3866
3867 2. The config.guess and config.sub files had not been updated when I moved to
3868 the latest autoconf.
3869
3870
3871 Version 3.5 15-Aug-01
3872 ---------------------
3873
3874 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3875 had been forgotten.
3876
3877 2. By using declared but undefined structures, we can avoid using "void"
3878 definitions in pcre.h while keeping the internal definitions of the structures
3879 private.
3880
3881 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3882 user point of view, this means that both static and shared libraries are built
3883 by default, but this can be individually controlled. More of the work of
3884 handling this static/shared cases is now inside libtool instead of PCRE's make
3885 file.
3886
3887 4. The pcretest utility is now installed along with pcregrep because it is
3888 useful for users (to test regexs) and by doing this, it automatically gets
3889 relinked by libtool. The documentation has been turned into a man page, so
3890 there are now .1, .txt, and .html versions in /doc.
3891
3892 5. Upgrades to pcregrep:
3893 (i) Added long-form option names like gnu grep.
3894 (ii) Added --help to list all options with an explanatory phrase.
3895 (iii) Added -r, --recursive to recurse into sub-directories.
3896 (iv) Added -f, --file to read patterns from a file.
3897
3898 6. pcre_exec() was referring to its "code" argument before testing that
3899 argument for NULL (and giving an error if it was NULL).
3900
3901 7. Upgraded Makefile.in to allow for compiling in a different directory from
3902 the source directory.
3903
3904 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3905 options bits, the pointer it was passed was to an int instead of to an unsigned
3906 long int. This mattered only on 64-bit systems.
3907
3908 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3909 generated) instead of pcre.in, which it its source. Also made the same change
3910 in several of the .c files.
3911
3912 10. A new release of gcc defines printf() as a macro, which broke pcretest
3913 because it had an ifdef in the middle of a string argument for printf(). Fixed
3914 by using separate calls to printf().
3915
3916 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3917 script, to force use of CR or LF instead of \n in the source. On non-Unix
3918 systems, the value can be set in config.h.
3919
3920 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3921 absolute limit. Changed the text of the error message to make this clear, and
3922 likewise updated the man page.
3923
3924 13. The limit of 99 on the number of capturing subpatterns has been removed.
3925 The new limit is 65535, which I hope will not be a "real" limit.
3926
3927
3928 Version 3.4 22-Aug-00
3929 ---------------------
3930
3931 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3932
3933 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3934
3935
3936 Version 3.3 01-Aug-00
3937 ---------------------
3938
3939 1. If an octal character was given, but the value was greater than \377, it
3940 was not getting masked to the least significant bits, as documented. This could
3941 lead to crashes in some systems.
3942
3943 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3944 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3945
3946 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3947 These just pass their arguments on to (pcre_free)(), but they are provided
3948 because some uses of PCRE bind it to non-C systems that can call its functions,
3949 but cannot call free() or pcre_free() directly.
3950
3951 4. Add "make test" as a synonym for "make check". Corrected some comments in
3952 the Makefile.
3953
3954 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3955 Makefile.
3956
3957 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3958 command called pgrep for grepping around the active processes.
3959
3960 7. Added the beginnings of support for UTF-8 character strings.
3961
3962 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3963 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3964 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3965 out for the ar command.)
3966
3967
3968 Version 3.2 12-May-00
3969 ---------------------
3970
3971 This is purely a bug fixing release.
3972
3973 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3974 of ZA. This was just one example of several cases that could provoke this bug,
3975 which was introduced by change 9 of version 2.00. The code for breaking
3976 infinite loops after an iteration that matches an empty string was't working
3977 correctly.
3978
3979 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3980 when matched against abbab (for example). After matching an empty string, it
3981 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3982 caused it to match further down the string than it should.
3983
3984 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3985 was there because it doesn't seem to be needed, and it causes trouble on some
3986 systems, as it is not a Standard C header. It has been removed.
3987
3988 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3989 were reported on the Macintosh. The changes were from
3990
3991 while ((c = *(++ptr)) != 0 && c != '\n');
3992 to
3993 while ((c = *(++ptr)) != 0 && c != '\n') ;
3994
3995 Totally extraordinary, but if that's what it takes...
3996
3997 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3998 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3999 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4000 assumes the way PCRE uses memmove() (always moving upwards).
4001
4002 6. PCRE is being used in one environment where strchr() is not available. There
4003 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4004 faster code anyway.
4005
4006
4007 Version 3.1 09-Feb-00
4008 ---------------------
4009
4010 The only change in this release is the fixing of some bugs in Makefile.in for
4011 the "install" target:
4012
4013 (1) It was failing to install pcreposix.h.
4014
4015 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4016
4017
4018 Version 3.0 01-Feb-00
4019 ---------------------
4020
4021 1. Add support for the /+ modifier to perltest (to output $` like it does in
4022 pcretest).
4023
4024 2. Add support for the /g modifier to perltest.
4025
4026 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4027 matches null strings.
4028
4029 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4030 pattern. Perl treats empty patterns specially - it reuses the most recent
4031 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4032 effect.
4033
4034 5. The POSIX interface was broken in that it was just handing over the POSIX
4035 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4036 required a bigger vector, with some working space on the end. This means that
4037 the POSIX wrapper now has to get and free some memory, and copy the results.
4038
4039 6. Added some simple autoconf support, placing the test data and the
4040 documentation in separate directories, re-organizing some of the
4041 information files, and making it build pcre-config (a GNU standard). Also added
4042 libtool support for building PCRE as a shared library, which is now the
4043 default.
4044
4045 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4046 09 are not valid octal constants. Single digits will be used for minor values
4047 less than 10.
4048
4049 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4050 existing programs that set these in the POSIX interface can use PCRE without
4051 modification.
4052
4053 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4054 return all that pcre_info() returns, plus additional data. The pcre_info()
4055 function is retained for compatibility, but is considered to be obsolete.
4056
4057 10. Added experimental recursion feature (?R) to handle one common case that
4058 Perl 5.6 will be able to do with (?p{...}).
4059
4060 11. Added support for POSIX character classes like [:alpha:], which Perl is
4061 adopting.
4062
4063
4064 Version 2.08 31-Aug-99
4065 ----------------------
4066
4067 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4068 trying to match at the startoffset position, but instead was moving forward to
4069 the next newline as if a previous match had failed.
4070
4071 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4072 and could get into a loop if a null string was matched other than at the start
4073 of the subject.
4074
4075 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4076 be distinguished at compile time, and for completeness also added PCRE_DATE.
4077
4078 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4079 in GnuWin32 environments.
4080
4081
4082 Version 2.07 29-Jul-99
4083 ----------------------
4084
4085 1. The documentation is now supplied in plain text form and HTML as well as in
4086 the form of man page sources.
4087
4088 2. C++ compilers don't like assigning (void *) values to other pointer types.
4089 In particular this affects malloc(). Although there is no problem in Standard
4090 C, I've put in casts to keep C++ compilers happy.
4091
4092 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4093 should be (const char *).
4094
4095 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4096 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4097 However, I haven't made this a standard facility. The documentation doesn't
4098 mention it, and the Makefile doesn't support it.
4099
4100 5. The Makefile now contains an "install" target, with editable destinations at
4101 the top of the file. The pcretest program is not installed.
4102
4103 6. pgrep -V now gives the PCRE version number and date.
4104
4105 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4106 causing the entire string to be ignored, instead of just the last character.
4107
4108 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4109 non-matching string, it can take a very, very long time, even for strings of
4110 quite modest length, because of the nested recursion. PCRE now does better in
4111 some of these cases. It does this by remembering the last required literal
4112 character in the pattern, and pre-searching the subject to ensure it is present
4113 before running the real match. In other words, it applies a heuristic to detect
4114 some types of certain failure quickly, and in the above example, if presented
4115 with a string that has no trailing " it gives "no match" very quickly.
4116
4117 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4118 other alternatives are tried instead.
4119
4120
4121 Version 2.06 09-Jun-99
4122 ----------------------
4123
4124 1. Change pcretest's output for amount of store used to show just the code
4125 space, because the remainder (the data block) varies in size between 32-bit and
4126 64-bit systems.
4127
4128 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4129 start matching at. This allows lookbehinds to work when searching for multiple
4130 occurrences in a string.
4131
4132 3. Added additional options to pcretest for testing multiple occurrences:
4133
4134 /+ outputs the rest of the string that follows a match
4135 /g loops for multiple occurrences, using the new startoffset argument
4136 /G loops for multiple occurrences by passing an incremented pointer
4137
4138 4. PCRE wasn't doing the "first character" optimization for patterns starting
4139 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4140 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4141 the letter 'x'. On long subject strings, this gives a significant speed-up.
4142
4143
4144 Version 2.05 21-Apr-99
4145 ----------------------
4146
4147 1. Changed the type of magic_number from int to long int so that it works
4148 properly on 16-bit systems.
4149
4150 2. Fixed a bug which caused patterns starting with .* not to work correctly
4151 when the subject string contained newline characters. PCRE was assuming
4152 anchoring for such patterns in all cases, which is not correct because .* will
4153 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4154 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4155 must be retried after every newline in the subject.
4156
4157
4158 Version 2.04 18-Feb-99
4159 ----------------------
4160
4161 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4162 computation of the store needed to hold the pattern was incorrect (too large).
4163 If such patterns were nested a few deep, this could multiply and become a real
4164 problem.
4165
4166 2. Added /M option to pcretest to show the memory requirement of a specific
4167 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4168
4169 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4170 compiled in such a way that the backtracking after subsequent failure was
4171 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4172 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4173
4174
4175 Version 2.03 02-Feb-99
4176 ----------------------
4177
4178 1. Fixed typo and small mistake in man page.
4179
4180 2. Added 4th condition (GPL supersedes if conflict) and created separate
4181 LICENCE file containing the conditions.
4182
4183 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4184 Perl, that is the internal \ allows the delimiter to be included in the
4185 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4186 the final delimiter, add \ to the end of the pattern (to test the error).
4187
4188 4. Added the convenience functions for extracting substrings after a successful
4189 match. Updated pcretest to make it able to test these functions.
4190
4191
4192 Version 2.02 14-Jan-99
4193 ----------------------
4194
4195 1. Initialized the working variables associated with each extraction so that
4196 their saving and restoring doesn't refer to uninitialized store.
4197
4198 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4199 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4200 fix the problem.
4201
4202 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4203 calls, and wasn't printing the correct value for compiling calls. Increased the
4204 default value of LOOPREPEAT, and the number of significant figures in the
4205 times.
4206
4207 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4208
4209 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4210 a building problem on Windows NT with a FAT file system.
4211
4212
4213 Version 2.01 21-Oct-98
4214 ----------------------
4215
4216 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4217 to character tables built by pcre_maketables() in the current locale. If NULL
4218 is passed, the default tables are used.
4219
4220
4221 Version 2.00 24-Sep-98
4222 ----------------------
4223
4224 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4225 it any more.
4226
4227 2. Allow quantification of (?>) groups, and make it work correctly.
4228
4229 3. The first character computation wasn't working for (?>) groups.
4230
4231 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4232 end of the subject) and add 5.005's \z, which really does match only at the
4233 very end of the subject.
4234
4235 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4236
4237 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4238 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4239 localized options. All options to pcre_study() were also removed.
4240
4241 7. Add other new features from 5.005:
4242
4243 $(?<= positive lookbehind
4244 $(?<! negative lookbehind
4245 (?imsx-imsx) added the unsetting capability
4246 such a setting is global if at outer level; local otherwise
4247 (?imsx-imsx:) non-capturing groups with option setting
4248 (?(cond)re|re) conditional pattern matching
4249
4250 A backreference to itself in a repeated group matches the previous
4251 captured string.
4252
4253 8. General tidying up of studying (both automatic and via "study")
4254 consequential on the addition of new assertions.
4255
4256 9. As in 5.005, unlimited repeated groups that could match an empty substring
4257 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4258 runtime if any iteration does actually match an empty substring.
4259
4260 10. Include the RunTest script in the distribution.
4261
4262 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4263 discrepancies, some of which were old and were also with respect to 5.004. They
4264 have now been fixed.
4265
4266
4267 Version 1.09 28-Apr-98
4268 ----------------------
4269
4270 1. A negated single character class followed by a quantifier with a minimum
4271 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4272 program crashes, or just wrong answers. This did not apply to negated classes
4273 containing more than one character, or to minima other than one.
4274
4275
4276 Version 1.08 27-Mar-98
4277 ----------------------
4278
4279 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4280
4281 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4282 latter must appear before anything that relies on it in the pattern.
4283
4284
4285 Version 1.07 16-Feb-98
4286 ----------------------
4287
4288 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4289 repeat of a potentially empty string).
4290
4291
4292 Version 1.06 23-Jan-98
4293 ----------------------
4294
4295 1. Added Markus Oberhumer's little patches for C++.
4296
4297 2. Literal strings longer than 255 characters were broken.
4298
4299
4300 Version 1.05 23-Dec-97
4301 ----------------------
4302
4303 1. Negated character classes containing more than one character were failing if
4304 PCRE_CASELESS was set at run time.
4305
4306
4307 Version 1.04 19-Dec-97
4308 ----------------------
4309
4310 1. Corrected the man page, where some "const" qualifiers had been omitted.
4311
4312 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4313 input syntax.
4314
4315 3. Fixed memory leak which occurred when a regex with back references was
4316 matched with an offsets vector that wasn't big enough. The temporary memory
4317 that is used in this case wasn't being freed if the match failed.
4318
4319 4. Tidied pcretest to ensure it frees memory that it gets.
4320
4321 5. Temporary memory was being obtained in the case where the passed offsets
4322 vector was exactly big enough.
4323
4324 6. Corrected definition of offsetof() from change 5 below.
4325
4326 7. I had screwed up change 6 below and broken the rules for the use of
4327 setjmp(). Now fixed.
4328
4329
4330 Version 1.03 18-Dec-97
4331 ----------------------
4332
4333 1. A erroneous regex with a missing opening parenthesis was correctly
4334 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4335 on some systems.
4336
4337 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4338 it was reported that one broken compiler failed on the former because "code" is
4339 also an independent variable.
4340
4341 3. The erroneous regex a[]b caused an array overrun reference.
4342
4343 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4344 fail on data ending with that character. (It was going on too far, and checking
4345 the next character, typically a binary zero.) This was specific to the
4346 optimized code for single-character negative classes.
4347
4348 5. Added a contributed patch from the TIN world which does the following:
4349
4350 + Add an undef for memmove, in case the the system defines a macro for it.
4351
4352 + Add a definition of offsetof(), in case there isn't one. (I don't know
4353 the reason behind this - offsetof() is part of the ANSI standard - but
4354 it does no harm).
4355
4356 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4357 most of the places where whitespace preceded '#'. I have given up and
4358 allowed the remaining 2 cases to be at the margin.
4359
4360 + Rename some variables in pcre to eliminate shadowing. This seems very
4361 pedantic, but does no harm, of course.
4362
4363 6. Moved the call to setjmp() into its own function, to get rid of warnings
4364 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4365
4366 7. Constructs such as \d{8,} were compiling into the equivalent of
4367 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4368 outcome, but in this particular case used more store than had been allocated,
4369 which caused the bug to be discovered because it threw up an internal error.
4370
4371 8. The debugging code in both pcre and pcretest for outputting the compiled
4372 form of a regex was going wrong in the case of back references followed by
4373 curly-bracketed repeats.
4374
4375
4376 Version 1.02 12-Dec-97
4377 ----------------------
4378
4379 1. Typos in pcre.3 and comments in the source fixed.
4380
4381 2. Applied a contributed patch to get rid of places where it used to remove
4382 'const' from variables, and fixed some signed/unsigned and uninitialized
4383 variable warnings.
4384
4385 3. Added the "runtest" target to Makefile.
4386
4387 4. Set default compiler flag to -O2 rather than just -O.
4388
4389
4390 Version 1.01 19-Nov-97
4391 ----------------------
4392
4393 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4394 like /([ab]*)*/, that is, for classes with more than one character in them.
4395
4396 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4397 as /((?>a*))*/ (a PCRE_EXTRA facility).
4398
4399
4400 Version 1.00 18-Nov-97
4401 ----------------------
4402
4403 1. Added compile-time macros to support systems such as SunOS4 which don't have
4404 memmove() or strerror() but have other things that can be used instead.
4405
4406 2. Arranged that "make clean" removes the executables.
4407
4408
4409 Version 0.99 27-Oct-97
4410 ----------------------
4411
4412 1. Fixed bug in code for optimizing classes with only one character. It was
4413 initializing a 32-byte map regardless, which could cause it to run off the end
4414 of the memory it had got.
4415
4416 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4417
4418
4419 Version 0.98 22-Oct-97
4420 ----------------------
4421
4422 1. Fixed bug in code for handling temporary memory usage when there are more
4423 back references than supplied space in the ovector. This could cause segfaults.
4424
4425
4426 Version 0.97 21-Oct-97
4427 ----------------------
4428
4429 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4430
4431 2. Optimized negated single characters not to use a bit map.
4432
4433 3. Brought error texts together as macro definitions; clarified some of them;
4434 fixed one that was wrong - it said "range out of order" when it meant "invalid
4435 escape sequence".
4436
4437 4. Changed some char * arguments to const char *.
4438
4439 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4440
4441 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4442 pcretest.
4443
4444
4445 Version 0.96 16-Oct-97
4446 ----------------------
4447
4448 1. Added a simple "pgrep" utility to the distribution.
4449
4450 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4451 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4452 where "ddd" means "one or more decimal digits".
4453
4454 3. Fixed serious bug. If a pattern had a back reference, but the call to
4455 pcre_exec() didn't supply a large enough ovector to record the related
4456 identifying subpattern, the match always failed. PCRE now remembers the number
4457 of the largest back reference, and gets some temporary memory in which to save
4458 the offsets during matching if necessary, in order to ensure that
4459 backreferences always work.
4460
4461 4. Increased the compatibility with Perl in a number of ways:
4462
4463 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4464 to request this handling. The option can be set at compile or exec time.
4465
4466 (b) $ matches before a terminating newline by default; an option
4467 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4468 mode). The option can be set at compile or exec time.
4469
4470 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4471 the same as Perl's. If the decimal number it represents is less than 10
4472 or there aren't that many previous left capturing parentheses, an octal
4473 escape is read. Inside a character class, it's always an octal escape,
4474 even if it is a single digit.
4475
4476 (d) An escaped but undefined alphabetic character is taken as a literal,
4477 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4478 escapes.
4479
4480 (e) {0} is now permitted. (The previous item is removed from the compiled
4481 pattern).
4482
4483 5. Changed all the names of code files so that the basic parts are no longer
4484 than 10 characters, and abolished the teeny "globals.c" file.
4485
4486 6. Changed the handling of character classes; they are now done with a 32-byte
4487 bit map always.
4488
4489 7. Added the -d and /D options to pcretest to make it possible to look at the
4490 internals of compilation without having to recompile pcre.
4491
4492
4493 Version 0.95 23-Sep-97
4494 ----------------------
4495
4496 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4497 \x20 at the start of a run of normal characters. These were being treated as
4498 real characters, instead of the source characters being re-checked.
4499
4500
4501 Version 0.94 18-Sep-97
4502 ----------------------
4503
4504 1. The functions are now thread-safe, with the caveat that the global variables
4505 containing pointers to malloc() and free() or alternative functions are the
4506 same for all threads.
4507
4508 2. Get pcre_study() to generate a bitmap of initial characters for non-
4509 anchored patterns when this is possible, and use it if passed to pcre_exec().
4510
4511
4512 Version 0.93 15-Sep-97
4513 ----------------------
4514
4515 1. /(b)|(:+)/ was computing an incorrect first character.
4516
4517 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4518 but not actually doing anything yet.
4519
4520 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4521 as Perl does (e.g. [-az] or [az-]).
4522
4523 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4524 all possible positions.
4525
4526 5. Split up into different modules to avoid including unneeded functions in a
4527 compiled binary. However, compile and exec are still in one module. The "study"
4528 function is split off.
4529
4530 6. The character tables are now in a separate module whose source is generated
4531 by an auxiliary program - but can then be edited by hand if required. There are
4532 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4533 toupper() in the code.
4534
4535 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4536 make them global. Abolish the function for setting them, as the caller can now
4537 set them directly.
4538
4539
4540 Version 0.92 11-Sep-97
4541 ----------------------
4542
4543 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4544 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4545
4546 2. Caseless matching was not working in character classes if the characters in
4547 the pattern were in upper case.
4548
4549 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4550
4551 4. Make PCRE_ANCHORED public and accept as a compile option.
4552
4553 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4554 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4555 pass them.
4556
4557 6. Give an error if bad option bits passed at compile or run time.
4558
4559 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4560 pcretest to cause it to pass that flag.
4561
4562 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4563 options, and the first character, if set.
4564
4565 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4566
4567
4568 Version 0.91 10-Sep-97
4569 ----------------------
4570
4571 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4572 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4573
4574 2. PCRE was looping on encountering an indefinitely repeated back reference to
4575 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4576 Perl does - treats the match as successful.
4577
4578 ****

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