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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 773 - (show annotations)
Wed Nov 30 18:10:27 2011 UTC (3 years, 5 months ago) by ph10
File size: 221985 byte(s)
Error occurred while calculating annotation data.
Expand compile workspace for very many forward references. This ups the limit 
by a factor of 100.
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
88 Version 8.20 21-Oct-2011
89 ------------------------
90
91 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
92 a POSIX class. After further experiments with Perl, which convinced me that
93 Perl has bugs and confusions, a closing square bracket is no longer allowed
94 in a POSIX name. This bug also affected patterns with classes that started
95 with full stops.
96
97 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
98 captured substring, but while checking the failing first alternative,
99 substring 1 is temporarily captured. If the output vector supplied to
100 pcre_exec() was not big enough for this capture, the yield of the function
101 was still zero ("insufficient space for captured substrings"). This cannot
102 be totally fixed without adding another stack variable, which seems a lot
103 of expense for a edge case. However, I have improved the situation in cases
104 such as /(a)(b)x|abc/ matched against "abc", where the return code
105 indicates that fewer than the maximum number of slots in the ovector have
106 been set.
107
108 3. Related to (2) above: when there are more back references in a pattern than
109 slots in the output vector, pcre_exec() uses temporary memory during
110 matching, and copies in the captures as far as possible afterwards. It was
111 using the entire output vector, but this conflicts with the specification
112 that only 2/3 is used for passing back captured substrings. Now it uses
113 only the first 2/3, for compatibility. This is, of course, another edge
114 case.
115
116 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
117 main code base, and can be used by building with --enable-jit. When this is
118 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
119 runtime --no-jit option is given.
120
121 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
122 ovector, the return from the function was zero, implying that there were
123 other matches that did not fit. The correct "exactly full" value is now
124 returned.
125
126 6. If a subpattern that was called recursively or as a subroutine contained
127 (*PRUNE) or any other control that caused it to give a non-standard return,
128 invalid errors such as "Error -26 (nested recursion at the same subject
129 position)" or even infinite loops could occur.
130
131 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
132 computing the minimum length on reaching *ACCEPT, and so ended up with the
133 wrong value of 1 rather than 0. Further investigation indicates that
134 computing a minimum subject length in the presence of *ACCEPT is difficult
135 (think back references, subroutine calls), and so I have changed the code
136 so that no minimum is registered for a pattern that contains *ACCEPT.
137
138 8. If (*THEN) was present in the first (true) branch of a conditional group,
139 it was not handled as intended. [But see 16 below.]
140
141 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
142 Sheri Pierce.
143
144 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
145 the first byte in a match must be "a".
146
147 11. Change 17 for 8.13 increased the recursion depth for patterns like
148 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
149 pattern contains any instances of (*THEN). If it does not, the old
150 optimizations are restored. It would be nice to do this on a per-group
151 basis, but at the moment that is not feasible.
152
153 12. In some environments, the output of pcretest -C is CRLF terminated. This
154 broke RunTest's code that checks for the link size. A single white space
155 character after the value is now allowed for.
156
157 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
158 For "fr", it uses the Windows-specific input and output files.
159
160 14. If (*THEN) appeared in a group that was called recursively or as a
161 subroutine, it did not work as intended. [But see next item.]
162
163 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
164 pattern fragments (but not containing any | characters). If A and B are
165 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
166 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
167 D. In other words, Perl considers parentheses that do not contain any |
168 characters to be part of a surrounding alternative, whereas PCRE was
169 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
170 differently. PCRE now behaves in the same way as Perl, except in the case
171 of subroutine/recursion calls such as (?1) which have in any case always
172 been different (but PCRE had them first :-).
173
174 16. Related to 15 above: Perl does not treat the | in a conditional group as
175 creating alternatives. Such a group is treated in the same way as an
176 ordinary group without any | characters when processing (*THEN). PCRE has
177 been changed to match Perl's behaviour.
178
179 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
180 RunGrepTest script failed.
181
182 18. Change 22 for version 13 caused atomic groups to use more stack. This is
183 inevitable for groups that contain captures, but it can lead to a lot of
184 stack use in large patterns. The old behaviour has been restored for atomic
185 groups that do not contain any capturing parentheses.
186
187 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
188 suppress the check for a minimum subject length at run time. (If it was
189 given to pcre_exec() or pcre_dfa_exec() it did work.)
190
191 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
192 fail to work when decoding hex characters in data strings in EBCDIC
193 environments.
194
195 21. It appears that in at least one Mac OS environment, the isxdigit() function
196 is implemented as a macro that evaluates to its argument more than once,
197 contravening the C 90 Standard (I haven't checked a later standard). There
198 was an instance in pcretest which caused it to go wrong when processing
199 \x{...} escapes in subject strings. The has been rewritten to avoid using
200 things like p++ in the argument of isxdigit().
201
202
203 Version 8.13 16-Aug-2011
204 ------------------------
205
206 1. The Unicode data tables have been updated to Unicode 6.0.0.
207
208 2. Two minor typos in pcre_internal.h have been fixed.
209
210 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
211 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
212 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
213
214 4. There were a number of related bugs in the code for matching backrefences
215 caselessly in UTF-8 mode when codes for the characters concerned were
216 different numbers of bytes. For example, U+023A and U+2C65 are an upper
217 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
218 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
219 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
220 2-byte code at the end of the subject (it thought there wasn't enough data
221 left).
222
223 5. Comprehensive information about what went wrong is now returned by
224 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
225 as the output vector has at least 2 elements. The offset of the start of
226 the failing character and a reason code are placed in the vector.
227
228 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
229 now returned is for the first byte of the failing character, instead of the
230 last byte inspected. This is an incompatible change, but I hope it is small
231 enough not to be a problem. It makes the returned offset consistent with
232 pcre_exec() and pcre_dfa_exec().
233
234 7. pcretest now gives a text phrase as well as the error number when
235 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
236 failure, the offset and reason code are output.
237
238 8. When \R was used with a maximizing quantifier it failed to skip backwards
239 over a \r\n pair if the subsequent match failed. Instead, it just skipped
240 back over a single character (\n). This seems wrong (because it treated the
241 two characters as a single entity when going forwards), conflicts with the
242 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
243 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
244 has been changed.
245
246 9. Some internal refactoring has changed the processing so that the handling
247 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
248 time (the PCRE_DOTALL option was changed this way some time ago: version
249 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
250 which was always a bit of a fudge. It also means that there is one less
251 argument for the match() function, which reduces its stack requirements
252 slightly. This change also fixes an incompatibility with Perl: the pattern
253 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
254
255 10. More internal refactoring has drastically reduced the number of recursive
256 calls to match() for possessively repeated groups such as (abc)++ when
257 using pcre_exec().
258
259 11. While implementing 10, a number of bugs in the handling of groups were
260 discovered and fixed:
261
262 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
263 (a|)*(?1) gave a compile-time internal error.
264 ((a|)+)+ did not notice that the outer group could match an empty string.
265 (^a|^)+ was not marked as anchored.
266 (.*a|.*)+ was not marked as matching at start or after a newline.
267
268 12. Yet more internal refactoring has removed another argument from the match()
269 function. Special calls to this function are now indicated by setting a
270 value in a variable in the "match data" data block.
271
272 13. Be more explicit in pcre_study() instead of relying on "default" for
273 opcodes that mean there is no starting character; this means that when new
274 ones are added and accidentally left out of pcre_study(), testing should
275 pick them up.
276
277 14. The -s option of pcretest has been documented for ages as being an old
278 synonym of -m (show memory usage). I have changed it to mean "force study
279 for every regex", that is, assume /S for every regex. This is similar to -i
280 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
281 using it. It makes it easier to run collections of tests with and without
282 study enabled, and thereby test pcre_study() more easily. All the standard
283 tests are now run with and without -s (but some patterns can be marked as
284 "never study" - see 20 below).
285
286 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
287 restoration of the capturing data to the outer values was not happening
288 correctly.
289
290 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
291 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
292 pattern had matched an empty string, and so incorrectly returned a no
293 match.
294
295 17. There was optimizing code for the last branch of non-capturing parentheses,
296 and also for the obeyed branch of a conditional subexpression, which used
297 tail recursion to cut down on stack usage. Unfortunately, now that there is
298 the possibility of (*THEN) occurring in these branches, tail recursion is
299 no longer possible because the return has to be checked for (*THEN). These
300 two optimizations have therefore been removed. [But see 8.20/11 above.]
301
302 18. If a pattern containing \R was studied, it was assumed that \R always
303 matched two bytes, thus causing the minimum subject length to be
304 incorrectly computed because \R can also match just one byte.
305
306 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
307 was incorrectly computed.
308
309 20. If /S is present twice on a test pattern in pcretest input, it now
310 *disables* studying, thereby overriding the use of -s on the command line
311 (see 14 above). This is necessary for one or two tests to keep the output
312 identical in both cases.
313
314 21. When (*ACCEPT) was used in an assertion that matched an empty string and
315 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
316
317 22. When an atomic group that contained a capturing parenthesis was
318 successfully matched, but the branch in which it appeared failed, the
319 capturing was not being forgotten if a higher numbered group was later
320 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
321 group 1 to "a", when in fact it should be unset. This applied to multi-
322 branched capturing and non-capturing groups, repeated or not, and also to
323 positive assertions (capturing in negative assertions does not happen
324 in PCRE) and also to nested atomic groups.
325
326 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
327 subject after a captured substring, to make it easier to tell which of a
328 number of identical substrings has been captured.
329
330 24. The way atomic groups are processed by pcre_exec() has been changed so that
331 if they are repeated, backtracking one repetition now resets captured
332 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
333 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
334 "aaa". Previously, it would have been "a". As part of this code
335 refactoring, the way recursive calls are handled has also been changed.
336
337 25. If an assertion condition captured any substrings, they were not passed
338 back unless some other capturing happened later. For example, if
339 (?(?=(a))a) was matched against "a", no capturing was returned.
340
341 26. When studying a pattern that contained subroutine calls or assertions,
342 the code for finding the minimum length of a possible match was handling
343 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
344 group 1 called group 2 while simultaneously a separate group 2 called group
345 1). A stack overflow occurred in this case. I have fixed this by limiting
346 the recursion depth to 10.
347
348 27. Updated RunTest.bat in the distribution to the version supplied by Tom
349 Fortmann. This supports explicit test numbers on the command line, and has
350 argument validation and error reporting.
351
352 28. An instance of \X with an unlimited repeat could fail if at any point the
353 first character it looked at was a mark character.
354
355 29. Some minor code refactoring concerning Unicode properties and scripts
356 should reduce the stack requirement of match() slightly.
357
358 30. Added the '=' option to pcretest to check the setting of unused capturing
359 slots at the end of the pattern, which are documented as being -1, but are
360 not included in the return count.
361
362 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
363 compiled something random. Now it gives a compile-time error (as does
364 Perl).
365
366 32. A *MARK encountered during the processing of a positive assertion is now
367 recorded and passed back (compatible with Perl).
368
369 33. If --only-matching or --colour was set on a pcregrep call whose pattern
370 had alternative anchored branches, the search for a second match in a line
371 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
372 matched the line "0102" twice. The same bug affected patterns that started
373 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
374 twice.
375
376 34. Previously, PCRE did not allow quantification of assertions. However, Perl
377 does, and because of capturing effects, quantifying parenthesized
378 assertions may at times be useful. Quantifiers are now allowed for
379 parenthesized assertions.
380
381 35. A minor code tidy in pcre_compile() when checking options for \R usage.
382
383 36. \g was being checked for fancy things in a character class, when it should
384 just be a literal "g".
385
386 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
387 appearance of a nested POSIX class supersedes an apparent external class.
388 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
389 unescaped square brackets may also appear as part of class names. For
390 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
391 more like Perl. (But see 8.20/1 above.)
392
393 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
394 was because it thought it was \N{name}, which is not supported).
395
396 39. Add minix to OS list not supporting the -S option in pcretest.
397
398 40. PCRE tries to detect cases of infinite recursion at compile time, but it
399 cannot analyze patterns in sufficient detail to catch mutual recursions
400 such as ((?1))((?2)). There is now a runtime test that gives an error if a
401 subgroup is called recursively as a subpattern for a second time at the
402 same position in the subject string. In previous releases this might have
403 been caught by the recursion limit, or it might have run out of stack.
404
405 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
406 happen only once. PCRE was, however incorrectly giving a compile time error
407 "recursive call could loop indefinitely" because it cannot analyze the
408 pattern in sufficient detail. The compile time test no longer happens when
409 PCRE is compiling a conditional subpattern, but actual runaway loops are
410 now caught at runtime (see 40 above).
411
412 42. It seems that Perl allows any characters other than a closing parenthesis
413 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
414 has been changed to be the same.
415
416 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
417 as not to get warnings when autogen.sh is called. Also changed
418 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
419
420 44. To help people who use pcregrep to scan files containing exceedingly long
421 lines, the following changes have been made:
422
423 (a) The default value of the buffer size parameter has been increased from
424 8K to 20K. (The actual buffer used is three times this size.)
425
426 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
427 PCRE is built.
428
429 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
430 to be set at run time.
431
432 (d) Numerical values in pcregrep options can be followed by K or M, for
433 example --buffer-size=50K.
434
435 (e) If a line being scanned overflows pcregrep's buffer, an error is now
436 given and the return code is set to 2.
437
438 45. Add a pointer to the latest mark to the callout data block.
439
440 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
441 partial match of an empty string instead of no match. This was specific to
442 the use of ".".
443
444 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
445 complete match instead of a partial match. This bug was dependent on both
446 the PCRE_UTF8 and PCRE_DOTALL options being set.
447
448 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
449 starting byte set, because \b was not being ignored.
450
451
452 Version 8.12 15-Jan-2011
453 ------------------------
454
455 1. Fixed some typos in the markup of the man pages, and wrote a script that
456 checks for such things as part of the documentation building process.
457
458 2. On a big-endian 64-bit system, pcregrep did not correctly process the
459 --match-limit and --recursion-limit options (added for 8.11). In
460 particular, this made one of the standard tests fail. (The integer value
461 went into the wrong half of a long int.)
462
463 3. If the --colour option was given to pcregrep with -v (invert match), it
464 did strange things, either producing crazy output, or crashing. It should,
465 of course, ignore a request for colour when reporting lines that do not
466 match.
467
468 4. Another pcregrep bug caused similar problems if --colour was specified with
469 -M (multiline) and the pattern match finished with a line ending.
470
471 5. In pcregrep, when a pattern that ended with a literal newline sequence was
472 matched in multiline mode, the following line was shown as part of the
473 match. This seems wrong, so I have changed it.
474
475 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
476 the check for further matches in the same line (so they could be coloured)
477 to overrun the end of the current line. If another match was found, it was
478 incorrectly shown (and then shown again when found in the next line).
479
480 7. If pcregrep was compiled under Windows, there was a reference to the
481 function pcregrep_exit() before it was defined. I am assuming this was
482 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
483 reported by a user. I've moved the definition above the reference.
484
485
486 Version 8.11 10-Dec-2010
487 ------------------------
488
489 1. (*THEN) was not working properly if there were untried alternatives prior
490 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
491 backtracked to try for "b" instead of moving to the next alternative branch
492 at the same level (in this case, to look for "c"). The Perl documentation
493 is clear that when (*THEN) is backtracked onto, it goes to the "next
494 alternative in the innermost enclosing group".
495
496 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
497 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
498 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
499 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
500 (*THEN).
501
502 3. If \s appeared in a character class, it removed the VT character from
503 the class, even if it had been included by some previous item, for example
504 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
505 of \s, but is part of the POSIX "space" class.)
506
507 4. A partial match never returns an empty string (because you can always
508 match an empty string at the end of the subject); however the checking for
509 an empty string was starting at the "start of match" point. This has been
510 changed to the "earliest inspected character" point, because the returned
511 data for a partial match starts at this character. This means that, for
512 example, /(?<=abc)def/ gives a partial match for the subject "abc"
513 (previously it gave "no match").
514
515 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
516 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
517 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
518 has an implication that the given string is incomplete (because a partial
519 match is preferred over a full match). For this reason, these items now
520 give a partial match in this situation. [Aside: previously, the one case
521 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
522 match rather than a full match, which was wrong by the old rules, but is
523 now correct.]
524
525 6. There was a bug in the handling of #-introduced comments, recognized when
526 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
527 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
528 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
529 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
530 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
531 places in pcre_compile().
532
533 7. Related to (6) above, when pcre_compile() was skipping #-introduced
534 comments when looking ahead for named forward references to subpatterns,
535 the only newline sequence it recognized was NL. It now handles newlines
536 according to the set newline convention.
537
538 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
539 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
540 cater for a lack of strerror(). These oversights have been fixed.
541
542 9. Added --match-limit and --recursion-limit to pcregrep.
543
544 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
545
546 11. When the -o option was used, pcregrep was setting a return code of 1, even
547 when matches were found, and --line-buffered was not being honoured.
548
549 12. Added an optional parentheses number to the -o and --only-matching options
550 of pcregrep.
551
552 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
553 can match an empty string. The code to do it in pcretest and pcredemo
554 needed fixing:
555
556 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
557 only one byte after an empty string match just before CRLF (this case
558 just got forgotten; "any" and "anycrlf" were OK).
559
560 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
561 mode when an empty string match preceded an ASCII character followed by
562 a non-ASCII character. (The code for advancing by one character rather
563 than one byte was nonsense.)
564
565 (c) The pcredemo.c sample program did not have any code at all to handle
566 the cases when CRLF is a valid newline sequence.
567
568 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
569 as a starting offset was within the subject string. There is now a new
570 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
571 negative or greater than the length of the string. In order to test this,
572 pcretest is extended to allow the setting of negative starting offsets.
573
574 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
575 starting offset points to the beginning of a UTF-8 character was
576 unnecessarily clumsy. I tidied it up.
577
578 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
579 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
580
581 17. Nobody had reported that the --include_dir option, which was added in
582 release 7.7 should have been called --include-dir (hyphen, not underscore)
583 for compatibility with GNU grep. I have changed it to --include-dir, but
584 left --include_dir as an undocumented synonym, and the same for
585 --exclude-dir, though that is not available in GNU grep, at least as of
586 release 2.5.4.
587
588 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
589 characters from a string of bytes) have been redefined so as not to use
590 loops, in order to improve performance in some environments. At the same
591 time, I abstracted some of the common code into auxiliary macros to save
592 repetition (this should not affect the compiled code).
593
594 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
595 compile-time error is now given if \c is not followed by an ASCII
596 character, that is, a byte less than 128. (In EBCDIC mode, the code is
597 different, and any byte value is allowed.)
598
599 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
600 START_OPTIMIZE option, which is now allowed at compile time - but just
601 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
602 to pcregrep and other applications that have no direct access to PCRE
603 options. The new /Y option in pcretest sets this option when calling
604 pcre_compile().
605
606 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
607 back references. Groups containing recursive back references were forced to
608 be atomic by that change, but in the case of named groups, the amount of
609 memory required was incorrectly computed, leading to "Failed: internal
610 error: code overflow". This has been fixed.
611
612 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
613 pcretest.c, to avoid build problems in some Borland environments.
614
615
616 Version 8.10 25-Jun-2010
617 ------------------------
618
619 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
620 THEN.
621
622 2. (*ACCEPT) was not working when inside an atomic group.
623
624 3. Inside a character class, \B is treated as a literal by default, but
625 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
626 causes the error). The code is unchanged, but I tidied the documentation.
627
628 4. Inside a character class, PCRE always treated \R and \X as literals,
629 whereas Perl faults them if its -w option is set. I have changed PCRE so
630 that it faults them when PCRE_EXTRA is set.
631
632 5. Added support for \N, which always matches any character other than
633 newline. (It is the same as "." when PCRE_DOTALL is not set.)
634
635 6. When compiling pcregrep with newer versions of gcc which may have
636 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
637 declared with attribute warn_unused_result" were given. Just casting the
638 result to (void) does not stop the warnings; a more elaborate fudge is
639 needed. I've used a macro to implement this.
640
641 7. Minor change to pcretest.c to avoid a compiler warning.
642
643 8. Added four artifical Unicode properties to help with an option to make
644 \s etc use properties (see next item). The new properties are: Xan
645 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
646
647 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
648 use Unicode properties. (*UCP) at the start of a pattern can be used to set
649 this option. Modified pcretest to add /W to test this facility. Added
650 REG_UCP to make it available via the POSIX interface.
651
652 10. Added --line-buffered to pcregrep.
653
654 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
655 studied, and the match started with a letter with a code point greater than
656 127 whose first byte was different to the first byte of the other case of
657 the letter, the other case of this starting letter was not recognized
658 (#976).
659
660 12. If a pattern that was studied started with a repeated Unicode property
661 test, for example, \p{Nd}+, there was the theoretical possibility of
662 setting up an incorrect bitmap of starting bytes, but fortunately it could
663 not have actually happened in practice until change 8 above was made (it
664 added property types that matched character-matching opcodes).
665
666 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
667 possible starting bytes for non-anchored patterns.
668
669 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
670 \R, and also a number of cases that involve Unicode properties, both
671 explicit and implicit when PCRE_UCP is set.
672
673 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
674 input, it could crash or give wrong results if characters with values
675 greater than 0xc0 were present in the subject string. (Detail: it assumed
676 UTF-8 input when processing these items.)
677
678 16. Added a lot of (int) casts to avoid compiler warnings in systems where
679 size_t is 64-bit (#991).
680
681 17. Added a check for running out of memory when PCRE is compiled with
682 --disable-stack-for-recursion (#990).
683
684 18. If the last data line in a file for pcretest does not have a newline on
685 the end, a newline was missing in the output.
686
687 19. The default pcre_chartables.c file recognizes only ASCII characters (values
688 less than 128) in its various bitmaps. However, there is a facility for
689 generating tables according to the current locale when PCRE is compiled. It
690 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
691 characters, are recognized by isspace() and therefore were getting set in
692 these tables, and indeed these tables seem to approximate to ISO 8859. This
693 caused a problem in UTF-8 mode when pcre_study() was used to create a list
694 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
695 which of course cannot start UTF-8 characters. I have changed the code so
696 that only real ASCII characters (less than 128) and the correct starting
697 bytes for UTF-8 encodings are set for characters greater than 127 when in
698 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
699 altogether.)
700
701 20. Added the /T option to pcretest so as to be able to run tests with non-
702 standard character tables, thus making it possible to include the tests
703 used for 19 above in the standard set of tests.
704
705 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
706 reference to a subpattern the other side of a comment that contains an
707 opening parenthesis caused either an internal compiling error, or a
708 reference to the wrong subpattern.
709
710
711 Version 8.02 19-Mar-2010
712 ------------------------
713
714 1. The Unicode data tables have been updated to Unicode 5.2.0.
715
716 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
717 configured.
718
719 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
720 original author of that file, following a query about its status.
721
722 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
723 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
724
725 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
726 quantifier applied to a forward-referencing subroutine call, could compile
727 incorrect code or give the error "internal error: previously-checked
728 referenced subpattern not found".
729
730 6. Both MS Visual Studio and Symbian OS have problems with initializing
731 variables to point to external functions. For these systems, therefore,
732 pcre_malloc etc. are now initialized to local functions that call the
733 relevant global functions.
734
735 7. There were two entries missing in the vectors called coptable and poptable
736 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
737 I've fixed the data, and added a kludgy way of testing at compile time that
738 the lengths are correct (equal to the number of opcodes).
739
740 8. Following on from 7, I added a similar kludge to check the length of the
741 eint vector in pcreposix.c.
742
743 9. Error texts for pcre_compile() are held as one long string to avoid too
744 much relocation at load time. To find a text, the string is searched,
745 counting zeros. There was no check for running off the end of the string,
746 which could happen if a new error number was added without updating the
747 string.
748
749 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
750
751 11. \K was not working if it appeared in an atomic group or in a group that
752 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
753 \K is "not well defined" if used in an assertion. PCRE now accepts it if
754 the assertion is positive, but not if it is negative.
755
756 12. Change 11 fortuitously reduced the size of the stack frame used in the
757 "match()" function of pcre_exec.c by one pointer. Forthcoming
758 implementation of support for (*MARK) will need an extra pointer on the
759 stack; I have reserved it now, so that the stack frame size does not
760 decrease.
761
762 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
763 item in branch that calls a recursion is a subroutine call - as in the
764 second branch in the above example - was incorrectly given the compile-
765 time error "recursive call could loop indefinitely" because pcre_compile()
766 was not correctly checking the subroutine for matching a non-empty string.
767
768 14. The checks for overrunning compiling workspace could trigger after an
769 overrun had occurred. This is a "should never occur" error, but it can be
770 triggered by pathological patterns such as hundreds of nested parentheses.
771 The checks now trigger 100 bytes before the end of the workspace.
772
773 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
774
775
776 Version 8.01 19-Jan-2010
777 ------------------------
778
779 1. If a pattern contained a conditional subpattern with only one branch (in
780 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
781 computed the wrong minimum data length (which is of course zero for such
782 subpatterns). This could cause incorrect "no match" results.
783
784 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
785 the pattern is reset in the first branch, pcre_compile() failed with
786 "internal error: code overflow at offset...". This happened only when
787 the reset was to the original external option setting. (An optimization
788 abstracts leading options settings into an external setting, which was the
789 cause of this.)
790
791 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
792 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
793 assertion pattern did not match (meaning that the assertion was true), it
794 was incorrectly treated as false if the SKIP had been reached during the
795 matching. This also applied to assertions used as conditions.
796
797 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
798 assertion subpattern, including such a pattern used as a condition,
799 unpredictable results occurred, instead of the error return
800 PCRE_ERROR_DFA_UITEM.
801
802 5. The C++ GlobalReplace function was not working like Perl for the special
803 situation when an empty string is matched. It now does the fancy magic
804 stuff that is necessary.
805
806 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
807 removed. (These were left over from very, very early versions of PCRE.)
808
809 7. Some cosmetic changes to the code to make life easier when compiling it
810 as part of something else:
811
812 (a) Change DEBUG to PCRE_DEBUG.
813
814 (b) In pcre_compile(), rename the member of the "branch_chain" structure
815 called "current" as "current_branch", to prevent a collision with the
816 Linux macro when compiled as a kernel module.
817
818 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
819 prevent a collision with the Linux macro when compiled as a kernel
820 module.
821
822 8. In pcre_compile() there are some checks for integer overflows that used to
823 cast potentially large values to (double). This has been changed to that
824 when building, a check for int64_t is made, and if it is found, it is used
825 instead, thus avoiding the use of floating point arithmetic. (There is no
826 other use of FP in PCRE.) If int64_t is not found, the fallback is to
827 double.
828
829 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
830 2005 (difference between two addresses compared to an unsigned value).
831
832 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
833 custom one, because of the following reported problem in Windows:
834
835 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
836 under Win32.
837 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
838 therefore missing the function definition.
839 - The compiler thus generates a "C" signature for the test function.
840 - The linker fails to find the "C" function.
841 - PCRE fails to configure if asked to do so against libbz2.
842
843 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
844 messages were output:
845
846 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
847 rerunning libtoolize, to keep the correct libtool macros in-tree.
848 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
849
850 I have done both of these things.
851
852 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
853 most of the time, it *can* run out if it is given a pattern that contains a
854 runaway infinite recursion. I updated the discussion in the pcrestack man
855 page.
856
857 13. Now that we have gone to the x.xx style of version numbers, the minor
858 version may start with zero. Using 08 or 09 is a bad idea because users
859 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
860 interpreted as invalid octal numbers. I've updated the previous comment in
861 configure.ac, and also added a check that gives an error if 08 or 09 are
862 used.
863
864 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
865 causing partial matching to fail when the end of the subject matched \W
866 in a UTF-8 pattern where \W was quantified with a minimum of 3.
867
868 15. There were some discrepancies between the declarations in pcre_internal.h
869 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
870 their definitions. The declarations used "const uschar *" and the
871 definitions used USPTR. Even though USPTR is normally defined as "const
872 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
873 reported that: "This difference in casting confuses some C++ compilers, for
874 example, SunCC recognizes above declarations as different functions and
875 generates broken code for hbpcre." I have changed the declarations to use
876 USPTR.
877
878 16. GNU libtool is named differently on some systems. The autogen.sh script now
879 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
880 (FreeBSD).
881
882 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
883 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
884 comment: "Figure out how to create a longlong from a string: strtoll and
885 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
886 instance, but it only takes 2 args instead of 3!"
887
888 18. A subtle bug concerned with back references has been fixed by a change of
889 specification, with a corresponding code fix. A pattern such as
890 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
891 refers, was giving matches when it shouldn't. For example, xa=xaaa would
892 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
893 same bug. Such groups have to be quantified to be useful, or contained
894 inside another quantified group. (If there's no repetition, the reference
895 can never match.) The problem arises because, having left the group and
896 moved on to the rest of the pattern, a later failure that backtracks into
897 the group uses the captured value from the final iteration of the group
898 rather than the correct earlier one. I have fixed this in PCRE by forcing
899 any group that contains a reference to itself to be an atomic group; that
900 is, there cannot be any backtracking into it once it has completed. This is
901 similar to recursive and subroutine calls.
902
903
904 Version 8.00 19-Oct-09
905 ----------------------
906
907 1. The table for translating pcre_compile() error codes into POSIX error codes
908 was out-of-date, and there was no check on the pcre_compile() error code
909 being within the table. This could lead to an OK return being given in
910 error.
911
912 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
913 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
914 in a Windows environment.
915
916 3. The pcregrep --count option prints the count for each file even when it is
917 zero, as does GNU grep. However, pcregrep was also printing all files when
918 --files-with-matches was added. Now, when both options are given, it prints
919 counts only for those files that have at least one match. (GNU grep just
920 prints the file name in this circumstance, but including the count seems
921 more useful - otherwise, why use --count?) Also ensured that the
922 combination -clh just lists non-zero counts, with no names.
923
924 4. The long form of the pcregrep -F option was incorrectly implemented as
925 --fixed_strings instead of --fixed-strings. This is an incompatible change,
926 but it seems right to fix it, and I didn't think it was worth preserving
927 the old behaviour.
928
929 5. The command line items --regex=pattern and --regexp=pattern were not
930 recognized by pcregrep, which required --regex pattern or --regexp pattern
931 (with a space rather than an '='). The man page documented the '=' forms,
932 which are compatible with GNU grep; these now work.
933
934 6. No libpcreposix.pc file was created for pkg-config; there was just
935 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
936
937 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
938 when UCP support is not needed, by modifying the Python script that
939 generates it from Unicode data files. This should not matter if the module
940 is correctly used as a library, but I received one complaint about 50K of
941 unwanted data. My guess is that the person linked everything into his
942 program rather than using a library. Anyway, it does no harm.
943
944 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
945 was a minimum greater than 1 for a wide character in a possessive
946 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
947 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
948 character. Chaos in the form of incorrect output or a compiling loop could
949 result.
950
951 9. The restrictions on what a pattern can contain when partial matching is
952 requested for pcre_exec() have been removed. All patterns can now be
953 partially matched by this function. In addition, if there are at least two
954 slots in the offset vector, the offset of the earliest inspected character
955 for the match and the offset of the end of the subject are set in them when
956 PCRE_ERROR_PARTIAL is returned.
957
958 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
959 synonymous with PCRE_PARTIAL, for backwards compatibility, and
960 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
961 and may be more useful for multi-segment matching.
962
963 11. Partial matching with pcre_exec() is now more intuitive. A partial match
964 used to be given if ever the end of the subject was reached; now it is
965 given only if matching could not proceed because another character was
966 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
967 string "Z", which now yields "no match" instead of "partial match". In the
968 case of pcre_dfa_exec(), "no match" is given if every matching path for the
969 final character ended with (*FAIL).
970
971 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
972 if the pattern had a "must contain" character that was already found in the
973 earlier partial match, unless partial matching was again requested. For
974 example, with the pattern /dog.(body)?/, the "must contain" character is
975 "g". If the first part-match was for the string "dog", restarting with
976 "sbody" failed. This bug has been fixed.
977
978 13. The string returned by pcre_dfa_exec() after a partial match has been
979 changed so that it starts at the first inspected character rather than the
980 first character of the match. This makes a difference only if the pattern
981 starts with a lookbehind assertion or \b or \B (\K is not supported by
982 pcre_dfa_exec()). It's an incompatible change, but it makes the two
983 matching functions compatible, and I think it's the right thing to do.
984
985 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
986 so that the demonstration program is easily available in environments where
987 PCRE has not been installed from source.
988
989 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
990 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
991 library.
992
993 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
994 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
995 is not the first non-POSIX option to be added. Clearly some people find
996 these options useful.
997
998 17. If a caller to the POSIX matching function regexec() passes a non-zero
999 value for nmatch with a NULL value for pmatch, the value of
1000 nmatch is forced to zero.
1001
1002 18. RunGrepTest did not have a test for the availability of the -u option of
1003 the diff command, as RunTest does. It now checks in the same way as
1004 RunTest, and also checks for the -b option.
1005
1006 19. If an odd number of negated classes containing just a single character
1007 interposed, within parentheses, between a forward reference to a named
1008 subpattern and the definition of the subpattern, compilation crashed with
1009 an internal error, complaining that it could not find the referenced
1010 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1011 [The bug was that it was starting one character too far in when skipping
1012 over the character class, thus treating the ] as data rather than
1013 terminating the class. This meant it could skip too much.]
1014
1015 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1016 /g option in pcretest when the pattern contains \K, which makes it possible
1017 to have an empty string match not at the start, even when the pattern is
1018 anchored. Updated pcretest and pcredemo to use this option.
1019
1020 21. If the maximum number of capturing subpatterns in a recursion was greater
1021 than the maximum at the outer level, the higher number was returned, but
1022 with unset values at the outer level. The correct (outer level) value is
1023 now given.
1024
1025 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1026 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1027 make it do so. The string so far is captured, making this feature
1028 compatible with Perl.
1029
1030 23. The tests have been re-organized, adding tests 11 and 12, to make it
1031 possible to check the Perl 5.10 features against Perl 5.10.
1032
1033 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1034 pattern matches a fixed length string. PCRE did not allow this; now it
1035 does. Neither allows recursion.
1036
1037 25. I finally figured out how to implement a request to provide the minimum
1038 length of subject string that was needed in order to match a given pattern.
1039 (It was back references and recursion that I had previously got hung up
1040 on.) This code has now been added to pcre_study(); it finds a lower bound
1041 to the length of subject needed. It is not necessarily the greatest lower
1042 bound, but using it to avoid searching strings that are too short does give
1043 some useful speed-ups. The value is available to calling programs via
1044 pcre_fullinfo().
1045
1046 26. While implementing 25, I discovered to my embarrassment that pcretest had
1047 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1048 study optimizations had never been tested with that matching function.
1049 Oops. What is worse, even when it was passed study data, there was a bug in
1050 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1051 were also very few tests of studied patterns with pcre_dfa_exec().
1052
1053 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1054 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1055 on the other side of the coin, they are no longer allowed to have different
1056 names, because these cannot be distinguished in PCRE, and this has caused
1057 confusion. (This is a difference from Perl.)
1058
1059 28. When duplicate subpattern names are present (necessarily with different
1060 numbers, as required by 27 above), and a test is made by name in a
1061 conditional pattern, either for a subpattern having been matched, or for
1062 recursion in such a pattern, all the associated numbered subpatterns are
1063 tested, and the overall condition is true if the condition is true for any
1064 one of them. This is the way Perl works, and is also more like the way
1065 testing by number works.
1066
1067
1068 Version 7.9 11-Apr-09
1069 ---------------------
1070
1071 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1072 (pcretest), all targets were linked against these libraries. This included
1073 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1074 libraries. This caused unwanted dependencies to be created. This problem
1075 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1076 pcretest is linked with readline.
1077
1078 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1079 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1080 moved outside it again, because FALSE and TRUE are already defined in AIX,
1081 but BOOL is not.
1082
1083 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1084 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1085
1086 4. The pcregrep documentation said spaces were inserted as well as colons (or
1087 hyphens) following file names and line numbers when outputting matching
1088 lines. This is not true; no spaces are inserted. I have also clarified the
1089 wording for the --colour (or --color) option.
1090
1091 5. In pcregrep, when --colour was used with -o, the list of matching strings
1092 was not coloured; this is different to GNU grep, so I have changed it to be
1093 the same.
1094
1095 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1096 each matching line was coloured. Now it goes on to look for further matches
1097 of any of the test patterns, which is the same behaviour as GNU grep.
1098
1099 7. A pattern that could match an empty string could cause pcregrep to loop; it
1100 doesn't make sense to accept an empty string match in pcregrep, so I have
1101 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1102 seems to be how GNU grep behaves.
1103
1104 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1105 start or after a newline", because the conditional assertion was not being
1106 correctly handled. The rule now is that both the assertion and what follows
1107 in the first alternative must satisfy the test.
1108
1109 9. If auto-callout was enabled in a pattern with a conditional group whose
1110 condition was an assertion, PCRE could crash during matching, both with
1111 pcre_exec() and pcre_dfa_exec().
1112
1113 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1114 used for matching.
1115
1116 11. Unicode property support in character classes was not working for
1117 characters (bytes) greater than 127 when not in UTF-8 mode.
1118
1119 12. Added the -M command line option to pcretest.
1120
1121 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1122
1123 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1124
1125 16. Added comments and documentation about mis-use of no_arg in the C++
1126 wrapper.
1127
1128 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1129 from Martin Jerabek that uses macro names for all relevant character and
1130 string constants.
1131
1132 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1133 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1134 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1135 these, but not everybody uses configure.
1136
1137 19. A conditional group that had only one branch was not being correctly
1138 recognized as an item that could match an empty string. This meant that an
1139 enclosing group might also not be so recognized, causing infinite looping
1140 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1141 with the subject "ab", where knowledge that the repeated group can match
1142 nothing is needed in order to break the loop.
1143
1144 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1145 exec(), but without supplying a callout function, matching went wrong.
1146
1147 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1148 leak if the size of the offset vector was greater than 30. When the vector
1149 is smaller, the saved offsets during recursion go onto a local stack
1150 vector, but for larger vectors malloc() is used. It was failing to free
1151 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1152 error, in fact).
1153
1154 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1155 heapframe that is used only when UTF-8 support is enabled. This caused no
1156 problem, but was untidy.
1157
1158 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1159 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1160 included within another project.
1161
1162 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1163 slightly modified by me:
1164
1165 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1166 not building pcregrep.
1167
1168 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1169 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1170
1171 25. Forward references, both numeric and by name, in patterns that made use of
1172 duplicate group numbers, could behave incorrectly or give incorrect errors,
1173 because when scanning forward to find the reference group, PCRE was not
1174 taking into account the duplicate group numbers. A pattern such as
1175 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1176
1177 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1178 the feature of a custom pointer more persuasive (as requested by a user).
1179
1180 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1181 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1182 pre-defined.
1183
1184 28. Added support for (*UTF8) at the start of a pattern.
1185
1186 29. Arrange for flags added by the "release type" setting in CMake to be shown
1187 in the configuration summary.
1188
1189
1190 Version 7.8 05-Sep-08
1191 ---------------------
1192
1193 1. Replaced UCP searching code with optimized version as implemented for Ad
1194 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1195 stage table and inline lookup instead of a function, giving speed ups of 2
1196 to 5 times on some simple patterns that I tested. Permission was given to
1197 distribute the MultiStage2.py script that generates the tables (it's not in
1198 the tarball, but is in the Subversion repository).
1199
1200 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1201 scripts.
1202
1203 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1204 a group with a zero qualifier. The result of the study could be incorrect,
1205 or the function might crash, depending on the pattern.
1206
1207 4. Caseless matching was not working for non-ASCII characters in back
1208 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1209 It now works when Unicode Property Support is available.
1210
1211 5. In pcretest, an escape such as \x{de} in the data was always generating
1212 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1213 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1214 truncation.
1215
1216 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1217
1218 7. Added two (int) casts to pcregrep when printing the difference of two
1219 pointers, in case they are 64-bit values.
1220
1221 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1222 test 2 if it fails.
1223
1224 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1225 and a #define of that name to empty if it is not externally set. This is to
1226 allow users of MSVC to set it if necessary.
1227
1228 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1229 the convenience functions in the pcre_get.c source file.
1230
1231 11. An option change at the start of a pattern that had top-level alternatives
1232 could cause overwriting and/or a crash. This command provoked a crash in
1233 some environments:
1234
1235 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1236
1237 This potential security problem was recorded as CVE-2008-2371.
1238
1239 12. For a pattern where the match had to start at the beginning or immediately
1240 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1241 pcre_dfa_exec() could read past the end of the passed subject if there was
1242 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1243 pcretest so that it places the subject at the end of its malloc-ed buffer.
1244
1245 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1246 exec() might read past the end of the data buffer in UTF-8 mode.
1247
1248 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1249 the data contained the byte 0x85 as part of a UTF-8 character within its
1250 first line. This applied both to normal and DFA matching.
1251
1252 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1253 /^[^d]*?$/8 failed to match "abc".
1254
1255 16. Added a missing copyright notice to pcrecpp_internal.h.
1256
1257 17. Make it more clear in the documentation that values returned from
1258 pcre_exec() in ovector are byte offsets, not character counts.
1259
1260 18. Tidied a few places to stop certain compilers from issuing warnings.
1261
1262 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1263 supplied by Stefan Weber. I made a further small update for 7.8 because
1264 there is a change of source arrangements: the pcre_searchfuncs.c module is
1265 replaced by pcre_ucd.c.
1266
1267
1268 Version 7.7 07-May-08
1269 ---------------------
1270
1271 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1272 a string to a long long, pretend we don't even have a long long." This is
1273 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1274
1275 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1276 pre-7.6 versions, which defined a global no_arg variable instead of putting
1277 it in the RE class. (See also #8 below.)
1278
1279 3. Remove a line of dead code, identified by coverity and reported by Nuno
1280 Lopes.
1281
1282 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1283
1284 (1) The include/exclude patterns were being applied to the whole pathnames
1285 of files, instead of just to the final components.
1286
1287 (2) If there was more than one level of directory, the subdirectories were
1288 skipped unless they satisfied the include/exclude conditions. This is
1289 inconsistent with GNU grep (and could even be seen as contrary to the
1290 pcregrep specification - which I improved to make it absolutely clear).
1291 The action now is always to scan all levels of directory, and just
1292 apply the include/exclude patterns to regular files.
1293
1294 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1295 --exclude_dir in the tests to avoid scanning .svn directories.
1296
1297 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1298 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1299 doesn't support NULs in patterns.
1300
1301 7. Added some missing "const"s to declarations of static tables in
1302 pcre_compile.c and pcre_dfa_exec.c.
1303
1304 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1305 caused by fix #2 above. (Subsequently also a second patch to fix the
1306 first patch. And a third patch - this was a messy problem.)
1307
1308 9. Applied Craig's patch to remove the use of push_back().
1309
1310 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1311 matching function regexec().
1312
1313 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1314 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1315 references. PCRE supports relative numbers with this syntax (I don't think
1316 Oniguruma does).
1317
1318 12. Previously, a group with a zero repeat such as (...){0} was completely
1319 omitted from the compiled regex. However, this means that if the group
1320 was called as a subroutine from elsewhere in the pattern, things went wrong
1321 (an internal error was given). Such groups are now left in the compiled
1322 pattern, with a new opcode that causes them to be skipped at execution
1323 time.
1324
1325 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1326 to the way PCRE behaves:
1327
1328 (a) A lone ] character is dis-allowed (Perl treats it as data).
1329
1330 (b) A back reference to an unmatched subpattern matches an empty string
1331 (Perl fails the current match path).
1332
1333 (c) A data ] in a character class must be notated as \] because if the
1334 first data character in a class is ], it defines an empty class. (In
1335 Perl it is not possible to have an empty class.) The empty class []
1336 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1337 The negative empty class [^] matches any one character, independently
1338 of the DOTALL setting.
1339
1340 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1341 non-existent subpattern following a character class starting with ']' and
1342 containing () gave an internal compiling error instead of "reference to
1343 non-existent subpattern". Fortunately, when the pattern did exist, the
1344 compiled code was correct. (When scanning forwards to check for the
1345 existencd of the subpattern, it was treating the data ']' as terminating
1346 the class, so got the count wrong. When actually compiling, the reference
1347 was subsequently set up correctly.)
1348
1349 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1350 it was being rejected as not supported by pcre_dfa_exec(), even though
1351 other assertions are supported. I have made pcre_dfa_exec() support
1352 (*FAIL).
1353
1354 16. The implementation of 13c above involved the invention of a new opcode,
1355 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1356 cannot be changed at match time, I realized I could make a small
1357 improvement to matching performance by compiling OP_ALLANY instead of
1358 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1359 on the OP_ANY path.
1360
1361 17. Compiling pcretest on Windows with readline support failed without the
1362 following two fixes: (1) Make the unistd.h include conditional on
1363 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1364
1365 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1366 ncurses library to be included for pcretest when ReadLine support is
1367 requested, but also to allow for it to be overridden. This patch came from
1368 Daniel Bergström.
1369
1370 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1371 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1372 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1373 spotting this.
1374
1375
1376 Version 7.6 28-Jan-08
1377 ---------------------
1378
1379 1. A character class containing a very large number of characters with
1380 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1381 overflow.
1382
1383 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1384 HAVE_LONG_LONG is not defined.
1385
1386 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1387 bring them up to date and include new features. This patch includes:
1388
1389 - Fixed PH's badly added libz and libbz2 support.
1390 - Fixed a problem with static linking.
1391 - Added pcredemo. [But later removed - see 7 below.]
1392 - Fixed dftables problem and added an option.
1393 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1394 HAVE_LONG_LONG.
1395 - Added readline support for pcretest.
1396 - Added an listing of the option settings after cmake has run.
1397
1398 4. A user submitted a patch to Makefile that makes it easy to create
1399 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1400 Makefile.am that cause it to include this special target, without
1401 affecting anything else. Note that the same mingw target plus all
1402 the other distribution libraries and programs are now supported
1403 when configuring with CMake (see 6 below) instead of with
1404 Configure/Make.
1405
1406 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1407 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1408 exported in the Windows port". It has not yet been confirmed that the patch
1409 solves the problem, but it does no harm.
1410
1411 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1412 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1413 with CMake, and also correct the comment about stack recursion.
1414
1415 7. Remove the automatic building of pcredemo from the ./configure system and
1416 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1417 of a program that users should build themselves after PCRE is installed, so
1418 building it automatically is not really right. What is more, it gave
1419 trouble in some build environments.
1420
1421 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1422
1423
1424 Version 7.5 10-Jan-08
1425 ---------------------
1426
1427 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1428 values in parens when parsing an RE using the C++ wrapper."
1429
1430 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1431 Characters greater than 255 were excluded from the class instead of being
1432 included.
1433
1434 3. The same bug as (2) above applied to negated POSIX classes such as
1435 [:^space:].
1436
1437 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1438 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1439 I have changed it.
1440
1441 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1442 first named subpattern) and a construct such as (?&a) would reference the
1443 first named subpattern whose name started with "a" (in other words, the
1444 length check was missing). Both these problems are fixed. "Subpattern name
1445 expected" is now given for (?&) (a zero-length name), and this patch also
1446 makes it give the same error for \k'' (previously it complained that that
1447 was a reference to a non-existent subpattern).
1448
1449 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1450 this is right because (?- can be followed by option settings as well as by
1451 digits. I have, however, made the messages clearer.
1452
1453 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1454 than the number used in the conditional) now cause a compile-time error.
1455 This is actually not compatible with Perl, which accepts such patterns, but
1456 treats the conditional as always being FALSE (as PCRE used to), but it
1457 seems to me that giving a diagnostic is better.
1458
1459 8. Change "alphameric" to the more common word "alphanumeric" in comments
1460 and messages.
1461
1462 9. Fix two occurrences of "backslash" in comments that should have been
1463 "backspace".
1464
1465 10. Remove two redundant lines of code that can never be obeyed (their function
1466 was moved elsewhere).
1467
1468 11. The program that makes PCRE's Unicode character property table had a bug
1469 which caused it to generate incorrect table entries for sequences of
1470 characters that have the same character type, but are in different scripts.
1471 It amalgamated them into a single range, with the script of the first of
1472 them. In other words, some characters were in the wrong script. There were
1473 thirteen such cases, affecting characters in the following ranges:
1474
1475 U+002b0 - U+002c1
1476 U+0060c - U+0060d
1477 U+0061e - U+00612
1478 U+0064b - U+0065e
1479 U+0074d - U+0076d
1480 U+01800 - U+01805
1481 U+01d00 - U+01d77
1482 U+01d9b - U+01dbf
1483 U+0200b - U+0200f
1484 U+030fc - U+030fe
1485 U+03260 - U+0327f
1486 U+0fb46 - U+0fbb1
1487 U+10450 - U+1049d
1488
1489 12. The -o option (show only the matching part of a line) for pcregrep was not
1490 compatible with GNU grep in that, if there was more than one match in a
1491 line, it showed only the first of them. It now behaves in the same way as
1492 GNU grep.
1493
1494 13. If the -o and -v options were combined for pcregrep, it printed a blank
1495 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1496 does the same. The return code can be used to tell if there were any
1497 non-matching lines.
1498
1499 14. Added --file-offsets and --line-offsets to pcregrep.
1500
1501 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1502 infinitely looping recursion. The bug was that positive lookaheads were not
1503 being skipped when checking for a possible empty match (negative lookaheads
1504 and both kinds of lookbehind were skipped).
1505
1506 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1507 inclusion of <windows.h> to before rather than after the definition of
1508 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1509
1510 17. Specifying a possessive quantifier with a specific limit for a Unicode
1511 character property caused pcre_compile() to compile bad code, which led at
1512 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1513 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1514 caused the error; without that there was no problem.
1515
1516 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1517
1518 19. Added --enable-pcretest-libreadline.
1519
1520 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1521 RE::GlobalReplace(). As a result, the number of replacements returned was
1522 double what it should be. I removed one of the increments, but Craig sent a
1523 later patch that removed the other one (the right fix) and added unit tests
1524 that check the return values (which was not done before).
1525
1526 21. Several CMake things:
1527
1528 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1529 the names libpcre and libpcreposix, not just pcre and pcreposix.
1530
1531 (2) The above change means that pcretest and pcregrep are now correctly
1532 linked with the newly-built libraries, not previously installed ones.
1533
1534 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1535
1536 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1537 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1538 UTF-8 newline character). The key issue is that the pattern starts .*;
1539 this means that the match must be either at the beginning, or after a
1540 newline. The bug was in the code for advancing after a failed match and
1541 checking that the new position followed a newline. It was not taking
1542 account of UTF-8 characters correctly.
1543
1544 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1545 character classes. PCRE was not treating the sequence [:...:] as a
1546 character class unless the ... were all letters. Perl, however, seems to
1547 allow any characters between [: and :], though of course it rejects as
1548 unknown any "names" that contain non-letters, because all the known class
1549 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1550 for example, whereas PCRE did not - it did not recognize a POSIX character
1551 class. This seemed a bit dangerous, so the code has been changed to be
1552 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1553 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1554 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1555 Perl does, and where it didn't before.
1556
1557 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1558 Windows environments %n is disabled by default.
1559
1560
1561 Version 7.4 21-Sep-07
1562 ---------------------
1563
1564 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1565 means that a class such as [\s] counted as "explicit reference to CR or
1566 LF". That isn't really right - the whole point of the change was to try to
1567 help when there was an actual mention of one of the two characters. So now
1568 the change happens only if \r or \n (or a literal CR or LF) character is
1569 encountered.
1570
1571 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1572 of both had grown to the point where there were only 3 bits left.
1573 Fortunately, there was spare space in the data structure, and so I have
1574 moved the internal flags into a new 16-bit field to free up more option
1575 bits.
1576
1577 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1578 but did not set the internal JCHANGED flag - either of these is enough to
1579 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1580 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1581 start sets both bits.
1582
1583 4. Added options (at build time, compile time, exec time) to change \R from
1584 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1585
1586 5. doc/pcresyntax.html was missing from the distribution.
1587
1588 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1589 compatibility, even though it is no longer used.
1590
1591 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1592 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1593 windows.h file is present (where different names are used). [This was
1594 reversed later after testing - see 16 below.]
1595
1596 8. Changed all #include <config.h> to #include "config.h". There were also
1597 some further <pcre.h> cases that I changed to "pcre.h".
1598
1599 9. When pcregrep was used with the --colour option, it missed the line ending
1600 sequence off the lines that it output.
1601
1602 10. It was pointed out to me that arrays of string pointers cause lots of
1603 relocations when a shared library is dynamically loaded. A technique of
1604 using a single long string with a table of offsets can drastically reduce
1605 these. I have refactored PCRE in four places to do this. The result is
1606 dramatic:
1607
1608 Originally: 290
1609 After changing UCP table: 187
1610 After changing error message table: 43
1611 After changing table of "verbs" 36
1612 After changing table of Posix names 22
1613
1614 Thanks to the folks working on Gregex for glib for this insight.
1615
1616 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1617 unicode-properties was also set.
1618
1619 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1620
1621 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1622 checked only for CRLF.
1623
1624 14. Added casts to pcretest.c to avoid compiler warnings.
1625
1626 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1627
1628 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1629 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1630 entirely. This removes changes made in 7 above.
1631
1632 17. The CMake files have been updated, and there is now more information about
1633 building with CMake in the NON-UNIX-USE document.
1634
1635
1636 Version 7.3 28-Aug-07
1637 ---------------------
1638
1639 1. In the rejigging of the build system that eventually resulted in 7.1, the
1640 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1641 brackets there is not right, since it causes compilers to look for an
1642 installed pcre.h, not the version that is in the source that is being
1643 compiled (which of course may be different). I have changed it back to:
1644
1645 #include "pcre.h"
1646
1647 I have a vague recollection that the change was concerned with compiling in
1648 different directories, but in the new build system, that is taken care of
1649 by the VPATH setting the Makefile.
1650
1651 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1652 when the subject happened to end in the byte 0x85 (e.g. if the last
1653 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1654 characters but of course it shouldn't be taken as a newline when it is part
1655 of another character. The bug was that, for an unlimited repeat of . in
1656 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1657 characters when looking for a newline.
1658
1659 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1660
1661 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1662 in debug output.
1663
1664 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1665 long printing in the pcrecpp unittest when running under MinGW.
1666
1667 6. ESC_K was left out of the EBCDIC table.
1668
1669 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1670 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1671 limit also applies to "virtual nesting" when a pattern is recursive, and in
1672 this case 1000 isn't so big. I have been able to remove this limit at the
1673 expense of backing off one optimization in certain circumstances. Normally,
1674 when pcre_exec() would call its internal match() function recursively and
1675 immediately return the result unconditionally, it uses a "tail recursion"
1676 feature to save stack. However, when a subpattern that can match an empty
1677 string has an unlimited repetition quantifier, it no longer makes this
1678 optimization. That gives it a stack frame in which to save the data for
1679 checking that an empty string has been matched. Previously this was taken
1680 from the 1000-entry workspace that had been reserved. So now there is no
1681 explicit limit, but more stack is used.
1682
1683 8. Applied Daniel's patches to solve problems with the import/export magic
1684 syntax that is required for Windows, and which was going wrong for the
1685 pcreposix and pcrecpp parts of the library. These were overlooked when this
1686 problem was solved for the main library.
1687
1688 9. There were some crude static tests to avoid integer overflow when computing
1689 the size of patterns that contain repeated groups with explicit upper
1690 limits. As the maximum quantifier is 65535, the maximum group length was
1691 set at 30,000 so that the product of these two numbers did not overflow a
1692 32-bit integer. However, it turns out that people want to use groups that
1693 are longer than 30,000 bytes (though not repeat them that many times).
1694 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1695 made it possible to implement the integer overflow checks in a much more
1696 dynamic way, which I have now done. The artificial limitation on group
1697 length has been removed - we now have only the limit on the total length of
1698 the compiled pattern, which depends on the LINK_SIZE setting.
1699
1700 10. Fixed a bug in the documentation for get/copy named substring when
1701 duplicate names are permitted. If none of the named substrings are set, the
1702 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1703 empty string.
1704
1705 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1706 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1707 because the ] is interpreted as the first data character and the
1708 terminating ] is not found. PCRE has been made compatible with Perl in this
1709 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1710 cause memory overwriting.
1711
1712 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1713 string has been matched (to stop an infinite loop). It was not recognizing
1714 a conditional subpattern that could match an empty string if that
1715 subpattern was within another subpattern. For example, it looped when
1716 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1717 condition was not nested. This bug has been fixed.
1718
1719 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1720 past the start of the subject in the presence of bytes with the top bit
1721 set, for example "\x8aBCD".
1722
1723 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1724 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1725
1726 14. Optimized (?!) to (*FAIL).
1727
1728 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1729 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1730 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1731 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1732 does: it's just the validity check that is more restrictive.
1733
1734 16. Inserted checks for integer overflows during escape sequence (backslash)
1735 processing, and also fixed erroneous offset values for syntax errors during
1736 backslash processing.
1737
1738 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1739 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1740
1741 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1742 caused an overrun.
1743
1744 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1745 something other than just ASCII characters) inside a group that had an
1746 unlimited repeat caused a loop at compile time (while checking to see
1747 whether the group could match an empty string).
1748
1749 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1750 [\P{Any}] did so. (Error in the code for printing property names.)
1751
1752 21. An orphan \E inside a character class could cause a crash.
1753
1754 22. A repeated capturing bracket such as (A)? could cause a wild memory
1755 reference during compilation.
1756
1757 23. There are several functions in pcre_compile() that scan along a compiled
1758 expression for various reasons (e.g. to see if it's fixed length for look
1759 behind). There were bugs in these functions when a repeated \p or \P was
1760 present in the pattern. These operators have additional parameters compared
1761 with \d, etc, and these were not being taken into account when moving along
1762 the compiled data. Specifically:
1763
1764 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1765 length.
1766
1767 (b) An item such as \pL+ within a repeated group could cause crashes or
1768 loops.
1769
1770 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1771 "reference to non-existent subpattern" error.
1772
1773 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1774
1775 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1776 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1777
1778 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1779
1780 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1781 character were causing crashes (broken optimization).
1782
1783 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1784 \p or \P) caused a compile-time loop.
1785
1786 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1787 break. For example, the unstudied pattern [\r\n]A does not match the string
1788 "\r\nA" because change 7.0/46 below moves the current point on by two
1789 characters after failing to match at the start. However, the pattern \nA
1790 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1791 the same is true. There doesn't seem any very clean way out of this, but
1792 what I have chosen to do makes the common cases work: PCRE now takes note
1793 of whether there can be an explicit match for \r or \n anywhere in the
1794 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1795 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1796 pattern has explicit CR or LF references.
1797
1798 29. Added (*CR) etc for changing newline setting at start of pattern.
1799
1800
1801 Version 7.2 19-Jun-07
1802 ---------------------
1803
1804 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1805 which is apparently normally available under Windows.
1806
1807 2. Re-jig the pcregrep tests with different newline settings in an attempt
1808 to make them independent of the local environment's newline setting.
1809
1810 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1811
1812 4. Some of the "internals" tests were previously cut out when the link size
1813 was not 2, because the output contained actual offsets. The recent new
1814 "Z" feature of pcretest means that these can be cut out, making the tests
1815 usable with all link sizes.
1816
1817 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1818 stack recursion. This gives a massive performance boost under BSD, but just
1819 a small improvement under Linux. However, it saves one field in the frame
1820 in all cases.
1821
1822 6. Added more features from the forthcoming Perl 5.10:
1823
1824 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1825 recursion call. It refers to the nth most recently opened parentheses.
1826
1827 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1828 to be opened parentheses.
1829
1830 (c) Conditions that refer to capturing parentheses can be specified
1831 relatively, for example, (?(-2)... or (?(+3)...
1832
1833 (d) \K resets the start of the current match so that everything before
1834 is not part of it.
1835
1836 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1837
1838 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1839 reference syntax.
1840
1841 (g) (?| introduces a group in which the numbering of parentheses in each
1842 alternative starts with the same number.
1843
1844 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1845
1846 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1847 PCRE_INFO_JCHANGED.
1848
1849 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1850 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1851 for detecting groups that can match an empty string.
1852
1853 9. A pattern with a very large number of alternatives (more than several
1854 hundred) was running out of internal workspace during the pre-compile
1855 phase, where pcre_compile() figures out how much memory will be needed. A
1856 bit of new cunning has reduced the workspace needed for groups with
1857 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1858 workspace instead of running out of the 4096 that are available.
1859
1860 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1861
1862 11. Applied patch from Google to remove an optimization that didn't quite work.
1863 The report of the bug said:
1864
1865 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1866 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1867 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1868
1869 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1870 it matched the wrong number of bytes.
1871
1872
1873 Version 7.1 24-Apr-07
1874 ---------------------
1875
1876 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1877 that is more "standard", making use of automake and other Autotools. There
1878 is some re-arrangement of the files and adjustment of comments consequent
1879 on this.
1880
1881 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1882 for recursive directory scanning broke on some systems because the files
1883 are not scanned in any specific order and on different systems the order
1884 was different. A call to "sort" has been inserted into RunGrepTest for the
1885 approprate test as a short-term fix. In the longer term there may be an
1886 alternative.
1887
1888 3. I had an email from Eric Raymond about problems translating some of PCRE's
1889 man pages to HTML (despite the fact that I distribute HTML pages, some
1890 people do their own conversions for various reasons). The problems
1891 concerned the use of low-level troff macros .br and .in. I have therefore
1892 removed all such uses from the man pages (some were redundant, some could
1893 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1894 HTML has been updated to handle .nf/.fi and to complain if it encounters
1895 .br or .in.
1896
1897 4. Updated comments in configure.ac that get placed in config.h.in and also
1898 arranged for config.h to be included in the distribution, with the name
1899 config.h.generic, for the benefit of those who have to compile without
1900 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1901
1902 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1903 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1904 makevp.bat for the current PCRE, using the additional files
1905 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1906
1907 6. A Windows user reported a minor discrepancy with test 2, which turned out
1908 to be caused by a trailing space on an input line that had got lost in his
1909 copy. The trailing space was an accident, so I've just removed it.
1910
1911 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1912 that is needed.
1913
1914 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1915 as "const" (a) because they are and (b) because it helps the PHP
1916 maintainers who have recently made a script to detect big data structures
1917 in the php code that should be moved to the .rodata section. I remembered
1918 to update Builducptable as well, so it won't revert if ucptable.h is ever
1919 re-created.
1920
1921 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1922 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1923 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1924 support is not required. This saves 1.5-2K of code, which is important in
1925 some applications.
1926
1927 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1928 so as not to refer to the tables, even though these functions will never be
1929 called when UTF-8 support is disabled. Otherwise there are problems with a
1930 shared library.
1931
1932 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1933
1934 (a) It was defining its arguments as char * instead of void *.
1935
1936 (b) It was assuming that all moves were upwards in memory; this was true
1937 a long time ago when I wrote it, but is no longer the case.
1938
1939 The emulated memove() is provided for those environments that have neither
1940 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1941 is clearly not the case, as these two bugs were recently reported.
1942
1943 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1944 and Detrail to create the HTML documentation, the .txt form of the man
1945 pages, and it removes trailing spaces from listed files. It also creates
1946 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1947 case, it wraps all the #defines with #ifndefs. This script should be run
1948 before "make dist".
1949
1950 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1951 with Unicode property support.
1952
1953 (a) For a maximizing quantifier, if the two different cases of the
1954 character were of different lengths in their UTF-8 codings (there are
1955 some cases like this - I found 11), and the matching function had to
1956 back up over a mixture of the two cases, it incorrectly assumed they
1957 were both the same length.
1958
1959 (b) When PCRE was configured to use the heap rather than the stack for
1960 recursion during matching, it was not correctly preserving the data for
1961 the other case of a UTF-8 character when checking ahead for a match
1962 while processing a minimizing repeat. If the check also involved
1963 matching a wide character, but failed, corruption could cause an
1964 erroneous result when trying to check for a repeat of the original
1965 character.
1966
1967 13. Some tidying changes to the testing mechanism:
1968
1969 (a) The RunTest script now detects the internal link size and whether there
1970 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1971 values substituted by "configure". (The RunGrepTest script already did
1972 this for UTF-8.) The configure.ac script no longer substitutes the
1973 relevant variables.
1974
1975 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1976 with length and offset values. This means that the output is different
1977 for different internal link sizes. Test 2 is skipped for link sizes
1978 other than 2 because of this, bypassing the problem. Unfortunately,
1979 there was also a test in test 3 (the locale tests) that used /B and
1980 failed for link sizes other than 2. Rather than cut the whole test out,
1981 I have added a new /Z option to pcretest that replaces the length and
1982 offset values with spaces. This is now used to make test 3 independent
1983 of link size. (Test 2 will be tidied up later.)
1984
1985 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1986 segmentation fault instead of returning the appropriate error message.
1987
1988 15. In multiline mode when the newline sequence was set to "any", the pattern
1989 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1990 This doesn't seem right; it now treats the CRLF combination as the line
1991 ending, and so does not match in that case. It's only a pattern such as ^$
1992 that would hit this one: something like ^ABC$ would have failed after \r
1993 and then tried again after \r\n.
1994
1995 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1996 in an attempt to make files that differ only in their line terminators
1997 compare equal. This works on Linux.
1998
1999 17. Under certain error circumstances pcregrep might try to free random memory
2000 as it exited. This is now fixed, thanks to valgrind.
2001
2002 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2003 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2004 was because its rules for how to advance for /g after matching an empty
2005 string at the end of a line did not allow for this case. They now check for
2006 it specially.
2007
2008 20. pcretest is supposed to handle patterns and data of any length, by
2009 extending its buffers when necessary. It was getting this wrong when the
2010 buffer for a data line had to be extended.
2011
2012 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2013 CRLF as a newline sequence.
2014
2015 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2016 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2017 I have nevertheless tidied it up.
2018
2019 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2020
2021 24. Added a man page for pcre-config.
2022
2023
2024 Version 7.0 19-Dec-06
2025 ---------------------
2026
2027 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2028 moving to gcc 4.1.1.
2029
2030 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2031 sys/time.h, which is documented as needed for this function. It doesn't
2032 seem to matter on Linux, but it showed up on some releases of OS X.
2033
2034 3. It seems that there are systems where bytes whose values are greater than
2035 127 match isprint() in the "C" locale. The "C" locale should be the
2036 default when a C program starts up. In most systems, only ASCII printing
2037 characters match isprint(). This difference caused the output from pcretest
2038 to vary, making some of the tests fail. I have changed pcretest so that:
2039
2040 (a) When it is outputting text in the compiled version of a pattern, bytes
2041 other than 32-126 are always shown as hex escapes.
2042
2043 (b) When it is outputting text that is a matched part of a subject string,
2044 it does the same, unless a different locale has been set for the match
2045 (using the /L modifier). In this case, it uses isprint() to decide.
2046
2047 4. Fixed a major bug that caused incorrect computation of the amount of memory
2048 required for a compiled pattern when options that changed within the
2049 pattern affected the logic of the preliminary scan that determines the
2050 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2051 that the computed length was too small. The symptoms of this bug were
2052 either the PCRE error "internal error: code overflow" from pcre_compile(),
2053 or a glibc crash with a message such as "pcretest: free(): invalid next
2054 size (fast)". Examples of patterns that provoked this bug (shown in
2055 pcretest format) are:
2056
2057 /(?-x: )/x
2058 /(?x)(?-x: \s*#\s*)/
2059 /((?i)[\x{c0}])/8
2060 /(?i:[\x{c0}])/8
2061
2062 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2063 is now done differently.
2064
2065 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2066 wrapper classes; (b) implement a new function in the C++ scanner that is
2067 more efficient than the old way of doing things because it avoids levels of
2068 recursion in the regex matching; (c) add a paragraph to the documentation
2069 for the FullMatch() function.
2070
2071 6. The escape sequence \n was being treated as whatever was defined as
2072 "newline". Not only was this contrary to the documentation, which states
2073 that \n is character 10 (hex 0A), but it also went horribly wrong when
2074 "newline" was defined as CRLF. This has been fixed.
2075
2076 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2077 was being set to -1 for the "end of line" case (supposedly a value that no
2078 character can have). Though this value is never used (the check for end of
2079 line is "zero bytes in current character"), it caused compiler complaints.
2080 I've changed it to 0xffffffff.
2081
2082 8. In pcre_version.c, the version string was being built by a sequence of
2083 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2084 string (as it is for production releases) called a macro with an empty
2085 argument. The C standard says the result of this is undefined. The gcc
2086 compiler treats it as an empty string (which was what was wanted) but it is
2087 reported that Visual C gives an error. The source has been hacked around to
2088 avoid this problem.
2089
2090 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2091 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2092 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2093 of them did).
2094
2095 10. Originally, pcretest opened its input and output without "b"; then I was
2096 told that "b" was needed in some environments, so it was added for release
2097 5.0 to both the input and output. (It makes no difference on Unix-like
2098 systems.) Later I was told that it is wrong for the input on Windows. I've
2099 now abstracted the modes into two macros, to make it easier to fiddle with
2100 them, and removed "b" from the input mode under Windows.
2101
2102 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2103
2104 12. Added -help and --help to pcretest as an official way of being reminded
2105 of the options.
2106
2107 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2108 and pcrecpp.cc because they annoy compilers at high warning levels.
2109
2110 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2111
2112 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2113 scripts are not C programs :-) and which was not noticed because it works
2114 on Linux.
2115
2116 16. pcretest is supposed to handle any length of pattern and data line (as one
2117 line or as a continued sequence of lines) by extending its input buffer if
2118 necessary. This feature was broken for very long pattern lines, leading to
2119 a string of junk being passed to pcre_compile() if the pattern was longer
2120 than about 50K.
2121
2122 17. I have done a major re-factoring of the way pcre_compile() computes the
2123 amount of memory needed for a compiled pattern. Previously, there was code
2124 that made a preliminary scan of the pattern in order to do this. That was
2125 OK when PCRE was new, but as the facilities have expanded, it has become
2126 harder and harder to keep it in step with the real compile phase, and there
2127 have been a number of bugs (see for example, 4 above). I have now found a
2128 cunning way of running the real compile function in a "fake" mode that
2129 enables it to compute how much memory it would need, while actually only
2130 ever using a few hundred bytes of working memory and without too many
2131 tests of the mode. This should make future maintenance and development
2132 easier. A side effect of this work is that the limit of 200 on the nesting
2133 depth of parentheses has been removed (though this was never a serious
2134 limitation, I suspect). However, there is a downside: pcre_compile() now
2135 runs more slowly than before (30% or more, depending on the pattern). I
2136 hope this isn't a big issue. There is no effect on runtime performance.
2137
2138 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2139 newline (only possible for the last line of a file) and it was a
2140 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2141
2142 19. Added additional timing features to pcretest. (1) The -tm option now times
2143 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2144 separate command line item, by a number that specifies the number of
2145 repeats to use when timing. The default is 50000; this gives better
2146 precision, but takes uncomfortably long for very large patterns.
2147
2148 20. Extended pcre_study() to be more clever in cases where a branch of a
2149 subpattern has no definite first character. For example, (a*|b*)[cd] would
2150 previously give no result from pcre_study(). Now it recognizes that the
2151 first character must be a, b, c, or d.
2152
2153 21. There was an incorrect error "recursive call could loop indefinitely" if
2154 a subpattern (or the entire pattern) that was being tested for matching an
2155 empty string contained only one non-empty item after a nested subpattern.
2156 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2157 incorrectly, because the \d was being skipped in the check.
2158
2159 22. The pcretest program now has a new pattern option /B and a command line
2160 option -b, which is equivalent to adding /B to every pattern. This causes
2161 it to show the compiled bytecode, without the additional information that
2162 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2163 is the same as /B/I).
2164
2165 23. A new optimization is now able automatically to treat some sequences such
2166 as a*b as a*+b. More specifically, if something simple (such as a character
2167 or a simple class like \d) has an unlimited quantifier, and is followed by
2168 something that cannot possibly match the quantified thing, the quantifier
2169 is automatically "possessified".
2170
2171 24. A recursive reference to a subpattern whose number was greater than 39
2172 went wrong under certain circumstances in UTF-8 mode. This bug could also
2173 have affected the operation of pcre_study().
2174
2175 25. Realized that a little bit of performance could be had by replacing
2176 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2177
2178 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2179
2180 27. Possessive quantifiers such as a++ were previously implemented by turning
2181 them into atomic groups such as ($>a+). Now they have their own opcodes,
2182 which improves performance. This includes the automatically created ones
2183 from 23 above.
2184
2185 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2186 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2187 the first matched character to be a colon. This applied both to named and
2188 numbered groups.
2189
2190 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2191
2192 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2193 building PCRE on Windows easier, so I have included it in the distribution.
2194
2195 31. There is now a check in pcretest against a ridiculously large number being
2196 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2197 loop, the loop is abandoned.
2198
2199 32. Forward references to subpatterns in conditions such as (?(2)...) where
2200 subpattern 2 is defined later cause pcre_compile() to search forwards in
2201 the pattern for the relevant set of parentheses. This search went wrong
2202 when there were unescaped parentheses in a character class, parentheses
2203 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2204
2205 33. "Subroutine" calls and backreferences were previously restricted to
2206 referencing subpatterns earlier in the regex. This restriction has now
2207 been removed.
2208
2209 34. Added a number of extra features that are going to be in Perl 5.10. On the
2210 whole, these are just syntactic alternatives for features that PCRE had
2211 previously implemented using the Python syntax or my own invention. The
2212 other formats are all retained for compatibility.
2213
2214 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2215 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2216 also .NET compatible.
2217
2218 (b) A recursion or subroutine call to a named group can now be defined as
2219 (?&name) as well as (?P>name).
2220
2221 (c) A backreference to a named group can now be defined as \k<name> or
2222 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2223 5.10, are also .NET compatible.
2224
2225 (d) A conditional reference to a named group can now use the syntax
2226 (?(<name>) or (?('name') as well as (?(name).
2227
2228 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2229 groups (named and numbered) that are never evaluated inline, but can be
2230 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2231 is always false. There may be only one alternative in such a group.
2232
2233 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2234 as the simple (?(R). The condition is true only if the most recent
2235 recursion is that of the given number or name. It does not search out
2236 through the entire recursion stack.
2237
2238 (g) The escape \gN or \g{N} has been added, where N is a positive or
2239 negative number, specifying an absolute or relative reference.
2240
2241 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2242 some "unreachable code" warnings.
2243
2244 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2245 things, this adds five new scripts.
2246
2247 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2248 There were also incompatibilities regarding the handling of \Q..\E inside
2249 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2250 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2251
2252 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2253 matches an empty string, and forcibly breaks the loop. There were bugs in
2254 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2255 against aaaa the result was just "a" rather than "aaaa", for example. Two
2256 separate and independent bugs (that affected different cases) have been
2257 fixed.
2258
2259 39. Refactored the code to abolish the use of different opcodes for small
2260 capturing bracket numbers. This is a tidy that I avoided doing when I
2261 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2262 The new approach is not only tidier, it makes it possible to reduce the
2263 memory needed to fix the previous bug (38).
2264
2265 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2266 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2267 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2268 mode.
2269
2270 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2271 report.
2272
2273 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2274 copy construction and assignment in the C++ wrapper.
2275
2276 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2277 couple of bugs that could have given wrong results in the "--newline=crlf"
2278 case.
2279
2280 44. Added a number of casts and did some reorganization of signed/unsigned int
2281 variables following suggestions from Dair Grant. Also renamed the variable
2282 "this" as "item" because it is a C++ keyword.
2283
2284 45. Arranged for dftables to add
2285
2286 #include "pcre_internal.h"
2287
2288 to pcre_chartables.c because without it, gcc 4.x may remove the array
2289 definition from the final binary if PCRE is built into a static library and
2290 dead code stripping is activated.
2291
2292 46. For an unanchored pattern, if a match attempt fails at the start of a
2293 newline sequence, and the newline setting is CRLF or ANY, and the next two
2294 characters are CRLF, advance by two characters instead of one.
2295
2296
2297 Version 6.7 04-Jul-06
2298 ---------------------
2299
2300 1. In order to handle tests when input lines are enormously long, pcretest has
2301 been re-factored so that it automatically extends its buffers when
2302 necessary. The code is crude, but this _is_ just a test program. The
2303 default size has been increased from 32K to 50K.
2304
2305 2. The code in pcre_study() was using the value of the re argument before
2306 testing it for NULL. (Of course, in any sensible call of the function, it
2307 won't be NULL.)
2308
2309 3. The memmove() emulation function in pcre_internal.h, which is used on
2310 systems that lack both memmove() and bcopy() - that is, hardly ever -
2311 was missing a "static" storage class specifier.
2312
2313 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2314 containing an extended class (one that cannot be represented by a bitmap
2315 because it contains high-valued characters or Unicode property items, e.g.
2316 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2317 pattern, but PCRE should not loop if you do not (it no longer does).
2318 [Detail: two cases were found: (a) a repeated subpattern containing an
2319 extended class; (b) a recursive reference to a subpattern that followed a
2320 previous extended class. It wasn't skipping over the extended class
2321 correctly when UTF-8 mode was not set.]
2322
2323 5. A negated single-character class was not being recognized as fixed-length
2324 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2325 compile error "lookbehind assertion is not fixed length".
2326
2327 6. The RunPerlTest auxiliary script was showing an unexpected difference
2328 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2329 write a Perl script that can interpret lines of an input file either as
2330 byte characters or as UTF-8, which is what "perltest" was being required to
2331 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2332 can't do is switch easily at run time between having the "use utf8;" pragma
2333 or not. In the end, I fudged it by using the RunPerlTest script to insert
2334 "use utf8;" explicitly for the UTF-8 tests.
2335
2336 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2337 the end of the subject string, contrary to the documentation and to what
2338 Perl does. This was true of both matching functions. Now it matches only at
2339 the start of the subject and immediately after *internal* newlines.
2340
2341 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2342 a pointer to an int instead of a pointer to an unsigned long int. This
2343 caused problems on 64-bit systems.
2344
2345 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2346 instance of the 'standard' template library not being so standard".
2347
2348 10. There was no check on the number of named subpatterns nor the maximum
2349 length of a subpattern name. The product of these values is used to compute
2350 the size of the memory block for a compiled pattern. By supplying a very
2351 long subpattern name and a large number of named subpatterns, the size
2352 computation could be caused to overflow. This is now prevented by limiting
2353 the length of names to 32 characters, and the number of named subpatterns
2354 to 10,000.
2355
2356 11. Subpatterns that are repeated with specific counts have to be replicated in
2357 the compiled pattern. The size of memory for this was computed from the
2358 length of the subpattern and the repeat count. The latter is limited to
2359 65535, but there was no limit on the former, meaning that integer overflow
2360 could in principle occur. The compiled length of a repeated subpattern is
2361 now limited to 30,000 bytes in order to prevent this.
2362
2363 12. Added the optional facility to have named substrings with the same name.
2364
2365 13. Added the ability to use a named substring as a condition, using the
2366 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2367 are numbers (not recommended). Forward references are permitted.
2368
2369 14. Added forward references in named backreferences (if you see what I mean).
2370
2371 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2372 pattern could run off the end of the subject. For example, the pattern
2373 "(?s)(.{1,5})"8 did this with the subject "ab".
2374
2375 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2376 PCRE_CASELESS was set when matching characters that were quantified with ?
2377 or *.
2378
2379 17. A character class other than a single negated character that had a minimum
2380 but no maximum quantifier - for example [ab]{6,} - was not handled
2381 correctly by pce_dfa_exec(). It would match only one character.
2382
2383 18. A valid (though odd) pattern that looked like a POSIX character
2384 class but used an invalid character after [ (for example [[,abc,]]) caused
2385 pcre_compile() to give the error "Failed: internal error: code overflow" or
2386 in some cases to crash with a glibc free() error. This could even happen if
2387 the pattern terminated after [[ but there just happened to be a sequence of
2388 letters, a binary zero, and a closing ] in the memory that followed.
2389
2390 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2391 over the years. Originally (before any Unicode support), just the bottom 8
2392 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2393 output from "man perlunicode" includes this:
2394
2395 The regular expression compiler produces polymorphic opcodes. That
2396 is, the pattern adapts to the data and automatically switches to
2397 the Unicode character scheme when presented with Unicode data--or
2398 instead uses a traditional byte scheme when presented with byte
2399 data.
2400
2401 Sadly, a wide octal escape does not cause a switch, and in a string with
2402 no other multibyte characters, these octal escapes are treated as before.
2403 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2404 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2405 Unicode string.
2406
2407 I have not perpetrated such confusion in PCRE. Up till now, it took just
2408 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2409 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2410 translate to the appropriate multibyte character.
2411
2412 29. Applied some refactoring to reduce the number of warnings from Microsoft
2413 and Borland compilers. This has included removing the fudge introduced
2414 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2415 a warning about an unused variable.
2416
2417 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2418 characters since release 4.0, because Perl (from release 5.004) does not.
2419 [Or at least, is documented not to: some releases seem to be in conflict
2420 with the documentation.] However, when a pattern was studied with
2421 pcre_study() and all its branches started with \s, PCRE still included VT
2422 as a possible starting character. Of course, this did no harm; it just
2423 caused an unnecessary match attempt.
2424
2425 22. Removed a now-redundant internal flag bit that recorded the fact that case
2426 dependency changed within the pattern. This was once needed for "required
2427 byte" processing, but is no longer used. This recovers a now-scarce options
2428 bit. Also moved the least significant internal flag bit to the most-
2429 significant bit of the word, which was not previously used (hangover from
2430 the days when it was an int rather than a uint) to free up another bit for
2431 the future.
2432
2433 23. Added support for CRLF line endings as well as CR and LF. As well as the
2434 default being selectable at build time, it can now be changed at runtime
2435 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2436 specify that it is scanning data with non-default line endings.
2437
2438 24. Changed the definition of CXXLINK to make it agree with the definition of
2439 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2440
2441 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2442 recursions. This makes a big different to stack usage for some patterns.
2443
2444 26. If a subpattern containing a named recursion or subroutine reference such
2445 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2446 the space required for the compiled pattern went wrong and gave too small a
2447 value. Depending on the environment, this could lead to "Failed: internal
2448 error: code overflow at offset 49" or "glibc detected double free or
2449 corruption" errors.
2450
2451 27. Applied patches from Google (a) to support the new newline modes and (b) to
2452 advance over multibyte UTF-8 characters in GlobalReplace.
2453
2454 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2455 difference for some implementation of PCRE in some Windows version.
2456
2457 29. Added some extra testing facilities to pcretest:
2458
2459 \q<number> in a data line sets the "match limit" value
2460 \Q<number> in a data line sets the "match recursion limt" value
2461 -S <number> sets the stack size, where <number> is in megabytes
2462
2463 The -S option isn't available for Windows.
2464
2465
2466 Version 6.6 06-Feb-06
2467 ---------------------
2468
2469 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2470 in pcreposix.h. I have copied the definition from pcre.h.
2471
2472 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2473 because pcre.h is no longer a built file.
2474
2475 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2476 not normally included in the compiled code.
2477
2478
2479 Version 6.5 01-Feb-06
2480 ---------------------
2481
2482 1. When using the partial match feature with pcre_dfa_exec(), it was not
2483 anchoring the second and subsequent partial matches at the new starting
2484 point. This could lead to incorrect results. For example, with the pattern
2485 /1234/, partially matching against "123" and then "a4" gave a match.
2486
2487 2. Changes to pcregrep:
2488
2489 (a) All non-match returns from pcre_exec() were being treated as failures
2490 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2491 error message is output. Some extra information is given for the
2492 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2493 probably the only errors that are likely to be caused by users (by
2494 specifying a regex that has nested indefinite repeats, for instance).
2495 If there are more than 20 of these errors, pcregrep is abandoned.
2496
2497 (b) A binary zero was treated as data while matching, but terminated the
2498 output line if it was written out. This has been fixed: binary zeroes
2499 are now no different to any other data bytes.
2500
2501 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2502 used to set a locale for matching. The --locale=xxxx long option has
2503 been added (no short equivalent) to specify a locale explicitly on the
2504 pcregrep command, overriding the environment variables.
2505
2506 (d) When -B was used with -n, some line numbers in the output were one less
2507 than they should have been.
2508
2509 (e) Added the -o (--only-matching) option.
2510
2511 (f) If -A or -C was used with -c (count only), some lines of context were
2512 accidentally printed for the final match.
2513
2514 (g) Added the -H (--with-filename) option.
2515
2516 (h) The combination of options -rh failed to suppress file names for files
2517 that were found from directory arguments.
2518
2519 (i) Added the -D (--devices) and -d (--directories) options.
2520
2521 (j) Added the -F (--fixed-strings) option.
2522
2523 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2524
2525 (l) Added the --colo(u)r option.
2526
2527 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2528 is not present by default.
2529
2530 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2531 items such as (?R) or (?1), when the recursion could match a number of
2532 alternatives. If it matched one of the alternatives, but subsequently,
2533 outside the recursion, there was a failure, the code tried to back up into
2534 the recursion. However, because of the way PCRE is implemented, this is not
2535 possible, and the result was an incorrect result from the match.
2536
2537 In order to prevent this happening, the specification of recursion has
2538 been changed so that all such subpatterns are automatically treated as
2539 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2540
2541 4. I had overlooked the fact that, in some locales, there are characters for
2542 which isalpha() is true but neither isupper() nor islower() are true. In
2543 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2544 and ordfeminine) are like this. This affected the treatment of \w and \W
2545 when they appeared in character classes, but not when they appeared outside
2546 a character class. The bit map for "word" characters is now created
2547 separately from the results of isalnum() instead of just taking it from the
2548 upper, lower, and digit maps. (Plus the underscore character, of course.)
2549
2550 5. The above bug also affected the handling of POSIX character classes such as
2551 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2552 permanent tables. Instead, the bit maps for such a class were previously
2553 created as the appropriate unions of the upper, lower, and digit bitmaps.
2554 Now they are created by subtraction from the [[:word:]] class, which has
2555 its own bitmap.
2556
2557 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2558 It is created by subtracting the vertical space characters (\x09, \x0a,
2559 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2560 subtraction was done in the overall bitmap for a character class, meaning
2561 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2562 be recognized. This bug has been fixed.
2563
2564 7. Patches from the folks at Google:
2565
2566 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2567 real life, but is still worth protecting against".
2568
2569 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2570 regular expressions".
2571
2572 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2573 have it.
2574
2575 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2576 "configure" and the latter not, in order to fix a problem somebody had
2577 with compiling the Arg class on HP-UX.
2578
2579 (e) Improve the error-handling of the C++ wrapper a little bit.
2580
2581 (f) New tests for checking recursion limiting.
2582
2583 8. The pcre_memmove() function, which is used only if the environment does not
2584 have a standard memmove() function (and is therefore rarely compiled),
2585 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2586 returning a result (though PCRE never actually uses the result).
2587
2588 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2589 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2590 returned instead of calling malloc() with an overflowing number that would
2591 most likely cause subsequent chaos.
2592
2593 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2594
2595 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2596 with this option is matched, the nmatch and pmatch options of regexec() are
2597 ignored.
2598
2599 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2600 provided in case anyone wants to the the POSIX interface with UTF-8
2601 strings.
2602
2603 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2604 C++ linking (needed for some HP-UX environments).
2605
2606 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2607 (unused parameter) and in the pcre_printint() function (omitted "default"
2608 switch label when the default is to do nothing).
2609
2610 15. Added some code to make it possible, when PCRE is compiled as a C++
2611 library, to replace subject pointers for pcre_exec() with a smart pointer
2612 class, thus making it possible to process discontinuous strings.
2613
2614 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2615 much the same function. They were added by different people who were trying
2616 to make PCRE easy to compile on non-Unix systems. It has been suggested
2617 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2618 for compiling on Windows systems. I have therefore replaced it with
2619 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2620 defaults to "extern" for C or "extern C" for C++, which works fine on
2621 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2622 SCOPE with something explicit in config.h. In addition:
2623
2624 (a) pcreposix.h still had just "extern" instead of either of these macros;
2625 I have replaced it with PCRE_DATA_SCOPE.
2626
2627 (b) Functions such as _pcre_xclass(), which are internal to the library,
2628 but external in the C sense, all had PCRE_EXPORT in their definitions.
2629 This is apparently wrong for the Windows case, so I have removed it.
2630 (It makes no difference on Unix-like systems.)
2631
2632 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2633 of recursive calls to match(). This is different to MATCH_LIMIT because
2634 that limits the total number of calls to match(), not all of which increase
2635 the depth of recursion. Limiting the recursion depth limits the amount of
2636 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2637 when PCRE is compiled, and changed at run time. A patch from Google adds
2638 this functionality to the C++ interface.
2639
2640 18. Changes to the handling of Unicode character properties:
2641
2642 (a) Updated the table to Unicode 4.1.0.
2643
2644 (b) Recognize characters that are not in the table as "Cn" (undefined).
2645
2646 (c) I revised the way the table is implemented to a much improved format
2647 which includes recognition of ranges. It now supports the ranges that
2648 are defined in UnicodeData.txt, and it also amalgamates other
2649 characters into ranges. This has reduced the number of entries in the
2650 table from around 16,000 to around 3,000, thus reducing its size
2651 considerably. I realized I did not need to use a tree structure after
2652 all - a binary chop search is just as efficient. Having reduced the
2653 number of entries, I extended their size from 6 bytes to 8 bytes to
2654 allow for more data.
2655
2656 (d) Added support for Unicode script names via properties such as \p{Han}.
2657
2658 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2659 matching that character.
2660
2661 20. When matching a repeated Unicode property with a minimum greater than zero,
2662 (for example \pL{2,}), PCRE could look past the end of the subject if it
2663 reached it while seeking the minimum number of characters. This could
2664 happen only if some of the characters were more than one byte long, because
2665 there is a check for at least the minimum number of bytes.
2666
2667 21. Refactored the implementation of \p and \P so as to be more general, to
2668 allow for more different types of property in future. This has changed the
2669 compiled form incompatibly. Anybody with saved compiled patterns that use
2670 \p or \P will have to recompile them.
2671
2672 22. Added "Any" and "L&" to the supported property types.
2673
2674 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2675 but give a compile time error if the value is greater than 0xff.
2676
2677 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2678 accidentally not being installed or uninstalled.
2679
2680 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2681 made were to insert the current release number. This seemed silly, because
2682 it made things harder for people building PCRE on systems that don't run
2683 "configure". I have turned pcre.h into a distributed file, no longer built
2684 by "configure", with the version identification directly included. There is
2685 no longer a pcre.h.in file.
2686
2687 However, this change necessitated a change to the pcre-config script as
2688 well. It is built from pcre-config.in, and one of the substitutions was the
2689 release number. I have updated configure.ac so that ./configure now finds
2690 the release number by grepping pcre.h.
2691
2692 26. Added the ability to run the tests under valgrind.
2693
2694
2695 Version 6.4 05-Sep-05
2696 ---------------------
2697
2698 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2699 "--" to be printed when multiple files were scanned, even when none of the
2700 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2701 consider it to be a bug, and have restored the previous behaviour.
2702
2703 2. A couple of code tidies to get rid of compiler warnings.
2704
2705 3. The pcretest program used to cheat by referring to symbols in the library
2706 whose names begin with _pcre_. These are internal symbols that are not
2707 really supposed to be visible externally, and in some environments it is
2708 possible to suppress them. The cheating is now confined to including
2709 certain files from the library's source, which is a bit cleaner.
2710
2711 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2712 file's purpose clearer.
2713
2714 5. Reorganized pcre_ucp_findchar().
2715
2716
2717 Version 6.3 15-Aug-05
2718 ---------------------
2719
2720 1. The file libpcre.pc.in did not have general read permission in the tarball.
2721
2722 2. There were some problems when building without C++ support:
2723
2724 (a) If C++ support was not built, "make install" and "make test" still
2725 tried to test it.
2726
2727 (b) There were problems when the value of CXX was explicitly set. Some
2728 changes have been made to try to fix these, and ...
2729
2730 (c) --disable-cpp can now be used to explicitly disable C++ support.
2731
2732 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2733 backslash in a target when C++ was disabled. This confuses some
2734 versions of "make", apparently. Using an intermediate variable solves
2735 this. (Same for CPP_LOBJ.)
2736
2737 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2738 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2739 necessary on certain architectures.
2740
2741 4. Added a setting of -export-symbols-regex to the link command to remove
2742 those symbols that are exported in the C sense, but actually are local
2743 within the library, and not documented. Their names all begin with
2744 "_pcre_". This is not a perfect job, because (a) we have to except some
2745 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2746 available (and never for static libraries). I have made a note to try to
2747 find a way round (a) in the future.
2748
2749
2750 Version 6.2 01-Aug-05
2751 ---------------------
2752
2753 1. There was no test for integer overflow of quantifier values. A construction
2754 such as {1111111111111111} would give undefined results. What is worse, if
2755 a minimum quantifier for a parenthesized subpattern overflowed and became
2756 negative, the calculation of the memory size went wrong. This could have
2757 led to memory overwriting.
2758
2759 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2760
2761 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2762 operating environments where this matters.
2763
2764 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2765 PCRE options from within the C++ wrapper.
2766
2767 5. Named capturing subpatterns were not being correctly counted when a pattern
2768 was compiled. This caused two problems: (a) If there were more than 100
2769 such subpatterns, the calculation of the memory needed for the whole
2770 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2771 back references of the form \12, where the number was greater than 9, were
2772 not recognized as back references, even though there were sufficient
2773 previous subpatterns.
2774
2775 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2776 versions of gcc, e.g. 2.95.4.
2777
2778
2779 Version 6.1 21-Jun-05
2780 ---------------------
2781
2782 1. There was one reference to the variable "posix" in pcretest.c that was not
2783 surrounded by "#if !defined NOPOSIX".
2784
2785 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2786 the cross-check on the old pcre_info() function, for the benefit of the
2787 cut-down version of PCRE that is currently imported into Exim.
2788
2789 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2790 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2791 patterns that start (?i) but I don't think that matters. The use of (?i) is
2792 just an example; this all applies to the other options as well.
2793
2794 4. Since libtool seems to echo the compile commands it is issuing, the output
2795 from "make" can be reduced a bit by putting "@" in front of each libtool
2796 compile command.
2797
2798 5. Patch from the folks at Google for configure.in to be a bit more thorough
2799 in checking for a suitable C++ installation before trying to compile the
2800 C++ stuff. This should fix a reported problem when a compiler was present,
2801 but no suitable headers.
2802
2803 6. The man pages all had just "PCRE" as their title. I have changed them to
2804 be the relevant file name. I have also arranged that these names are
2805 retained in the file doc/pcre.txt, which is a concatenation in text format
2806 of all the man pages except the little individual ones for each function.
2807
2808 7. The NON-UNIX-USE file had not been updated for the different set of source
2809 files that come with release 6. I also added a few comments about the C++
2810 wrapper.
2811
2812
2813 Version 6.0 07-Jun-05
2814 ---------------------
2815
2816 1. Some minor internal re-organization to help with my DFA experiments.
2817
2818 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2819 didn't matter for the library itself when fully configured, but did matter
2820 when compiling without UCP support, or within Exim, where the ucp files are
2821 not imported.
2822
2823 3. Refactoring of the library code to split up the various functions into
2824 different source modules. The addition of the new DFA matching code (see
2825 below) to a single monolithic source would have made it really too
2826 unwieldy, quite apart from causing all the code to be include in a
2827 statically linked application, when only some functions are used. This is
2828 relevant even without the DFA addition now that patterns can be compiled in
2829 one application and matched in another.
2830
2831 The downside of splitting up is that there have to be some external
2832 functions and data tables that are used internally in different modules of
2833 the library but which are not part of the API. These have all had their
2834 names changed to start with "_pcre_" so that they are unlikely to clash
2835 with other external names.
2836
2837 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2838 a different (DFA) algorithm. Although it is slower than the original
2839 function, it does have some advantages for certain types of matching
2840 problem.
2841
2842 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2843 including restarting after a partial match.
2844
2845 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2846 defined when compiling for Windows was sent to me. I have put it into the
2847 code, though I have no means of testing or verifying it.
2848
2849 7. Added the pcre_refcount() auxiliary function.
2850
2851 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2852 match before or at the first newline in the subject string. In pcretest,
2853 the /f option on a pattern can be used to set this.
2854
2855 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2856 would behave wrongly. This has been present in PCRE since release 4.0.
2857
2858 10. A number of changes to the pcregrep command:
2859
2860 (a) Refactored how -x works; insert ^(...)$ instead of setting
2861 PCRE_ANCHORED and checking the length, in preparation for adding
2862 something similar for -w.
2863
2864 (b) Added the -w (match as a word) option.
2865
2866 (c) Refactored the way lines are read and buffered so as to have more
2867 than one at a time available.
2868
2869 (d) Implemented a pcregrep test script.
2870
2871 (e) Added the -M (multiline match) option. This allows patterns to match
2872 over several lines of the subject. The buffering ensures that at least
2873 8K, or the rest of the document (whichever is the shorter) is available
2874 for matching (and similarly the previous 8K for lookbehind assertions).
2875
2876 (f) Changed the --help output so that it now says
2877
2878 -w, --word-regex(p)
2879
2880 instead of two lines, one with "regex" and the other with "regexp"
2881 because that confused at least one person since the short forms are the
2882 same. (This required a bit of code, as the output is generated
2883 automatically from a table. It wasn't just a text change.)
2884
2885 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2886 option but starts with a hyphen. Could be a pattern or a path name
2887 starting with a hyphen, for instance.
2888
2889 (h) "-" can be given as a file name to represent stdin.
2890
2891 (i) When file names are being printed, "(standard input)" is used for
2892 the standard input, for compatibility with GNU grep. Previously
2893 "<stdin>" was used.
2894
2895 (j) The option --label=xxx can be used to supply a name to be used for
2896 stdin when file names are being printed. There is no short form.
2897
2898 (k) Re-factored the options decoding logic because we are going to add
2899 two more options that take data. Such options can now be given in four
2900 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2901
2902 (l) Added the -A, -B, and -C options for requesting that lines of context
2903 around matches be printed.
2904
2905 (m) Added the -L option to print the names of files that do not contain
2906 any matching lines, that is, the complement of -l.
2907
2908 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2909 continue to scan other files.
2910
2911 (o) The -s option was incorrectly implemented. For compatibility with other
2912 greps, it now suppresses the error message for a non-existent or non-
2913 accessible file (but not the return code). There is a new option called
2914 -q that suppresses the output of matching lines, which was what -s was
2915 previously doing.
2916
2917 (p) Added --include and --exclude options to specify files for inclusion
2918 and exclusion when recursing.
2919
2920 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2921 Hopefully, it now does.
2922
2923 12. Missing cast in pcre_study().
2924
2925 13. Added an "uninstall" target to the makefile.
2926
2927 14. Replaced "extern" in the function prototypes in Makefile.in with
2928 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2929 world, but is set differently for Windows.
2930
2931 15. Added a second compiling function called pcre_compile2(). The only
2932 difference is that it has an extra argument, which is a pointer to an
2933 integer error code. When there is a compile-time failure, this is set
2934 non-zero, in addition to the error test pointer being set to point to an
2935 error message. The new argument may be NULL if no error number is required
2936 (but then you may as well call pcre_compile(), which is now just a
2937 wrapper). This facility is provided because some applications need a
2938 numeric error indication, but it has also enabled me to tidy up the way
2939 compile-time errors are handled in the POSIX wrapper.
2940
2941 16. Added VPATH=.libs to the makefile; this should help when building with one
2942 prefix path and installing with another. (Or so I'm told by someone who
2943 knows more about this stuff than I do.)
2944
2945 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2946 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2947 match everything, including newlines. This is not POSIX-compatible, but
2948 somebody wanted the feature. From pcretest it can be activated by using
2949 both the P and the s flags.
2950
2951 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2952
2953 19. libpcre.pc was being incorrectly installed as executable.
2954
2955 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2956 it now also looks for '\r' so that it will work unmodified on Windows.
2957
2958 21. Added Google's contributed C++ wrapper to the distribution.
2959
2960 22. Added some untidy missing memory free() calls in pcretest, to keep
2961 Electric Fence happy when testing.
2962
2963
2964
2965 Version 5.0 13-Sep-04
2966 ---------------------
2967
2968 1. Internal change: literal characters are no longer packed up into items
2969 containing multiple characters in a single byte-string. Each character
2970 is now matched using a separate opcode. However, there may be more than one
2971 byte in the character in UTF-8 mode.
2972
2973 2. The pcre_callout_block structure has two new fields: pattern_position and
2974 next_item_length. These contain the offset in the pattern to the next match
2975 item, and its length, respectively.
2976
2977 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2978 insertion of callouts before each pattern item. Added the /C option to
2979 pcretest to make use of this.
2980
2981 4. On the advice of a Windows user, the lines
2982
2983 #if defined(_WIN32) || defined(WIN32)
2984 _setmode( _fileno( stdout ), 0x8000 );
2985 #endif /* defined(_WIN32) || defined(WIN32) */
2986
2987 have been added to the source of pcretest. This apparently does useful
2988 magic in relation to line terminators.
2989
2990 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2991 for the benefit of those environments where the "b" makes a difference.
2992
2993 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2994 to know about it. I have put a hack into configure.in that adds in code
2995 to set GCC=yes if CC=icc. This seems to end up at a point in the
2996 generated configure script that is early enough to affect the setting of
2997 compiler options, which is what is needed, but I have no means of testing
2998 whether it really works. (The user who reported this had patched the
2999 generated configure script, which of course I cannot do.)
3000
3001 LATER: After change 22 below (new libtool files), the configure script
3002 seems to know about icc (and also ecc). Therefore, I have commented out
3003 this hack in configure.in.
3004
3005 7. Added support for pkg-config (2 patches were sent in).
3006
3007 8. Negated POSIX character classes that used a combination of internal tables
3008 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3009 [[:^ascii]]. Typically, they would match almost any characters. The other
3010 POSIX classes were not broken in this way.
3011
3012 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3013 to find the match, as PCRE was deluded into thinking that the match had to
3014 start at the start point or following a newline. The same bug applied to
3015 patterns with negative forward assertions or any backward assertions
3016 preceding ".*" at the start, unless the pattern required a fixed first
3017 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3018
3019 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3020 starting at the last subject character, bytes beyond the end of the subject
3021 string were read.
3022
3023 11. Renamed the variable "class" as "classbits" to make life easier for C++
3024 users. (Previously there was a macro definition, but it apparently wasn't
3025 enough.)
3026
3027 12. Added the new field "tables" to the extra data so that tables can be passed
3028 in at exec time, or the internal tables can be re-selected. This allows
3029 a compiled regex to be saved and re-used at a later time by a different
3030 program that might have everything at different addresses.
3031
3032 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3033 -R library as well as a -L library.
3034
3035 14. The debugging options of pcretest (-d on the command line or D on a
3036 pattern) showed incorrect output for anything following an extended class
3037 that contained multibyte characters and which was followed by a quantifier.
3038
3039 15. Added optional support for general category Unicode character properties
3040 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3041 support. It adds about 90K to the size of the library. The meanings of the
3042 inbuilt class escapes such as \d and \s have NOT been changed.
3043
3044 16. Updated pcredemo.c to include calls to free() to release the memory for the
3045 compiled pattern.
3046
3047 17. The generated file chartables.c was being created in the source directory
3048 instead of in the building directory. This caused the build to fail if the
3049 source directory was different from the building directory, and was
3050 read-only.
3051
3052 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3053 file. No doubt somebody will tell me if they don't make sense... Also added
3054 Dan Mooney's comments about building on OpenVMS.
3055
3056 19. Added support for partial matching via the PCRE_PARTIAL option for
3057 pcre_exec() and the \P data escape in pcretest.
3058
3059 20. Extended pcretest with 3 new pattern features:
3060
3061 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3062 write the compiled pattern to the file whose name is "rest-of-line".
3063 This is a straight binary dump of the data, with the saved pointer to
3064 the character tables forced to be NULL. The study data, if any, is
3065 written too. After writing, pcretest reads a new pattern.
3066
3067 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3068 compiled pattern from the given file. There must not be any
3069 occurrences of "<" in the file name (pretty unlikely); if there are,
3070 pcretest will instead treat the initial "<" as a pattern delimiter.
3071 After reading in the pattern, pcretest goes on to read data lines as
3072 usual.
3073
3074 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3075 and 16-bit fields in a compiled pattern, to simulate a pattern that
3076 was compiled on a host of opposite endianness.
3077
3078 21. The pcre-exec() function can now cope with patterns that were compiled on
3079 hosts of opposite endianness, with this restriction:
3080
3081 As for any compiled expression that is saved and used later, the tables
3082 pointer field cannot be preserved; the extra_data field in the arguments
3083 to pcre_exec() should be used to pass in a tables address if a value
3084 other than the default internal tables were used at compile time.
3085
3086 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3087 now diagnosed as an error. Previously, most of the time, a negative number
3088 would have been treated as zero, but if in addition "ovector" was passed as
3089 NULL, a crash could occur.
3090
3091 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3092 new versions from the libtool 1.5 distribution (the last one is a copy of
3093 a file called libtool.m4). This seems to have fixed the need to patch
3094 "configure" to support Darwin 1.3 (which I used to do). However, I still
3095 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3096 workstation).
3097
3098 24. Changed the PCRE licence to be the more standard "BSD" licence.
3099
3100
3101 Version 4.5 01-Dec-03
3102 ---------------------
3103
3104 1. There has been some re-arrangement of the code for the match() function so
3105 that it can be compiled in a version that does not call itself recursively.
3106 Instead, it keeps those local variables that need separate instances for
3107 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3108 needs to "recurse". Keeping track of where control must go is done by means
3109 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3110 hide most of the details from the main code, and operates only if
3111 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3112 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3113 operating.
3114
3115 To make it easier for callers to provide specially tailored get/free
3116 functions for this usage, two new functions, pcre_stack_malloc, and
3117 pcre_stack_free, are used. They are always called in strict stacking order,
3118 and the size of block requested is always the same.
3119
3120 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3121 PCRE has been compiled to use the stack or the heap for recursion. The
3122 -C option of pcretest uses this to show which version is compiled.
3123
3124 A new data escape \S, is added to pcretest; it causes the amounts of store
3125 obtained and freed by both kinds of malloc/free at match time to be added
3126 to the output.
3127
3128 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3129 what's available on my current Linux desktop machine.
3130
3131 3. When matching a UTF-8 string, the test for a valid string at the start has
3132 been extended. If start_offset is not zero, PCRE now checks that it points
3133 to a byte that is the start of a UTF-8 character. If not, it returns
3134 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3135 this is necessary because there may be backward assertions in the pattern.
3136 When matching the same subject several times, it may save resources to use
3137 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3138
3139 4. The code for checking the validity of UTF-8 strings has been tightened so
3140 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3141 containing "overlong sequences".
3142
3143 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3144 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3145 should have been "&", but it just so happened that all the cases this let
3146 through by mistake were picked up later in the function.
3147
3148 6. I had used a variable called "isblank" - this is a C99 function, causing
3149 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3150
3151 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3152 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3153 so the version is shown; (c) stop "make test" echoing "./RunTest".
3154
3155 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3156
3157 9. The prototype for memmove() for systems that don't have it was using
3158 size_t, but the inclusion of the header that defines size_t was later. I've
3159 moved the #includes for the C headers earlier to avoid this.
3160
3161 10. Added some adjustments to the code to make it easier to compiler on certain
3162 special systems:
3163
3164 (a) Some "const" qualifiers were missing.
3165 (b) Added the macro EXPORT before all exported functions; by default this
3166 is defined to be empty.
3167 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3168 that it reads its output file name as an argument instead of writing
3169 to the standard output and assuming this can be redirected.
3170
3171 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3172 class containing characters with values greater than 255, PCRE compilation
3173 went into a loop.
3174
3175 12. A recursive reference to a subpattern that was within another subpattern
3176 that had a minimum quantifier of zero caused PCRE to crash. For example,
3177 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3178 recursion. If the recursively-called subpattern itself had a zero repeat,
3179 that was OK.
3180
3181 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3182 buffer into which it was copied (for escape processing) was still set at
3183 1024, so long lines caused crashes.
3184
3185 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3186 "internal error: code overflow...". This applied to any character class
3187 that was followed by a possessive quantifier.
3188
3189 15. Modified the Makefile to add libpcre.la as a prerequisite for
3190 libpcreposix.la because I was told this is needed for a parallel build to
3191 work.
3192
3193 16. If a pattern that contained .* following optional items at the start was
3194 studied, the wrong optimizing data was generated, leading to matching
3195 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3196 matching string must start with a or b or c. The correct conclusion for
3197 this pattern is that a match can start with any character.
3198
3199
3200 Version 4.4 13-Aug-03
3201 ---------------------
3202
3203 1. In UTF-8 mode, a character class containing characters with values between
3204 127 and 255 was not handled correctly if the compiled pattern was studied.
3205 In fixing this, I have also improved the studying algorithm for such
3206 classes (slightly).
3207
3208 2. Three internal functions had redundant arguments passed to them. Removal
3209 might give a very teeny performance improvement.
3210
3211 3. Documentation bug: the value of the capture_top field in a callout is *one
3212 more than* the number of the hightest numbered captured substring.
3213
3214 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3215 in incorrectly linking with a previously installed version. They now link
3216 explicitly with libpcre.la.
3217
3218 5. configure.in no longer needs to recognize Cygwin specially.
3219
3220 6. A problem in pcre.in for Windows platforms is fixed.
3221
3222 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3223 pcretest, it used to include the size of the study block as part of its
3224 output. Unfortunately, the structure contains a field that has a different
3225 size on different hardware architectures. This meant that the tests that
3226 showed this size failed. As the block is currently always of a fixed size,
3227 this information isn't actually particularly useful in pcretest output, so
3228 I have just removed it.
3229
3230 8. Three pre-processor statements accidentally did not start in column 1.
3231 Sadly, there are *still* compilers around that complain, even though
3232 standard C has not required this for well over a decade. Sigh.
3233
3234 9. In pcretest, the code for checking callouts passed small integers in the
3235 callout_data field, which is a void * field. However, some picky compilers
3236 complained about the casts involved for this on 64-bit systems. Now
3237 pcretest passes the address of the small integer instead, which should get
3238 rid of the warnings.
3239
3240 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3241 both compile and run time, and gives an error if an invalid UTF-8 sequence
3242 is found. There is a option for disabling this check in cases where the
3243 string is known to be correct and/or the maximum performance is wanted.
3244
3245 11. In response to a bug report, I changed one line in Makefile.in from
3246
3247 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3248 to
3249 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3250
3251 to look similar to other lines, but I have no way of telling whether this
3252 is the right thing to do, as I do not use Windows. No doubt I'll get told
3253 if it's wrong...
3254
3255
3256 Version 4.3 21-May-03
3257 ---------------------
3258
3259 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3260 Makefile.
3261
3262 2. Some refactoring to improve the quality of the code:
3263
3264 (i) The utf8_table... variables are now declared "const".
3265
3266 (ii) The code for \cx, which used the "case flipping" table to upper case
3267 lower case letters, now just substracts 32. This is ASCII-specific,
3268 but the whole concept of \cx is ASCII-specific, so it seems
3269 reasonable.
3270
3271 (iii) PCRE was using its character types table to recognize decimal and
3272 hexadecimal digits in the pattern. This is silly, because it handles
3273 only 0-9, a-f, and A-F, but the character types table is locale-
3274 specific, which means strange things might happen. A private
3275 table is now used for this - though it costs 256 bytes, a table is
3276 much faster than multiple explicit tests. Of course, the standard
3277 character types table is still used for matching digits in subject
3278 strings against \d.
3279
3280 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3281 ending in _t are). So I've renamed it as ESC_tee.
3282
3283 3. The first argument for regexec() in the POSIX wrapper should have been
3284 defined as "const".
3285
3286 4. Changed pcretest to use malloc() for its buffers so that they can be
3287 Electric Fenced for debugging.
3288
3289 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3290 to read one or more bytes before the start of the subject string. Often this
3291 had no effect on PCRE's behaviour, but in some circumstances it could
3292 provoke a segmentation fault.
3293
3294 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3295 to try to read one or more bytes before the start of the subject string.
3296
3297 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3298 UTF-8 support could misbehave in various ways if the subject string
3299 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3300 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3301 back over UTF-8 characters.)
3302
3303
3304 Version 4.2 14-Apr-03
3305 ---------------------
3306
3307 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3308
3309 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3310 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3311 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3312 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3313 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3314 and BUILD_EXEEXT
3315 Note: automatic setting of the BUILD variables is not yet working
3316 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3317 compile-time but not at link-time
3318 [LINK]: use for linking executables only
3319 make different versions for Windows and non-Windows
3320 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3321 libraries
3322 [LINK_FOR_BUILD]: new variable
3323 [OBJEXT]: use throughout
3324 [EXEEXT]: use throughout
3325 <winshared>: new target
3326 <wininstall>: new target
3327 <dftables.o>: use native compiler
3328 <dftables>: use native linker
3329 <install>: handle Windows platform correctly
3330 <clean>: ditto
3331 <check>: ditto
3332 copy DLL to top builddir before testing
3333
3334 As part of these changes, -no-undefined was removed again. This was reported
3335 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3336 in any case.
3337
3338 3. Some tidies to get rid of compiler warnings:
3339
3340 . In the match_data structure, match_limit was an unsigned long int, whereas
3341 match_call_count was an int. I've made them both unsigned long ints.
3342
3343 . In pcretest the fact that a const uschar * doesn't automatically cast to
3344 a void * provoked a warning.
3345
3346 . Turning on some more compiler warnings threw up some "shadow" variables
3347 and a few more missing casts.
3348
3349 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3350 option, a class that contained a single character with a value between 128
3351 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3352
3353 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3354 option, a class that contained several characters, but with at least one
3355 whose value was between 128 and 255 caused PCRE to crash.
3356
3357
3358 Version 4.1 12-Mar-03
3359 ---------------------
3360
3361 1. Compiling with gcc -pedantic found a couple of places where casts were
3362 needed, and a string in dftables.c that was longer than standard compilers are
3363 required to support.
3364
3365 2. Compiling with Sun's compiler found a few more places where the code could
3366 be tidied up in order to avoid warnings.
3367
3368 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3369 first of these names is deprecated in the latest Autoconf in favour of the name
3370 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3371 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3372 analogy I have changed it to CFLAGS_FOR_BUILD.
3373
3374 4. Added -no-undefined to the linking command in the Makefile, because this is
3375 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3376 linking step for the pcreposix library.
3377
3378 5. PCRE was failing to diagnose the case of two named groups with the same
3379 name.
3380
3381 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3382 literal character that is needed in the subject for a match, and scans along to
3383 ensure that it is present before embarking on the full matching process. This
3384 saves time in cases of nested unlimited repeats that are never going to match.
3385 Problem: the scan can take a lot of time if the subject is very long (e.g.
3386 megabytes), thus penalizing straightforward matches. It is now done only if the
3387 amount of subject to be scanned is less than 1000 bytes.
3388
3389 7. A lesser problem with the same optimization is that it was recording the
3390 first character of an anchored pattern as "needed", thus provoking a search
3391 right along the subject, even when the first match of the pattern was going to
3392 fail. The "needed" character is now not set for anchored patterns, unless it
3393 follows something in the pattern that is of non-fixed length. Thus, it still
3394 fulfils its original purpose of finding quick non-matches in cases of nested
3395 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3396
3397
3398 Version 4.0 17-Feb-03
3399 ---------------------
3400
3401 1. If a comment in an extended regex that started immediately after a meta-item
3402 extended to the end of string, PCRE compiled incorrect data. This could lead to
3403 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3404
3405 2. Moved to autoconf 2.53 and libtool 1.4.2.
3406
3407 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3408 the special perltest8 script is no longer needed - all the tests can be run
3409 from a single perltest script.
3410
3411 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3412 by \s. It has now been removed in PCRE. This means it isn't recognized as
3413 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3414 class [:space:] *does* include VT, thereby creating a mess.
3415
3416 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3417 space and tab.
3418
3419 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3420 its new features into the main test script, reducing the number of scripts.
3421
3422 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3423 were backward compatible, and made the (?i) apply to the whole pattern, as if
3424 /i were given. Now it behaves more logically, and applies the option setting
3425 only to what follows. PCRE has been changed to follow suit. However, if it
3426 finds options settings right at the start of the pattern, it extracts them into
3427 the global options, as before. Thus, they show up in the info data.
3428
3429 8. Added support for the \Q...\E escape sequence. Characters in between are
3430 treated as literals. This is slightly different from Perl in that $ and @ are
3431 also handled as literals inside the quotes. In Perl, they will cause variable
3432 interpolation. Note the following examples:
3433
3434 Pattern PCRE matches Perl matches
3435
3436 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3437 \Qabc\$xyz\E abc\$xyz abc\$xyz
3438 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3439
3440 For compatibility with Perl, \Q...\E sequences are recognized inside character
3441 classes as well as outside them.
3442
3443 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3444 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3445 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3446 signed/unsigned warnings.
3447
3448 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3449 option for pcretest, so I've replaced it by a simple function that does just
3450 that job.
3451
3452 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3453 "pcregrep -".
3454
3455 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3456 Java package. This provides some syntactic sugar for simple cases of what my
3457 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3458 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3459 item, you can use this simplified notation. Note that only makes sense with
3460 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3461 greediness, whatever the setting of the PCRE_UNGREEDY option.
3462
3463 13. A change of greediness default within a pattern was not taking effect at
3464 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3465 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3466 was abstracted outside.
3467
3468 14. PCRE now supports the \G assertion. It is true when the current matching
3469 position is at the start point of the match. This differs from \A when the
3470 starting offset is non-zero. Used with the /g option of pcretest (or similar
3471 code), it works in the same way as it does for Perl's /g option. If all
3472 alternatives of a regex begin with \G, the expression is anchored to the start
3473 match position, and the "anchored" flag is set in the compiled expression.
3474
3475 15. Some bugs concerning the handling of certain option changes within patterns
3476 have been fixed. These applied to options other than (?ims). For example,
3477 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3478 been the other way round. Some of this was related to change 7 above.
3479
3480 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3481 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3482 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3483 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3484
3485 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3486 mode. Unlike ".", it always matches newline, whatever the setting of
3487 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3488 assertions. Perl allows it, but it doesn't (in general) work because it can't
3489 calculate the length of the lookbehind. At least, that's the case for Perl
3490 5.8.0 - I've been told they are going to document that it doesn't work in
3491 future.
3492
3493 18. Added an error diagnosis for escapes that PCRE does not support: these are
3494 \L, \l, \N, \P, \p, \U, \u, and \X.
3495
3496 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3497 reading past the end of the pattern in cases such as /[abcd/.
3498
3499 20. PCRE was getting more memory than necessary for patterns with classes that
3500 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3501
3502 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3503 compiling PCRE for use with Virtual Pascal.
3504
3505 22. Small fix to the Makefile to make it work properly if the build is done
3506 outside the source tree.
3507
3508 23. Added a new extension: a condition to go with recursion. If a conditional
3509 subpattern starts with (?(R) the "true" branch is used if recursion has
3510 happened, whereas the "false" branch is used only at the top level.
3511
3512 24. When there was a very long string of literal characters (over 255 bytes
3513 without UTF support, over 250 bytes with UTF support), the computation of how
3514 much memory was required could be incorrect, leading to segfaults or other
3515 strange effects.
3516
3517 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3518 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3519 there was a subsequent back reference to those brackets. This meant that, for
3520 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3521 possible to check for precisely this case. All we can do is abandon the
3522 optimization if .* occurs inside capturing brackets when there are any back
3523 references whatsoever. (See below for a better fix that came later.)
3524
3525 26. The handling of the optimization for finding the first character of a
3526 non-anchored pattern, and for finding a character that is required later in the
3527 match were failing in some cases. This didn't break the matching; it just
3528 failed to optimize when it could. The way this is done has been re-implemented.
3529
3530 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3531
3532 28. Added a new feature that provides some of the functionality that Perl
3533 provides with (?{...}). The facility is termed a "callout". The way it is done
3534 in PCRE is for the caller to provide an optional function, by setting
3535 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3536 global variable. By default it is unset, which disables all calling out. To get
3537 the function called, the regex must include (?C) at appropriate points. This
3538 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3539 This provides a means of identifying different callout points. When PCRE
3540 reaches such a point in the regex, if pcre_callout has been set, the external
3541 function is called. It is provided with data in a structure called
3542 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3543 matching continues; if it returns a non-zero value, the match at the current
3544 point fails. However, backtracking will occur if possible. [This was changed
3545 later and other features added - see item 49 below.]
3546
3547 29. pcretest is upgraded to test the callout functionality. It provides a
3548 callout function that displays information. By default, it shows the start of
3549 the match and the current position in the text. There are some new data escapes
3550 to vary what happens:
3551
3552 \C+ in addition, show current contents of captured substrings
3553 \C- do not supply a callout function
3554 \C!n return 1 when callout number n is reached
3555 \C!n!m return 1 when callout number n is reached for the mth time
3556
3557 30. If pcregrep was called with the -l option and just a single file name, it
3558 output "<stdin>" if a match was found, instead of the file name.
3559
3560 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3561 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3562 pcre_exec(). This saves a malloc/free per call. The default value of
3563 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3564 when configuring.
3565
3566 32. The default maximum size of a compiled pattern is 64K. There have been a
3567 few cases of people hitting this limit. The code now uses macros to handle the
3568 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3569 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3570 configuring. Tests 2 and 5 work only with 2-byte links because they output
3571 debugging information about compiled patterns.
3572
3573 33. Internal code re-arrangements:
3574
3575 (a) Moved the debugging function for printing out a compiled regex into
3576 its own source file (printint.c) and used #include to pull it into
3577 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3578 separate copies.
3579
3580 (b) Defined the list of op-code names for debugging as a macro in
3581 internal.h so that it is next to the definition of the opcodes.
3582
3583 (c) Defined a table of op-code lengths for simpler skipping along compiled
3584 code. This is again a macro in internal.h so that it is next to the
3585 definition of the opcodes.
3586
3587 34. Added support for recursive calls to individual subpatterns, along the
3588 lines of Robin Houston's patch (but implemented somewhat differently).
3589
3590 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3591 allow it to read and process whole directories in Win32. This code was
3592 contributed by Lionel Fourquaux; it has not been tested by me.
3593
3594 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3595 used to name a group. Names consist of alphanumerics and underscores, and must
3596 be unique. Back references use the syntax (?P=name) and recursive calls use
3597 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3598 numbers. The function pcre_fullinfo() can be used after compilation to extract
3599 a name/number map. There are three relevant calls:
3600
3601 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3602 PCRE_INFO_NAMECOUNT yields the number of entries
3603 PCRE_INFO_NAMETABLE yields a pointer to the map.
3604
3605 The map is a vector of fixed-size entries. The size of each entry depends on
3606 the length of the longest name used. The first two bytes of each entry are the
3607 group number, most significant byte first. There follows the corresponding
3608 name, zero terminated. The names are in alphabetical order.
3609
3610 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3611 case instead of 255. Making it the same both with and without UTF-8 support
3612 means that the same test output works with both.
3613
3614 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3615 calling malloc() with a zero argument.
3616
3617 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3618 optimization. I've improved things by keeping a bitmap of backreferences with
3619 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3620 fact referenced, the optimization can be applied. It is unlikely that a
3621 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3622 the match to follow \n) will appear inside brackets with a number greater than
3623 31, but if it does, any back reference > 31 suppresses the optimization.
3624
3625 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3626 of disabling numbered capturing parentheses. Any opening parenthesis that is
3627 not followed by ? behaves as if it were followed by ?: but named parentheses
3628 can still be used for capturing (and they will acquire numbers in the usual
3629 way).
3630
3631 41. Redesigned the return codes from the match() function into yes/no/error so
3632 that errors can be passed back from deep inside the nested calls. A malloc
3633 failure while inside a recursive subpattern call now causes the
3634 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3635
3636 42. It is now possible to set a limit on the number of times the match()
3637 function is called in a call to pcre_exec(). This facility makes it possible to
3638 limit the amount of recursion and backtracking, though not in a directly
3639 obvious way, because the match() function is used in a number of different
3640 circumstances. The count starts from zero for each position in the subject
3641 string (for non-anchored patterns). The default limit is, for compatibility, a
3642 large number, namely 10 000 000. You can change this in two ways:
3643
3644 (a) When configuring PCRE before making, you can use --with-match-limit=n
3645 to set a default value for the compiled library.
3646
3647 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3648 a different value is set. See 45 below.
3649
3650 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3651
3652 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3653 of things that can be changed at compile time. The first argument specifies
3654 what is wanted and the second points to where the information is to be placed.
3655 The current list of available information is:
3656
3657 PCRE_CONFIG_UTF8
3658
3659 The output is an integer that is set to one if UTF-8 support is available;
3660 otherwise it is set to zero.
3661
3662 PCRE_CONFIG_NEWLINE
3663
3664 The output is an integer that it set to the value of the code that is used for
3665 newline. It is either LF (10) or CR (13).
3666
3667 PCRE_CONFIG_LINK_SIZE
3668
3669 The output is an integer that contains the number of bytes used for internal
3670 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3671
3672 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3673
3674 The output is an integer that contains the threshold above which the POSIX
3675 interface uses malloc() for output vectors. See item 31 above.
3676
3677 PCRE_CONFIG_MATCH_LIMIT
3678
3679 The output is an unsigned integer that contains the default limit of the number
3680 of match() calls in a pcre_exec() execution. See 42 above.
3681
3682 44. pcretest has been upgraded by the addition of the -C option. This causes it
3683 to extract all the available output from the new pcre_config() function, and to
3684 output it. The program then exits immediately.
3685
3686 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3687 order to support additional features. One way would have been to define
3688 pcre_exec2() (for example) with extra arguments, but this would not have been
3689 extensible, and would also have required all calls to the original function to
3690 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3691 is used for passing in "extra" data from pcre_study().
3692
3693 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3694 contains the following fields:
3695
3696 flags a bitmap indicating which of the following fields are set
3697 study_data opaque data from pcre_study()
3698 match_limit a way of specifying a limit on match() calls for a specific
3699 call to pcre_exec()
3700 callout_data data for callouts (see 49 below)
3701
3702 The flag bits are also defined in pcre.h, and are
3703
3704 PCRE_EXTRA_STUDY_DATA
3705 PCRE_EXTRA_MATCH_LIMIT
3706 PCRE_EXTRA_CALLOUT_DATA
3707
3708 The pcre_study() function now returns one of these new pcre_extra blocks, with
3709 the actual study data pointed to by the study_data field, and the
3710 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3711 before. That is, this change is entirely upwards-compatible and requires no
3712 change to existing code.
3713
3714 If you want to pass in additional data to pcre_exec(), you can either place it
3715 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3716 block.
3717
3718 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3719 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3720 times with different match limits, until it finds the minimum value needed for
3721 pcre_exec() to complete. The value is then output. This can be instructive; for
3722 most simple matches the number is quite small, but for pathological cases it
3723 gets very large very quickly.
3724
3725 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3726 returns the size of the data block pointed to by the study_data field in a
3727 pcre_extra block, that is, the value that was passed as the argument to
3728 pcre_malloc() when PCRE was getting memory in which to place the information
3729 created by pcre_study(). The fourth argument should point to a size_t variable.
3730 pcretest has been extended so that this information is shown after a successful
3731 pcre_study() call when information about the compiled regex is being displayed.
3732
3733 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3734 because what follows is always an absolute path. (Later: it turns out that this
3735 is more than cosmetic for MinGW, because it doesn't like empty path
3736 components.)
3737
3738 49. Some changes have been made to the callout feature (see 28 above):
3739
3740 (i) A callout function now has three choices for what it returns:
3741
3742 0 => success, carry on matching
3743 > 0 => failure at this point, but backtrack if possible
3744 < 0 => serious error, return this value from pcre_exec()
3745
3746 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3747 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3748 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3749 use by callout functions. It will never be used by PCRE itself.
3750
3751 (ii) The pcre_extra structure (see 45 above) has a void * field called
3752 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3753 pcre_callout_block structure has a field of the same name. The contents of
3754 the field passed in the pcre_extra structure are passed to the callout
3755 function in the corresponding field in the callout block. This makes it
3756 easier to use the same callout-containing regex from multiple threads. For
3757 testing, the pcretest program has a new data escape
3758
3759 \C*n pass the number n (may be negative) as callout_data
3760
3761 If the callout function in pcretest receives a non-zero value as
3762 callout_data, it returns that value.
3763
3764 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3765 there were some redundant $(CFLAGS) in commands that are now specified as
3766 $(LINK), which already includes $(CFLAGS).
3767
3768 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3769 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3770 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3771 one-byte characters throughout. Note that case-insensitive matching applies
3772 only to characters whose values are less than 256. PCRE doesn't support the
3773 notion of cases for higher-valued characters.
3774
3775 (i) A character class whose characters are all within 0-255 is handled as
3776 a bit map, and the map is inverted for negative classes. Previously, a
3777 character > 255 always failed to match such a class; however it should
3778 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3779
3780 (ii) A negated character class with a single character < 255 is coded as
3781 "not this character" (OP_NOT). This wasn't working properly when the test
3782 character was multibyte, either singly or repeated.
3783
3784 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3785 mode, for example: \x{100}{2,3}.
3786
3787 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3788 singly or repeated) now correctly test multibyte characters. However,
3789 PCRE doesn't recognize any characters with values greater than 255 as
3790 digits, spaces, or word characters. Such characters always match \D, \S,
3791 and \W, and never match \d, \s, or \w.
3792
3793 (v) Classes may now contain characters and character ranges with values
3794 greater than 255. For example: [ab\x{100}-\x{400}].
3795
3796 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3797 PCRE in UTF-8 mode.
3798
3799 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3800 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3801 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3802 value.)
3803
3804 53. The single man page has become too large. I have therefore split it up into
3805 a number of separate man pages. These also give rise to individual HTML pages;
3806 these are now put in a separate directory, and there is an index.html page that
3807 lists them all. Some hyperlinking between the pages has been installed.
3808
3809 54. Added convenience functions for handling named capturing parentheses.
3810
3811 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3812 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3813 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3814 are faulted.
3815
3816 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3817 calling configure. These values are used when compiling the dftables.c program
3818 which is run to generate the source of the default character tables. They
3819 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3820 you will need to set these values.
3821
3822 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3823
3824
3825 Version 3.9 02-Jan-02
3826 ---------------------
3827
3828 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3829
3830 2. If --disable-static was given, the building process failed when trying to
3831 build pcretest and pcregrep. (For some reason it was using libtool to compile
3832 them, which is not right, as they aren't part of the library.)
3833
3834
3835 Version 3.8 18-Dec-01
3836 ---------------------
3837
3838 1. The experimental UTF-8 code was completely screwed up. It was packing the
3839 bytes in the wrong order. How dumb can you get?
3840
3841
3842 Version 3.7 29-Oct-01
3843 ---------------------
3844
3845 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3846 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3847 this didn't happen under Solaris 8, where I normally test things.
3848
3849 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3850 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3851 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3852 this fix an hour or so after the initial 3.7 release.)
3853
3854
3855 Version 3.6 23-Oct-01
3856 ---------------------
3857
3858 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3859 offsets passed as NULL with zero offset count.
3860
3861 2. The config.guess and config.sub files had not been updated when I moved to
3862 the latest autoconf.
3863
3864
3865 Version 3.5 15-Aug-01
3866 ---------------------
3867
3868 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3869 had been forgotten.
3870
3871 2. By using declared but undefined structures, we can avoid using "void"
3872 definitions in pcre.h while keeping the internal definitions of the structures
3873 private.
3874
3875 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3876 user point of view, this means that both static and shared libraries are built
3877 by default, but this can be individually controlled. More of the work of
3878 handling this static/shared cases is now inside libtool instead of PCRE's make
3879 file.
3880
3881 4. The pcretest utility is now installed along with pcregrep because it is
3882 useful for users (to test regexs) and by doing this, it automatically gets
3883 relinked by libtool. The documentation has been turned into a man page, so
3884 there are now .1, .txt, and .html versions in /doc.
3885
3886 5. Upgrades to pcregrep:
3887 (i) Added long-form option names like gnu grep.
3888 (ii) Added --help to list all options with an explanatory phrase.
3889 (iii) Added -r, --recursive to recurse into sub-directories.
3890 (iv) Added -f, --file to read patterns from a file.
3891
3892 6. pcre_exec() was referring to its "code" argument before testing that
3893 argument for NULL (and giving an error if it was NULL).
3894
3895 7. Upgraded Makefile.in to allow for compiling in a different directory from
3896 the source directory.
3897
3898 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3899 options bits, the pointer it was passed was to an int instead of to an unsigned
3900 long int. This mattered only on 64-bit systems.
3901
3902 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3903 generated) instead of pcre.in, which it its source. Also made the same change
3904 in several of the .c files.
3905
3906 10. A new release of gcc defines printf() as a macro, which broke pcretest
3907 because it had an ifdef in the middle of a string argument for printf(). Fixed
3908 by using separate calls to printf().
3909
3910 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3911 script, to force use of CR or LF instead of \n in the source. On non-Unix
3912 systems, the value can be set in config.h.
3913
3914 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3915 absolute limit. Changed the text of the error message to make this clear, and
3916 likewise updated the man page.
3917
3918 13. The limit of 99 on the number of capturing subpatterns has been removed.
3919 The new limit is 65535, which I hope will not be a "real" limit.
3920
3921
3922 Version 3.4 22-Aug-00
3923 ---------------------
3924
3925 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3926
3927 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3928
3929
3930 Version 3.3 01-Aug-00
3931 ---------------------
3932
3933 1. If an octal character was given, but the value was greater than \377, it
3934 was not getting masked to the least significant bits, as documented. This could
3935 lead to crashes in some systems.
3936
3937 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3938 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3939
3940 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3941 These just pass their arguments on to (pcre_free)(), but they are provided
3942 because some uses of PCRE bind it to non-C systems that can call its functions,
3943 but cannot call free() or pcre_free() directly.
3944
3945 4. Add "make test" as a synonym for "make check". Corrected some comments in
3946 the Makefile.
3947
3948 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3949 Makefile.
3950
3951 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3952 command called pgrep for grepping around the active processes.
3953
3954 7. Added the beginnings of support for UTF-8 character strings.
3955
3956 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3957 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3958 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3959 out for the ar command.)
3960
3961
3962 Version 3.2 12-May-00
3963 ---------------------
3964
3965 This is purely a bug fixing release.
3966
3967 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3968 of ZA. This was just one example of several cases that could provoke this bug,
3969 which was introduced by change 9 of version 2.00. The code for breaking
3970 infinite loops after an iteration that matches an empty string was't working
3971 correctly.
3972
3973 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3974 when matched against abbab (for example). After matching an empty string, it
3975 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3976 caused it to match further down the string than it should.
3977
3978 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3979 was there because it doesn't seem to be needed, and it causes trouble on some
3980 systems, as it is not a Standard C header. It has been removed.
3981
3982 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3983 were reported on the Macintosh. The changes were from
3984
3985 while ((c = *(++ptr)) != 0 && c != '\n');
3986 to
3987 while ((c = *(++ptr)) != 0 && c != '\n') ;
3988
3989 Totally extraordinary, but if that's what it takes...
3990
3991 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3992 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3993 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3994 assumes the way PCRE uses memmove() (always moving upwards).
3995
3996 6. PCRE is being used in one environment where strchr() is not available. There
3997 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3998 faster code anyway.
3999
4000
4001 Version 3.1 09-Feb-00
4002 ---------------------
4003
4004 The only change in this release is the fixing of some bugs in Makefile.in for
4005 the "install" target:
4006
4007 (1) It was failing to install pcreposix.h.
4008
4009 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4010
4011
4012 Version 3.0 01-Feb-00
4013 ---------------------
4014
4015 1. Add support for the /+ modifier to perltest (to output $` like it does in
4016 pcretest).
4017
4018 2. Add support for the /g modifier to perltest.
4019
4020 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4021 matches null strings.
4022
4023 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4024 pattern. Perl treats empty patterns specially - it reuses the most recent
4025 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4026 effect.
4027
4028 5. The POSIX interface was broken in that it was just handing over the POSIX
4029 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4030 required a bigger vector, with some working space on the end. This means that
4031 the POSIX wrapper now has to get and free some memory, and copy the results.
4032
4033 6. Added some simple autoconf support, placing the test data and the
4034 documentation in separate directories, re-organizing some of the
4035 information files, and making it build pcre-config (a GNU standard). Also added
4036 libtool support for building PCRE as a shared library, which is now the
4037 default.
4038
4039 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4040 09 are not valid octal constants. Single digits will be used for minor values
4041 less than 10.
4042
4043 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4044 existing programs that set these in the POSIX interface can use PCRE without
4045 modification.
4046
4047 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4048 return all that pcre_info() returns, plus additional data. The pcre_info()
4049 function is retained for compatibility, but is considered to be obsolete.
4050
4051 10. Added experimental recursion feature (?R) to handle one common case that
4052 Perl 5.6 will be able to do with (?p{...}).
4053
4054 11. Added support for POSIX character classes like [:alpha:], which Perl is
4055 adopting.
4056
4057
4058 Version 2.08 31-Aug-99
4059 ----------------------
4060
4061 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4062 trying to match at the startoffset position, but instead was moving forward to
4063 the next newline as if a previous match had failed.
4064
4065 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4066 and could get into a loop if a null string was matched other than at the start
4067 of the subject.
4068
4069 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4070 be distinguished at compile time, and for completeness also added PCRE_DATE.
4071
4072 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4073 in GnuWin32 environments.
4074
4075
4076 Version 2.07 29-Jul-99
4077 ----------------------
4078
4079 1. The documentation is now supplied in plain text form and HTML as well as in
4080 the form of man page sources.
4081
4082 2. C++ compilers don't like assigning (void *) values to other pointer types.
4083 In particular this affects malloc(). Although there is no problem in Standard
4084 C, I've put in casts to keep C++ compilers happy.
4085
4086 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4087 should be (const char *).
4088
4089 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4090 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4091 However, I haven't made this a standard facility. The documentation doesn't
4092 mention it, and the Makefile doesn't support it.
4093
4094 5. The Makefile now contains an "install" target, with editable destinations at
4095 the top of the file. The pcretest program is not installed.
4096
4097 6. pgrep -V now gives the PCRE version number and date.
4098
4099 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4100 causing the entire string to be ignored, instead of just the last character.
4101
4102 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4103 non-matching string, it can take a very, very long time, even for strings of
4104 quite modest length, because of the nested recursion. PCRE now does better in
4105 some of these cases. It does this by remembering the last required literal
4106 character in the pattern, and pre-searching the subject to ensure it is present
4107 before running the real match. In other words, it applies a heuristic to detect
4108 some types of certain failure quickly, and in the above example, if presented
4109 with a string that has no trailing " it gives "no match" very quickly.
4110
4111 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4112 other alternatives are tried instead.
4113
4114
4115 Version 2.06 09-Jun-99
4116 ----------------------
4117
4118 1. Change pcretest's output for amount of store used to show just the code
4119 space, because the remainder (the data block) varies in size between 32-bit and
4120 64-bit systems.
4121
4122 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4123 start matching at. This allows lookbehinds to work when searching for multiple
4124 occurrences in a string.
4125
4126 3. Added additional options to pcretest for testing multiple occurrences:
4127
4128 /+ outputs the rest of the string that follows a match
4129 /g loops for multiple occurrences, using the new startoffset argument
4130 /G loops for multiple occurrences by passing an incremented pointer
4131
4132 4. PCRE wasn't doing the "first character" optimization for patterns starting
4133 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4134 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4135 the letter 'x'. On long subject strings, this gives a significant speed-up.
4136
4137
4138 Version 2.05 21-Apr-99
4139 ----------------------
4140
4141 1. Changed the type of magic_number from int to long int so that it works
4142 properly on 16-bit systems.
4143
4144 2. Fixed a bug which caused patterns starting with .* not to work correctly
4145 when the subject string contained newline characters. PCRE was assuming
4146 anchoring for such patterns in all cases, which is not correct because .* will
4147 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4148 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4149 must be retried after every newline in the subject.
4150
4151
4152 Version 2.04 18-Feb-99
4153 ----------------------
4154
4155 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4156 computation of the store needed to hold the pattern was incorrect (too large).
4157 If such patterns were nested a few deep, this could multiply and become a real
4158 problem.
4159
4160 2. Added /M option to pcretest to show the memory requirement of a specific
4161 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4162
4163 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4164 compiled in such a way that the backtracking after subsequent failure was
4165 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4166 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4167
4168
4169 Version 2.03 02-Feb-99
4170 ----------------------
4171
4172 1. Fixed typo and small mistake in man page.
4173
4174 2. Added 4th condition (GPL supersedes if conflict) and created separate
4175 LICENCE file containing the conditions.
4176
4177 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4178 Perl, that is the internal \ allows the delimiter to be included in the
4179 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4180 the final delimiter, add \ to the end of the pattern (to test the error).
4181
4182 4. Added the convenience functions for extracting substrings after a successful
4183 match. Updated pcretest to make it able to test these functions.
4184
4185
4186 Version 2.02 14-Jan-99
4187 ----------------------
4188
4189 1. Initialized the working variables associated with each extraction so that
4190 their saving and restoring doesn't refer to uninitialized store.
4191
4192 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4193 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4194 fix the problem.
4195
4196 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4197 calls, and wasn't printing the correct value for compiling calls. Increased the
4198 default value of LOOPREPEAT, and the number of significant figures in the
4199 times.
4200
4201 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4202
4203 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4204 a building problem on Windows NT with a FAT file system.
4205
4206
4207 Version 2.01 21-Oct-98
4208 ----------------------
4209
4210 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4211 to character tables built by pcre_maketables() in the current locale. If NULL
4212 is passed, the default tables are used.
4213
4214
4215 Version 2.00 24-Sep-98
4216 ----------------------
4217
4218 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4219 it any more.
4220
4221 2. Allow quantification of (?>) groups, and make it work correctly.
4222
4223 3. The first character computation wasn't working for (?>) groups.
4224
4225 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4226 end of the subject) and add 5.005's \z, which really does match only at the
4227 very end of the subject.
4228
4229 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4230
4231 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4232 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4233 localized options. All options to pcre_study() were also removed.
4234
4235 7. Add other new features from 5.005:
4236
4237 $(?<= positive lookbehind
4238 $(?<! negative lookbehind
4239 (?imsx-imsx) added the unsetting capability
4240 such a setting is global if at outer level; local otherwise
4241 (?imsx-imsx:) non-capturing groups with option setting
4242 (?(cond)re|re) conditional pattern matching
4243
4244 A backreference to itself in a repeated group matches the previous
4245 captured string.
4246
4247 8. General tidying up of studying (both automatic and via "study")
4248 consequential on the addition of new assertions.
4249
4250 9. As in 5.005, unlimited repeated groups that could match an empty substring
4251 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4252 runtime if any iteration does actually match an empty substring.
4253
4254 10. Include the RunTest script in the distribution.
4255
4256 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4257 discrepancies, some of which were old and were also with respect to 5.004. They
4258 have now been fixed.
4259
4260
4261 Version 1.09 28-Apr-98
4262 ----------------------
4263
4264 1. A negated single character class followed by a quantifier with a minimum
4265 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4266 program crashes, or just wrong answers. This did not apply to negated classes
4267 containing more than one character, or to minima other than one.
4268
4269
4270 Version 1.08 27-Mar-98
4271 ----------------------
4272
4273 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4274
4275 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4276 latter must appear before anything that relies on it in the pattern.
4277
4278
4279 Version 1.07 16-Feb-98
4280 ----------------------
4281
4282 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4283 repeat of a potentially empty string).
4284
4285
4286 Version 1.06 23-Jan-98
4287 ----------------------
4288
4289 1. Added Markus Oberhumer's little patches for C++.
4290
4291 2. Literal strings longer than 255 characters were broken.
4292
4293
4294 Version 1.05 23-Dec-97
4295 ----------------------
4296
4297 1. Negated character classes containing more than one character were failing if
4298 PCRE_CASELESS was set at run time.
4299
4300
4301 Version 1.04 19-Dec-97
4302 ----------------------
4303
4304 1. Corrected the man page, where some "const" qualifiers had been omitted.
4305
4306 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4307 input syntax.
4308
4309 3. Fixed memory leak which occurred when a regex with back references was
4310 matched with an offsets vector that wasn't big enough. The temporary memory
4311 that is used in this case wasn't being freed if the match failed.
4312
4313 4. Tidied pcretest to ensure it frees memory that it gets.
4314
4315 5. Temporary memory was being obtained in the case where the passed offsets
4316 vector was exactly big enough.
4317
4318 6. Corrected definition of offsetof() from change 5 below.
4319
4320 7. I had screwed up change 6 below and broken the rules for the use of
4321 setjmp(). Now fixed.
4322
4323
4324 Version 1.03 18-Dec-97
4325 ----------------------
4326
4327 1. A erroneous regex with a missing opening parenthesis was correctly
4328 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4329 on some systems.
4330
4331 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4332 it was reported that one broken compiler failed on the former because "code" is
4333 also an independent variable.
4334
4335 3. The erroneous regex a[]b caused an array overrun reference.
4336
4337 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4338 fail on data ending with that character. (It was going on too far, and checking
4339 the next character, typically a binary zero.) This was specific to the
4340 optimized code for single-character negative classes.
4341
4342 5. Added a contributed patch from the TIN world which does the following:
4343
4344 + Add an undef for memmove, in case the the system defines a macro for it.
4345
4346 + Add a definition of offsetof(), in case there isn't one. (I don't know
4347 the reason behind this - offsetof() is part of the ANSI standard - but
4348 it does no harm).
4349
4350 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4351 most of the places where whitespace preceded '#'. I have given up and
4352 allowed the remaining 2 cases to be at the margin.
4353
4354 + Rename some variables in pcre to eliminate shadowing. This seems very
4355 pedantic, but does no harm, of course.
4356
4357 6. Moved the call to setjmp() into its own function, to get rid of warnings
4358 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4359
4360 7. Constructs such as \d{8,} were compiling into the equivalent of
4361 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4362 outcome, but in this particular case used more store than had been allocated,
4363 which caused the bug to be discovered because it threw up an internal error.
4364
4365 8. The debugging code in both pcre and pcretest for outputting the compiled
4366 form of a regex was going wrong in the case of back references followed by
4367 curly-bracketed repeats.
4368
4369
4370 Version 1.02 12-Dec-97
4371 ----------------------
4372
4373 1. Typos in pcre.3 and comments in the source fixed.
4374
4375 2. Applied a contributed patch to get rid of places where it used to remove
4376 'const' from variables, and fixed some signed/unsigned and uninitialized
4377 variable warnings.
4378
4379 3. Added the "runtest" target to Makefile.
4380
4381 4. Set default compiler flag to -O2 rather than just -O.
4382
4383
4384 Version 1.01 19-Nov-97
4385 ----------------------
4386
4387 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4388 like /([ab]*)*/, that is, for classes with more than one character in them.
4389
4390 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4391 as /((?>a*))*/ (a PCRE_EXTRA facility).
4392
4393
4394 Version 1.00 18-Nov-97
4395 ----------------------
4396
4397 1. Added compile-time macros to support systems such as SunOS4 which don't have
4398 memmove() or strerror() but have other things that can be used instead.
4399
4400 2. Arranged that "make clean" removes the executables.
4401
4402
4403 Version 0.99 27-Oct-97
4404 ----------------------
4405
4406 1. Fixed bug in code for optimizing classes with only one character. It was
4407 initializing a 32-byte map regardless, which could cause it to run off the end
4408 of the memory it had got.
4409
4410 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4411
4412
4413 Version 0.98 22-Oct-97
4414 ----------------------
4415
4416 1. Fixed bug in code for handling temporary memory usage when there are more
4417 back references than supplied space in the ovector. This could cause segfaults.
4418
4419
4420 Version 0.97 21-Oct-97
4421 ----------------------
4422
4423 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4424
4425 2. Optimized negated single characters not to use a bit map.
4426
4427 3. Brought error texts together as macro definitions; clarified some of them;
4428 fixed one that was wrong - it said "range out of order" when it meant "invalid
4429 escape sequence".
4430
4431 4. Changed some char * arguments to const char *.
4432
4433 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4434
4435 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4436 pcretest.
4437
4438
4439 Version 0.96 16-Oct-97
4440 ----------------------
4441
4442 1. Added a simple "pgrep" utility to the distribution.
4443
4444 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4445 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4446 where "ddd" means "one or more decimal digits".
4447
4448 3. Fixed serious bug. If a pattern had a back reference, but the call to
4449 pcre_exec() didn't supply a large enough ovector to record the related
4450 identifying subpattern, the match always failed. PCRE now remembers the number
4451 of the largest back reference, and gets some temporary memory in which to save
4452 the offsets during matching if necessary, in order to ensure that
4453 backreferences always work.
4454
4455 4. Increased the compatibility with Perl in a number of ways:
4456
4457 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4458 to request this handling. The option can be set at compile or exec time.
4459
4460 (b) $ matches before a terminating newline by default; an option
4461 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4462 mode). The option can be set at compile or exec time.
4463
4464 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4465 the same as Perl's. If the decimal number it represents is less than 10
4466 or there aren't that many previous left capturing parentheses, an octal
4467 escape is read. Inside a character class, it's always an octal escape,
4468 even if it is a single digit.
4469
4470 (d) An escaped but undefined alphabetic character is taken as a literal,
4471 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4472 escapes.
4473
4474 (e) {0} is now permitted. (The previous item is removed from the compiled
4475 pattern).
4476
4477 5. Changed all the names of code files so that the basic parts are no longer
4478 than 10 characters, and abolished the teeny "globals.c" file.
4479
4480 6. Changed the handling of character classes; they are now done with a 32-byte
4481 bit map always.
4482
4483 7. Added the -d and /D options to pcretest to make it possible to look at the
4484 internals of compilation without having to recompile pcre.
4485
4486
4487 Version 0.95 23-Sep-97
4488 ----------------------
4489
4490 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4491 \x20 at the start of a run of normal characters. These were being treated as
4492 real characters, instead of the source characters being re-checked.
4493
4494
4495 Version 0.94 18-Sep-97
4496 ----------------------
4497
4498 1. The functions are now thread-safe, with the caveat that the global variables
4499 containing pointers to malloc() and free() or alternative functions are the
4500 same for all threads.
4501
4502 2. Get pcre_study() to generate a bitmap of initial characters for non-
4503 anchored patterns when this is possible, and use it if passed to pcre_exec().
4504
4505
4506 Version 0.93 15-Sep-97
4507 ----------------------
4508
4509 1. /(b)|(:+)/ was computing an incorrect first character.
4510
4511 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4512 but not actually doing anything yet.
4513
4514 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4515 as Perl does (e.g. [-az] or [az-]).
4516
4517 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4518 all possible positions.
4519
4520 5. Split up into different modules to avoid including unneeded functions in a
4521 compiled binary. However, compile and exec are still in one module. The "study"
4522 function is split off.
4523
4524 6. The character tables are now in a separate module whose source is generated
4525 by an auxiliary program - but can then be edited by hand if required. There are
4526 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4527 toupper() in the code.
4528
4529 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4530 make them global. Abolish the function for setting them, as the caller can now
4531 set them directly.
4532
4533
4534 Version 0.92 11-Sep-97
4535 ----------------------
4536
4537 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4538 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4539
4540 2. Caseless matching was not working in character classes if the characters in
4541 the pattern were in upper case.
4542
4543 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4544
4545 4. Make PCRE_ANCHORED public and accept as a compile option.
4546
4547 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4548 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4549 pass them.
4550
4551 6. Give an error if bad option bits passed at compile or run time.
4552
4553 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4554 pcretest to cause it to pass that flag.
4555
4556 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4557 options, and the first character, if set.
4558
4559 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4560
4561
4562 Version 0.91 10-Sep-97
4563 ----------------------
4564
4565 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4566 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4567
4568 2. PCRE was looping on encountering an indefinitely repeated back reference to
4569 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4570 Perl does - treats the match as successful.
4571
4572 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5