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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 760 - (show annotations) (download)
Tue Nov 22 11:23:43 2011 UTC (2 years, 8 months ago) by ph10
File size: 220383 byte(s)
Test for workspace overflow with forward reference data.

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

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12