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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 753 - (show annotations)
Sat Nov 19 17:03:35 2011 UTC (3 years, 6 months ago) by ph10
File size: 219969 byte(s)
Error occurred while calculating annotation data.
Remove -L from pcre-config when install is into /usr/lib (fixes 
cross-compiling).
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
54 Version 8.20 21-Oct-2011
55 ------------------------
56
57 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
58 a POSIX class. After further experiments with Perl, which convinced me that
59 Perl has bugs and confusions, a closing square bracket is no longer allowed
60 in a POSIX name. This bug also affected patterns with classes that started
61 with full stops.
62
63 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
64 captured substring, but while checking the failing first alternative,
65 substring 1 is temporarily captured. If the output vector supplied to
66 pcre_exec() was not big enough for this capture, the yield of the function
67 was still zero ("insufficient space for captured substrings"). This cannot
68 be totally fixed without adding another stack variable, which seems a lot
69 of expense for a edge case. However, I have improved the situation in cases
70 such as /(a)(b)x|abc/ matched against "abc", where the return code
71 indicates that fewer than the maximum number of slots in the ovector have
72 been set.
73
74 3. Related to (2) above: when there are more back references in a pattern than
75 slots in the output vector, pcre_exec() uses temporary memory during
76 matching, and copies in the captures as far as possible afterwards. It was
77 using the entire output vector, but this conflicts with the specification
78 that only 2/3 is used for passing back captured substrings. Now it uses
79 only the first 2/3, for compatibility. This is, of course, another edge
80 case.
81
82 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
83 main code base, and can be used by building with --enable-jit. When this is
84 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
85 runtime --no-jit option is given.
86
87 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
88 ovector, the return from the function was zero, implying that there were
89 other matches that did not fit. The correct "exactly full" value is now
90 returned.
91
92 6. If a subpattern that was called recursively or as a subroutine contained
93 (*PRUNE) or any other control that caused it to give a non-standard return,
94 invalid errors such as "Error -26 (nested recursion at the same subject
95 position)" or even infinite loops could occur.
96
97 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
98 computing the minimum length on reaching *ACCEPT, and so ended up with the
99 wrong value of 1 rather than 0. Further investigation indicates that
100 computing a minimum subject length in the presence of *ACCEPT is difficult
101 (think back references, subroutine calls), and so I have changed the code
102 so that no minimum is registered for a pattern that contains *ACCEPT.
103
104 8. If (*THEN) was present in the first (true) branch of a conditional group,
105 it was not handled as intended. [But see 16 below.]
106
107 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
108 Sheri Pierce.
109
110 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
111 the first byte in a match must be "a".
112
113 11. Change 17 for 8.13 increased the recursion depth for patterns like
114 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
115 pattern contains any instances of (*THEN). If it does not, the old
116 optimizations are restored. It would be nice to do this on a per-group
117 basis, but at the moment that is not feasible.
118
119 12. In some environments, the output of pcretest -C is CRLF terminated. This
120 broke RunTest's code that checks for the link size. A single white space
121 character after the value is now allowed for.
122
123 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
124 For "fr", it uses the Windows-specific input and output files.
125
126 14. If (*THEN) appeared in a group that was called recursively or as a
127 subroutine, it did not work as intended. [But see next item.]
128
129 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
130 pattern fragments (but not containing any | characters). If A and B are
131 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
132 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
133 D. In other words, Perl considers parentheses that do not contain any |
134 characters to be part of a surrounding alternative, whereas PCRE was
135 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
136 differently. PCRE now behaves in the same way as Perl, except in the case
137 of subroutine/recursion calls such as (?1) which have in any case always
138 been different (but PCRE had them first :-).
139
140 16. Related to 15 above: Perl does not treat the | in a conditional group as
141 creating alternatives. Such a group is treated in the same way as an
142 ordinary group without any | characters when processing (*THEN). PCRE has
143 been changed to match Perl's behaviour.
144
145 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
146 RunGrepTest script failed.
147
148 18. Change 22 for version 13 caused atomic groups to use more stack. This is
149 inevitable for groups that contain captures, but it can lead to a lot of
150 stack use in large patterns. The old behaviour has been restored for atomic
151 groups that do not contain any capturing parentheses.
152
153 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
154 suppress the check for a minimum subject length at run time. (If it was
155 given to pcre_exec() or pcre_dfa_exec() it did work.)
156
157 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
158 fail to work when decoding hex characters in data strings in EBCDIC
159 environments.
160
161 21. It appears that in at least one Mac OS environment, the isxdigit() function
162 is implemented as a macro that evaluates to its argument more than once,
163 contravening the C 90 Standard (I haven't checked a later standard). There
164 was an instance in pcretest which caused it to go wrong when processing
165 \x{...} escapes in subject strings. The has been rewritten to avoid using
166 things like p++ in the argument of isxdigit().
167
168
169 Version 8.13 16-Aug-2011
170 ------------------------
171
172 1. The Unicode data tables have been updated to Unicode 6.0.0.
173
174 2. Two minor typos in pcre_internal.h have been fixed.
175
176 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
177 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
178 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
179
180 4. There were a number of related bugs in the code for matching backrefences
181 caselessly in UTF-8 mode when codes for the characters concerned were
182 different numbers of bytes. For example, U+023A and U+2C65 are an upper
183 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
184 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
185 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
186 2-byte code at the end of the subject (it thought there wasn't enough data
187 left).
188
189 5. Comprehensive information about what went wrong is now returned by
190 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
191 as the output vector has at least 2 elements. The offset of the start of
192 the failing character and a reason code are placed in the vector.
193
194 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
195 now returned is for the first byte of the failing character, instead of the
196 last byte inspected. This is an incompatible change, but I hope it is small
197 enough not to be a problem. It makes the returned offset consistent with
198 pcre_exec() and pcre_dfa_exec().
199
200 7. pcretest now gives a text phrase as well as the error number when
201 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
202 failure, the offset and reason code are output.
203
204 8. When \R was used with a maximizing quantifier it failed to skip backwards
205 over a \r\n pair if the subsequent match failed. Instead, it just skipped
206 back over a single character (\n). This seems wrong (because it treated the
207 two characters as a single entity when going forwards), conflicts with the
208 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
209 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
210 has been changed.
211
212 9. Some internal refactoring has changed the processing so that the handling
213 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
214 time (the PCRE_DOTALL option was changed this way some time ago: version
215 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
216 which was always a bit of a fudge. It also means that there is one less
217 argument for the match() function, which reduces its stack requirements
218 slightly. This change also fixes an incompatibility with Perl: the pattern
219 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
220
221 10. More internal refactoring has drastically reduced the number of recursive
222 calls to match() for possessively repeated groups such as (abc)++ when
223 using pcre_exec().
224
225 11. While implementing 10, a number of bugs in the handling of groups were
226 discovered and fixed:
227
228 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
229 (a|)*(?1) gave a compile-time internal error.
230 ((a|)+)+ did not notice that the outer group could match an empty string.
231 (^a|^)+ was not marked as anchored.
232 (.*a|.*)+ was not marked as matching at start or after a newline.
233
234 12. Yet more internal refactoring has removed another argument from the match()
235 function. Special calls to this function are now indicated by setting a
236 value in a variable in the "match data" data block.
237
238 13. Be more explicit in pcre_study() instead of relying on "default" for
239 opcodes that mean there is no starting character; this means that when new
240 ones are added and accidentally left out of pcre_study(), testing should
241 pick them up.
242
243 14. The -s option of pcretest has been documented for ages as being an old
244 synonym of -m (show memory usage). I have changed it to mean "force study
245 for every regex", that is, assume /S for every regex. This is similar to -i
246 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
247 using it. It makes it easier to run collections of tests with and without
248 study enabled, and thereby test pcre_study() more easily. All the standard
249 tests are now run with and without -s (but some patterns can be marked as
250 "never study" - see 20 below).
251
252 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
253 restoration of the capturing data to the outer values was not happening
254 correctly.
255
256 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
257 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
258 pattern had matched an empty string, and so incorrectly returned a no
259 match.
260
261 17. There was optimizing code for the last branch of non-capturing parentheses,
262 and also for the obeyed branch of a conditional subexpression, which used
263 tail recursion to cut down on stack usage. Unfortunately, now that there is
264 the possibility of (*THEN) occurring in these branches, tail recursion is
265 no longer possible because the return has to be checked for (*THEN). These
266 two optimizations have therefore been removed. [But see 8.20/11 above.]
267
268 18. If a pattern containing \R was studied, it was assumed that \R always
269 matched two bytes, thus causing the minimum subject length to be
270 incorrectly computed because \R can also match just one byte.
271
272 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
273 was incorrectly computed.
274
275 20. If /S is present twice on a test pattern in pcretest input, it now
276 *disables* studying, thereby overriding the use of -s on the command line
277 (see 14 above). This is necessary for one or two tests to keep the output
278 identical in both cases.
279
280 21. When (*ACCEPT) was used in an assertion that matched an empty string and
281 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
282
283 22. When an atomic group that contained a capturing parenthesis was
284 successfully matched, but the branch in which it appeared failed, the
285 capturing was not being forgotten if a higher numbered group was later
286 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
287 group 1 to "a", when in fact it should be unset. This applied to multi-
288 branched capturing and non-capturing groups, repeated or not, and also to
289 positive assertions (capturing in negative assertions does not happen
290 in PCRE) and also to nested atomic groups.
291
292 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
293 subject after a captured substring, to make it easier to tell which of a
294 number of identical substrings has been captured.
295
296 24. The way atomic groups are processed by pcre_exec() has been changed so that
297 if they are repeated, backtracking one repetition now resets captured
298 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
299 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
300 "aaa". Previously, it would have been "a". As part of this code
301 refactoring, the way recursive calls are handled has also been changed.
302
303 25. If an assertion condition captured any substrings, they were not passed
304 back unless some other capturing happened later. For example, if
305 (?(?=(a))a) was matched against "a", no capturing was returned.
306
307 26. When studying a pattern that contained subroutine calls or assertions,
308 the code for finding the minimum length of a possible match was handling
309 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
310 group 1 called group 2 while simultaneously a separate group 2 called group
311 1). A stack overflow occurred in this case. I have fixed this by limiting
312 the recursion depth to 10.
313
314 27. Updated RunTest.bat in the distribution to the version supplied by Tom
315 Fortmann. This supports explicit test numbers on the command line, and has
316 argument validation and error reporting.
317
318 28. An instance of \X with an unlimited repeat could fail if at any point the
319 first character it looked at was a mark character.
320
321 29. Some minor code refactoring concerning Unicode properties and scripts
322 should reduce the stack requirement of match() slightly.
323
324 30. Added the '=' option to pcretest to check the setting of unused capturing
325 slots at the end of the pattern, which are documented as being -1, but are
326 not included in the return count.
327
328 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
329 compiled something random. Now it gives a compile-time error (as does
330 Perl).
331
332 32. A *MARK encountered during the processing of a positive assertion is now
333 recorded and passed back (compatible with Perl).
334
335 33. If --only-matching or --colour was set on a pcregrep call whose pattern
336 had alternative anchored branches, the search for a second match in a line
337 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
338 matched the line "0102" twice. The same bug affected patterns that started
339 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
340 twice.
341
342 34. Previously, PCRE did not allow quantification of assertions. However, Perl
343 does, and because of capturing effects, quantifying parenthesized
344 assertions may at times be useful. Quantifiers are now allowed for
345 parenthesized assertions.
346
347 35. A minor code tidy in pcre_compile() when checking options for \R usage.
348
349 36. \g was being checked for fancy things in a character class, when it should
350 just be a literal "g".
351
352 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
353 appearance of a nested POSIX class supersedes an apparent external class.
354 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
355 unescaped square brackets may also appear as part of class names. For
356 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
357 more like Perl. (But see 8.20/1 above.)
358
359 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
360 was because it thought it was \N{name}, which is not supported).
361
362 39. Add minix to OS list not supporting the -S option in pcretest.
363
364 40. PCRE tries to detect cases of infinite recursion at compile time, but it
365 cannot analyze patterns in sufficient detail to catch mutual recursions
366 such as ((?1))((?2)). There is now a runtime test that gives an error if a
367 subgroup is called recursively as a subpattern for a second time at the
368 same position in the subject string. In previous releases this might have
369 been caught by the recursion limit, or it might have run out of stack.
370
371 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
372 happen only once. PCRE was, however incorrectly giving a compile time error
373 "recursive call could loop indefinitely" because it cannot analyze the
374 pattern in sufficient detail. The compile time test no longer happens when
375 PCRE is compiling a conditional subpattern, but actual runaway loops are
376 now caught at runtime (see 40 above).
377
378 42. It seems that Perl allows any characters other than a closing parenthesis
379 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
380 has been changed to be the same.
381
382 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
383 as not to get warnings when autogen.sh is called. Also changed
384 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
385
386 44. To help people who use pcregrep to scan files containing exceedingly long
387 lines, the following changes have been made:
388
389 (a) The default value of the buffer size parameter has been increased from
390 8K to 20K. (The actual buffer used is three times this size.)
391
392 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
393 PCRE is built.
394
395 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
396 to be set at run time.
397
398 (d) Numerical values in pcregrep options can be followed by K or M, for
399 example --buffer-size=50K.
400
401 (e) If a line being scanned overflows pcregrep's buffer, an error is now
402 given and the return code is set to 2.
403
404 45. Add a pointer to the latest mark to the callout data block.
405
406 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
407 partial match of an empty string instead of no match. This was specific to
408 the use of ".".
409
410 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
411 complete match instead of a partial match. This bug was dependent on both
412 the PCRE_UTF8 and PCRE_DOTALL options being set.
413
414 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
415 starting byte set, because \b was not being ignored.
416
417
418 Version 8.12 15-Jan-2011
419 ------------------------
420
421 1. Fixed some typos in the markup of the man pages, and wrote a script that
422 checks for such things as part of the documentation building process.
423
424 2. On a big-endian 64-bit system, pcregrep did not correctly process the
425 --match-limit and --recursion-limit options (added for 8.11). In
426 particular, this made one of the standard tests fail. (The integer value
427 went into the wrong half of a long int.)
428
429 3. If the --colour option was given to pcregrep with -v (invert match), it
430 did strange things, either producing crazy output, or crashing. It should,
431 of course, ignore a request for colour when reporting lines that do not
432 match.
433
434 4. Another pcregrep bug caused similar problems if --colour was specified with
435 -M (multiline) and the pattern match finished with a line ending.
436
437 5. In pcregrep, when a pattern that ended with a literal newline sequence was
438 matched in multiline mode, the following line was shown as part of the
439 match. This seems wrong, so I have changed it.
440
441 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
442 the check for further matches in the same line (so they could be coloured)
443 to overrun the end of the current line. If another match was found, it was
444 incorrectly shown (and then shown again when found in the next line).
445
446 7. If pcregrep was compiled under Windows, there was a reference to the
447 function pcregrep_exit() before it was defined. I am assuming this was
448 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
449 reported by a user. I've moved the definition above the reference.
450
451
452 Version 8.11 10-Dec-2010
453 ------------------------
454
455 1. (*THEN) was not working properly if there were untried alternatives prior
456 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
457 backtracked to try for "b" instead of moving to the next alternative branch
458 at the same level (in this case, to look for "c"). The Perl documentation
459 is clear that when (*THEN) is backtracked onto, it goes to the "next
460 alternative in the innermost enclosing group".
461
462 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
463 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
464 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
465 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
466 (*THEN).
467
468 3. If \s appeared in a character class, it removed the VT character from
469 the class, even if it had been included by some previous item, for example
470 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
471 of \s, but is part of the POSIX "space" class.)
472
473 4. A partial match never returns an empty string (because you can always
474 match an empty string at the end of the subject); however the checking for
475 an empty string was starting at the "start of match" point. This has been
476 changed to the "earliest inspected character" point, because the returned
477 data for a partial match starts at this character. This means that, for
478 example, /(?<=abc)def/ gives a partial match for the subject "abc"
479 (previously it gave "no match").
480
481 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
482 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
483 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
484 has an implication that the given string is incomplete (because a partial
485 match is preferred over a full match). For this reason, these items now
486 give a partial match in this situation. [Aside: previously, the one case
487 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
488 match rather than a full match, which was wrong by the old rules, but is
489 now correct.]
490
491 6. There was a bug in the handling of #-introduced comments, recognized when
492 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
493 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
494 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
495 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
496 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
497 places in pcre_compile().
498
499 7. Related to (6) above, when pcre_compile() was skipping #-introduced
500 comments when looking ahead for named forward references to subpatterns,
501 the only newline sequence it recognized was NL. It now handles newlines
502 according to the set newline convention.
503
504 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
505 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
506 cater for a lack of strerror(). These oversights have been fixed.
507
508 9. Added --match-limit and --recursion-limit to pcregrep.
509
510 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
511
512 11. When the -o option was used, pcregrep was setting a return code of 1, even
513 when matches were found, and --line-buffered was not being honoured.
514
515 12. Added an optional parentheses number to the -o and --only-matching options
516 of pcregrep.
517
518 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
519 can match an empty string. The code to do it in pcretest and pcredemo
520 needed fixing:
521
522 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
523 only one byte after an empty string match just before CRLF (this case
524 just got forgotten; "any" and "anycrlf" were OK).
525
526 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
527 mode when an empty string match preceded an ASCII character followed by
528 a non-ASCII character. (The code for advancing by one character rather
529 than one byte was nonsense.)
530
531 (c) The pcredemo.c sample program did not have any code at all to handle
532 the cases when CRLF is a valid newline sequence.
533
534 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
535 as a starting offset was within the subject string. There is now a new
536 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
537 negative or greater than the length of the string. In order to test this,
538 pcretest is extended to allow the setting of negative starting offsets.
539
540 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
541 starting offset points to the beginning of a UTF-8 character was
542 unnecessarily clumsy. I tidied it up.
543
544 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
545 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
546
547 17. Nobody had reported that the --include_dir option, which was added in
548 release 7.7 should have been called --include-dir (hyphen, not underscore)
549 for compatibility with GNU grep. I have changed it to --include-dir, but
550 left --include_dir as an undocumented synonym, and the same for
551 --exclude-dir, though that is not available in GNU grep, at least as of
552 release 2.5.4.
553
554 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
555 characters from a string of bytes) have been redefined so as not to use
556 loops, in order to improve performance in some environments. At the same
557 time, I abstracted some of the common code into auxiliary macros to save
558 repetition (this should not affect the compiled code).
559
560 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
561 compile-time error is now given if \c is not followed by an ASCII
562 character, that is, a byte less than 128. (In EBCDIC mode, the code is
563 different, and any byte value is allowed.)
564
565 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
566 START_OPTIMIZE option, which is now allowed at compile time - but just
567 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
568 to pcregrep and other applications that have no direct access to PCRE
569 options. The new /Y option in pcretest sets this option when calling
570 pcre_compile().
571
572 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
573 back references. Groups containing recursive back references were forced to
574 be atomic by that change, but in the case of named groups, the amount of
575 memory required was incorrectly computed, leading to "Failed: internal
576 error: code overflow". This has been fixed.
577
578 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
579 pcretest.c, to avoid build problems in some Borland environments.
580
581
582 Version 8.10 25-Jun-2010
583 ------------------------
584
585 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
586 THEN.
587
588 2. (*ACCEPT) was not working when inside an atomic group.
589
590 3. Inside a character class, \B is treated as a literal by default, but
591 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
592 causes the error). The code is unchanged, but I tidied the documentation.
593
594 4. Inside a character class, PCRE always treated \R and \X as literals,
595 whereas Perl faults them if its -w option is set. I have changed PCRE so
596 that it faults them when PCRE_EXTRA is set.
597
598 5. Added support for \N, which always matches any character other than
599 newline. (It is the same as "." when PCRE_DOTALL is not set.)
600
601 6. When compiling pcregrep with newer versions of gcc which may have
602 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
603 declared with attribute warn_unused_result" were given. Just casting the
604 result to (void) does not stop the warnings; a more elaborate fudge is
605 needed. I've used a macro to implement this.
606
607 7. Minor change to pcretest.c to avoid a compiler warning.
608
609 8. Added four artifical Unicode properties to help with an option to make
610 \s etc use properties (see next item). The new properties are: Xan
611 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
612
613 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
614 use Unicode properties. (*UCP) at the start of a pattern can be used to set
615 this option. Modified pcretest to add /W to test this facility. Added
616 REG_UCP to make it available via the POSIX interface.
617
618 10. Added --line-buffered to pcregrep.
619
620 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
621 studied, and the match started with a letter with a code point greater than
622 127 whose first byte was different to the first byte of the other case of
623 the letter, the other case of this starting letter was not recognized
624 (#976).
625
626 12. If a pattern that was studied started with a repeated Unicode property
627 test, for example, \p{Nd}+, there was the theoretical possibility of
628 setting up an incorrect bitmap of starting bytes, but fortunately it could
629 not have actually happened in practice until change 8 above was made (it
630 added property types that matched character-matching opcodes).
631
632 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
633 possible starting bytes for non-anchored patterns.
634
635 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
636 \R, and also a number of cases that involve Unicode properties, both
637 explicit and implicit when PCRE_UCP is set.
638
639 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
640 input, it could crash or give wrong results if characters with values
641 greater than 0xc0 were present in the subject string. (Detail: it assumed
642 UTF-8 input when processing these items.)
643
644 16. Added a lot of (int) casts to avoid compiler warnings in systems where
645 size_t is 64-bit (#991).
646
647 17. Added a check for running out of memory when PCRE is compiled with
648 --disable-stack-for-recursion (#990).
649
650 18. If the last data line in a file for pcretest does not have a newline on
651 the end, a newline was missing in the output.
652
653 19. The default pcre_chartables.c file recognizes only ASCII characters (values
654 less than 128) in its various bitmaps. However, there is a facility for
655 generating tables according to the current locale when PCRE is compiled. It
656 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
657 characters, are recognized by isspace() and therefore were getting set in
658 these tables, and indeed these tables seem to approximate to ISO 8859. This
659 caused a problem in UTF-8 mode when pcre_study() was used to create a list
660 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
661 which of course cannot start UTF-8 characters. I have changed the code so
662 that only real ASCII characters (less than 128) and the correct starting
663 bytes for UTF-8 encodings are set for characters greater than 127 when in
664 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
665 altogether.)
666
667 20. Added the /T option to pcretest so as to be able to run tests with non-
668 standard character tables, thus making it possible to include the tests
669 used for 19 above in the standard set of tests.
670
671 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
672 reference to a subpattern the other side of a comment that contains an
673 opening parenthesis caused either an internal compiling error, or a
674 reference to the wrong subpattern.
675
676
677 Version 8.02 19-Mar-2010
678 ------------------------
679
680 1. The Unicode data tables have been updated to Unicode 5.2.0.
681
682 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
683 configured.
684
685 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
686 original author of that file, following a query about its status.
687
688 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
689 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
690
691 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
692 quantifier applied to a forward-referencing subroutine call, could compile
693 incorrect code or give the error "internal error: previously-checked
694 referenced subpattern not found".
695
696 6. Both MS Visual Studio and Symbian OS have problems with initializing
697 variables to point to external functions. For these systems, therefore,
698 pcre_malloc etc. are now initialized to local functions that call the
699 relevant global functions.
700
701 7. There were two entries missing in the vectors called coptable and poptable
702 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
703 I've fixed the data, and added a kludgy way of testing at compile time that
704 the lengths are correct (equal to the number of opcodes).
705
706 8. Following on from 7, I added a similar kludge to check the length of the
707 eint vector in pcreposix.c.
708
709 9. Error texts for pcre_compile() are held as one long string to avoid too
710 much relocation at load time. To find a text, the string is searched,
711 counting zeros. There was no check for running off the end of the string,
712 which could happen if a new error number was added without updating the
713 string.
714
715 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
716
717 11. \K was not working if it appeared in an atomic group or in a group that
718 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
719 \K is "not well defined" if used in an assertion. PCRE now accepts it if
720 the assertion is positive, but not if it is negative.
721
722 12. Change 11 fortuitously reduced the size of the stack frame used in the
723 "match()" function of pcre_exec.c by one pointer. Forthcoming
724 implementation of support for (*MARK) will need an extra pointer on the
725 stack; I have reserved it now, so that the stack frame size does not
726 decrease.
727
728 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
729 item in branch that calls a recursion is a subroutine call - as in the
730 second branch in the above example - was incorrectly given the compile-
731 time error "recursive call could loop indefinitely" because pcre_compile()
732 was not correctly checking the subroutine for matching a non-empty string.
733
734 14. The checks for overrunning compiling workspace could trigger after an
735 overrun had occurred. This is a "should never occur" error, but it can be
736 triggered by pathological patterns such as hundreds of nested parentheses.
737 The checks now trigger 100 bytes before the end of the workspace.
738
739 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
740
741
742 Version 8.01 19-Jan-2010
743 ------------------------
744
745 1. If a pattern contained a conditional subpattern with only one branch (in
746 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
747 computed the wrong minimum data length (which is of course zero for such
748 subpatterns). This could cause incorrect "no match" results.
749
750 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
751 the pattern is reset in the first branch, pcre_compile() failed with
752 "internal error: code overflow at offset...". This happened only when
753 the reset was to the original external option setting. (An optimization
754 abstracts leading options settings into an external setting, which was the
755 cause of this.)
756
757 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
758 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
759 assertion pattern did not match (meaning that the assertion was true), it
760 was incorrectly treated as false if the SKIP had been reached during the
761 matching. This also applied to assertions used as conditions.
762
763 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
764 assertion subpattern, including such a pattern used as a condition,
765 unpredictable results occurred, instead of the error return
766 PCRE_ERROR_DFA_UITEM.
767
768 5. The C++ GlobalReplace function was not working like Perl for the special
769 situation when an empty string is matched. It now does the fancy magic
770 stuff that is necessary.
771
772 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
773 removed. (These were left over from very, very early versions of PCRE.)
774
775 7. Some cosmetic changes to the code to make life easier when compiling it
776 as part of something else:
777
778 (a) Change DEBUG to PCRE_DEBUG.
779
780 (b) In pcre_compile(), rename the member of the "branch_chain" structure
781 called "current" as "current_branch", to prevent a collision with the
782 Linux macro when compiled as a kernel module.
783
784 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
785 prevent a collision with the Linux macro when compiled as a kernel
786 module.
787
788 8. In pcre_compile() there are some checks for integer overflows that used to
789 cast potentially large values to (double). This has been changed to that
790 when building, a check for int64_t is made, and if it is found, it is used
791 instead, thus avoiding the use of floating point arithmetic. (There is no
792 other use of FP in PCRE.) If int64_t is not found, the fallback is to
793 double.
794
795 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
796 2005 (difference between two addresses compared to an unsigned value).
797
798 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
799 custom one, because of the following reported problem in Windows:
800
801 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
802 under Win32.
803 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
804 therefore missing the function definition.
805 - The compiler thus generates a "C" signature for the test function.
806 - The linker fails to find the "C" function.
807 - PCRE fails to configure if asked to do so against libbz2.
808
809 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
810 messages were output:
811
812 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
813 rerunning libtoolize, to keep the correct libtool macros in-tree.
814 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
815
816 I have done both of these things.
817
818 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
819 most of the time, it *can* run out if it is given a pattern that contains a
820 runaway infinite recursion. I updated the discussion in the pcrestack man
821 page.
822
823 13. Now that we have gone to the x.xx style of version numbers, the minor
824 version may start with zero. Using 08 or 09 is a bad idea because users
825 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
826 interpreted as invalid octal numbers. I've updated the previous comment in
827 configure.ac, and also added a check that gives an error if 08 or 09 are
828 used.
829
830 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
831 causing partial matching to fail when the end of the subject matched \W
832 in a UTF-8 pattern where \W was quantified with a minimum of 3.
833
834 15. There were some discrepancies between the declarations in pcre_internal.h
835 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
836 their definitions. The declarations used "const uschar *" and the
837 definitions used USPTR. Even though USPTR is normally defined as "const
838 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
839 reported that: "This difference in casting confuses some C++ compilers, for
840 example, SunCC recognizes above declarations as different functions and
841 generates broken code for hbpcre." I have changed the declarations to use
842 USPTR.
843
844 16. GNU libtool is named differently on some systems. The autogen.sh script now
845 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
846 (FreeBSD).
847
848 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
849 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
850 comment: "Figure out how to create a longlong from a string: strtoll and
851 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
852 instance, but it only takes 2 args instead of 3!"
853
854 18. A subtle bug concerned with back references has been fixed by a change of
855 specification, with a corresponding code fix. A pattern such as
856 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
857 refers, was giving matches when it shouldn't. For example, xa=xaaa would
858 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
859 same bug. Such groups have to be quantified to be useful, or contained
860 inside another quantified group. (If there's no repetition, the reference
861 can never match.) The problem arises because, having left the group and
862 moved on to the rest of the pattern, a later failure that backtracks into
863 the group uses the captured value from the final iteration of the group
864 rather than the correct earlier one. I have fixed this in PCRE by forcing
865 any group that contains a reference to itself to be an atomic group; that
866 is, there cannot be any backtracking into it once it has completed. This is
867 similar to recursive and subroutine calls.
868
869
870 Version 8.00 19-Oct-09
871 ----------------------
872
873 1. The table for translating pcre_compile() error codes into POSIX error codes
874 was out-of-date, and there was no check on the pcre_compile() error code
875 being within the table. This could lead to an OK return being given in
876 error.
877
878 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
879 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
880 in a Windows environment.
881
882 3. The pcregrep --count option prints the count for each file even when it is
883 zero, as does GNU grep. However, pcregrep was also printing all files when
884 --files-with-matches was added. Now, when both options are given, it prints
885 counts only for those files that have at least one match. (GNU grep just
886 prints the file name in this circumstance, but including the count seems
887 more useful - otherwise, why use --count?) Also ensured that the
888 combination -clh just lists non-zero counts, with no names.
889
890 4. The long form of the pcregrep -F option was incorrectly implemented as
891 --fixed_strings instead of --fixed-strings. This is an incompatible change,
892 but it seems right to fix it, and I didn't think it was worth preserving
893 the old behaviour.
894
895 5. The command line items --regex=pattern and --regexp=pattern were not
896 recognized by pcregrep, which required --regex pattern or --regexp pattern
897 (with a space rather than an '='). The man page documented the '=' forms,
898 which are compatible with GNU grep; these now work.
899
900 6. No libpcreposix.pc file was created for pkg-config; there was just
901 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
902
903 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
904 when UCP support is not needed, by modifying the Python script that
905 generates it from Unicode data files. This should not matter if the module
906 is correctly used as a library, but I received one complaint about 50K of
907 unwanted data. My guess is that the person linked everything into his
908 program rather than using a library. Anyway, it does no harm.
909
910 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
911 was a minimum greater than 1 for a wide character in a possessive
912 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
913 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
914 character. Chaos in the form of incorrect output or a compiling loop could
915 result.
916
917 9. The restrictions on what a pattern can contain when partial matching is
918 requested for pcre_exec() have been removed. All patterns can now be
919 partially matched by this function. In addition, if there are at least two
920 slots in the offset vector, the offset of the earliest inspected character
921 for the match and the offset of the end of the subject are set in them when
922 PCRE_ERROR_PARTIAL is returned.
923
924 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
925 synonymous with PCRE_PARTIAL, for backwards compatibility, and
926 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
927 and may be more useful for multi-segment matching.
928
929 11. Partial matching with pcre_exec() is now more intuitive. A partial match
930 used to be given if ever the end of the subject was reached; now it is
931 given only if matching could not proceed because another character was
932 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
933 string "Z", which now yields "no match" instead of "partial match". In the
934 case of pcre_dfa_exec(), "no match" is given if every matching path for the
935 final character ended with (*FAIL).
936
937 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
938 if the pattern had a "must contain" character that was already found in the
939 earlier partial match, unless partial matching was again requested. For
940 example, with the pattern /dog.(body)?/, the "must contain" character is
941 "g". If the first part-match was for the string "dog", restarting with
942 "sbody" failed. This bug has been fixed.
943
944 13. The string returned by pcre_dfa_exec() after a partial match has been
945 changed so that it starts at the first inspected character rather than the
946 first character of the match. This makes a difference only if the pattern
947 starts with a lookbehind assertion or \b or \B (\K is not supported by
948 pcre_dfa_exec()). It's an incompatible change, but it makes the two
949 matching functions compatible, and I think it's the right thing to do.
950
951 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
952 so that the demonstration program is easily available in environments where
953 PCRE has not been installed from source.
954
955 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
956 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
957 library.
958
959 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
960 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
961 is not the first non-POSIX option to be added. Clearly some people find
962 these options useful.
963
964 17. If a caller to the POSIX matching function regexec() passes a non-zero
965 value for nmatch with a NULL value for pmatch, the value of
966 nmatch is forced to zero.
967
968 18. RunGrepTest did not have a test for the availability of the -u option of
969 the diff command, as RunTest does. It now checks in the same way as
970 RunTest, and also checks for the -b option.
971
972 19. If an odd number of negated classes containing just a single character
973 interposed, within parentheses, between a forward reference to a named
974 subpattern and the definition of the subpattern, compilation crashed with
975 an internal error, complaining that it could not find the referenced
976 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
977 [The bug was that it was starting one character too far in when skipping
978 over the character class, thus treating the ] as data rather than
979 terminating the class. This meant it could skip too much.]
980
981 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
982 /g option in pcretest when the pattern contains \K, which makes it possible
983 to have an empty string match not at the start, even when the pattern is
984 anchored. Updated pcretest and pcredemo to use this option.
985
986 21. If the maximum number of capturing subpatterns in a recursion was greater
987 than the maximum at the outer level, the higher number was returned, but
988 with unset values at the outer level. The correct (outer level) value is
989 now given.
990
991 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
992 PCRE did not set those parentheses (unlike Perl). I have now found a way to
993 make it do so. The string so far is captured, making this feature
994 compatible with Perl.
995
996 23. The tests have been re-organized, adding tests 11 and 12, to make it
997 possible to check the Perl 5.10 features against Perl 5.10.
998
999 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1000 pattern matches a fixed length string. PCRE did not allow this; now it
1001 does. Neither allows recursion.
1002
1003 25. I finally figured out how to implement a request to provide the minimum
1004 length of subject string that was needed in order to match a given pattern.
1005 (It was back references and recursion that I had previously got hung up
1006 on.) This code has now been added to pcre_study(); it finds a lower bound
1007 to the length of subject needed. It is not necessarily the greatest lower
1008 bound, but using it to avoid searching strings that are too short does give
1009 some useful speed-ups. The value is available to calling programs via
1010 pcre_fullinfo().
1011
1012 26. While implementing 25, I discovered to my embarrassment that pcretest had
1013 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1014 study optimizations had never been tested with that matching function.
1015 Oops. What is worse, even when it was passed study data, there was a bug in
1016 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1017 were also very few tests of studied patterns with pcre_dfa_exec().
1018
1019 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1020 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1021 on the other side of the coin, they are no longer allowed to have different
1022 names, because these cannot be distinguished in PCRE, and this has caused
1023 confusion. (This is a difference from Perl.)
1024
1025 28. When duplicate subpattern names are present (necessarily with different
1026 numbers, as required by 27 above), and a test is made by name in a
1027 conditional pattern, either for a subpattern having been matched, or for
1028 recursion in such a pattern, all the associated numbered subpatterns are
1029 tested, and the overall condition is true if the condition is true for any
1030 one of them. This is the way Perl works, and is also more like the way
1031 testing by number works.
1032
1033
1034 Version 7.9 11-Apr-09
1035 ---------------------
1036
1037 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1038 (pcretest), all targets were linked against these libraries. This included
1039 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1040 libraries. This caused unwanted dependencies to be created. This problem
1041 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1042 pcretest is linked with readline.
1043
1044 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1045 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1046 moved outside it again, because FALSE and TRUE are already defined in AIX,
1047 but BOOL is not.
1048
1049 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1050 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1051
1052 4. The pcregrep documentation said spaces were inserted as well as colons (or
1053 hyphens) following file names and line numbers when outputting matching
1054 lines. This is not true; no spaces are inserted. I have also clarified the
1055 wording for the --colour (or --color) option.
1056
1057 5. In pcregrep, when --colour was used with -o, the list of matching strings
1058 was not coloured; this is different to GNU grep, so I have changed it to be
1059 the same.
1060
1061 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1062 each matching line was coloured. Now it goes on to look for further matches
1063 of any of the test patterns, which is the same behaviour as GNU grep.
1064
1065 7. A pattern that could match an empty string could cause pcregrep to loop; it
1066 doesn't make sense to accept an empty string match in pcregrep, so I have
1067 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1068 seems to be how GNU grep behaves.
1069
1070 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1071 start or after a newline", because the conditional assertion was not being
1072 correctly handled. The rule now is that both the assertion and what follows
1073 in the first alternative must satisfy the test.
1074
1075 9. If auto-callout was enabled in a pattern with a conditional group whose
1076 condition was an assertion, PCRE could crash during matching, both with
1077 pcre_exec() and pcre_dfa_exec().
1078
1079 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1080 used for matching.
1081
1082 11. Unicode property support in character classes was not working for
1083 characters (bytes) greater than 127 when not in UTF-8 mode.
1084
1085 12. Added the -M command line option to pcretest.
1086
1087 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1088
1089 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1090
1091 16. Added comments and documentation about mis-use of no_arg in the C++
1092 wrapper.
1093
1094 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1095 from Martin Jerabek that uses macro names for all relevant character and
1096 string constants.
1097
1098 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1099 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1100 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1101 these, but not everybody uses configure.
1102
1103 19. A conditional group that had only one branch was not being correctly
1104 recognized as an item that could match an empty string. This meant that an
1105 enclosing group might also not be so recognized, causing infinite looping
1106 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1107 with the subject "ab", where knowledge that the repeated group can match
1108 nothing is needed in order to break the loop.
1109
1110 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1111 exec(), but without supplying a callout function, matching went wrong.
1112
1113 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1114 leak if the size of the offset vector was greater than 30. When the vector
1115 is smaller, the saved offsets during recursion go onto a local stack
1116 vector, but for larger vectors malloc() is used. It was failing to free
1117 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1118 error, in fact).
1119
1120 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1121 heapframe that is used only when UTF-8 support is enabled. This caused no
1122 problem, but was untidy.
1123
1124 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1125 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1126 included within another project.
1127
1128 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1129 slightly modified by me:
1130
1131 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1132 not building pcregrep.
1133
1134 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1135 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1136
1137 25. Forward references, both numeric and by name, in patterns that made use of
1138 duplicate group numbers, could behave incorrectly or give incorrect errors,
1139 because when scanning forward to find the reference group, PCRE was not
1140 taking into account the duplicate group numbers. A pattern such as
1141 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1142
1143 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1144 the feature of a custom pointer more persuasive (as requested by a user).
1145
1146 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1147 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1148 pre-defined.
1149
1150 28. Added support for (*UTF8) at the start of a pattern.
1151
1152 29. Arrange for flags added by the "release type" setting in CMake to be shown
1153 in the configuration summary.
1154
1155
1156 Version 7.8 05-Sep-08
1157 ---------------------
1158
1159 1. Replaced UCP searching code with optimized version as implemented for Ad
1160 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1161 stage table and inline lookup instead of a function, giving speed ups of 2
1162 to 5 times on some simple patterns that I tested. Permission was given to
1163 distribute the MultiStage2.py script that generates the tables (it's not in
1164 the tarball, but is in the Subversion repository).
1165
1166 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1167 scripts.
1168
1169 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1170 a group with a zero qualifier. The result of the study could be incorrect,
1171 or the function might crash, depending on the pattern.
1172
1173 4. Caseless matching was not working for non-ASCII characters in back
1174 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1175 It now works when Unicode Property Support is available.
1176
1177 5. In pcretest, an escape such as \x{de} in the data was always generating
1178 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1179 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1180 truncation.
1181
1182 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1183
1184 7. Added two (int) casts to pcregrep when printing the difference of two
1185 pointers, in case they are 64-bit values.
1186
1187 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1188 test 2 if it fails.
1189
1190 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1191 and a #define of that name to empty if it is not externally set. This is to
1192 allow users of MSVC to set it if necessary.
1193
1194 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1195 the convenience functions in the pcre_get.c source file.
1196
1197 11. An option change at the start of a pattern that had top-level alternatives
1198 could cause overwriting and/or a crash. This command provoked a crash in
1199 some environments:
1200
1201 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1202
1203 This potential security problem was recorded as CVE-2008-2371.
1204
1205 12. For a pattern where the match had to start at the beginning or immediately
1206 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1207 pcre_dfa_exec() could read past the end of the passed subject if there was
1208 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1209 pcretest so that it places the subject at the end of its malloc-ed buffer.
1210
1211 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1212 exec() might read past the end of the data buffer in UTF-8 mode.
1213
1214 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1215 the data contained the byte 0x85 as part of a UTF-8 character within its
1216 first line. This applied both to normal and DFA matching.
1217
1218 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1219 /^[^d]*?$/8 failed to match "abc".
1220
1221 16. Added a missing copyright notice to pcrecpp_internal.h.
1222
1223 17. Make it more clear in the documentation that values returned from
1224 pcre_exec() in ovector are byte offsets, not character counts.
1225
1226 18. Tidied a few places to stop certain compilers from issuing warnings.
1227
1228 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1229 supplied by Stefan Weber. I made a further small update for 7.8 because
1230 there is a change of source arrangements: the pcre_searchfuncs.c module is
1231 replaced by pcre_ucd.c.
1232
1233
1234 Version 7.7 07-May-08
1235 ---------------------
1236
1237 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1238 a string to a long long, pretend we don't even have a long long." This is
1239 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1240
1241 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1242 pre-7.6 versions, which defined a global no_arg variable instead of putting
1243 it in the RE class. (See also #8 below.)
1244
1245 3. Remove a line of dead code, identified by coverity and reported by Nuno
1246 Lopes.
1247
1248 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1249
1250 (1) The include/exclude patterns were being applied to the whole pathnames
1251 of files, instead of just to the final components.
1252
1253 (2) If there was more than one level of directory, the subdirectories were
1254 skipped unless they satisfied the include/exclude conditions. This is
1255 inconsistent with GNU grep (and could even be seen as contrary to the
1256 pcregrep specification - which I improved to make it absolutely clear).
1257 The action now is always to scan all levels of directory, and just
1258 apply the include/exclude patterns to regular files.
1259
1260 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1261 --exclude_dir in the tests to avoid scanning .svn directories.
1262
1263 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1264 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1265 doesn't support NULs in patterns.
1266
1267 7. Added some missing "const"s to declarations of static tables in
1268 pcre_compile.c and pcre_dfa_exec.c.
1269
1270 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1271 caused by fix #2 above. (Subsequently also a second patch to fix the
1272 first patch. And a third patch - this was a messy problem.)
1273
1274 9. Applied Craig's patch to remove the use of push_back().
1275
1276 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1277 matching function regexec().
1278
1279 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1280 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1281 references. PCRE supports relative numbers with this syntax (I don't think
1282 Oniguruma does).
1283
1284 12. Previously, a group with a zero repeat such as (...){0} was completely
1285 omitted from the compiled regex. However, this means that if the group
1286 was called as a subroutine from elsewhere in the pattern, things went wrong
1287 (an internal error was given). Such groups are now left in the compiled
1288 pattern, with a new opcode that causes them to be skipped at execution
1289 time.
1290
1291 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1292 to the way PCRE behaves:
1293
1294 (a) A lone ] character is dis-allowed (Perl treats it as data).
1295
1296 (b) A back reference to an unmatched subpattern matches an empty string
1297 (Perl fails the current match path).
1298
1299 (c) A data ] in a character class must be notated as \] because if the
1300 first data character in a class is ], it defines an empty class. (In
1301 Perl it is not possible to have an empty class.) The empty class []
1302 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1303 The negative empty class [^] matches any one character, independently
1304 of the DOTALL setting.
1305
1306 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1307 non-existent subpattern following a character class starting with ']' and
1308 containing () gave an internal compiling error instead of "reference to
1309 non-existent subpattern". Fortunately, when the pattern did exist, the
1310 compiled code was correct. (When scanning forwards to check for the
1311 existencd of the subpattern, it was treating the data ']' as terminating
1312 the class, so got the count wrong. When actually compiling, the reference
1313 was subsequently set up correctly.)
1314
1315 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1316 it was being rejected as not supported by pcre_dfa_exec(), even though
1317 other assertions are supported. I have made pcre_dfa_exec() support
1318 (*FAIL).
1319
1320 16. The implementation of 13c above involved the invention of a new opcode,
1321 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1322 cannot be changed at match time, I realized I could make a small
1323 improvement to matching performance by compiling OP_ALLANY instead of
1324 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1325 on the OP_ANY path.
1326
1327 17. Compiling pcretest on Windows with readline support failed without the
1328 following two fixes: (1) Make the unistd.h include conditional on
1329 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1330
1331 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1332 ncurses library to be included for pcretest when ReadLine support is
1333 requested, but also to allow for it to be overridden. This patch came from
1334 Daniel Bergström.
1335
1336 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1337 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1338 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1339 spotting this.
1340
1341
1342 Version 7.6 28-Jan-08
1343 ---------------------
1344
1345 1. A character class containing a very large number of characters with
1346 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1347 overflow.
1348
1349 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1350 HAVE_LONG_LONG is not defined.
1351
1352 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1353 bring them up to date and include new features. This patch includes:
1354
1355 - Fixed PH's badly added libz and libbz2 support.
1356 - Fixed a problem with static linking.
1357 - Added pcredemo. [But later removed - see 7 below.]
1358 - Fixed dftables problem and added an option.
1359 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1360 HAVE_LONG_LONG.
1361 - Added readline support for pcretest.
1362 - Added an listing of the option settings after cmake has run.
1363
1364 4. A user submitted a patch to Makefile that makes it easy to create
1365 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1366 Makefile.am that cause it to include this special target, without
1367 affecting anything else. Note that the same mingw target plus all
1368 the other distribution libraries and programs are now supported
1369 when configuring with CMake (see 6 below) instead of with
1370 Configure/Make.
1371
1372 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1373 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1374 exported in the Windows port". It has not yet been confirmed that the patch
1375 solves the problem, but it does no harm.
1376
1377 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1378 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1379 with CMake, and also correct the comment about stack recursion.
1380
1381 7. Remove the automatic building of pcredemo from the ./configure system and
1382 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1383 of a program that users should build themselves after PCRE is installed, so
1384 building it automatically is not really right. What is more, it gave
1385 trouble in some build environments.
1386
1387 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1388
1389
1390 Version 7.5 10-Jan-08
1391 ---------------------
1392
1393 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1394 values in parens when parsing an RE using the C++ wrapper."
1395
1396 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1397 Characters greater than 255 were excluded from the class instead of being
1398 included.
1399
1400 3. The same bug as (2) above applied to negated POSIX classes such as
1401 [:^space:].
1402
1403 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1404 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1405 I have changed it.
1406
1407 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1408 first named subpattern) and a construct such as (?&a) would reference the
1409 first named subpattern whose name started with "a" (in other words, the
1410 length check was missing). Both these problems are fixed. "Subpattern name
1411 expected" is now given for (?&) (a zero-length name), and this patch also
1412 makes it give the same error for \k'' (previously it complained that that
1413 was a reference to a non-existent subpattern).
1414
1415 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1416 this is right because (?- can be followed by option settings as well as by
1417 digits. I have, however, made the messages clearer.
1418
1419 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1420 than the number used in the conditional) now cause a compile-time error.
1421 This is actually not compatible with Perl, which accepts such patterns, but
1422 treats the conditional as always being FALSE (as PCRE used to), but it
1423 seems to me that giving a diagnostic is better.
1424
1425 8. Change "alphameric" to the more common word "alphanumeric" in comments
1426 and messages.
1427
1428 9. Fix two occurrences of "backslash" in comments that should have been
1429 "backspace".
1430
1431 10. Remove two redundant lines of code that can never be obeyed (their function
1432 was moved elsewhere).
1433
1434 11. The program that makes PCRE's Unicode character property table had a bug
1435 which caused it to generate incorrect table entries for sequences of
1436 characters that have the same character type, but are in different scripts.
1437 It amalgamated them into a single range, with the script of the first of
1438 them. In other words, some characters were in the wrong script. There were
1439 thirteen such cases, affecting characters in the following ranges:
1440
1441 U+002b0 - U+002c1
1442 U+0060c - U+0060d
1443 U+0061e - U+00612
1444 U+0064b - U+0065e
1445 U+0074d - U+0076d
1446 U+01800 - U+01805
1447 U+01d00 - U+01d77
1448 U+01d9b - U+01dbf
1449 U+0200b - U+0200f
1450 U+030fc - U+030fe
1451 U+03260 - U+0327f
1452 U+0fb46 - U+0fbb1
1453 U+10450 - U+1049d
1454
1455 12. The -o option (show only the matching part of a line) for pcregrep was not
1456 compatible with GNU grep in that, if there was more than one match in a
1457 line, it showed only the first of them. It now behaves in the same way as
1458 GNU grep.
1459
1460 13. If the -o and -v options were combined for pcregrep, it printed a blank
1461 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1462 does the same. The return code can be used to tell if there were any
1463 non-matching lines.
1464
1465 14. Added --file-offsets and --line-offsets to pcregrep.
1466
1467 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1468 infinitely looping recursion. The bug was that positive lookaheads were not
1469 being skipped when checking for a possible empty match (negative lookaheads
1470 and both kinds of lookbehind were skipped).
1471
1472 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1473 inclusion of <windows.h> to before rather than after the definition of
1474 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1475
1476 17. Specifying a possessive quantifier with a specific limit for a Unicode
1477 character property caused pcre_compile() to compile bad code, which led at
1478 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1479 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1480 caused the error; without that there was no problem.
1481
1482 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1483
1484 19. Added --enable-pcretest-libreadline.
1485
1486 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1487 RE::GlobalReplace(). As a result, the number of replacements returned was
1488 double what it should be. I removed one of the increments, but Craig sent a
1489 later patch that removed the other one (the right fix) and added unit tests
1490 that check the return values (which was not done before).
1491
1492 21. Several CMake things:
1493
1494 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1495 the names libpcre and libpcreposix, not just pcre and pcreposix.
1496
1497 (2) The above change means that pcretest and pcregrep are now correctly
1498 linked with the newly-built libraries, not previously installed ones.
1499
1500 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1501
1502 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1503 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1504 UTF-8 newline character). The key issue is that the pattern starts .*;
1505 this means that the match must be either at the beginning, or after a
1506 newline. The bug was in the code for advancing after a failed match and
1507 checking that the new position followed a newline. It was not taking
1508 account of UTF-8 characters correctly.
1509
1510 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1511 character classes. PCRE was not treating the sequence [:...:] as a
1512 character class unless the ... were all letters. Perl, however, seems to
1513 allow any characters between [: and :], though of course it rejects as
1514 unknown any "names" that contain non-letters, because all the known class
1515 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1516 for example, whereas PCRE did not - it did not recognize a POSIX character
1517 class. This seemed a bit dangerous, so the code has been changed to be
1518 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1519 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1520 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1521 Perl does, and where it didn't before.
1522
1523 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1524 Windows environments %n is disabled by default.
1525
1526
1527 Version 7.4 21-Sep-07
1528 ---------------------
1529
1530 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1531 means that a class such as [\s] counted as "explicit reference to CR or
1532 LF". That isn't really right - the whole point of the change was to try to
1533 help when there was an actual mention of one of the two characters. So now
1534 the change happens only if \r or \n (or a literal CR or LF) character is
1535 encountered.
1536
1537 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1538 of both had grown to the point where there were only 3 bits left.
1539 Fortunately, there was spare space in the data structure, and so I have
1540 moved the internal flags into a new 16-bit field to free up more option
1541 bits.
1542
1543 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1544 but did not set the internal JCHANGED flag - either of these is enough to
1545 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1546 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1547 start sets both bits.
1548
1549 4. Added options (at build time, compile time, exec time) to change \R from
1550 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1551
1552 5. doc/pcresyntax.html was missing from the distribution.
1553
1554 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1555 compatibility, even though it is no longer used.
1556
1557 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1558 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1559 windows.h file is present (where different names are used). [This was
1560 reversed later after testing - see 16 below.]
1561
1562 8. Changed all #include <config.h> to #include "config.h". There were also
1563 some further <pcre.h> cases that I changed to "pcre.h".
1564
1565 9. When pcregrep was used with the --colour option, it missed the line ending
1566 sequence off the lines that it output.
1567
1568 10. It was pointed out to me that arrays of string pointers cause lots of
1569 relocations when a shared library is dynamically loaded. A technique of
1570 using a single long string with a table of offsets can drastically reduce
1571 these. I have refactored PCRE in four places to do this. The result is
1572 dramatic:
1573
1574 Originally: 290
1575 After changing UCP table: 187
1576 After changing error message table: 43
1577 After changing table of "verbs" 36
1578 After changing table of Posix names 22
1579
1580 Thanks to the folks working on Gregex for glib for this insight.
1581
1582 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1583 unicode-properties was also set.
1584
1585 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1586
1587 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1588 checked only for CRLF.
1589
1590 14. Added casts to pcretest.c to avoid compiler warnings.
1591
1592 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1593
1594 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1595 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1596 entirely. This removes changes made in 7 above.
1597
1598 17. The CMake files have been updated, and there is now more information about
1599 building with CMake in the NON-UNIX-USE document.
1600
1601
1602 Version 7.3 28-Aug-07
1603 ---------------------
1604
1605 1. In the rejigging of the build system that eventually resulted in 7.1, the
1606 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1607 brackets there is not right, since it causes compilers to look for an
1608 installed pcre.h, not the version that is in the source that is being
1609 compiled (which of course may be different). I have changed it back to:
1610
1611 #include "pcre.h"
1612
1613 I have a vague recollection that the change was concerned with compiling in
1614 different directories, but in the new build system, that is taken care of
1615 by the VPATH setting the Makefile.
1616
1617 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1618 when the subject happened to end in the byte 0x85 (e.g. if the last
1619 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1620 characters but of course it shouldn't be taken as a newline when it is part
1621 of another character. The bug was that, for an unlimited repeat of . in
1622 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1623 characters when looking for a newline.
1624
1625 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1626
1627 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1628 in debug output.
1629
1630 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1631 long printing in the pcrecpp unittest when running under MinGW.
1632
1633 6. ESC_K was left out of the EBCDIC table.
1634
1635 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1636 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1637 limit also applies to "virtual nesting" when a pattern is recursive, and in
1638 this case 1000 isn't so big. I have been able to remove this limit at the
1639 expense of backing off one optimization in certain circumstances. Normally,
1640 when pcre_exec() would call its internal match() function recursively and
1641 immediately return the result unconditionally, it uses a "tail recursion"
1642 feature to save stack. However, when a subpattern that can match an empty
1643 string has an unlimited repetition quantifier, it no longer makes this
1644 optimization. That gives it a stack frame in which to save the data for
1645 checking that an empty string has been matched. Previously this was taken
1646 from the 1000-entry workspace that had been reserved. So now there is no
1647 explicit limit, but more stack is used.
1648
1649 8. Applied Daniel's patches to solve problems with the import/export magic
1650 syntax that is required for Windows, and which was going wrong for the
1651 pcreposix and pcrecpp parts of the library. These were overlooked when this
1652 problem was solved for the main library.
1653
1654 9. There were some crude static tests to avoid integer overflow when computing
1655 the size of patterns that contain repeated groups with explicit upper
1656 limits. As the maximum quantifier is 65535, the maximum group length was
1657 set at 30,000 so that the product of these two numbers did not overflow a
1658 32-bit integer. However, it turns out that people want to use groups that
1659 are longer than 30,000 bytes (though not repeat them that many times).
1660 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1661 made it possible to implement the integer overflow checks in a much more
1662 dynamic way, which I have now done. The artificial limitation on group
1663 length has been removed - we now have only the limit on the total length of
1664 the compiled pattern, which depends on the LINK_SIZE setting.
1665
1666 10. Fixed a bug in the documentation for get/copy named substring when
1667 duplicate names are permitted. If none of the named substrings are set, the
1668 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1669 empty string.
1670
1671 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1672 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1673 because the ] is interpreted as the first data character and the
1674 terminating ] is not found. PCRE has been made compatible with Perl in this
1675 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1676 cause memory overwriting.
1677
1678 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1679 string has been matched (to stop an infinite loop). It was not recognizing
1680 a conditional subpattern that could match an empty string if that
1681 subpattern was within another subpattern. For example, it looped when
1682 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1683 condition was not nested. This bug has been fixed.
1684
1685 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1686 past the start of the subject in the presence of bytes with the top bit
1687 set, for example "\x8aBCD".
1688
1689 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1690 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1691
1692 14. Optimized (?!) to (*FAIL).
1693
1694 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1695 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1696 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1697 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1698 does: it's just the validity check that is more restrictive.
1699
1700 16. Inserted checks for integer overflows during escape sequence (backslash)
1701 processing, and also fixed erroneous offset values for syntax errors during
1702 backslash processing.
1703
1704 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1705 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1706
1707 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1708 caused an overrun.
1709
1710 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1711 something other than just ASCII characters) inside a group that had an
1712 unlimited repeat caused a loop at compile time (while checking to see
1713 whether the group could match an empty string).
1714
1715 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1716 [\P{Any}] did so. (Error in the code for printing property names.)
1717
1718 21. An orphan \E inside a character class could cause a crash.
1719
1720 22. A repeated capturing bracket such as (A)? could cause a wild memory
1721 reference during compilation.
1722
1723 23. There are several functions in pcre_compile() that scan along a compiled
1724 expression for various reasons (e.g. to see if it's fixed length for look
1725 behind). There were bugs in these functions when a repeated \p or \P was
1726 present in the pattern. These operators have additional parameters compared
1727 with \d, etc, and these were not being taken into account when moving along
1728 the compiled data. Specifically:
1729
1730 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1731 length.
1732
1733 (b) An item such as \pL+ within a repeated group could cause crashes or
1734 loops.
1735
1736 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1737 "reference to non-existent subpattern" error.
1738
1739 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1740
1741 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1742 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1743
1744 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1745
1746 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1747 character were causing crashes (broken optimization).
1748
1749 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1750 \p or \P) caused a compile-time loop.
1751
1752 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1753 break. For example, the unstudied pattern [\r\n]A does not match the string
1754 "\r\nA" because change 7.0/46 below moves the current point on by two
1755 characters after failing to match at the start. However, the pattern \nA
1756 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1757 the same is true. There doesn't seem any very clean way out of this, but
1758 what I have chosen to do makes the common cases work: PCRE now takes note
1759 of whether there can be an explicit match for \r or \n anywhere in the
1760 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1761 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1762 pattern has explicit CR or LF references.
1763
1764 29. Added (*CR) etc for changing newline setting at start of pattern.
1765
1766
1767 Version 7.2 19-Jun-07
1768 ---------------------
1769
1770 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1771 which is apparently normally available under Windows.
1772
1773 2. Re-jig the pcregrep tests with different newline settings in an attempt
1774 to make them independent of the local environment's newline setting.
1775
1776 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1777
1778 4. Some of the "internals" tests were previously cut out when the link size
1779 was not 2, because the output contained actual offsets. The recent new
1780 "Z" feature of pcretest means that these can be cut out, making the tests
1781 usable with all link sizes.
1782
1783 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1784 stack recursion. This gives a massive performance boost under BSD, but just
1785 a small improvement under Linux. However, it saves one field in the frame
1786 in all cases.
1787
1788 6. Added more features from the forthcoming Perl 5.10:
1789
1790 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1791 recursion call. It refers to the nth most recently opened parentheses.
1792
1793 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1794 to be opened parentheses.
1795
1796 (c) Conditions that refer to capturing parentheses can be specified
1797 relatively, for example, (?(-2)... or (?(+3)...
1798
1799 (d) \K resets the start of the current match so that everything before
1800 is not part of it.
1801
1802 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1803
1804 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1805 reference syntax.
1806
1807 (g) (?| introduces a group in which the numbering of parentheses in each
1808 alternative starts with the same number.
1809
1810 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1811
1812 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1813 PCRE_INFO_JCHANGED.
1814
1815 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1816 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1817 for detecting groups that can match an empty string.
1818
1819 9. A pattern with a very large number of alternatives (more than several
1820 hundred) was running out of internal workspace during the pre-compile
1821 phase, where pcre_compile() figures out how much memory will be needed. A
1822 bit of new cunning has reduced the workspace needed for groups with
1823 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1824 workspace instead of running out of the 4096 that are available.
1825
1826 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1827
1828 11. Applied patch from Google to remove an optimization that didn't quite work.
1829 The report of the bug said:
1830
1831 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1832 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1833 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1834
1835 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1836 it matched the wrong number of bytes.
1837
1838
1839 Version 7.1 24-Apr-07
1840 ---------------------
1841
1842 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1843 that is more "standard", making use of automake and other Autotools. There
1844 is some re-arrangement of the files and adjustment of comments consequent
1845 on this.
1846
1847 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1848 for recursive directory scanning broke on some systems because the files
1849 are not scanned in any specific order and on different systems the order
1850 was different. A call to "sort" has been inserted into RunGrepTest for the
1851 approprate test as a short-term fix. In the longer term there may be an
1852 alternative.
1853
1854 3. I had an email from Eric Raymond about problems translating some of PCRE's
1855 man pages to HTML (despite the fact that I distribute HTML pages, some
1856 people do their own conversions for various reasons). The problems
1857 concerned the use of low-level troff macros .br and .in. I have therefore
1858 removed all such uses from the man pages (some were redundant, some could
1859 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1860 HTML has been updated to handle .nf/.fi and to complain if it encounters
1861 .br or .in.
1862
1863 4. Updated comments in configure.ac that get placed in config.h.in and also
1864 arranged for config.h to be included in the distribution, with the name
1865 config.h.generic, for the benefit of those who have to compile without
1866 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1867
1868 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1869 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1870 makevp.bat for the current PCRE, using the additional files
1871 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1872
1873 6. A Windows user reported a minor discrepancy with test 2, which turned out
1874 to be caused by a trailing space on an input line that had got lost in his
1875 copy. The trailing space was an accident, so I've just removed it.
1876
1877 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1878 that is needed.
1879
1880 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1881 as "const" (a) because they are and (b) because it helps the PHP
1882 maintainers who have recently made a script to detect big data structures
1883 in the php code that should be moved to the .rodata section. I remembered
1884 to update Builducptable as well, so it won't revert if ucptable.h is ever
1885 re-created.
1886
1887 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1888 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1889 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1890 support is not required. This saves 1.5-2K of code, which is important in
1891 some applications.
1892
1893 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1894 so as not to refer to the tables, even though these functions will never be
1895 called when UTF-8 support is disabled. Otherwise there are problems with a
1896 shared library.
1897
1898 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1899
1900 (a) It was defining its arguments as char * instead of void *.
1901
1902 (b) It was assuming that all moves were upwards in memory; this was true
1903 a long time ago when I wrote it, but is no longer the case.
1904
1905 The emulated memove() is provided for those environments that have neither
1906 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1907 is clearly not the case, as these two bugs were recently reported.
1908
1909 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1910 and Detrail to create the HTML documentation, the .txt form of the man
1911 pages, and it removes trailing spaces from listed files. It also creates
1912 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1913 case, it wraps all the #defines with #ifndefs. This script should be run
1914 before "make dist".
1915
1916 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1917 with Unicode property support.
1918
1919 (a) For a maximizing quantifier, if the two different cases of the
1920 character were of different lengths in their UTF-8 codings (there are
1921 some cases like this - I found 11), and the matching function had to
1922 back up over a mixture of the two cases, it incorrectly assumed they
1923 were both the same length.
1924
1925 (b) When PCRE was configured to use the heap rather than the stack for
1926 recursion during matching, it was not correctly preserving the data for
1927 the other case of a UTF-8 character when checking ahead for a match
1928 while processing a minimizing repeat. If the check also involved
1929 matching a wide character, but failed, corruption could cause an
1930 erroneous result when trying to check for a repeat of the original
1931 character.
1932
1933 13. Some tidying changes to the testing mechanism:
1934
1935 (a) The RunTest script now detects the internal link size and whether there
1936 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1937 values substituted by "configure". (The RunGrepTest script already did
1938 this for UTF-8.) The configure.ac script no longer substitutes the
1939 relevant variables.
1940
1941 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1942 with length and offset values. This means that the output is different
1943 for different internal link sizes. Test 2 is skipped for link sizes
1944 other than 2 because of this, bypassing the problem. Unfortunately,
1945 there was also a test in test 3 (the locale tests) that used /B and
1946 failed for link sizes other than 2. Rather than cut the whole test out,
1947 I have added a new /Z option to pcretest that replaces the length and
1948 offset values with spaces. This is now used to make test 3 independent
1949 of link size. (Test 2 will be tidied up later.)
1950
1951 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
1952 segmentation fault instead of returning the appropriate error message.
1953
1954 15. In multiline mode when the newline sequence was set to "any", the pattern
1955 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
1956 This doesn't seem right; it now treats the CRLF combination as the line
1957 ending, and so does not match in that case. It's only a pattern such as ^$
1958 that would hit this one: something like ^ABC$ would have failed after \r
1959 and then tried again after \r\n.
1960
1961 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
1962 in an attempt to make files that differ only in their line terminators
1963 compare equal. This works on Linux.
1964
1965 17. Under certain error circumstances pcregrep might try to free random memory
1966 as it exited. This is now fixed, thanks to valgrind.
1967
1968 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
1969 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
1970 was because its rules for how to advance for /g after matching an empty
1971 string at the end of a line did not allow for this case. They now check for
1972 it specially.
1973
1974 20. pcretest is supposed to handle patterns and data of any length, by
1975 extending its buffers when necessary. It was getting this wrong when the
1976 buffer for a data line had to be extended.
1977
1978 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
1979 CRLF as a newline sequence.
1980
1981 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
1982 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
1983 I have nevertheless tidied it up.
1984
1985 23. Added some casts to kill warnings from HP-UX ia64 compiler.
1986
1987 24. Added a man page for pcre-config.
1988
1989
1990 Version 7.0 19-Dec-06
1991 ---------------------
1992
1993 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
1994 moving to gcc 4.1.1.
1995
1996 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
1997 sys/time.h, which is documented as needed for this function. It doesn't
1998 seem to matter on Linux, but it showed up on some releases of OS X.
1999
2000 3. It seems that there are systems where bytes whose values are greater than
2001 127 match isprint() in the "C" locale. The "C" locale should be the
2002 default when a C program starts up. In most systems, only ASCII printing
2003 characters match isprint(). This difference caused the output from pcretest
2004 to vary, making some of the tests fail. I have changed pcretest so that:
2005
2006 (a) When it is outputting text in the compiled version of a pattern, bytes
2007 other than 32-126 are always shown as hex escapes.
2008
2009 (b) When it is outputting text that is a matched part of a subject string,
2010 it does the same, unless a different locale has been set for the match
2011 (using the /L modifier). In this case, it uses isprint() to decide.
2012
2013 4. Fixed a major bug that caused incorrect computation of the amount of memory
2014 required for a compiled pattern when options that changed within the
2015 pattern affected the logic of the preliminary scan that determines the
2016 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2017 that the computed length was too small. The symptoms of this bug were
2018 either the PCRE error "internal error: code overflow" from pcre_compile(),
2019 or a glibc crash with a message such as "pcretest: free(): invalid next
2020 size (fast)". Examples of patterns that provoked this bug (shown in
2021 pcretest format) are:
2022
2023 /(?-x: )/x
2024 /(?x)(?-x: \s*#\s*)/
2025 /((?i)[\x{c0}])/8
2026 /(?i:[\x{c0}])/8
2027
2028 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2029 is now done differently.
2030
2031 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2032 wrapper classes; (b) implement a new function in the C++ scanner that is
2033 more efficient than the old way of doing things because it avoids levels of
2034 recursion in the regex matching; (c) add a paragraph to the documentation
2035 for the FullMatch() function.
2036
2037 6. The escape sequence \n was being treated as whatever was defined as
2038 "newline". Not only was this contrary to the documentation, which states
2039 that \n is character 10 (hex 0A), but it also went horribly wrong when
2040 "newline" was defined as CRLF. This has been fixed.
2041
2042 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2043 was being set to -1 for the "end of line" case (supposedly a value that no
2044 character can have). Though this value is never used (the check for end of
2045 line is "zero bytes in current character"), it caused compiler complaints.
2046 I've changed it to 0xffffffff.
2047
2048 8. In pcre_version.c, the version string was being built by a sequence of
2049 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2050 string (as it is for production releases) called a macro with an empty
2051 argument. The C standard says the result of this is undefined. The gcc
2052 compiler treats it as an empty string (which was what was wanted) but it is
2053 reported that Visual C gives an error. The source has been hacked around to
2054 avoid this problem.
2055
2056 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2057 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2058 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2059 of them did).
2060
2061 10. Originally, pcretest opened its input and output without "b"; then I was
2062 told that "b" was needed in some environments, so it was added for release
2063 5.0 to both the input and output. (It makes no difference on Unix-like
2064 systems.) Later I was told that it is wrong for the input on Windows. I've
2065 now abstracted the modes into two macros, to make it easier to fiddle with
2066 them, and removed "b" from the input mode under Windows.
2067
2068 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2069
2070 12. Added -help and --help to pcretest as an official way of being reminded
2071 of the options.
2072
2073 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2074 and pcrecpp.cc because they annoy compilers at high warning levels.
2075
2076 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2077
2078 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2079 scripts are not C programs :-) and which was not noticed because it works
2080 on Linux.
2081
2082 16. pcretest is supposed to handle any length of pattern and data line (as one
2083 line or as a continued sequence of lines) by extending its input buffer if
2084 necessary. This feature was broken for very long pattern lines, leading to
2085 a string of junk being passed to pcre_compile() if the pattern was longer
2086 than about 50K.
2087
2088 17. I have done a major re-factoring of the way pcre_compile() computes the
2089 amount of memory needed for a compiled pattern. Previously, there was code
2090 that made a preliminary scan of the pattern in order to do this. That was
2091 OK when PCRE was new, but as the facilities have expanded, it has become
2092 harder and harder to keep it in step with the real compile phase, and there
2093 have been a number of bugs (see for example, 4 above). I have now found a
2094 cunning way of running the real compile function in a "fake" mode that
2095 enables it to compute how much memory it would need, while actually only
2096 ever using a few hundred bytes of working memory and without too many
2097 tests of the mode. This should make future maintenance and development
2098 easier. A side effect of this work is that the limit of 200 on the nesting
2099 depth of parentheses has been removed (though this was never a serious
2100 limitation, I suspect). However, there is a downside: pcre_compile() now
2101 runs more slowly than before (30% or more, depending on the pattern). I
2102 hope this isn't a big issue. There is no effect on runtime performance.
2103
2104 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2105 newline (only possible for the last line of a file) and it was a
2106 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2107
2108 19. Added additional timing features to pcretest. (1) The -tm option now times
2109 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2110 separate command line item, by a number that specifies the number of
2111 repeats to use when timing. The default is 50000; this gives better
2112 precision, but takes uncomfortably long for very large patterns.
2113
2114 20. Extended pcre_study() to be more clever in cases where a branch of a
2115 subpattern has no definite first character. For example, (a*|b*)[cd] would
2116 previously give no result from pcre_study(). Now it recognizes that the
2117 first character must be a, b, c, or d.
2118
2119 21. There was an incorrect error "recursive call could loop indefinitely" if
2120 a subpattern (or the entire pattern) that was being tested for matching an
2121 empty string contained only one non-empty item after a nested subpattern.
2122 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2123 incorrectly, because the \d was being skipped in the check.
2124
2125 22. The pcretest program now has a new pattern option /B and a command line
2126 option -b, which is equivalent to adding /B to every pattern. This causes
2127 it to show the compiled bytecode, without the additional information that
2128 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2129 is the same as /B/I).
2130
2131 23. A new optimization is now able automatically to treat some sequences such
2132 as a*b as a*+b. More specifically, if something simple (such as a character
2133 or a simple class like \d) has an unlimited quantifier, and is followed by
2134 something that cannot possibly match the quantified thing, the quantifier
2135 is automatically "possessified".
2136
2137 24. A recursive reference to a subpattern whose number was greater than 39
2138 went wrong under certain circumstances in UTF-8 mode. This bug could also
2139 have affected the operation of pcre_study().
2140
2141 25. Realized that a little bit of performance could be had by replacing
2142 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2143
2144 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2145
2146 27. Possessive quantifiers such as a++ were previously implemented by turning
2147 them into atomic groups such as ($>a+). Now they have their own opcodes,
2148 which improves performance. This includes the automatically created ones
2149 from 23 above.
2150
2151 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2152 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2153 the first matched character to be a colon. This applied both to named and
2154 numbered groups.
2155
2156 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2157
2158 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2159 building PCRE on Windows easier, so I have included it in the distribution.
2160
2161 31. There is now a check in pcretest against a ridiculously large number being
2162 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2163 loop, the loop is abandoned.
2164
2165 32. Forward references to subpatterns in conditions such as (?(2)...) where
2166 subpattern 2 is defined later cause pcre_compile() to search forwards in
2167 the pattern for the relevant set of parentheses. This search went wrong
2168 when there were unescaped parentheses in a character class, parentheses
2169 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2170
2171 33. "Subroutine" calls and backreferences were previously restricted to
2172 referencing subpatterns earlier in the regex. This restriction has now
2173 been removed.
2174
2175 34. Added a number of extra features that are going to be in Perl 5.10. On the
2176 whole, these are just syntactic alternatives for features that PCRE had
2177 previously implemented using the Python syntax or my own invention. The
2178 other formats are all retained for compatibility.
2179
2180 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2181 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2182 also .NET compatible.
2183
2184 (b) A recursion or subroutine call to a named group can now be defined as
2185 (?&name) as well as (?P>name).
2186
2187 (c) A backreference to a named group can now be defined as \k<name> or
2188 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2189 5.10, are also .NET compatible.
2190
2191 (d) A conditional reference to a named group can now use the syntax
2192 (?(<name>) or (?('name') as well as (?(name).
2193
2194 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2195 groups (named and numbered) that are never evaluated inline, but can be
2196 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2197 is always false. There may be only one alternative in such a group.
2198
2199 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2200 as the simple (?(R). The condition is true only if the most recent
2201 recursion is that of the given number or name. It does not search out
2202 through the entire recursion stack.
2203
2204 (g) The escape \gN or \g{N} has been added, where N is a positive or
2205 negative number, specifying an absolute or relative reference.
2206
2207 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2208 some "unreachable code" warnings.
2209
2210 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2211 things, this adds five new scripts.
2212
2213 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2214 There were also incompatibilities regarding the handling of \Q..\E inside
2215 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2216 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2217
2218 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2219 matches an empty string, and forcibly breaks the loop. There were bugs in
2220 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2221 against aaaa the result was just "a" rather than "aaaa", for example. Two
2222 separate and independent bugs (that affected different cases) have been
2223 fixed.
2224
2225 39. Refactored the code to abolish the use of different opcodes for small
2226 capturing bracket numbers. This is a tidy that I avoided doing when I
2227 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2228 The new approach is not only tidier, it makes it possible to reduce the
2229 memory needed to fix the previous bug (38).
2230
2231 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2232 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2233 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2234 mode.
2235
2236 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2237 report.
2238
2239 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2240 copy construction and assignment in the C++ wrapper.
2241
2242 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2243 couple of bugs that could have given wrong results in the "--newline=crlf"
2244 case.
2245
2246 44. Added a number of casts and did some reorganization of signed/unsigned int
2247 variables following suggestions from Dair Grant. Also renamed the variable
2248 "this" as "item" because it is a C++ keyword.
2249
2250 45. Arranged for dftables to add
2251
2252 #include "pcre_internal.h"
2253
2254 to pcre_chartables.c because without it, gcc 4.x may remove the array
2255 definition from the final binary if PCRE is built into a static library and
2256 dead code stripping is activated.
2257
2258 46. For an unanchored pattern, if a match attempt fails at the start of a
2259 newline sequence, and the newline setting is CRLF or ANY, and the next two
2260 characters are CRLF, advance by two characters instead of one.
2261
2262
2263 Version 6.7 04-Jul-06
2264 ---------------------
2265
2266 1. In order to handle tests when input lines are enormously long, pcretest has
2267 been re-factored so that it automatically extends its buffers when
2268 necessary. The code is crude, but this _is_ just a test program. The
2269 default size has been increased from 32K to 50K.
2270
2271 2. The code in pcre_study() was using the value of the re argument before
2272 testing it for NULL. (Of course, in any sensible call of the function, it
2273 won't be NULL.)
2274
2275 3. The memmove() emulation function in pcre_internal.h, which is used on
2276 systems that lack both memmove() and bcopy() - that is, hardly ever -
2277 was missing a "static" storage class specifier.
2278
2279 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2280 containing an extended class (one that cannot be represented by a bitmap
2281 because it contains high-valued characters or Unicode property items, e.g.
2282 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2283 pattern, but PCRE should not loop if you do not (it no longer does).
2284 [Detail: two cases were found: (a) a repeated subpattern containing an
2285 extended class; (b) a recursive reference to a subpattern that followed a
2286 previous extended class. It wasn't skipping over the extended class
2287 correctly when UTF-8 mode was not set.]
2288
2289 5. A negated single-character class was not being recognized as fixed-length
2290 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2291 compile error "lookbehind assertion is not fixed length".
2292
2293 6. The RunPerlTest auxiliary script was showing an unexpected difference
2294 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2295 write a Perl script that can interpret lines of an input file either as
2296 byte characters or as UTF-8, which is what "perltest" was being required to
2297 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2298 can't do is switch easily at run time between having the "use utf8;" pragma
2299 or not. In the end, I fudged it by using the RunPerlTest script to insert
2300 "use utf8;" explicitly for the UTF-8 tests.
2301
2302 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2303 the end of the subject string, contrary to the documentation and to what
2304 Perl does. This was true of both matching functions. Now it matches only at
2305 the start of the subject and immediately after *internal* newlines.
2306
2307 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2308 a pointer to an int instead of a pointer to an unsigned long int. This
2309 caused problems on 64-bit systems.
2310
2311 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2312 instance of the 'standard' template library not being so standard".
2313
2314 10. There was no check on the number of named subpatterns nor the maximum
2315 length of a subpattern name. The product of these values is used to compute
2316 the size of the memory block for a compiled pattern. By supplying a very
2317 long subpattern name and a large number of named subpatterns, the size
2318 computation could be caused to overflow. This is now prevented by limiting
2319 the length of names to 32 characters, and the number of named subpatterns
2320 to 10,000.
2321
2322 11. Subpatterns that are repeated with specific counts have to be replicated in
2323 the compiled pattern. The size of memory for this was computed from the
2324 length of the subpattern and the repeat count. The latter is limited to
2325 65535, but there was no limit on the former, meaning that integer overflow
2326 could in principle occur. The compiled length of a repeated subpattern is
2327 now limited to 30,000 bytes in order to prevent this.
2328
2329 12. Added the optional facility to have named substrings with the same name.
2330
2331 13. Added the ability to use a named substring as a condition, using the
2332 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2333 are numbers (not recommended). Forward references are permitted.
2334
2335 14. Added forward references in named backreferences (if you see what I mean).
2336
2337 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2338 pattern could run off the end of the subject. For example, the pattern
2339 "(?s)(.{1,5})"8 did this with the subject "ab".
2340
2341 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2342 PCRE_CASELESS was set when matching characters that were quantified with ?
2343 or *.
2344
2345 17. A character class other than a single negated character that had a minimum
2346 but no maximum quantifier - for example [ab]{6,} - was not handled
2347 correctly by pce_dfa_exec(). It would match only one character.
2348
2349 18. A valid (though odd) pattern that looked like a POSIX character
2350 class but used an invalid character after [ (for example [[,abc,]]) caused
2351 pcre_compile() to give the error "Failed: internal error: code overflow" or
2352 in some cases to crash with a glibc free() error. This could even happen if
2353 the pattern terminated after [[ but there just happened to be a sequence of
2354 letters, a binary zero, and a closing ] in the memory that followed.
2355
2356 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2357 over the years. Originally (before any Unicode support), just the bottom 8
2358 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2359 output from "man perlunicode" includes this:
2360
2361 The regular expression compiler produces polymorphic opcodes. That
2362 is, the pattern adapts to the data and automatically switches to
2363 the Unicode character scheme when presented with Unicode data--or
2364 instead uses a traditional byte scheme when presented with byte
2365 data.
2366
2367 Sadly, a wide octal escape does not cause a switch, and in a string with
2368 no other multibyte characters, these octal escapes are treated as before.
2369 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2370 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2371 Unicode string.
2372
2373 I have not perpetrated such confusion in PCRE. Up till now, it took just
2374 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2375 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2376 translate to the appropriate multibyte character.
2377
2378 29. Applied some refactoring to reduce the number of warnings from Microsoft
2379 and Borland compilers. This has included removing the fudge introduced
2380 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2381 a warning about an unused variable.
2382
2383 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2384 characters since release 4.0, because Perl (from release 5.004) does not.
2385 [Or at least, is documented not to: some releases seem to be in conflict
2386 with the documentation.] However, when a pattern was studied with
2387 pcre_study() and all its branches started with \s, PCRE still included VT
2388 as a possible starting character. Of course, this did no harm; it just
2389 caused an unnecessary match attempt.
2390
2391 22. Removed a now-redundant internal flag bit that recorded the fact that case
2392 dependency changed within the pattern. This was once needed for "required
2393 byte" processing, but is no longer used. This recovers a now-scarce options
2394 bit. Also moved the least significant internal flag bit to the most-
2395 significant bit of the word, which was not previously used (hangover from
2396 the days when it was an int rather than a uint) to free up another bit for
2397 the future.
2398
2399 23. Added support for CRLF line endings as well as CR and LF. As well as the
2400 default being selectable at build time, it can now be changed at runtime
2401 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2402 specify that it is scanning data with non-default line endings.
2403
2404 24. Changed the definition of CXXLINK to make it agree with the definition of
2405 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2406
2407 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2408 recursions. This makes a big different to stack usage for some patterns.
2409
2410 26. If a subpattern containing a named recursion or subroutine reference such
2411 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2412 the space required for the compiled pattern went wrong and gave too small a
2413 value. Depending on the environment, this could lead to "Failed: internal
2414 error: code overflow at offset 49" or "glibc detected double free or
2415 corruption" errors.
2416
2417 27. Applied patches from Google (a) to support the new newline modes and (b) to
2418 advance over multibyte UTF-8 characters in GlobalReplace.
2419
2420 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2421 difference for some implementation of PCRE in some Windows version.
2422
2423 29. Added some extra testing facilities to pcretest:
2424
2425 \q<number> in a data line sets the "match limit" value
2426 \Q<number> in a data line sets the "match recursion limt" value
2427 -S <number> sets the stack size, where <number> is in megabytes
2428
2429 The -S option isn't available for Windows.
2430
2431
2432 Version 6.6 06-Feb-06
2433 ---------------------
2434
2435 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2436 in pcreposix.h. I have copied the definition from pcre.h.
2437
2438 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2439 because pcre.h is no longer a built file.
2440
2441 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2442 not normally included in the compiled code.
2443
2444
2445 Version 6.5 01-Feb-06
2446 ---------------------
2447
2448 1. When using the partial match feature with pcre_dfa_exec(), it was not
2449 anchoring the second and subsequent partial matches at the new starting
2450 point. This could lead to incorrect results. For example, with the pattern
2451 /1234/, partially matching against "123" and then "a4" gave a match.
2452
2453 2. Changes to pcregrep:
2454
2455 (a) All non-match returns from pcre_exec() were being treated as failures
2456 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2457 error message is output. Some extra information is given for the
2458 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2459 probably the only errors that are likely to be caused by users (by
2460 specifying a regex that has nested indefinite repeats, for instance).
2461 If there are more than 20 of these errors, pcregrep is abandoned.
2462
2463 (b) A binary zero was treated as data while matching, but terminated the
2464 output line if it was written out. This has been fixed: binary zeroes
2465 are now no different to any other data bytes.
2466
2467 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2468 used to set a locale for matching. The --locale=xxxx long option has
2469 been added (no short equivalent) to specify a locale explicitly on the
2470 pcregrep command, overriding the environment variables.
2471
2472 (d) When -B was used with -n, some line numbers in the output were one less
2473 than they should have been.
2474
2475 (e) Added the -o (--only-matching) option.
2476
2477 (f) If -A or -C was used with -c (count only), some lines of context were
2478 accidentally printed for the final match.
2479
2480 (g) Added the -H (--with-filename) option.
2481
2482 (h) The combination of options -rh failed to suppress file names for files
2483 that were found from directory arguments.
2484
2485 (i) Added the -D (--devices) and -d (--directories) options.
2486
2487 (j) Added the -F (--fixed-strings) option.
2488
2489 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2490
2491 (l) Added the --colo(u)r option.
2492
2493 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2494 is not present by default.
2495
2496 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2497 items such as (?R) or (?1), when the recursion could match a number of
2498 alternatives. If it matched one of the alternatives, but subsequently,
2499 outside the recursion, there was a failure, the code tried to back up into
2500 the recursion. However, because of the way PCRE is implemented, this is not
2501 possible, and the result was an incorrect result from the match.
2502
2503 In order to prevent this happening, the specification of recursion has
2504 been changed so that all such subpatterns are automatically treated as
2505 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2506
2507 4. I had overlooked the fact that, in some locales, there are characters for
2508 which isalpha() is true but neither isupper() nor islower() are true. In
2509 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2510 and ordfeminine) are like this. This affected the treatment of \w and \W
2511 when they appeared in character classes, but not when they appeared outside
2512 a character class. The bit map for "word" characters is now created
2513 separately from the results of isalnum() instead of just taking it from the
2514 upper, lower, and digit maps. (Plus the underscore character, of course.)
2515
2516 5. The above bug also affected the handling of POSIX character classes such as
2517 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2518 permanent tables. Instead, the bit maps for such a class were previously
2519 created as the appropriate unions of the upper, lower, and digit bitmaps.
2520 Now they are created by subtraction from the [[:word:]] class, which has
2521 its own bitmap.
2522
2523 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2524 It is created by subtracting the vertical space characters (\x09, \x0a,
2525 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2526 subtraction was done in the overall bitmap for a character class, meaning
2527 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2528 be recognized. This bug has been fixed.
2529
2530 7. Patches from the folks at Google:
2531
2532 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2533 real life, but is still worth protecting against".
2534
2535 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2536 regular expressions".
2537
2538 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2539 have it.
2540
2541 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2542 "configure" and the latter not, in order to fix a problem somebody had
2543 with compiling the Arg class on HP-UX.
2544
2545 (e) Improve the error-handling of the C++ wrapper a little bit.
2546
2547 (f) New tests for checking recursion limiting.
2548
2549 8. The pcre_memmove() function, which is used only if the environment does not
2550 have a standard memmove() function (and is therefore rarely compiled),
2551 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2552 returning a result (though PCRE never actually uses the result).
2553
2554 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2555 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2556 returned instead of calling malloc() with an overflowing number that would
2557 most likely cause subsequent chaos.
2558
2559 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2560
2561 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2562 with this option is matched, the nmatch and pmatch options of regexec() are
2563 ignored.
2564
2565 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2566 provided in case anyone wants to the the POSIX interface with UTF-8
2567 strings.
2568
2569 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2570 C++ linking (needed for some HP-UX environments).
2571
2572 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2573 (unused parameter) and in the pcre_printint() function (omitted "default"
2574 switch label when the default is to do nothing).
2575
2576 15. Added some code to make it possible, when PCRE is compiled as a C++
2577 library, to replace subject pointers for pcre_exec() with a smart pointer
2578 class, thus making it possible to process discontinuous strings.
2579
2580 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2581 much the same function. They were added by different people who were trying
2582 to make PCRE easy to compile on non-Unix systems. It has been suggested
2583 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2584 for compiling on Windows systems. I have therefore replaced it with
2585 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2586 defaults to "extern" for C or "extern C" for C++, which works fine on
2587 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2588 SCOPE with something explicit in config.h. In addition:
2589
2590 (a) pcreposix.h still had just "extern" instead of either of these macros;
2591 I have replaced it with PCRE_DATA_SCOPE.
2592
2593 (b) Functions such as _pcre_xclass(), which are internal to the library,
2594 but external in the C sense, all had PCRE_EXPORT in their definitions.
2595 This is apparently wrong for the Windows case, so I have removed it.
2596 (It makes no difference on Unix-like systems.)
2597
2598 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2599 of recursive calls to match(). This is different to MATCH_LIMIT because
2600 that limits the total number of calls to match(), not all of which increase
2601 the depth of recursion. Limiting the recursion depth limits the amount of
2602 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2603 when PCRE is compiled, and changed at run time. A patch from Google adds
2604 this functionality to the C++ interface.
2605
2606 18. Changes to the handling of Unicode character properties:
2607
2608 (a) Updated the table to Unicode 4.1.0.
2609
2610 (b) Recognize characters that are not in the table as "Cn" (undefined).
2611
2612 (c) I revised the way the table is implemented to a much improved format
2613 which includes recognition of ranges. It now supports the ranges that
2614 are defined in UnicodeData.txt, and it also amalgamates other
2615 characters into ranges. This has reduced the number of entries in the
2616 table from around 16,000 to around 3,000, thus reducing its size
2617 considerably. I realized I did not need to use a tree structure after
2618 all - a binary chop search is just as efficient. Having reduced the
2619 number of entries, I extended their size from 6 bytes to 8 bytes to
2620 allow for more data.
2621
2622 (d) Added support for Unicode script names via properties such as \p{Han}.
2623
2624 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2625 matching that character.
2626
2627 20. When matching a repeated Unicode property with a minimum greater than zero,
2628 (for example \pL{2,}), PCRE could look past the end of the subject if it
2629 reached it while seeking the minimum number of characters. This could
2630 happen only if some of the characters were more than one byte long, because
2631 there is a check for at least the minimum number of bytes.
2632
2633 21. Refactored the implementation of \p and \P so as to be more general, to
2634 allow for more different types of property in future. This has changed the
2635 compiled form incompatibly. Anybody with saved compiled patterns that use
2636 \p or \P will have to recompile them.
2637
2638 22. Added "Any" and "L&" to the supported property types.
2639
2640 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2641 but give a compile time error if the value is greater than 0xff.
2642
2643 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2644 accidentally not being installed or uninstalled.
2645
2646 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2647 made were to insert the current release number. This seemed silly, because
2648 it made things harder for people building PCRE on systems that don't run
2649 "configure". I have turned pcre.h into a distributed file, no longer built
2650 by "configure", with the version identification directly included. There is
2651 no longer a pcre.h.in file.
2652
2653 However, this change necessitated a change to the pcre-config script as
2654 well. It is built from pcre-config.in, and one of the substitutions was the
2655 release number. I have updated configure.ac so that ./configure now finds
2656 the release number by grepping pcre.h.
2657
2658 26. Added the ability to run the tests under valgrind.
2659
2660
2661 Version 6.4 05-Sep-05
2662 ---------------------
2663
2664 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2665 "--" to be printed when multiple files were scanned, even when none of the
2666 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2667 consider it to be a bug, and have restored the previous behaviour.
2668
2669 2. A couple of code tidies to get rid of compiler warnings.
2670
2671 3. The pcretest program used to cheat by referring to symbols in the library
2672 whose names begin with _pcre_. These are internal symbols that are not
2673 really supposed to be visible externally, and in some environments it is
2674 possible to suppress them. The cheating is now confined to including
2675 certain files from the library's source, which is a bit cleaner.
2676
2677 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2678 file's purpose clearer.
2679
2680 5. Reorganized pcre_ucp_findchar().
2681
2682
2683 Version 6.3 15-Aug-05
2684 ---------------------
2685
2686 1. The file libpcre.pc.in did not have general read permission in the tarball.
2687
2688 2. There were some problems when building without C++ support:
2689
2690 (a) If C++ support was not built, "make install" and "make test" still
2691 tried to test it.
2692
2693 (b) There were problems when the value of CXX was explicitly set. Some
2694 changes have been made to try to fix these, and ...
2695
2696 (c) --disable-cpp can now be used to explicitly disable C++ support.
2697
2698 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2699 backslash in a target when C++ was disabled. This confuses some
2700 versions of "make", apparently. Using an intermediate variable solves
2701 this. (Same for CPP_LOBJ.)
2702
2703 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2704 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2705 necessary on certain architectures.
2706
2707 4. Added a setting of -export-symbols-regex to the link command to remove
2708 those symbols that are exported in the C sense, but actually are local
2709 within the library, and not documented. Their names all begin with
2710 "_pcre_". This is not a perfect job, because (a) we have to except some
2711 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2712 available (and never for static libraries). I have made a note to try to
2713 find a way round (a) in the future.
2714
2715
2716 Version 6.2 01-Aug-05
2717 ---------------------
2718
2719 1. There was no test for integer overflow of quantifier values. A construction
2720 such as {1111111111111111} would give undefined results. What is worse, if
2721 a minimum quantifier for a parenthesized subpattern overflowed and became
2722 negative, the calculation of the memory size went wrong. This could have
2723 led to memory overwriting.
2724
2725 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2726
2727 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2728 operating environments where this matters.
2729
2730 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2731 PCRE options from within the C++ wrapper.
2732
2733 5. Named capturing subpatterns were not being correctly counted when a pattern
2734 was compiled. This caused two problems: (a) If there were more than 100
2735 such subpatterns, the calculation of the memory needed for the whole
2736 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2737 back references of the form \12, where the number was greater than 9, were
2738 not recognized as back references, even though there were sufficient
2739 previous subpatterns.
2740
2741 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2742 versions of gcc, e.g. 2.95.4.
2743
2744
2745 Version 6.1 21-Jun-05
2746 ---------------------
2747
2748 1. There was one reference to the variable "posix" in pcretest.c that was not
2749 surrounded by "#if !defined NOPOSIX".
2750
2751 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2752 the cross-check on the old pcre_info() function, for the benefit of the
2753 cut-down version of PCRE that is currently imported into Exim.
2754
2755 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2756 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2757 patterns that start (?i) but I don't think that matters. The use of (?i) is
2758 just an example; this all applies to the other options as well.
2759
2760 4. Since libtool seems to echo the compile commands it is issuing, the output
2761 from "make" can be reduced a bit by putting "@" in front of each libtool
2762 compile command.
2763
2764 5. Patch from the folks at Google for configure.in to be a bit more thorough
2765 in checking for a suitable C++ installation before trying to compile the
2766 C++ stuff. This should fix a reported problem when a compiler was present,
2767 but no suitable headers.
2768
2769 6. The man pages all had just "PCRE" as their title. I have changed them to
2770 be the relevant file name. I have also arranged that these names are
2771 retained in the file doc/pcre.txt, which is a concatenation in text format
2772 of all the man pages except the little individual ones for each function.
2773
2774 7. The NON-UNIX-USE file had not been updated for the different set of source
2775 files that come with release 6. I also added a few comments about the C++
2776 wrapper.
2777
2778
2779 Version 6.0 07-Jun-05
2780 ---------------------
2781
2782 1. Some minor internal re-organization to help with my DFA experiments.
2783
2784 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2785 didn't matter for the library itself when fully configured, but did matter
2786 when compiling without UCP support, or within Exim, where the ucp files are
2787 not imported.
2788
2789 3. Refactoring of the library code to split up the various functions into
2790 different source modules. The addition of the new DFA matching code (see
2791 below) to a single monolithic source would have made it really too
2792 unwieldy, quite apart from causing all the code to be include in a
2793 statically linked application, when only some functions are used. This is
2794 relevant even without the DFA addition now that patterns can be compiled in
2795 one application and matched in another.
2796
2797 The downside of splitting up is that there have to be some external
2798 functions and data tables that are used internally in different modules of
2799 the library but which are not part of the API. These have all had their
2800 names changed to start with "_pcre_" so that they are unlikely to clash
2801 with other external names.
2802
2803 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2804 a different (DFA) algorithm. Although it is slower than the original
2805 function, it does have some advantages for certain types of matching
2806 problem.
2807
2808 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2809 including restarting after a partial match.
2810
2811 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2812 defined when compiling for Windows was sent to me. I have put it into the
2813 code, though I have no means of testing or verifying it.
2814
2815 7. Added the pcre_refcount() auxiliary function.
2816
2817 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2818 match before or at the first newline in the subject string. In pcretest,
2819 the /f option on a pattern can be used to set this.
2820
2821 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2822 would behave wrongly. This has been present in PCRE since release 4.0.
2823
2824 10. A number of changes to the pcregrep command:
2825
2826 (a) Refactored how -x works; insert ^(...)$ instead of setting
2827 PCRE_ANCHORED and checking the length, in preparation for adding
2828 something similar for -w.
2829
2830 (b) Added the -w (match as a word) option.
2831
2832 (c) Refactored the way lines are read and buffered so as to have more
2833 than one at a time available.
2834
2835 (d) Implemented a pcregrep test script.
2836
2837 (e) Added the -M (multiline match) option. This allows patterns to match
2838 over several lines of the subject. The buffering ensures that at least
2839 8K, or the rest of the document (whichever is the shorter) is available
2840 for matching (and similarly the previous 8K for lookbehind assertions).
2841
2842 (f) Changed the --help output so that it now says
2843
2844 -w, --word-regex(p)
2845
2846 instead of two lines, one with "regex" and the other with "regexp"
2847 because that confused at least one person since the short forms are the
2848 same. (This required a bit of code, as the output is generated
2849 automatically from a table. It wasn't just a text change.)
2850
2851 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2852 option but starts with a hyphen. Could be a pattern or a path name
2853 starting with a hyphen, for instance.
2854
2855 (h) "-" can be given as a file name to represent stdin.
2856
2857 (i) When file names are being printed, "(standard input)" is used for
2858 the standard input, for compatibility with GNU grep. Previously
2859 "<stdin>" was used.
2860
2861 (j) The option --label=xxx can be used to supply a name to be used for
2862 stdin when file names are being printed. There is no short form.
2863
2864 (k) Re-factored the options decoding logic because we are going to add
2865 two more options that take data. Such options can now be given in four
2866 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2867
2868 (l) Added the -A, -B, and -C options for requesting that lines of context
2869 around matches be printed.
2870
2871 (m) Added the -L option to print the names of files that do not contain
2872 any matching lines, that is, the complement of -l.
2873
2874 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2875 continue to scan other files.
2876
2877 (o) The -s option was incorrectly implemented. For compatibility with other
2878 greps, it now suppresses the error message for a non-existent or non-
2879 accessible file (but not the return code). There is a new option called
2880 -q that suppresses the output of matching lines, which was what -s was
2881 previously doing.
2882
2883 (p) Added --include and --exclude options to specify files for inclusion
2884 and exclusion when recursing.
2885
2886 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2887 Hopefully, it now does.
2888
2889 12. Missing cast in pcre_study().
2890
2891 13. Added an "uninstall" target to the makefile.
2892
2893 14. Replaced "extern" in the function prototypes in Makefile.in with
2894 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2895 world, but is set differently for Windows.
2896
2897 15. Added a second compiling function called pcre_compile2(). The only
2898 difference is that it has an extra argument, which is a pointer to an
2899 integer error code. When there is a compile-time failure, this is set
2900 non-zero, in addition to the error test pointer being set to point to an
2901 error message. The new argument may be NULL if no error number is required
2902 (but then you may as well call pcre_compile(), which is now just a
2903 wrapper). This facility is provided because some applications need a
2904 numeric error indication, but it has also enabled me to tidy up the way
2905 compile-time errors are handled in the POSIX wrapper.
2906
2907 16. Added VPATH=.libs to the makefile; this should help when building with one
2908 prefix path and installing with another. (Or so I'm told by someone who
2909 knows more about this stuff than I do.)
2910
2911 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2912 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2913 match everything, including newlines. This is not POSIX-compatible, but
2914 somebody wanted the feature. From pcretest it can be activated by using
2915 both the P and the s flags.
2916
2917 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2918
2919 19. libpcre.pc was being incorrectly installed as executable.
2920
2921 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2922 it now also looks for '\r' so that it will work unmodified on Windows.
2923
2924 21. Added Google's contributed C++ wrapper to the distribution.
2925
2926 22. Added some untidy missing memory free() calls in pcretest, to keep
2927 Electric Fence happy when testing.
2928
2929
2930
2931 Version 5.0 13-Sep-04
2932 ---------------------
2933
2934 1. Internal change: literal characters are no longer packed up into items
2935 containing multiple characters in a single byte-string. Each character
2936 is now matched using a separate opcode. However, there may be more than one
2937 byte in the character in UTF-8 mode.
2938
2939 2. The pcre_callout_block structure has two new fields: pattern_position and
2940 next_item_length. These contain the offset in the pattern to the next match
2941 item, and its length, respectively.
2942
2943 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2944 insertion of callouts before each pattern item. Added the /C option to
2945 pcretest to make use of this.
2946
2947 4. On the advice of a Windows user, the lines
2948
2949 #if defined(_WIN32) || defined(WIN32)
2950 _setmode( _fileno( stdout ), 0x8000 );
2951 #endif /* defined(_WIN32) || defined(WIN32) */
2952
2953 have been added to the source of pcretest. This apparently does useful
2954 magic in relation to line terminators.
2955
2956 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
2957 for the benefit of those environments where the "b" makes a difference.
2958
2959 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
2960 to know about it. I have put a hack into configure.in that adds in code
2961 to set GCC=yes if CC=icc. This seems to end up at a point in the
2962 generated configure script that is early enough to affect the setting of
2963 compiler options, which is what is needed, but I have no means of testing
2964 whether it really works. (The user who reported this had patched the
2965 generated configure script, which of course I cannot do.)
2966
2967 LATER: After change 22 below (new libtool files), the configure script
2968 seems to know about icc (and also ecc). Therefore, I have commented out
2969 this hack in configure.in.
2970
2971 7. Added support for pkg-config (2 patches were sent in).
2972
2973 8. Negated POSIX character classes that used a combination of internal tables
2974 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
2975 [[:^ascii]]. Typically, they would match almost any characters. The other
2976 POSIX classes were not broken in this way.
2977
2978 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
2979 to find the match, as PCRE was deluded into thinking that the match had to
2980 start at the start point or following a newline. The same bug applied to
2981 patterns with negative forward assertions or any backward assertions
2982 preceding ".*" at the start, unless the pattern required a fixed first
2983 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
2984
2985 10. In UTF-8 mode, when moving forwards in the subject after a failed match
2986 starting at the last subject character, bytes beyond the end of the subject
2987 string were read.
2988
2989 11. Renamed the variable "class" as "classbits" to make life easier for C++
2990 users. (Previously there was a macro definition, but it apparently wasn't
2991 enough.)
2992
2993 12. Added the new field "tables" to the extra data so that tables can be passed
2994 in at exec time, or the internal tables can be re-selected. This allows
2995 a compiled regex to be saved and re-used at a later time by a different
2996 program that might have everything at different addresses.
2997
2998 13. Modified the pcre-config script so that, when run on Solaris, it shows a
2999 -R library as well as a -L library.
3000
3001 14. The debugging options of pcretest (-d on the command line or D on a
3002 pattern) showed incorrect output for anything following an extended class
3003 that contained multibyte characters and which was followed by a quantifier.
3004
3005 15. Added optional support for general category Unicode character properties
3006 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3007 support. It adds about 90K to the size of the library. The meanings of the
3008 inbuilt class escapes such as \d and \s have NOT been changed.
3009
3010 16. Updated pcredemo.c to include calls to free() to release the memory for the
3011 compiled pattern.
3012
3013 17. The generated file chartables.c was being created in the source directory
3014 instead of in the building directory. This caused the build to fail if the
3015 source directory was different from the building directory, and was
3016 read-only.
3017
3018 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3019 file. No doubt somebody will tell me if they don't make sense... Also added
3020 Dan Mooney's comments about building on OpenVMS.
3021
3022 19. Added support for partial matching via the PCRE_PARTIAL option for
3023 pcre_exec() and the \P data escape in pcretest.
3024
3025 20. Extended pcretest with 3 new pattern features:
3026
3027 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3028 write the compiled pattern to the file whose name is "rest-of-line".
3029 This is a straight binary dump of the data, with the saved pointer to
3030 the character tables forced to be NULL. The study data, if any, is
3031 written too. After writing, pcretest reads a new pattern.
3032
3033 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3034 compiled pattern from the given file. There must not be any
3035 occurrences of "<" in the file name (pretty unlikely); if there are,
3036 pcretest will instead treat the initial "<" as a pattern delimiter.
3037 After reading in the pattern, pcretest goes on to read data lines as
3038 usual.
3039
3040 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3041 and 16-bit fields in a compiled pattern, to simulate a pattern that
3042 was compiled on a host of opposite endianness.
3043
3044 21. The pcre-exec() function can now cope with patterns that were compiled on
3045 hosts of opposite endianness, with this restriction:
3046
3047 As for any compiled expression that is saved and used later, the tables
3048 pointer field cannot be preserved; the extra_data field in the arguments
3049 to pcre_exec() should be used to pass in a tables address if a value
3050 other than the default internal tables were used at compile time.
3051
3052 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3053 now diagnosed as an error. Previously, most of the time, a negative number
3054 would have been treated as zero, but if in addition "ovector" was passed as
3055 NULL, a crash could occur.
3056
3057 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3058 new versions from the libtool 1.5 distribution (the last one is a copy of
3059 a file called libtool.m4). This seems to have fixed the need to patch
3060 "configure" to support Darwin 1.3 (which I used to do). However, I still
3061 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3062 workstation).
3063
3064 24. Changed the PCRE licence to be the more standard "BSD" licence.
3065
3066
3067 Version 4.5 01-Dec-03
3068 ---------------------
3069
3070 1. There has been some re-arrangement of the code for the match() function so
3071 that it can be compiled in a version that does not call itself recursively.
3072 Instead, it keeps those local variables that need separate instances for
3073 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3074 needs to "recurse". Keeping track of where control must go is done by means
3075 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3076 hide most of the details from the main code, and operates only if
3077 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3078 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3079 operating.
3080
3081 To make it easier for callers to provide specially tailored get/free
3082 functions for this usage, two new functions, pcre_stack_malloc, and
3083 pcre_stack_free, are used. They are always called in strict stacking order,
3084 and the size of block requested is always the same.
3085
3086 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3087 PCRE has been compiled to use the stack or the heap for recursion. The
3088 -C option of pcretest uses this to show which version is compiled.
3089
3090 A new data escape \S, is added to pcretest; it causes the amounts of store
3091 obtained and freed by both kinds of malloc/free at match time to be added
3092 to the output.
3093
3094 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3095 what's available on my current Linux desktop machine.
3096
3097 3. When matching a UTF-8 string, the test for a valid string at the start has
3098 been extended. If start_offset is not zero, PCRE now checks that it points
3099 to a byte that is the start of a UTF-8 character. If not, it returns
3100 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3101 this is necessary because there may be backward assertions in the pattern.
3102 When matching the same subject several times, it may save resources to use
3103 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3104
3105 4. The code for checking the validity of UTF-8 strings has been tightened so
3106 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3107 containing "overlong sequences".
3108
3109 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3110 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3111 should have been "&", but it just so happened that all the cases this let
3112 through by mistake were picked up later in the function.
3113
3114 6. I had used a variable called "isblank" - this is a C99 function, causing
3115 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3116
3117 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3118 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3119 so the version is shown; (c) stop "make test" echoing "./RunTest".
3120
3121 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3122
3123 9. The prototype for memmove() for systems that don't have it was using
3124 size_t, but the inclusion of the header that defines size_t was later. I've
3125 moved the #includes for the C headers earlier to avoid this.
3126
3127 10. Added some adjustments to the code to make it easier to compiler on certain
3128 special systems:
3129
3130 (a) Some "const" qualifiers were missing.
3131 (b) Added the macro EXPORT before all exported functions; by default this
3132 is defined to be empty.
3133 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3134 that it reads its output file name as an argument instead of writing
3135 to the standard output and assuming this can be redirected.
3136
3137 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3138 class containing characters with values greater than 255, PCRE compilation
3139 went into a loop.
3140
3141 12. A recursive reference to a subpattern that was within another subpattern
3142 that had a minimum quantifier of zero caused PCRE to crash. For example,
3143 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3144 recursion. If the recursively-called subpattern itself had a zero repeat,
3145 that was OK.
3146
3147 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3148 buffer into which it was copied (for escape processing) was still set at
3149 1024, so long lines caused crashes.
3150
3151 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3152 "internal error: code overflow...". This applied to any character class
3153 that was followed by a possessive quantifier.
3154
3155 15. Modified the Makefile to add libpcre.la as a prerequisite for
3156 libpcreposix.la because I was told this is needed for a parallel build to
3157 work.
3158
3159 16. If a pattern that contained .* following optional items at the start was
3160 studied, the wrong optimizing data was generated, leading to matching
3161 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3162 matching string must start with a or b or c. The correct conclusion for
3163 this pattern is that a match can start with any character.
3164
3165
3166 Version 4.4 13-Aug-03
3167 ---------------------
3168
3169 1. In UTF-8 mode, a character class containing characters with values between
3170 127 and 255 was not handled correctly if the compiled pattern was studied.
3171 In fixing this, I have also improved the studying algorithm for such
3172 classes (slightly).
3173
3174 2. Three internal functions had redundant arguments passed to them. Removal
3175 might give a very teeny performance improvement.
3176
3177 3. Documentation bug: the value of the capture_top field in a callout is *one
3178 more than* the number of the hightest numbered captured substring.
3179
3180 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3181 in incorrectly linking with a previously installed version. They now link
3182 explicitly with libpcre.la.
3183
3184 5. configure.in no longer needs to recognize Cygwin specially.
3185
3186 6. A problem in pcre.in for Windows platforms is fixed.
3187
3188 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3189 pcretest, it used to include the size of the study block as part of its
3190 output. Unfortunately, the structure contains a field that has a different
3191 size on different hardware architectures. This meant that the tests that
3192 showed this size failed. As the block is currently always of a fixed size,
3193 this information isn't actually particularly useful in pcretest output, so
3194 I have just removed it.
3195
3196 8. Three pre-processor statements accidentally did not start in column 1.
3197 Sadly, there are *still* compilers around that complain, even though
3198 standard C has not required this for well over a decade. Sigh.
3199
3200 9. In pcretest, the code for checking callouts passed small integers in the
3201 callout_data field, which is a void * field. However, some picky compilers
3202 complained about the casts involved for this on 64-bit systems. Now
3203 pcretest passes the address of the small integer instead, which should get
3204 rid of the warnings.
3205
3206 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3207 both compile and run time, and gives an error if an invalid UTF-8 sequence
3208 is found. There is a option for disabling this check in cases where the
3209 string is known to be correct and/or the maximum performance is wanted.
3210
3211 11. In response to a bug report, I changed one line in Makefile.in from
3212
3213 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3214 to
3215 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3216
3217 to look similar to other lines, but I have no way of telling whether this
3218 is the right thing to do, as I do not use Windows. No doubt I'll get told
3219 if it's wrong...
3220
3221
3222 Version 4.3 21-May-03
3223 ---------------------
3224
3225 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3226 Makefile.
3227
3228 2. Some refactoring to improve the quality of the code:
3229
3230 (i) The utf8_table... variables are now declared "const".
3231
3232 (ii) The code for \cx, which used the "case flipping" table to upper case
3233 lower case letters, now just substracts 32. This is ASCII-specific,
3234 but the whole concept of \cx is ASCII-specific, so it seems
3235 reasonable.
3236
3237 (iii) PCRE was using its character types table to recognize decimal and
3238 hexadecimal digits in the pattern. This is silly, because it handles
3239 only 0-9, a-f, and A-F, but the character types table is locale-
3240 specific, which means strange things might happen. A private
3241 table is now used for this - though it costs 256 bytes, a table is
3242 much faster than multiple explicit tests. Of course, the standard
3243 character types table is still used for matching digits in subject
3244 strings against \d.
3245
3246 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3247 ending in _t are). So I've renamed it as ESC_tee.
3248
3249 3. The first argument for regexec() in the POSIX wrapper should have been
3250 defined as "const".
3251
3252 4. Changed pcretest to use malloc() for its buffers so that they can be
3253 Electric Fenced for debugging.
3254
3255 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3256 to read one or more bytes before the start of the subject string. Often this
3257 had no effect on PCRE's behaviour, but in some circumstances it could
3258 provoke a segmentation fault.
3259
3260 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3261 to try to read one or more bytes before the start of the subject string.
3262
3263 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3264 UTF-8 support could misbehave in various ways if the subject string
3265 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3266 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3267 back over UTF-8 characters.)
3268
3269
3270 Version 4.2 14-Apr-03
3271 ---------------------
3272
3273 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3274
3275 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3276 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3277 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3278 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3279 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3280 and BUILD_EXEEXT
3281 Note: automatic setting of the BUILD variables is not yet working
3282 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3283 compile-time but not at link-time
3284 [LINK]: use for linking executables only
3285 make different versions for Windows and non-Windows
3286 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3287 libraries
3288 [LINK_FOR_BUILD]: new variable
3289 [OBJEXT]: use throughout
3290 [EXEEXT]: use throughout
3291 <winshared>: new target
3292 <wininstall>: new target
3293 <dftables.o>: use native compiler
3294 <dftables>: use native linker
3295 <install>: handle Windows platform correctly
3296 <clean>: ditto
3297 <check>: ditto
3298 copy DLL to top builddir before testing
3299
3300 As part of these changes, -no-undefined was removed again. This was reported
3301 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3302 in any case.
3303
3304 3. Some tidies to get rid of compiler warnings:
3305
3306 . In the match_data structure, match_limit was an unsigned long int, whereas
3307 match_call_count was an int. I've made them both unsigned long ints.
3308
3309 . In pcretest the fact that a const uschar * doesn't automatically cast to
3310 a void * provoked a warning.
3311
3312 . Turning on some more compiler warnings threw up some "shadow" variables
3313 and a few more missing casts.
3314
3315 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3316 option, a class that contained a single character with a value between 128
3317 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3318
3319 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3320 option, a class that contained several characters, but with at least one
3321 whose value was between 128 and 255 caused PCRE to crash.
3322
3323
3324 Version 4.1 12-Mar-03
3325 ---------------------
3326
3327 1. Compiling with gcc -pedantic found a couple of places where casts were
3328 needed, and a string in dftables.c that was longer than standard compilers are
3329 required to support.
3330
3331 2. Compiling with Sun's compiler found a few more places where the code could
3332 be tidied up in order to avoid warnings.
3333
3334 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3335 first of these names is deprecated in the latest Autoconf in favour of the name
3336 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3337 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3338 analogy I have changed it to CFLAGS_FOR_BUILD.
3339
3340 4. Added -no-undefined to the linking command in the Makefile, because this is
3341 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3342 linking step for the pcreposix library.
3343
3344 5. PCRE was failing to diagnose the case of two named groups with the same
3345 name.
3346
3347 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3348 literal character that is needed in the subject for a match, and scans along to
3349 ensure that it is present before embarking on the full matching process. This
3350 saves time in cases of nested unlimited repeats that are never going to match.
3351 Problem: the scan can take a lot of time if the subject is very long (e.g.
3352 megabytes), thus penalizing straightforward matches. It is now done only if the
3353 amount of subject to be scanned is less than 1000 bytes.
3354
3355 7. A lesser problem with the same optimization is that it was recording the
3356 first character of an anchored pattern as "needed", thus provoking a search
3357 right along the subject, even when the first match of the pattern was going to
3358 fail. The "needed" character is now not set for anchored patterns, unless it
3359 follows something in the pattern that is of non-fixed length. Thus, it still
3360 fulfils its original purpose of finding quick non-matches in cases of nested
3361 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3362
3363
3364 Version 4.0 17-Feb-03
3365 ---------------------
3366
3367 1. If a comment in an extended regex that started immediately after a meta-item
3368 extended to the end of string, PCRE compiled incorrect data. This could lead to
3369 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3370
3371 2. Moved to autoconf 2.53 and libtool 1.4.2.
3372
3373 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3374 the special perltest8 script is no longer needed - all the tests can be run
3375 from a single perltest script.
3376
3377 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3378 by \s. It has now been removed in PCRE. This means it isn't recognized as
3379 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3380 class [:space:] *does* include VT, thereby creating a mess.
3381
3382 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3383 space and tab.
3384
3385 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3386 its new features into the main test script, reducing the number of scripts.
3387
3388 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3389 were backward compatible, and made the (?i) apply to the whole pattern, as if
3390 /i were given. Now it behaves more logically, and applies the option setting
3391 only to what follows. PCRE has been changed to follow suit. However, if it
3392 finds options settings right at the start of the pattern, it extracts them into
3393 the global options, as before. Thus, they show up in the info data.
3394
3395 8. Added support for the \Q...\E escape sequence. Characters in between are
3396 treated as literals. This is slightly different from Perl in that $ and @ are
3397 also handled as literals inside the quotes. In Perl, they will cause variable
3398 interpolation. Note the following examples:
3399
3400 Pattern PCRE matches Perl matches
3401
3402 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3403 \Qabc\$xyz\E abc\$xyz abc\$xyz
3404 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3405
3406 For compatibility with Perl, \Q...\E sequences are recognized inside character
3407 classes as well as outside them.
3408
3409 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3410 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3411 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3412 signed/unsigned warnings.
3413
3414 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3415 option for pcretest, so I've replaced it by a simple function that does just
3416 that job.
3417
3418 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3419 "pcregrep -".
3420
3421 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3422 Java package. This provides some syntactic sugar for simple cases of what my
3423 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3424 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3425 item, you can use this simplified notation. Note that only makes sense with
3426 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3427 greediness, whatever the setting of the PCRE_UNGREEDY option.
3428
3429 13. A change of greediness default within a pattern was not taking effect at
3430 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3431 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3432 was abstracted outside.
3433
3434 14. PCRE now supports the \G assertion. It is true when the current matching
3435 position is at the start point of the match. This differs from \A when the
3436 starting offset is non-zero. Used with the /g option of pcretest (or similar
3437 code), it works in the same way as it does for Perl's /g option. If all
3438 alternatives of a regex begin with \G, the expression is anchored to the start
3439 match position, and the "anchored" flag is set in the compiled expression.
3440
3441 15. Some bugs concerning the handling of certain option changes within patterns
3442 have been fixed. These applied to options other than (?ims). For example,
3443 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3444 been the other way round. Some of this was related to change 7 above.
3445
3446 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3447 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3448 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3449 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3450
3451 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3452 mode. Unlike ".", it always matches newline, whatever the setting of
3453 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3454 assertions. Perl allows it, but it doesn't (in general) work because it can't
3455 calculate the length of the lookbehind. At least, that's the case for Perl
3456 5.8.0 - I've been told they are going to document that it doesn't work in
3457 future.
3458
3459 18. Added an error diagnosis for escapes that PCRE does not support: these are
3460 \L, \l, \N, \P, \p, \U, \u, and \X.
3461
3462 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3463 reading past the end of the pattern in cases such as /[abcd/.
3464
3465 20. PCRE was getting more memory than necessary for patterns with classes that
3466 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3467
3468 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3469 compiling PCRE for use with Virtual Pascal.
3470
3471 22. Small fix to the Makefile to make it work properly if the build is done
3472 outside the source tree.
3473
3474 23. Added a new extension: a condition to go with recursion. If a conditional
3475 subpattern starts with (?(R) the "true" branch is used if recursion has
3476 happened, whereas the "false" branch is used only at the top level.
3477
3478 24. When there was a very long string of literal characters (over 255 bytes
3479 without UTF support, over 250 bytes with UTF support), the computation of how
3480 much memory was required could be incorrect, leading to segfaults or other
3481 strange effects.
3482
3483 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3484 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3485 there was a subsequent back reference to those brackets. This meant that, for
3486 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3487 possible to check for precisely this case. All we can do is abandon the
3488 optimization if .* occurs inside capturing brackets when there are any back
3489 references whatsoever. (See below for a better fix that came later.)
3490
3491 26. The handling of the optimization for finding the first character of a
3492 non-anchored pattern, and for finding a character that is required later in the
3493 match were failing in some cases. This didn't break the matching; it just
3494 failed to optimize when it could. The way this is done has been re-implemented.
3495
3496 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3497
3498 28. Added a new feature that provides some of the functionality that Perl
3499 provides with (?{...}). The facility is termed a "callout". The way it is done
3500 in PCRE is for the caller to provide an optional function, by setting
3501 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3502 global variable. By default it is unset, which disables all calling out. To get
3503 the function called, the regex must include (?C) at appropriate points. This
3504 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3505 This provides a means of identifying different callout points. When PCRE
3506 reaches such a point in the regex, if pcre_callout has been set, the external
3507 function is called. It is provided with data in a structure called
3508 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3509 matching continues; if it returns a non-zero value, the match at the current
3510 point fails. However, backtracking will occur if possible. [This was changed
3511 later and other features added - see item 49 below.]
3512
3513 29. pcretest is upgraded to test the callout functionality. It provides a
3514 callout function that displays information. By default, it shows the start of
3515 the match and the current position in the text. There are some new data escapes
3516 to vary what happens:
3517
3518 \C+ in addition, show current contents of captured substrings
3519 \C- do not supply a callout function
3520 \C!n return 1 when callout number n is reached
3521 \C!n!m return 1 when callout number n is reached for the mth time
3522
3523 30. If pcregrep was called with the -l option and just a single file name, it
3524 output "<stdin>" if a match was found, instead of the file name.
3525
3526 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3527 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3528 pcre_exec(). This saves a malloc/free per call. The default value of
3529 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3530 when configuring.
3531
3532 32. The default maximum size of a compiled pattern is 64K. There have been a
3533 few cases of people hitting this limit. The code now uses macros to handle the
3534 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3535 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3536 configuring. Tests 2 and 5 work only with 2-byte links because they output
3537 debugging information about compiled patterns.
3538
3539 33. Internal code re-arrangements:
3540
3541 (a) Moved the debugging function for printing out a compiled regex into
3542 its own source file (printint.c) and used #include to pull it into
3543 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3544 separate copies.
3545
3546 (b) Defined the list of op-code names for debugging as a macro in
3547 internal.h so that it is next to the definition of the opcodes.
3548
3549 (c) Defined a table of op-code lengths for simpler skipping along compiled
3550 code. This is again a macro in internal.h so that it is next to the
3551 definition of the opcodes.
3552
3553 34. Added support for recursive calls to individual subpatterns, along the
3554 lines of Robin Houston's patch (but implemented somewhat differently).
3555
3556 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3557 allow it to read and process whole directories in Win32. This code was
3558 contributed by Lionel Fourquaux; it has not been tested by me.
3559
3560 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3561 used to name a group. Names consist of alphanumerics and underscores, and must
3562 be unique. Back references use the syntax (?P=name) and recursive calls use
3563 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3564 numbers. The function pcre_fullinfo() can be used after compilation to extract
3565 a name/number map. There are three relevant calls:
3566
3567 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3568 PCRE_INFO_NAMECOUNT yields the number of entries
3569 PCRE_INFO_NAMETABLE yields a pointer to the map.
3570
3571 The map is a vector of fixed-size entries. The size of each entry depends on
3572 the length of the longest name used. The first two bytes of each entry are the
3573 group number, most significant byte first. There follows the corresponding
3574 name, zero terminated. The names are in alphabetical order.
3575
3576 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3577 case instead of 255. Making it the same both with and without UTF-8 support
3578 means that the same test output works with both.
3579
3580 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3581 calling malloc() with a zero argument.
3582
3583 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3584 optimization. I've improved things by keeping a bitmap of backreferences with
3585 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3586 fact referenced, the optimization can be applied. It is unlikely that a
3587 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3588 the match to follow \n) will appear inside brackets with a number greater than
3589 31, but if it does, any back reference > 31 suppresses the optimization.
3590
3591 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3592 of disabling numbered capturing parentheses. Any opening parenthesis that is
3593 not followed by ? behaves as if it were followed by ?: but named parentheses
3594 can still be used for capturing (and they will acquire numbers in the usual
3595 way).
3596
3597 41. Redesigned the return codes from the match() function into yes/no/error so
3598 that errors can be passed back from deep inside the nested calls. A malloc
3599 failure while inside a recursive subpattern call now causes the
3600 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3601
3602 42. It is now possible to set a limit on the number of times the match()
3603 function is called in a call to pcre_exec(). This facility makes it possible to
3604 limit the amount of recursion and backtracking, though not in a directly
3605 obvious way, because the match() function is used in a number of different
3606 circumstances. The count starts from zero for each position in the subject
3607 string (for non-anchored patterns). The default limit is, for compatibility, a
3608 large number, namely 10 000 000. You can change this in two ways:
3609
3610 (a) When configuring PCRE before making, you can use --with-match-limit=n
3611 to set a default value for the compiled library.
3612
3613 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3614 a different value is set. See 45 below.
3615
3616 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3617
3618 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3619 of things that can be changed at compile time. The first argument specifies
3620 what is wanted and the second points to where the information is to be placed.
3621 The current list of available information is:
3622
3623 PCRE_CONFIG_UTF8
3624
3625 The output is an integer that is set to one if UTF-8 support is available;
3626 otherwise it is set to zero.
3627
3628 PCRE_CONFIG_NEWLINE
3629
3630 The output is an integer that it set to the value of the code that is used for
3631 newline. It is either LF (10) or CR (13).
3632
3633 PCRE_CONFIG_LINK_SIZE
3634
3635 The output is an integer that contains the number of bytes used for internal
3636 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3637
3638 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3639
3640 The output is an integer that contains the threshold above which the POSIX
3641 interface uses malloc() for output vectors. See item 31 above.
3642
3643 PCRE_CONFIG_MATCH_LIMIT
3644
3645 The output is an unsigned integer that contains the default limit of the number
3646 of match() calls in a pcre_exec() execution. See 42 above.
3647
3648 44. pcretest has been upgraded by the addition of the -C option. This causes it
3649 to extract all the available output from the new pcre_config() function, and to
3650 output it. The program then exits immediately.
3651
3652 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3653 order to support additional features. One way would have been to define
3654 pcre_exec2() (for example) with extra arguments, but this would not have been
3655 extensible, and would also have required all calls to the original function to
3656 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3657 is used for passing in "extra" data from pcre_study().
3658
3659 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3660 contains the following fields:
3661
3662 flags a bitmap indicating which of the following fields are set
3663 study_data opaque data from pcre_study()
3664 match_limit a way of specifying a limit on match() calls for a specific
3665 call to pcre_exec()
3666 callout_data data for callouts (see 49 below)
3667
3668 The flag bits are also defined in pcre.h, and are
3669
3670 PCRE_EXTRA_STUDY_DATA
3671 PCRE_EXTRA_MATCH_LIMIT
3672 PCRE_EXTRA_CALLOUT_DATA
3673
3674 The pcre_study() function now returns one of these new pcre_extra blocks, with
3675 the actual study data pointed to by the study_data field, and the
3676 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3677 before. That is, this change is entirely upwards-compatible and requires no
3678 change to existing code.
3679
3680 If you want to pass in additional data to pcre_exec(), you can either place it
3681 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3682 block.
3683
3684 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3685 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3686 times with different match limits, until it finds the minimum value needed for
3687 pcre_exec() to complete. The value is then output. This can be instructive; for
3688 most simple matches the number is quite small, but for pathological cases it
3689 gets very large very quickly.
3690
3691 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3692 returns the size of the data block pointed to by the study_data field in a
3693 pcre_extra block, that is, the value that was passed as the argument to
3694 pcre_malloc() when PCRE was getting memory in which to place the information
3695 created by pcre_study(). The fourth argument should point to a size_t variable.
3696 pcretest has been extended so that this information is shown after a successful
3697 pcre_study() call when information about the compiled regex is being displayed.
3698
3699 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3700 because what follows is always an absolute path. (Later: it turns out that this
3701 is more than cosmetic for MinGW, because it doesn't like empty path
3702 components.)
3703
3704 49. Some changes have been made to the callout feature (see 28 above):
3705
3706 (i) A callout function now has three choices for what it returns:
3707
3708 0 => success, carry on matching
3709 > 0 => failure at this point, but backtrack if possible
3710 < 0 => serious error, return this value from pcre_exec()
3711
3712 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3713 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3714 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3715 use by callout functions. It will never be used by PCRE itself.
3716
3717 (ii) The pcre_extra structure (see 45 above) has a void * field called
3718 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3719 pcre_callout_block structure has a field of the same name. The contents of
3720 the field passed in the pcre_extra structure are passed to the callout
3721 function in the corresponding field in the callout block. This makes it
3722 easier to use the same callout-containing regex from multiple threads. For
3723 testing, the pcretest program has a new data escape
3724
3725 \C*n pass the number n (may be negative) as callout_data
3726
3727 If the callout function in pcretest receives a non-zero value as
3728 callout_data, it returns that value.
3729
3730 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3731 there were some redundant $(CFLAGS) in commands that are now specified as
3732 $(LINK), which already includes $(CFLAGS).
3733
3734 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3735 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3736 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3737 one-byte characters throughout. Note that case-insensitive matching applies
3738 only to characters whose values are less than 256. PCRE doesn't support the
3739 notion of cases for higher-valued characters.
3740
3741 (i) A character class whose characters are all within 0-255 is handled as
3742 a bit map, and the map is inverted for negative classes. Previously, a
3743 character > 255 always failed to match such a class; however it should
3744 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3745
3746 (ii) A negated character class with a single character < 255 is coded as
3747 "not this character" (OP_NOT). This wasn't working properly when the test
3748 character was multibyte, either singly or repeated.
3749
3750 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3751 mode, for example: \x{100}{2,3}.
3752
3753 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3754 singly or repeated) now correctly test multibyte characters. However,
3755 PCRE doesn't recognize any characters with values greater than 255 as
3756 digits, spaces, or word characters. Such characters always match \D, \S,
3757 and \W, and never match \d, \s, or \w.
3758
3759 (v) Classes may now contain characters and character ranges with values
3760 greater than 255. For example: [ab\x{100}-\x{400}].
3761
3762 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3763 PCRE in UTF-8 mode.
3764
3765 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3766 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3767 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3768 value.)
3769
3770 53. The single man page has become too large. I have therefore split it up into
3771 a number of separate man pages. These also give rise to individual HTML pages;
3772 these are now put in a separate directory, and there is an index.html page that
3773 lists them all. Some hyperlinking between the pages has been installed.
3774
3775 54. Added convenience functions for handling named capturing parentheses.
3776
3777 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3778 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3779 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3780 are faulted.
3781
3782 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3783 calling configure. These values are used when compiling the dftables.c program
3784 which is run to generate the source of the default character tables. They
3785 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3786 you will need to set these values.
3787
3788 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3789
3790
3791 Version 3.9 02-Jan-02
3792 ---------------------
3793
3794 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3795
3796 2. If --disable-static was given, the building process failed when trying to
3797 build pcretest and pcregrep. (For some reason it was using libtool to compile
3798 them, which is not right, as they aren't part of the library.)
3799
3800
3801 Version 3.8 18-Dec-01
3802 ---------------------
3803
3804 1. The experimental UTF-8 code was completely screwed up. It was packing the
3805 bytes in the wrong order. How dumb can you get?
3806
3807
3808 Version 3.7 29-Oct-01
3809 ---------------------
3810
3811 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3812 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3813 this didn't happen under Solaris 8, where I normally test things.
3814
3815 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3816 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3817 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3818 this fix an hour or so after the initial 3.7 release.)
3819
3820
3821 Version 3.6 23-Oct-01
3822 ---------------------
3823
3824 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3825 offsets passed as NULL with zero offset count.
3826
3827 2. The config.guess and config.sub files had not been updated when I moved to
3828 the latest autoconf.
3829
3830
3831 Version 3.5 15-Aug-01
3832 ---------------------
3833
3834 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3835 had been forgotten.
3836
3837 2. By using declared but undefined structures, we can avoid using "void"
3838 definitions in pcre.h while keeping the internal definitions of the structures
3839 private.
3840
3841 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3842 user point of view, this means that both static and shared libraries are built
3843 by default, but this can be individually controlled. More of the work of
3844 handling this static/shared cases is now inside libtool instead of PCRE's make
3845 file.
3846
3847 4. The pcretest utility is now installed along with pcregrep because it is
3848 useful for users (to test regexs) and by doing this, it automatically gets
3849 relinked by libtool. The documentation has been turned into a man page, so
3850 there are now .1, .txt, and .html versions in /doc.
3851
3852 5. Upgrades to pcregrep:
3853 (i) Added long-form option names like gnu grep.
3854 (ii) Added --help to list all options with an explanatory phrase.
3855 (iii) Added -r, --recursive to recurse into sub-directories.
3856 (iv) Added -f, --file to read patterns from a file.
3857
3858 6. pcre_exec() was referring to its "code" argument before testing that
3859 argument for NULL (and giving an error if it was NULL).
3860
3861 7. Upgraded Makefile.in to allow for compiling in a different directory from
3862 the source directory.
3863
3864 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3865 options bits, the pointer it was passed was to an int instead of to an unsigned
3866 long int. This mattered only on 64-bit systems.
3867
3868 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3869 generated) instead of pcre.in, which it its source. Also made the same change
3870 in several of the .c files.
3871
3872 10. A new release of gcc defines printf() as a macro, which broke pcretest
3873 because it had an ifdef in the middle of a string argument for printf(). Fixed
3874 by using separate calls to printf().
3875
3876 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3877 script, to force use of CR or LF instead of \n in the source. On non-Unix
3878 systems, the value can be set in config.h.
3879
3880 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3881 absolute limit. Changed the text of the error message to make this clear, and
3882 likewise updated the man page.
3883
3884 13. The limit of 99 on the number of capturing subpatterns has been removed.
3885 The new limit is 65535, which I hope will not be a "real" limit.
3886
3887
3888 Version 3.4 22-Aug-00
3889 ---------------------
3890
3891 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3892
3893 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3894
3895
3896 Version 3.3 01-Aug-00
3897 ---------------------
3898
3899 1. If an octal character was given, but the value was greater than \377, it
3900 was not getting masked to the least significant bits, as documented. This could
3901 lead to crashes in some systems.
3902
3903 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3904 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3905
3906 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3907 These just pass their arguments on to (pcre_free)(), but they are provided
3908 because some uses of PCRE bind it to non-C systems that can call its functions,
3909 but cannot call free() or pcre_free() directly.
3910
3911 4. Add "make test" as a synonym for "make check". Corrected some comments in
3912 the Makefile.
3913
3914 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3915 Makefile.
3916
3917 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3918 command called pgrep for grepping around the active processes.
3919
3920 7. Added the beginnings of support for UTF-8 character strings.
3921
3922 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3923 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3924 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3925 out for the ar command.)
3926
3927
3928 Version 3.2 12-May-00
3929 ---------------------
3930
3931 This is purely a bug fixing release.
3932
3933 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3934 of ZA. This was just one example of several cases that could provoke this bug,
3935 which was introduced by change 9 of version 2.00. The code for breaking
3936 infinite loops after an iteration that matches an empty string was't working
3937 correctly.
3938
3939 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3940 when matched against abbab (for example). After matching an empty string, it
3941 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3942 caused it to match further down the string than it should.
3943
3944 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3945 was there because it doesn't seem to be needed, and it causes trouble on some
3946 systems, as it is not a Standard C header. It has been removed.
3947
3948 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3949 were reported on the Macintosh. The changes were from
3950
3951 while ((c = *(++ptr)) != 0 && c != '\n');
3952 to
3953 while ((c = *(++ptr)) != 0 && c != '\n') ;
3954
3955 Totally extraordinary, but if that's what it takes...
3956
3957 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
3958 available. Added HAVE_BCOPY and an autoconf test for it; if neither
3959 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
3960 assumes the way PCRE uses memmove() (always moving upwards).
3961
3962 6. PCRE is being used in one environment where strchr() is not available. There
3963 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
3964 faster code anyway.
3965
3966
3967 Version 3.1 09-Feb-00
3968 ---------------------
3969
3970 The only change in this release is the fixing of some bugs in Makefile.in for
3971 the "install" target:
3972
3973 (1) It was failing to install pcreposix.h.
3974
3975 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
3976
3977
3978 Version 3.0 01-Feb-00
3979 ---------------------
3980
3981 1. Add support for the /+ modifier to perltest (to output $` like it does in
3982 pcretest).
3983
3984 2. Add support for the /g modifier to perltest.
3985
3986 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
3987 matches null strings.
3988
3989 4. Fix perltest so that it doesn't do unwanted things when fed an empty
3990 pattern. Perl treats empty patterns specially - it reuses the most recent
3991 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
3992 effect.
3993
3994 5. The POSIX interface was broken in that it was just handing over the POSIX
3995 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
3996 required a bigger vector, with some working space on the end. This means that
3997 the POSIX wrapper now has to get and free some memory, and copy the results.
3998
3999 6. Added some simple autoconf support, placing the test data and the
4000 documentation in separate directories, re-organizing some of the
4001 information files, and making it build pcre-config (a GNU standard). Also added
4002 libtool support for building PCRE as a shared library, which is now the
4003 default.
4004
4005 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4006 09 are not valid octal constants. Single digits will be used for minor values
4007 less than 10.
4008
4009 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4010 existing programs that set these in the POSIX interface can use PCRE without
4011 modification.
4012
4013 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4014 return all that pcre_info() returns, plus additional data. The pcre_info()
4015 function is retained for compatibility, but is considered to be obsolete.
4016
4017 10. Added experimental recursion feature (?R) to handle one common case that
4018 Perl 5.6 will be able to do with (?p{...}).
4019
4020 11. Added support for POSIX character classes like [:alpha:], which Perl is
4021 adopting.
4022
4023
4024 Version 2.08 31-Aug-99
4025 ----------------------
4026
4027 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4028 trying to match at the startoffset position, but instead was moving forward to
4029 the next newline as if a previous match had failed.
4030
4031 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4032 and could get into a loop if a null string was matched other than at the start
4033 of the subject.
4034
4035 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4036 be distinguished at compile time, and for completeness also added PCRE_DATE.
4037
4038 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4039 in GnuWin32 environments.
4040
4041
4042 Version 2.07 29-Jul-99
4043 ----------------------
4044
4045 1. The documentation is now supplied in plain text form and HTML as well as in
4046 the form of man page sources.
4047
4048 2. C++ compilers don't like assigning (void *) values to other pointer types.
4049 In particular this affects malloc(). Although there is no problem in Standard
4050 C, I've put in casts to keep C++ compilers happy.
4051
4052 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4053 should be (const char *).
4054
4055 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4056 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4057 However, I haven't made this a standard facility. The documentation doesn't
4058 mention it, and the Makefile doesn't support it.
4059
4060 5. The Makefile now contains an "install" target, with editable destinations at
4061 the top of the file. The pcretest program is not installed.
4062
4063 6. pgrep -V now gives the PCRE version number and date.
4064
4065 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4066 causing the entire string to be ignored, instead of just the last character.
4067
4068 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4069 non-matching string, it can take a very, very long time, even for strings of
4070 quite modest length, because of the nested recursion. PCRE now does better in
4071 some of these cases. It does this by remembering the last required literal
4072 character in the pattern, and pre-searching the subject to ensure it is present
4073 before running the real match. In other words, it applies a heuristic to detect
4074 some types of certain failure quickly, and in the above example, if presented
4075 with a string that has no trailing " it gives "no match" very quickly.
4076
4077 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4078 other alternatives are tried instead.
4079
4080
4081 Version 2.06 09-Jun-99
4082 ----------------------
4083
4084 1. Change pcretest's output for amount of store used to show just the code
4085 space, because the remainder (the data block) varies in size between 32-bit and
4086 64-bit systems.
4087
4088 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4089 start matching at. This allows lookbehinds to work when searching for multiple
4090 occurrences in a string.
4091
4092 3. Added additional options to pcretest for testing multiple occurrences:
4093
4094 /+ outputs the rest of the string that follows a match
4095 /g loops for multiple occurrences, using the new startoffset argument
4096 /G loops for multiple occurrences by passing an incremented pointer
4097
4098 4. PCRE wasn't doing the "first character" optimization for patterns starting
4099 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4100 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4101 the letter 'x'. On long subject strings, this gives a significant speed-up.
4102
4103
4104 Version 2.05 21-Apr-99
4105 ----------------------
4106
4107 1. Changed the type of magic_number from int to long int so that it works
4108 properly on 16-bit systems.
4109
4110 2. Fixed a bug which caused patterns starting with .* not to work correctly
4111 when the subject string contained newline characters. PCRE was assuming
4112 anchoring for such patterns in all cases, which is not correct because .* will
4113 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4114 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4115 must be retried after every newline in the subject.
4116
4117
4118 Version 2.04 18-Feb-99
4119 ----------------------
4120
4121 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4122 computation of the store needed to hold the pattern was incorrect (too large).
4123 If such patterns were nested a few deep, this could multiply and become a real
4124 problem.
4125
4126 2. Added /M option to pcretest to show the memory requirement of a specific
4127 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4128
4129 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4130 compiled in such a way that the backtracking after subsequent failure was
4131 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4132 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4133
4134
4135 Version 2.03 02-Feb-99
4136 ----------------------
4137
4138 1. Fixed typo and small mistake in man page.
4139
4140 2. Added 4th condition (GPL supersedes if conflict) and created separate
4141 LICENCE file containing the conditions.
4142
4143 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4144 Perl, that is the internal \ allows the delimiter to be included in the
4145 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4146 the final delimiter, add \ to the end of the pattern (to test the error).
4147
4148 4. Added the convenience functions for extracting substrings after a successful
4149 match. Updated pcretest to make it able to test these functions.
4150
4151
4152 Version 2.02 14-Jan-99
4153 ----------------------
4154
4155 1. Initialized the working variables associated with each extraction so that
4156 their saving and restoring doesn't refer to uninitialized store.
4157
4158 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4159 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4160 fix the problem.
4161
4162 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4163 calls, and wasn't printing the correct value for compiling calls. Increased the
4164 default value of LOOPREPEAT, and the number of significant figures in the
4165 times.
4166
4167 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4168
4169 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4170 a building problem on Windows NT with a FAT file system.
4171
4172
4173 Version 2.01 21-Oct-98
4174 ----------------------
4175
4176 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4177 to character tables built by pcre_maketables() in the current locale. If NULL
4178 is passed, the default tables are used.
4179
4180
4181 Version 2.00 24-Sep-98
4182 ----------------------
4183
4184 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4185 it any more.
4186
4187 2. Allow quantification of (?>) groups, and make it work correctly.
4188
4189 3. The first character computation wasn't working for (?>) groups.
4190
4191 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4192 end of the subject) and add 5.005's \z, which really does match only at the
4193 very end of the subject.
4194
4195 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4196
4197 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4198 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4199 localized options. All options to pcre_study() were also removed.
4200
4201 7. Add other new features from 5.005:
4202
4203 $(?<= positive lookbehind
4204 $(?<! negative lookbehind
4205 (?imsx-imsx) added the unsetting capability
4206 such a setting is global if at outer level; local otherwise
4207 (?imsx-imsx:) non-capturing groups with option setting
4208 (?(cond)re|re) conditional pattern matching
4209
4210 A backreference to itself in a repeated group matches the previous
4211 captured string.
4212
4213 8. General tidying up of studying (both automatic and via "study")
4214 consequential on the addition of new assertions.
4215
4216 9. As in 5.005, unlimited repeated groups that could match an empty substring
4217 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4218 runtime if any iteration does actually match an empty substring.
4219
4220 10. Include the RunTest script in the distribution.
4221
4222 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4223 discrepancies, some of which were old and were also with respect to 5.004. They
4224 have now been fixed.
4225
4226
4227 Version 1.09 28-Apr-98
4228 ----------------------
4229
4230 1. A negated single character class followed by a quantifier with a minimum
4231 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4232 program crashes, or just wrong answers. This did not apply to negated classes
4233 containing more than one character, or to minima other than one.
4234
4235
4236 Version 1.08 27-Mar-98
4237 ----------------------
4238
4239 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4240
4241 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4242 latter must appear before anything that relies on it in the pattern.
4243
4244
4245 Version 1.07 16-Feb-98
4246 ----------------------
4247
4248 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4249 repeat of a potentially empty string).
4250
4251
4252 Version 1.06 23-Jan-98
4253 ----------------------
4254
4255 1. Added Markus Oberhumer's little patches for C++.
4256
4257 2. Literal strings longer than 255 characters were broken.
4258
4259
4260 Version 1.05 23-Dec-97
4261 ----------------------
4262
4263 1. Negated character classes containing more than one character were failing if
4264 PCRE_CASELESS was set at run time.
4265
4266
4267 Version 1.04 19-Dec-97
4268 ----------------------
4269
4270 1. Corrected the man page, where some "const" qualifiers had been omitted.
4271
4272 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4273 input syntax.
4274
4275 3. Fixed memory leak which occurred when a regex with back references was
4276 matched with an offsets vector that wasn't big enough. The temporary memory
4277 that is used in this case wasn't being freed if the match failed.
4278
4279 4. Tidied pcretest to ensure it frees memory that it gets.
4280
4281 5. Temporary memory was being obtained in the case where the passed offsets
4282 vector was exactly big enough.
4283
4284 6. Corrected definition of offsetof() from change 5 below.
4285
4286 7. I had screwed up change 6 below and broken the rules for the use of
4287 setjmp(). Now fixed.
4288
4289
4290 Version 1.03 18-Dec-97
4291 ----------------------
4292
4293 1. A erroneous regex with a missing opening parenthesis was correctly
4294 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4295 on some systems.
4296
4297 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4298 it was reported that one broken compiler failed on the former because "code" is
4299 also an independent variable.
4300
4301 3. The erroneous regex a[]b caused an array overrun reference.
4302
4303 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4304 fail on data ending with that character. (It was going on too far, and checking
4305 the next character, typically a binary zero.) This was specific to the
4306 optimized code for single-character negative classes.
4307
4308 5. Added a contributed patch from the TIN world which does the following:
4309
4310 + Add an undef for memmove, in case the the system defines a macro for it.
4311
4312 + Add a definition of offsetof(), in case there isn't one. (I don't know
4313 the reason behind this - offsetof() is part of the ANSI standard - but
4314 it does no harm).
4315
4316 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4317 most of the places where whitespace preceded '#'. I have given up and
4318 allowed the remaining 2 cases to be at the margin.
4319
4320 + Rename some variables in pcre to eliminate shadowing. This seems very
4321 pedantic, but does no harm, of course.
4322
4323 6. Moved the call to setjmp() into its own function, to get rid of warnings
4324 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4325
4326 7. Constructs such as \d{8,} were compiling into the equivalent of
4327 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4328 outcome, but in this particular case used more store than had been allocated,
4329 which caused the bug to be discovered because it threw up an internal error.
4330
4331 8. The debugging code in both pcre and pcretest for outputting the compiled
4332 form of a regex was going wrong in the case of back references followed by
4333 curly-bracketed repeats.
4334
4335
4336 Version 1.02 12-Dec-97
4337 ----------------------
4338
4339 1. Typos in pcre.3 and comments in the source fixed.
4340
4341 2. Applied a contributed patch to get rid of places where it used to remove
4342 'const' from variables, and fixed some signed/unsigned and uninitialized
4343 variable warnings.
4344
4345 3. Added the "runtest" target to Makefile.
4346
4347 4. Set default compiler flag to -O2 rather than just -O.
4348
4349
4350 Version 1.01 19-Nov-97
4351 ----------------------
4352
4353 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4354 like /([ab]*)*/, that is, for classes with more than one character in them.
4355
4356 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4357 as /((?>a*))*/ (a PCRE_EXTRA facility).
4358
4359
4360 Version 1.00 18-Nov-97
4361 ----------------------
4362
4363 1. Added compile-time macros to support systems such as SunOS4 which don't have
4364 memmove() or strerror() but have other things that can be used instead.
4365
4366 2. Arranged that "make clean" removes the executables.
4367
4368
4369 Version 0.99 27-Oct-97
4370 ----------------------
4371
4372 1. Fixed bug in code for optimizing classes with only one character. It was
4373 initializing a 32-byte map regardless, which could cause it to run off the end
4374 of the memory it had got.
4375
4376 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4377
4378
4379 Version 0.98 22-Oct-97
4380 ----------------------
4381
4382 1. Fixed bug in code for handling temporary memory usage when there are more
4383 back references than supplied space in the ovector. This could cause segfaults.
4384
4385
4386 Version 0.97 21-Oct-97
4387 ----------------------
4388
4389 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4390
4391 2. Optimized negated single characters not to use a bit map.
4392
4393 3. Brought error texts together as macro definitions; clarified some of them;
4394 fixed one that was wrong - it said "range out of order" when it meant "invalid
4395 escape sequence".
4396
4397 4. Changed some char * arguments to const char *.
4398
4399 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4400
4401 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4402 pcretest.
4403
4404
4405 Version 0.96 16-Oct-97
4406 ----------------------
4407
4408 1. Added a simple "pgrep" utility to the distribution.
4409
4410 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4411 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4412 where "ddd" means "one or more decimal digits".
4413
4414 3. Fixed serious bug. If a pattern had a back reference, but the call to
4415 pcre_exec() didn't supply a large enough ovector to record the related
4416 identifying subpattern, the match always failed. PCRE now remembers the number
4417 of the largest back reference, and gets some temporary memory in which to save
4418 the offsets during matching if necessary, in order to ensure that
4419 backreferences always work.
4420
4421 4. Increased the compatibility with Perl in a number of ways:
4422
4423 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4424 to request this handling. The option can be set at compile or exec time.
4425
4426 (b) $ matches before a terminating newline by default; an option
4427 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4428 mode). The option can be set at compile or exec time.
4429
4430 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4431 the same as Perl's. If the decimal number it represents is less than 10
4432 or there aren't that many previous left capturing parentheses, an octal
4433 escape is read. Inside a character class, it's always an octal escape,
4434 even if it is a single digit.
4435
4436 (d) An escaped but undefined alphabetic character is taken as a literal,
4437 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4438 escapes.
4439
4440 (e) {0} is now permitted. (The previous item is removed from the compiled
4441 pattern).
4442
4443 5. Changed all the names of code files so that the basic parts are no longer
4444 than 10 characters, and abolished the teeny "globals.c" file.
4445
4446 6. Changed the handling of character classes; they are now done with a 32-byte
4447 bit map always.
4448
4449 7. Added the -d and /D options to pcretest to make it possible to look at the
4450 internals of compilation without having to recompile pcre.
4451
4452
4453 Version 0.95 23-Sep-97
4454 ----------------------
4455
4456 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4457 \x20 at the start of a run of normal characters. These were being treated as
4458 real characters, instead of the source characters being re-checked.
4459
4460
4461 Version 0.94 18-Sep-97
4462 ----------------------
4463
4464 1. The functions are now thread-safe, with the caveat that the global variables
4465 containing pointers to malloc() and free() or alternative functions are the
4466 same for all threads.
4467
4468 2. Get pcre_study() to generate a bitmap of initial characters for non-
4469 anchored patterns when this is possible, and use it if passed to pcre_exec().
4470
4471
4472 Version 0.93 15-Sep-97
4473 ----------------------
4474
4475 1. /(b)|(:+)/ was computing an incorrect first character.
4476
4477 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4478 but not actually doing anything yet.
4479
4480 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4481 as Perl does (e.g. [-az] or [az-]).
4482
4483 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4484 all possible positions.
4485
4486 5. Split up into different modules to avoid including unneeded functions in a
4487 compiled binary. However, compile and exec are still in one module. The "study"
4488 function is split off.
4489
4490 6. The character tables are now in a separate module whose source is generated
4491 by an auxiliary program - but can then be edited by hand if required. There are
4492 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4493 toupper() in the code.
4494
4495 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4496 make them global. Abolish the function for setting them, as the caller can now
4497 set them directly.
4498
4499
4500 Version 0.92 11-Sep-97
4501 ----------------------
4502
4503 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4504 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4505
4506 2. Caseless matching was not working in character classes if the characters in
4507 the pattern were in upper case.
4508
4509 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4510
4511 4. Make PCRE_ANCHORED public and accept as a compile option.
4512
4513 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4514 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4515 pass them.
4516
4517 6. Give an error if bad option bits passed at compile or run time.
4518
4519 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4520 pcretest to cause it to pass that flag.
4521
4522 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4523 options, and the first character, if set.
4524
4525 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4526
4527
4528 Version 0.91 10-Sep-97
4529 ----------------------
4530
4531 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4532 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4533
4534 2. PCRE was looping on encountering an indefinitely repeated back reference to
4535 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4536 Perl does - treats the match as successful.
4537
4538 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5