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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 780 - (show annotations) (download)
Fri Dec 2 11:39:21 2011 UTC (2 years, 10 months ago) by ph10
File size: 222637 byte(s)
Preserve the size of JIT compiled code. Add PCRE_INFO_JITSIZE and use it for
the /M option in pcretest. 

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.21
5 ------------
6
7 1. Updating the JIT compiler.
8
9 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10 are added as well.
11
12 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14 calling _pcre_jit_exec. Some extra comments are added.
15
16 4. Mark settings inside atomic groups that do not contain any capturing
17 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18 was introduced by change 18 for 8.20.
19
20 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21 ECMA-262 standard.
22
23 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25 This bug was probably introduced by change 9 of 8.13.
26
27 7. While fixing 6 above, I noticed that a number of other items were being
28 incorrectly rejected as "not fixed length". This arose partly because newer
29 opcodes had not been added to the fixed-length checking code. I have (a)
30 corrected the bug and added tests for these items, and (b) arranged for an
31 error to occur if an unknown opcode is encountered while checking for fixed
32 length instead of just assuming "not fixed length". The items that were
33 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36
37 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38 being incorrectly compiled and would have given unpredicatble results.
39
40 9. A possessively repeated subpattern with minimum repeat count greater than
41 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
42 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
43 the first (A) could occur when it should not.
44
45 10. Add a cast and remove a redundant test from the code.
46
47 11. JIT should use pcre_malloc/pcre_free for allocation.
48
49 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
50 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
51 is anything other than /usr, -L is still shown).
52
53 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
54
55 14. Perl does not support \N without a following name in a [] class; PCRE now
56 also gives an error.
57
58 15. If a forward reference was repeated with an upper limit of around 2000,
59 it caused the error "internal error: overran compiling workspace". The
60 maximum number of forward references (including repeats) was limited by the
61 internal workspace, and dependent on the LINK_SIZE. The code has been
62 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
63 the default depends on LINK_SIZE. There is a new upper limit (for safety)
64 of around 200,000 forward references. While doing this, I also speeded up
65 the filling in of repeated forward references.
66
67 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
68 incorrectly expecting the subject to contain another "a" after the start.
69
70 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
71 in the match, the SKIP should be ignored. This was not happening; instead
72 the SKIP was being treated as NOMATCH. For patterns such as
73 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
74 tested.
75
76 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
77 now much more compatible with Perl, in particular in cases where the result
78 is a non-match for a non-anchored pattern. For example, if
79 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
80 "m", where previously it did not return a name. A side effect of this
81 change is that for partial matches, the last encountered mark name is
82 returned, as for non matches. A number of tests that were previously not
83 Perl-compatible have been moved into the Perl-compatible test files. The
84 refactoring has had the pleasing side effect of removing one argument from
85 the match() function, thus reducing its stack requirements.
86
87 19. If the /S+ option was used in pcretest to study a pattern using JIT,
88 subsequent uses of /S (without +) incorrectly behaved like /S+.
89
90 21. Retrieve executable code size support for the JIT compiler and fixing
91 some warnings.
92
93 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
94 not work when the shorter character appeared right at the end of the
95 subject string.
96
97 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
98 systems.
99
100 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
101 output it when the /M option is used in pcretest.
102
103
104 Version 8.20 21-Oct-2011
105 ------------------------
106
107 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
108 a POSIX class. After further experiments with Perl, which convinced me that
109 Perl has bugs and confusions, a closing square bracket is no longer allowed
110 in a POSIX name. This bug also affected patterns with classes that started
111 with full stops.
112
113 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
114 captured substring, but while checking the failing first alternative,
115 substring 1 is temporarily captured. If the output vector supplied to
116 pcre_exec() was not big enough for this capture, the yield of the function
117 was still zero ("insufficient space for captured substrings"). This cannot
118 be totally fixed without adding another stack variable, which seems a lot
119 of expense for a edge case. However, I have improved the situation in cases
120 such as /(a)(b)x|abc/ matched against "abc", where the return code
121 indicates that fewer than the maximum number of slots in the ovector have
122 been set.
123
124 3. Related to (2) above: when there are more back references in a pattern than
125 slots in the output vector, pcre_exec() uses temporary memory during
126 matching, and copies in the captures as far as possible afterwards. It was
127 using the entire output vector, but this conflicts with the specification
128 that only 2/3 is used for passing back captured substrings. Now it uses
129 only the first 2/3, for compatibility. This is, of course, another edge
130 case.
131
132 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
133 main code base, and can be used by building with --enable-jit. When this is
134 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
135 runtime --no-jit option is given.
136
137 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
138 ovector, the return from the function was zero, implying that there were
139 other matches that did not fit. The correct "exactly full" value is now
140 returned.
141
142 6. If a subpattern that was called recursively or as a subroutine contained
143 (*PRUNE) or any other control that caused it to give a non-standard return,
144 invalid errors such as "Error -26 (nested recursion at the same subject
145 position)" or even infinite loops could occur.
146
147 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
148 computing the minimum length on reaching *ACCEPT, and so ended up with the
149 wrong value of 1 rather than 0. Further investigation indicates that
150 computing a minimum subject length in the presence of *ACCEPT is difficult
151 (think back references, subroutine calls), and so I have changed the code
152 so that no minimum is registered for a pattern that contains *ACCEPT.
153
154 8. If (*THEN) was present in the first (true) branch of a conditional group,
155 it was not handled as intended. [But see 16 below.]
156
157 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
158 Sheri Pierce.
159
160 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
161 the first byte in a match must be "a".
162
163 11. Change 17 for 8.13 increased the recursion depth for patterns like
164 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
165 pattern contains any instances of (*THEN). If it does not, the old
166 optimizations are restored. It would be nice to do this on a per-group
167 basis, but at the moment that is not feasible.
168
169 12. In some environments, the output of pcretest -C is CRLF terminated. This
170 broke RunTest's code that checks for the link size. A single white space
171 character after the value is now allowed for.
172
173 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
174 For "fr", it uses the Windows-specific input and output files.
175
176 14. If (*THEN) appeared in a group that was called recursively or as a
177 subroutine, it did not work as intended. [But see next item.]
178
179 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
180 pattern fragments (but not containing any | characters). If A and B are
181 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
182 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
183 D. In other words, Perl considers parentheses that do not contain any |
184 characters to be part of a surrounding alternative, whereas PCRE was
185 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
186 differently. PCRE now behaves in the same way as Perl, except in the case
187 of subroutine/recursion calls such as (?1) which have in any case always
188 been different (but PCRE had them first :-).
189
190 16. Related to 15 above: Perl does not treat the | in a conditional group as
191 creating alternatives. Such a group is treated in the same way as an
192 ordinary group without any | characters when processing (*THEN). PCRE has
193 been changed to match Perl's behaviour.
194
195 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
196 RunGrepTest script failed.
197
198 18. Change 22 for version 13 caused atomic groups to use more stack. This is
199 inevitable for groups that contain captures, but it can lead to a lot of
200 stack use in large patterns. The old behaviour has been restored for atomic
201 groups that do not contain any capturing parentheses.
202
203 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
204 suppress the check for a minimum subject length at run time. (If it was
205 given to pcre_exec() or pcre_dfa_exec() it did work.)
206
207 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
208 fail to work when decoding hex characters in data strings in EBCDIC
209 environments.
210
211 21. It appears that in at least one Mac OS environment, the isxdigit() function
212 is implemented as a macro that evaluates to its argument more than once,
213 contravening the C 90 Standard (I haven't checked a later standard). There
214 was an instance in pcretest which caused it to go wrong when processing
215 \x{...} escapes in subject strings. The has been rewritten to avoid using
216 things like p++ in the argument of isxdigit().
217
218
219 Version 8.13 16-Aug-2011
220 ------------------------
221
222 1. The Unicode data tables have been updated to Unicode 6.0.0.
223
224 2. Two minor typos in pcre_internal.h have been fixed.
225
226 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
227 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
228 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
229
230 4. There were a number of related bugs in the code for matching backrefences
231 caselessly in UTF-8 mode when codes for the characters concerned were
232 different numbers of bytes. For example, U+023A and U+2C65 are an upper
233 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
234 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
235 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
236 2-byte code at the end of the subject (it thought there wasn't enough data
237 left).
238
239 5. Comprehensive information about what went wrong is now returned by
240 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
241 as the output vector has at least 2 elements. The offset of the start of
242 the failing character and a reason code are placed in the vector.
243
244 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
245 now returned is for the first byte of the failing character, instead of the
246 last byte inspected. This is an incompatible change, but I hope it is small
247 enough not to be a problem. It makes the returned offset consistent with
248 pcre_exec() and pcre_dfa_exec().
249
250 7. pcretest now gives a text phrase as well as the error number when
251 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
252 failure, the offset and reason code are output.
253
254 8. When \R was used with a maximizing quantifier it failed to skip backwards
255 over a \r\n pair if the subsequent match failed. Instead, it just skipped
256 back over a single character (\n). This seems wrong (because it treated the
257 two characters as a single entity when going forwards), conflicts with the
258 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
259 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
260 has been changed.
261
262 9. Some internal refactoring has changed the processing so that the handling
263 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
264 time (the PCRE_DOTALL option was changed this way some time ago: version
265 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
266 which was always a bit of a fudge. It also means that there is one less
267 argument for the match() function, which reduces its stack requirements
268 slightly. This change also fixes an incompatibility with Perl: the pattern
269 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
270
271 10. More internal refactoring has drastically reduced the number of recursive
272 calls to match() for possessively repeated groups such as (abc)++ when
273 using pcre_exec().
274
275 11. While implementing 10, a number of bugs in the handling of groups were
276 discovered and fixed:
277
278 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
279 (a|)*(?1) gave a compile-time internal error.
280 ((a|)+)+ did not notice that the outer group could match an empty string.
281 (^a|^)+ was not marked as anchored.
282 (.*a|.*)+ was not marked as matching at start or after a newline.
283
284 12. Yet more internal refactoring has removed another argument from the match()
285 function. Special calls to this function are now indicated by setting a
286 value in a variable in the "match data" data block.
287
288 13. Be more explicit in pcre_study() instead of relying on "default" for
289 opcodes that mean there is no starting character; this means that when new
290 ones are added and accidentally left out of pcre_study(), testing should
291 pick them up.
292
293 14. The -s option of pcretest has been documented for ages as being an old
294 synonym of -m (show memory usage). I have changed it to mean "force study
295 for every regex", that is, assume /S for every regex. This is similar to -i
296 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
297 using it. It makes it easier to run collections of tests with and without
298 study enabled, and thereby test pcre_study() more easily. All the standard
299 tests are now run with and without -s (but some patterns can be marked as
300 "never study" - see 20 below).
301
302 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
303 restoration of the capturing data to the outer values was not happening
304 correctly.
305
306 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
307 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
308 pattern had matched an empty string, and so incorrectly returned a no
309 match.
310
311 17. There was optimizing code for the last branch of non-capturing parentheses,
312 and also for the obeyed branch of a conditional subexpression, which used
313 tail recursion to cut down on stack usage. Unfortunately, now that there is
314 the possibility of (*THEN) occurring in these branches, tail recursion is
315 no longer possible because the return has to be checked for (*THEN). These
316 two optimizations have therefore been removed. [But see 8.20/11 above.]
317
318 18. If a pattern containing \R was studied, it was assumed that \R always
319 matched two bytes, thus causing the minimum subject length to be
320 incorrectly computed because \R can also match just one byte.
321
322 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
323 was incorrectly computed.
324
325 20. If /S is present twice on a test pattern in pcretest input, it now
326 *disables* studying, thereby overriding the use of -s on the command line
327 (see 14 above). This is necessary for one or two tests to keep the output
328 identical in both cases.
329
330 21. When (*ACCEPT) was used in an assertion that matched an empty string and
331 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
332
333 22. When an atomic group that contained a capturing parenthesis was
334 successfully matched, but the branch in which it appeared failed, the
335 capturing was not being forgotten if a higher numbered group was later
336 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
337 group 1 to "a", when in fact it should be unset. This applied to multi-
338 branched capturing and non-capturing groups, repeated or not, and also to
339 positive assertions (capturing in negative assertions does not happen
340 in PCRE) and also to nested atomic groups.
341
342 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
343 subject after a captured substring, to make it easier to tell which of a
344 number of identical substrings has been captured.
345
346 24. The way atomic groups are processed by pcre_exec() has been changed so that
347 if they are repeated, backtracking one repetition now resets captured
348 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
349 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
350 "aaa". Previously, it would have been "a". As part of this code
351 refactoring, the way recursive calls are handled has also been changed.
352
353 25. If an assertion condition captured any substrings, they were not passed
354 back unless some other capturing happened later. For example, if
355 (?(?=(a))a) was matched against "a", no capturing was returned.
356
357 26. When studying a pattern that contained subroutine calls or assertions,
358 the code for finding the minimum length of a possible match was handling
359 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
360 group 1 called group 2 while simultaneously a separate group 2 called group
361 1). A stack overflow occurred in this case. I have fixed this by limiting
362 the recursion depth to 10.
363
364 27. Updated RunTest.bat in the distribution to the version supplied by Tom
365 Fortmann. This supports explicit test numbers on the command line, and has
366 argument validation and error reporting.
367
368 28. An instance of \X with an unlimited repeat could fail if at any point the
369 first character it looked at was a mark character.
370
371 29. Some minor code refactoring concerning Unicode properties and scripts
372 should reduce the stack requirement of match() slightly.
373
374 30. Added the '=' option to pcretest to check the setting of unused capturing
375 slots at the end of the pattern, which are documented as being -1, but are
376 not included in the return count.
377
378 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
379 compiled something random. Now it gives a compile-time error (as does
380 Perl).
381
382 32. A *MARK encountered during the processing of a positive assertion is now
383 recorded and passed back (compatible with Perl).
384
385 33. If --only-matching or --colour was set on a pcregrep call whose pattern
386 had alternative anchored branches, the search for a second match in a line
387 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
388 matched the line "0102" twice. The same bug affected patterns that started
389 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
390 twice.
391
392 34. Previously, PCRE did not allow quantification of assertions. However, Perl
393 does, and because of capturing effects, quantifying parenthesized
394 assertions may at times be useful. Quantifiers are now allowed for
395 parenthesized assertions.
396
397 35. A minor code tidy in pcre_compile() when checking options for \R usage.
398
399 36. \g was being checked for fancy things in a character class, when it should
400 just be a literal "g".
401
402 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
403 appearance of a nested POSIX class supersedes an apparent external class.
404 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
405 unescaped square brackets may also appear as part of class names. For
406 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
407 more like Perl. (But see 8.20/1 above.)
408
409 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
410 was because it thought it was \N{name}, which is not supported).
411
412 39. Add minix to OS list not supporting the -S option in pcretest.
413
414 40. PCRE tries to detect cases of infinite recursion at compile time, but it
415 cannot analyze patterns in sufficient detail to catch mutual recursions
416 such as ((?1))((?2)). There is now a runtime test that gives an error if a
417 subgroup is called recursively as a subpattern for a second time at the
418 same position in the subject string. In previous releases this might have
419 been caught by the recursion limit, or it might have run out of stack.
420
421 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
422 happen only once. PCRE was, however incorrectly giving a compile time error
423 "recursive call could loop indefinitely" because it cannot analyze the
424 pattern in sufficient detail. The compile time test no longer happens when
425 PCRE is compiling a conditional subpattern, but actual runaway loops are
426 now caught at runtime (see 40 above).
427
428 42. It seems that Perl allows any characters other than a closing parenthesis
429 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
430 has been changed to be the same.
431
432 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
433 as not to get warnings when autogen.sh is called. Also changed
434 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
435
436 44. To help people who use pcregrep to scan files containing exceedingly long
437 lines, the following changes have been made:
438
439 (a) The default value of the buffer size parameter has been increased from
440 8K to 20K. (The actual buffer used is three times this size.)
441
442 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
443 PCRE is built.
444
445 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
446 to be set at run time.
447
448 (d) Numerical values in pcregrep options can be followed by K or M, for
449 example --buffer-size=50K.
450
451 (e) If a line being scanned overflows pcregrep's buffer, an error is now
452 given and the return code is set to 2.
453
454 45. Add a pointer to the latest mark to the callout data block.
455
456 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
457 partial match of an empty string instead of no match. This was specific to
458 the use of ".".
459
460 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
461 complete match instead of a partial match. This bug was dependent on both
462 the PCRE_UTF8 and PCRE_DOTALL options being set.
463
464 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
465 starting byte set, because \b was not being ignored.
466
467
468 Version 8.12 15-Jan-2011
469 ------------------------
470
471 1. Fixed some typos in the markup of the man pages, and wrote a script that
472 checks for such things as part of the documentation building process.
473
474 2. On a big-endian 64-bit system, pcregrep did not correctly process the
475 --match-limit and --recursion-limit options (added for 8.11). In
476 particular, this made one of the standard tests fail. (The integer value
477 went into the wrong half of a long int.)
478
479 3. If the --colour option was given to pcregrep with -v (invert match), it
480 did strange things, either producing crazy output, or crashing. It should,
481 of course, ignore a request for colour when reporting lines that do not
482 match.
483
484 4. Another pcregrep bug caused similar problems if --colour was specified with
485 -M (multiline) and the pattern match finished with a line ending.
486
487 5. In pcregrep, when a pattern that ended with a literal newline sequence was
488 matched in multiline mode, the following line was shown as part of the
489 match. This seems wrong, so I have changed it.
490
491 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
492 the check for further matches in the same line (so they could be coloured)
493 to overrun the end of the current line. If another match was found, it was
494 incorrectly shown (and then shown again when found in the next line).
495
496 7. If pcregrep was compiled under Windows, there was a reference to the
497 function pcregrep_exit() before it was defined. I am assuming this was
498 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
499 reported by a user. I've moved the definition above the reference.
500
501
502 Version 8.11 10-Dec-2010
503 ------------------------
504
505 1. (*THEN) was not working properly if there were untried alternatives prior
506 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
507 backtracked to try for "b" instead of moving to the next alternative branch
508 at the same level (in this case, to look for "c"). The Perl documentation
509 is clear that when (*THEN) is backtracked onto, it goes to the "next
510 alternative in the innermost enclosing group".
511
512 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
513 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
514 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
515 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
516 (*THEN).
517
518 3. If \s appeared in a character class, it removed the VT character from
519 the class, even if it had been included by some previous item, for example
520 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
521 of \s, but is part of the POSIX "space" class.)
522
523 4. A partial match never returns an empty string (because you can always
524 match an empty string at the end of the subject); however the checking for
525 an empty string was starting at the "start of match" point. This has been
526 changed to the "earliest inspected character" point, because the returned
527 data for a partial match starts at this character. This means that, for
528 example, /(?<=abc)def/ gives a partial match for the subject "abc"
529 (previously it gave "no match").
530
531 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
532 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
533 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
534 has an implication that the given string is incomplete (because a partial
535 match is preferred over a full match). For this reason, these items now
536 give a partial match in this situation. [Aside: previously, the one case
537 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
538 match rather than a full match, which was wrong by the old rules, but is
539 now correct.]
540
541 6. There was a bug in the handling of #-introduced comments, recognized when
542 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
543 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
544 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
545 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
546 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
547 places in pcre_compile().
548
549 7. Related to (6) above, when pcre_compile() was skipping #-introduced
550 comments when looking ahead for named forward references to subpatterns,
551 the only newline sequence it recognized was NL. It now handles newlines
552 according to the set newline convention.
553
554 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
555 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
556 cater for a lack of strerror(). These oversights have been fixed.
557
558 9. Added --match-limit and --recursion-limit to pcregrep.
559
560 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
561
562 11. When the -o option was used, pcregrep was setting a return code of 1, even
563 when matches were found, and --line-buffered was not being honoured.
564
565 12. Added an optional parentheses number to the -o and --only-matching options
566 of pcregrep.
567
568 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
569 can match an empty string. The code to do it in pcretest and pcredemo
570 needed fixing:
571
572 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
573 only one byte after an empty string match just before CRLF (this case
574 just got forgotten; "any" and "anycrlf" were OK).
575
576 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
577 mode when an empty string match preceded an ASCII character followed by
578 a non-ASCII character. (The code for advancing by one character rather
579 than one byte was nonsense.)
580
581 (c) The pcredemo.c sample program did not have any code at all to handle
582 the cases when CRLF is a valid newline sequence.
583
584 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
585 as a starting offset was within the subject string. There is now a new
586 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
587 negative or greater than the length of the string. In order to test this,
588 pcretest is extended to allow the setting of negative starting offsets.
589
590 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
591 starting offset points to the beginning of a UTF-8 character was
592 unnecessarily clumsy. I tidied it up.
593
594 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
595 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
596
597 17. Nobody had reported that the --include_dir option, which was added in
598 release 7.7 should have been called --include-dir (hyphen, not underscore)
599 for compatibility with GNU grep. I have changed it to --include-dir, but
600 left --include_dir as an undocumented synonym, and the same for
601 --exclude-dir, though that is not available in GNU grep, at least as of
602 release 2.5.4.
603
604 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
605 characters from a string of bytes) have been redefined so as not to use
606 loops, in order to improve performance in some environments. At the same
607 time, I abstracted some of the common code into auxiliary macros to save
608 repetition (this should not affect the compiled code).
609
610 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
611 compile-time error is now given if \c is not followed by an ASCII
612 character, that is, a byte less than 128. (In EBCDIC mode, the code is
613 different, and any byte value is allowed.)
614
615 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
616 START_OPTIMIZE option, which is now allowed at compile time - but just
617 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
618 to pcregrep and other applications that have no direct access to PCRE
619 options. The new /Y option in pcretest sets this option when calling
620 pcre_compile().
621
622 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
623 back references. Groups containing recursive back references were forced to
624 be atomic by that change, but in the case of named groups, the amount of
625 memory required was incorrectly computed, leading to "Failed: internal
626 error: code overflow". This has been fixed.
627
628 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
629 pcretest.c, to avoid build problems in some Borland environments.
630
631
632 Version 8.10 25-Jun-2010
633 ------------------------
634
635 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
636 THEN.
637
638 2. (*ACCEPT) was not working when inside an atomic group.
639
640 3. Inside a character class, \B is treated as a literal by default, but
641 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
642 causes the error). The code is unchanged, but I tidied the documentation.
643
644 4. Inside a character class, PCRE always treated \R and \X as literals,
645 whereas Perl faults them if its -w option is set. I have changed PCRE so
646 that it faults them when PCRE_EXTRA is set.
647
648 5. Added support for \N, which always matches any character other than
649 newline. (It is the same as "." when PCRE_DOTALL is not set.)
650
651 6. When compiling pcregrep with newer versions of gcc which may have
652 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
653 declared with attribute warn_unused_result" were given. Just casting the
654 result to (void) does not stop the warnings; a more elaborate fudge is
655 needed. I've used a macro to implement this.
656
657 7. Minor change to pcretest.c to avoid a compiler warning.
658
659 8. Added four artifical Unicode properties to help with an option to make
660 \s etc use properties (see next item). The new properties are: Xan
661 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
662
663 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
664 use Unicode properties. (*UCP) at the start of a pattern can be used to set
665 this option. Modified pcretest to add /W to test this facility. Added
666 REG_UCP to make it available via the POSIX interface.
667
668 10. Added --line-buffered to pcregrep.
669
670 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
671 studied, and the match started with a letter with a code point greater than
672 127 whose first byte was different to the first byte of the other case of
673 the letter, the other case of this starting letter was not recognized
674 (#976).
675
676 12. If a pattern that was studied started with a repeated Unicode property
677 test, for example, \p{Nd}+, there was the theoretical possibility of
678 setting up an incorrect bitmap of starting bytes, but fortunately it could
679 not have actually happened in practice until change 8 above was made (it
680 added property types that matched character-matching opcodes).
681
682 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
683 possible starting bytes for non-anchored patterns.
684
685 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
686 \R, and also a number of cases that involve Unicode properties, both
687 explicit and implicit when PCRE_UCP is set.
688
689 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
690 input, it could crash or give wrong results if characters with values
691 greater than 0xc0 were present in the subject string. (Detail: it assumed
692 UTF-8 input when processing these items.)
693
694 16. Added a lot of (int) casts to avoid compiler warnings in systems where
695 size_t is 64-bit (#991).
696
697 17. Added a check for running out of memory when PCRE is compiled with
698 --disable-stack-for-recursion (#990).
699
700 18. If the last data line in a file for pcretest does not have a newline on
701 the end, a newline was missing in the output.
702
703 19. The default pcre_chartables.c file recognizes only ASCII characters (values
704 less than 128) in its various bitmaps. However, there is a facility for
705 generating tables according to the current locale when PCRE is compiled. It
706 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
707 characters, are recognized by isspace() and therefore were getting set in
708 these tables, and indeed these tables seem to approximate to ISO 8859. This
709 caused a problem in UTF-8 mode when pcre_study() was used to create a list
710 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
711 which of course cannot start UTF-8 characters. I have changed the code so
712 that only real ASCII characters (less than 128) and the correct starting
713 bytes for UTF-8 encodings are set for characters greater than 127 when in
714 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
715 altogether.)
716
717 20. Added the /T option to pcretest so as to be able to run tests with non-
718 standard character tables, thus making it possible to include the tests
719 used for 19 above in the standard set of tests.
720
721 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
722 reference to a subpattern the other side of a comment that contains an
723 opening parenthesis caused either an internal compiling error, or a
724 reference to the wrong subpattern.
725
726
727 Version 8.02 19-Mar-2010
728 ------------------------
729
730 1. The Unicode data tables have been updated to Unicode 5.2.0.
731
732 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
733 configured.
734
735 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
736 original author of that file, following a query about its status.
737
738 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
739 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
740
741 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
742 quantifier applied to a forward-referencing subroutine call, could compile
743 incorrect code or give the error "internal error: previously-checked
744 referenced subpattern not found".
745
746 6. Both MS Visual Studio and Symbian OS have problems with initializing
747 variables to point to external functions. For these systems, therefore,
748 pcre_malloc etc. are now initialized to local functions that call the
749 relevant global functions.
750
751 7. There were two entries missing in the vectors called coptable and poptable
752 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
753 I've fixed the data, and added a kludgy way of testing at compile time that
754 the lengths are correct (equal to the number of opcodes).
755
756 8. Following on from 7, I added a similar kludge to check the length of the
757 eint vector in pcreposix.c.
758
759 9. Error texts for pcre_compile() are held as one long string to avoid too
760 much relocation at load time. To find a text, the string is searched,
761 counting zeros. There was no check for running off the end of the string,
762 which could happen if a new error number was added without updating the
763 string.
764
765 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
766
767 11. \K was not working if it appeared in an atomic group or in a group that
768 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
769 \K is "not well defined" if used in an assertion. PCRE now accepts it if
770 the assertion is positive, but not if it is negative.
771
772 12. Change 11 fortuitously reduced the size of the stack frame used in the
773 "match()" function of pcre_exec.c by one pointer. Forthcoming
774 implementation of support for (*MARK) will need an extra pointer on the
775 stack; I have reserved it now, so that the stack frame size does not
776 decrease.
777
778 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
779 item in branch that calls a recursion is a subroutine call - as in the
780 second branch in the above example - was incorrectly given the compile-
781 time error "recursive call could loop indefinitely" because pcre_compile()
782 was not correctly checking the subroutine for matching a non-empty string.
783
784 14. The checks for overrunning compiling workspace could trigger after an
785 overrun had occurred. This is a "should never occur" error, but it can be
786 triggered by pathological patterns such as hundreds of nested parentheses.
787 The checks now trigger 100 bytes before the end of the workspace.
788
789 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
790
791
792 Version 8.01 19-Jan-2010
793 ------------------------
794
795 1. If a pattern contained a conditional subpattern with only one branch (in
796 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
797 computed the wrong minimum data length (which is of course zero for such
798 subpatterns). This could cause incorrect "no match" results.
799
800 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
801 the pattern is reset in the first branch, pcre_compile() failed with
802 "internal error: code overflow at offset...". This happened only when
803 the reset was to the original external option setting. (An optimization
804 abstracts leading options settings into an external setting, which was the
805 cause of this.)
806
807 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
808 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
809 assertion pattern did not match (meaning that the assertion was true), it
810 was incorrectly treated as false if the SKIP had been reached during the
811 matching. This also applied to assertions used as conditions.
812
813 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
814 assertion subpattern, including such a pattern used as a condition,
815 unpredictable results occurred, instead of the error return
816 PCRE_ERROR_DFA_UITEM.
817
818 5. The C++ GlobalReplace function was not working like Perl for the special
819 situation when an empty string is matched. It now does the fancy magic
820 stuff that is necessary.
821
822 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
823 removed. (These were left over from very, very early versions of PCRE.)
824
825 7. Some cosmetic changes to the code to make life easier when compiling it
826 as part of something else:
827
828 (a) Change DEBUG to PCRE_DEBUG.
829
830 (b) In pcre_compile(), rename the member of the "branch_chain" structure
831 called "current" as "current_branch", to prevent a collision with the
832 Linux macro when compiled as a kernel module.
833
834 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
835 prevent a collision with the Linux macro when compiled as a kernel
836 module.
837
838 8. In pcre_compile() there are some checks for integer overflows that used to
839 cast potentially large values to (double). This has been changed to that
840 when building, a check for int64_t is made, and if it is found, it is used
841 instead, thus avoiding the use of floating point arithmetic. (There is no
842 other use of FP in PCRE.) If int64_t is not found, the fallback is to
843 double.
844
845 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
846 2005 (difference between two addresses compared to an unsigned value).
847
848 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
849 custom one, because of the following reported problem in Windows:
850
851 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
852 under Win32.
853 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
854 therefore missing the function definition.
855 - The compiler thus generates a "C" signature for the test function.
856 - The linker fails to find the "C" function.
857 - PCRE fails to configure if asked to do so against libbz2.
858
859 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
860 messages were output:
861
862 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
863 rerunning libtoolize, to keep the correct libtool macros in-tree.
864 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
865
866 I have done both of these things.
867
868 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
869 most of the time, it *can* run out if it is given a pattern that contains a
870 runaway infinite recursion. I updated the discussion in the pcrestack man
871 page.
872
873 13. Now that we have gone to the x.xx style of version numbers, the minor
874 version may start with zero. Using 08 or 09 is a bad idea because users
875 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
876 interpreted as invalid octal numbers. I've updated the previous comment in
877 configure.ac, and also added a check that gives an error if 08 or 09 are
878 used.
879
880 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
881 causing partial matching to fail when the end of the subject matched \W
882 in a UTF-8 pattern where \W was quantified with a minimum of 3.
883
884 15. There were some discrepancies between the declarations in pcre_internal.h
885 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
886 their definitions. The declarations used "const uschar *" and the
887 definitions used USPTR. Even though USPTR is normally defined as "const
888 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
889 reported that: "This difference in casting confuses some C++ compilers, for
890 example, SunCC recognizes above declarations as different functions and
891 generates broken code for hbpcre." I have changed the declarations to use
892 USPTR.
893
894 16. GNU libtool is named differently on some systems. The autogen.sh script now
895 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
896 (FreeBSD).
897
898 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
899 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
900 comment: "Figure out how to create a longlong from a string: strtoll and
901 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
902 instance, but it only takes 2 args instead of 3!"
903
904 18. A subtle bug concerned with back references has been fixed by a change of
905 specification, with a corresponding code fix. A pattern such as
906 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
907 refers, was giving matches when it shouldn't. For example, xa=xaaa would
908 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
909 same bug. Such groups have to be quantified to be useful, or contained
910 inside another quantified group. (If there's no repetition, the reference
911 can never match.) The problem arises because, having left the group and
912 moved on to the rest of the pattern, a later failure that backtracks into
913 the group uses the captured value from the final iteration of the group
914 rather than the correct earlier one. I have fixed this in PCRE by forcing
915 any group that contains a reference to itself to be an atomic group; that
916 is, there cannot be any backtracking into it once it has completed. This is
917 similar to recursive and subroutine calls.
918
919
920 Version 8.00 19-Oct-09
921 ----------------------
922
923 1. The table for translating pcre_compile() error codes into POSIX error codes
924 was out-of-date, and there was no check on the pcre_compile() error code
925 being within the table. This could lead to an OK return being given in
926 error.
927
928 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
929 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
930 in a Windows environment.
931
932 3. The pcregrep --count option prints the count for each file even when it is
933 zero, as does GNU grep. However, pcregrep was also printing all files when
934 --files-with-matches was added. Now, when both options are given, it prints
935 counts only for those files that have at least one match. (GNU grep just
936 prints the file name in this circumstance, but including the count seems
937 more useful - otherwise, why use --count?) Also ensured that the
938 combination -clh just lists non-zero counts, with no names.
939
940 4. The long form of the pcregrep -F option was incorrectly implemented as
941 --fixed_strings instead of --fixed-strings. This is an incompatible change,
942 but it seems right to fix it, and I didn't think it was worth preserving
943 the old behaviour.
944
945 5. The command line items --regex=pattern and --regexp=pattern were not
946 recognized by pcregrep, which required --regex pattern or --regexp pattern
947 (with a space rather than an '='). The man page documented the '=' forms,
948 which are compatible with GNU grep; these now work.
949
950 6. No libpcreposix.pc file was created for pkg-config; there was just
951 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
952
953 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
954 when UCP support is not needed, by modifying the Python script that
955 generates it from Unicode data files. This should not matter if the module
956 is correctly used as a library, but I received one complaint about 50K of
957 unwanted data. My guess is that the person linked everything into his
958 program rather than using a library. Anyway, it does no harm.
959
960 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
961 was a minimum greater than 1 for a wide character in a possessive
962 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
963 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
964 character. Chaos in the form of incorrect output or a compiling loop could
965 result.
966
967 9. The restrictions on what a pattern can contain when partial matching is
968 requested for pcre_exec() have been removed. All patterns can now be
969 partially matched by this function. In addition, if there are at least two
970 slots in the offset vector, the offset of the earliest inspected character
971 for the match and the offset of the end of the subject are set in them when
972 PCRE_ERROR_PARTIAL is returned.
973
974 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
975 synonymous with PCRE_PARTIAL, for backwards compatibility, and
976 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
977 and may be more useful for multi-segment matching.
978
979 11. Partial matching with pcre_exec() is now more intuitive. A partial match
980 used to be given if ever the end of the subject was reached; now it is
981 given only if matching could not proceed because another character was
982 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
983 string "Z", which now yields "no match" instead of "partial match". In the
984 case of pcre_dfa_exec(), "no match" is given if every matching path for the
985 final character ended with (*FAIL).
986
987 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
988 if the pattern had a "must contain" character that was already found in the
989 earlier partial match, unless partial matching was again requested. For
990 example, with the pattern /dog.(body)?/, the "must contain" character is
991 "g". If the first part-match was for the string "dog", restarting with
992 "sbody" failed. This bug has been fixed.
993
994 13. The string returned by pcre_dfa_exec() after a partial match has been
995 changed so that it starts at the first inspected character rather than the
996 first character of the match. This makes a difference only if the pattern
997 starts with a lookbehind assertion or \b or \B (\K is not supported by
998 pcre_dfa_exec()). It's an incompatible change, but it makes the two
999 matching functions compatible, and I think it's the right thing to do.
1000
1001 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1002 so that the demonstration program is easily available in environments where
1003 PCRE has not been installed from source.
1004
1005 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1006 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1007 library.
1008
1009 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1010 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1011 is not the first non-POSIX option to be added. Clearly some people find
1012 these options useful.
1013
1014 17. If a caller to the POSIX matching function regexec() passes a non-zero
1015 value for nmatch with a NULL value for pmatch, the value of
1016 nmatch is forced to zero.
1017
1018 18. RunGrepTest did not have a test for the availability of the -u option of
1019 the diff command, as RunTest does. It now checks in the same way as
1020 RunTest, and also checks for the -b option.
1021
1022 19. If an odd number of negated classes containing just a single character
1023 interposed, within parentheses, between a forward reference to a named
1024 subpattern and the definition of the subpattern, compilation crashed with
1025 an internal error, complaining that it could not find the referenced
1026 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1027 [The bug was that it was starting one character too far in when skipping
1028 over the character class, thus treating the ] as data rather than
1029 terminating the class. This meant it could skip too much.]
1030
1031 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1032 /g option in pcretest when the pattern contains \K, which makes it possible
1033 to have an empty string match not at the start, even when the pattern is
1034 anchored. Updated pcretest and pcredemo to use this option.
1035
1036 21. If the maximum number of capturing subpatterns in a recursion was greater
1037 than the maximum at the outer level, the higher number was returned, but
1038 with unset values at the outer level. The correct (outer level) value is
1039 now given.
1040
1041 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1042 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1043 make it do so. The string so far is captured, making this feature
1044 compatible with Perl.
1045
1046 23. The tests have been re-organized, adding tests 11 and 12, to make it
1047 possible to check the Perl 5.10 features against Perl 5.10.
1048
1049 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1050 pattern matches a fixed length string. PCRE did not allow this; now it
1051 does. Neither allows recursion.
1052
1053 25. I finally figured out how to implement a request to provide the minimum
1054 length of subject string that was needed in order to match a given pattern.
1055 (It was back references and recursion that I had previously got hung up
1056 on.) This code has now been added to pcre_study(); it finds a lower bound
1057 to the length of subject needed. It is not necessarily the greatest lower
1058 bound, but using it to avoid searching strings that are too short does give
1059 some useful speed-ups. The value is available to calling programs via
1060 pcre_fullinfo().
1061
1062 26. While implementing 25, I discovered to my embarrassment that pcretest had
1063 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1064 study optimizations had never been tested with that matching function.
1065 Oops. What is worse, even when it was passed study data, there was a bug in
1066 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1067 were also very few tests of studied patterns with pcre_dfa_exec().
1068
1069 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1070 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1071 on the other side of the coin, they are no longer allowed to have different
1072 names, because these cannot be distinguished in PCRE, and this has caused
1073 confusion. (This is a difference from Perl.)
1074
1075 28. When duplicate subpattern names are present (necessarily with different
1076 numbers, as required by 27 above), and a test is made by name in a
1077 conditional pattern, either for a subpattern having been matched, or for
1078 recursion in such a pattern, all the associated numbered subpatterns are
1079 tested, and the overall condition is true if the condition is true for any
1080 one of them. This is the way Perl works, and is also more like the way
1081 testing by number works.
1082
1083
1084 Version 7.9 11-Apr-09
1085 ---------------------
1086
1087 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1088 (pcretest), all targets were linked against these libraries. This included
1089 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1090 libraries. This caused unwanted dependencies to be created. This problem
1091 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1092 pcretest is linked with readline.
1093
1094 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1095 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1096 moved outside it again, because FALSE and TRUE are already defined in AIX,
1097 but BOOL is not.
1098
1099 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1100 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1101
1102 4. The pcregrep documentation said spaces were inserted as well as colons (or
1103 hyphens) following file names and line numbers when outputting matching
1104 lines. This is not true; no spaces are inserted. I have also clarified the
1105 wording for the --colour (or --color) option.
1106
1107 5. In pcregrep, when --colour was used with -o, the list of matching strings
1108 was not coloured; this is different to GNU grep, so I have changed it to be
1109 the same.
1110
1111 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1112 each matching line was coloured. Now it goes on to look for further matches
1113 of any of the test patterns, which is the same behaviour as GNU grep.
1114
1115 7. A pattern that could match an empty string could cause pcregrep to loop; it
1116 doesn't make sense to accept an empty string match in pcregrep, so I have
1117 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1118 seems to be how GNU grep behaves.
1119
1120 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1121 start or after a newline", because the conditional assertion was not being
1122 correctly handled. The rule now is that both the assertion and what follows
1123 in the first alternative must satisfy the test.
1124
1125 9. If auto-callout was enabled in a pattern with a conditional group whose
1126 condition was an assertion, PCRE could crash during matching, both with
1127 pcre_exec() and pcre_dfa_exec().
1128
1129 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1130 used for matching.
1131
1132 11. Unicode property support in character classes was not working for
1133 characters (bytes) greater than 127 when not in UTF-8 mode.
1134
1135 12. Added the -M command line option to pcretest.
1136
1137 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1138
1139 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1140
1141 16. Added comments and documentation about mis-use of no_arg in the C++
1142 wrapper.
1143
1144 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1145 from Martin Jerabek that uses macro names for all relevant character and
1146 string constants.
1147
1148 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1149 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1150 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1151 these, but not everybody uses configure.
1152
1153 19. A conditional group that had only one branch was not being correctly
1154 recognized as an item that could match an empty string. This meant that an
1155 enclosing group might also not be so recognized, causing infinite looping
1156 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1157 with the subject "ab", where knowledge that the repeated group can match
1158 nothing is needed in order to break the loop.
1159
1160 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1161 exec(), but without supplying a callout function, matching went wrong.
1162
1163 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1164 leak if the size of the offset vector was greater than 30. When the vector
1165 is smaller, the saved offsets during recursion go onto a local stack
1166 vector, but for larger vectors malloc() is used. It was failing to free
1167 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1168 error, in fact).
1169
1170 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1171 heapframe that is used only when UTF-8 support is enabled. This caused no
1172 problem, but was untidy.
1173
1174 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1175 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1176 included within another project.
1177
1178 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1179 slightly modified by me:
1180
1181 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1182 not building pcregrep.
1183
1184 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1185 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1186
1187 25. Forward references, both numeric and by name, in patterns that made use of
1188 duplicate group numbers, could behave incorrectly or give incorrect errors,
1189 because when scanning forward to find the reference group, PCRE was not
1190 taking into account the duplicate group numbers. A pattern such as
1191 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1192
1193 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1194 the feature of a custom pointer more persuasive (as requested by a user).
1195
1196 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1197 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1198 pre-defined.
1199
1200 28. Added support for (*UTF8) at the start of a pattern.
1201
1202 29. Arrange for flags added by the "release type" setting in CMake to be shown
1203 in the configuration summary.
1204
1205
1206 Version 7.8 05-Sep-08
1207 ---------------------
1208
1209 1. Replaced UCP searching code with optimized version as implemented for Ad
1210 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1211 stage table and inline lookup instead of a function, giving speed ups of 2
1212 to 5 times on some simple patterns that I tested. Permission was given to
1213 distribute the MultiStage2.py script that generates the tables (it's not in
1214 the tarball, but is in the Subversion repository).
1215
1216 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1217 scripts.
1218
1219 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1220 a group with a zero qualifier. The result of the study could be incorrect,
1221 or the function might crash, depending on the pattern.
1222
1223 4. Caseless matching was not working for non-ASCII characters in back
1224 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1225 It now works when Unicode Property Support is available.
1226
1227 5. In pcretest, an escape such as \x{de} in the data was always generating
1228 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1229 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1230 truncation.
1231
1232 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1233
1234 7. Added two (int) casts to pcregrep when printing the difference of two
1235 pointers, in case they are 64-bit values.
1236
1237 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1238 test 2 if it fails.
1239
1240 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1241 and a #define of that name to empty if it is not externally set. This is to
1242 allow users of MSVC to set it if necessary.
1243
1244 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1245 the convenience functions in the pcre_get.c source file.
1246
1247 11. An option change at the start of a pattern that had top-level alternatives
1248 could cause overwriting and/or a crash. This command provoked a crash in
1249 some environments:
1250
1251 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1252
1253 This potential security problem was recorded as CVE-2008-2371.
1254
1255 12. For a pattern where the match had to start at the beginning or immediately
1256 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1257 pcre_dfa_exec() could read past the end of the passed subject if there was
1258 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1259 pcretest so that it places the subject at the end of its malloc-ed buffer.
1260
1261 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1262 exec() might read past the end of the data buffer in UTF-8 mode.
1263
1264 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1265 the data contained the byte 0x85 as part of a UTF-8 character within its
1266 first line. This applied both to normal and DFA matching.
1267
1268 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1269 /^[^d]*?$/8 failed to match "abc".
1270
1271 16. Added a missing copyright notice to pcrecpp_internal.h.
1272
1273 17. Make it more clear in the documentation that values returned from
1274 pcre_exec() in ovector are byte offsets, not character counts.
1275
1276 18. Tidied a few places to stop certain compilers from issuing warnings.
1277
1278 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1279 supplied by Stefan Weber. I made a further small update for 7.8 because
1280 there is a change of source arrangements: the pcre_searchfuncs.c module is
1281 replaced by pcre_ucd.c.
1282
1283
1284 Version 7.7 07-May-08
1285 ---------------------
1286
1287 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1288 a string to a long long, pretend we don't even have a long long." This is
1289 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1290
1291 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1292 pre-7.6 versions, which defined a global no_arg variable instead of putting
1293 it in the RE class. (See also #8 below.)
1294
1295 3. Remove a line of dead code, identified by coverity and reported by Nuno
1296 Lopes.
1297
1298 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1299
1300 (1) The include/exclude patterns were being applied to the whole pathnames
1301 of files, instead of just to the final components.
1302
1303 (2) If there was more than one level of directory, the subdirectories were
1304 skipped unless they satisfied the include/exclude conditions. This is
1305 inconsistent with GNU grep (and could even be seen as contrary to the
1306 pcregrep specification - which I improved to make it absolutely clear).
1307 The action now is always to scan all levels of directory, and just
1308 apply the include/exclude patterns to regular files.
1309
1310 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1311 --exclude_dir in the tests to avoid scanning .svn directories.
1312
1313 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1314 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1315 doesn't support NULs in patterns.
1316
1317 7. Added some missing "const"s to declarations of static tables in
1318 pcre_compile.c and pcre_dfa_exec.c.
1319
1320 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1321 caused by fix #2 above. (Subsequently also a second patch to fix the
1322 first patch. And a third patch - this was a messy problem.)
1323
1324 9. Applied Craig's patch to remove the use of push_back().
1325
1326 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1327 matching function regexec().
1328
1329 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1330 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1331 references. PCRE supports relative numbers with this syntax (I don't think
1332 Oniguruma does).
1333
1334 12. Previously, a group with a zero repeat such as (...){0} was completely
1335 omitted from the compiled regex. However, this means that if the group
1336 was called as a subroutine from elsewhere in the pattern, things went wrong
1337 (an internal error was given). Such groups are now left in the compiled
1338 pattern, with a new opcode that causes them to be skipped at execution
1339 time.
1340
1341 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1342 to the way PCRE behaves:
1343
1344 (a) A lone ] character is dis-allowed (Perl treats it as data).
1345
1346 (b) A back reference to an unmatched subpattern matches an empty string
1347 (Perl fails the current match path).
1348
1349 (c) A data ] in a character class must be notated as \] because if the
1350 first data character in a class is ], it defines an empty class. (In
1351 Perl it is not possible to have an empty class.) The empty class []
1352 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1353 The negative empty class [^] matches any one character, independently
1354 of the DOTALL setting.
1355
1356 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1357 non-existent subpattern following a character class starting with ']' and
1358 containing () gave an internal compiling error instead of "reference to
1359 non-existent subpattern". Fortunately, when the pattern did exist, the
1360 compiled code was correct. (When scanning forwards to check for the
1361 existencd of the subpattern, it was treating the data ']' as terminating
1362 the class, so got the count wrong. When actually compiling, the reference
1363 was subsequently set up correctly.)
1364
1365 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1366 it was being rejected as not supported by pcre_dfa_exec(), even though
1367 other assertions are supported. I have made pcre_dfa_exec() support
1368 (*FAIL).
1369
1370 16. The implementation of 13c above involved the invention of a new opcode,
1371 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1372 cannot be changed at match time, I realized I could make a small
1373 improvement to matching performance by compiling OP_ALLANY instead of
1374 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1375 on the OP_ANY path.
1376
1377 17. Compiling pcretest on Windows with readline support failed without the
1378 following two fixes: (1) Make the unistd.h include conditional on
1379 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1380
1381 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1382 ncurses library to be included for pcretest when ReadLine support is
1383 requested, but also to allow for it to be overridden. This patch came from
1384 Daniel Bergström.
1385
1386 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1387 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1388 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1389 spotting this.
1390
1391
1392 Version 7.6 28-Jan-08
1393 ---------------------
1394
1395 1. A character class containing a very large number of characters with
1396 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1397 overflow.
1398
1399 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1400 HAVE_LONG_LONG is not defined.
1401
1402 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1403 bring them up to date and include new features. This patch includes:
1404
1405 - Fixed PH's badly added libz and libbz2 support.
1406 - Fixed a problem with static linking.
1407 - Added pcredemo. [But later removed - see 7 below.]
1408 - Fixed dftables problem and added an option.
1409 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1410 HAVE_LONG_LONG.
1411 - Added readline support for pcretest.
1412 - Added an listing of the option settings after cmake has run.
1413
1414 4. A user submitted a patch to Makefile that makes it easy to create
1415 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1416 Makefile.am that cause it to include this special target, without
1417 affecting anything else. Note that the same mingw target plus all
1418 the other distribution libraries and programs are now supported
1419 when configuring with CMake (see 6 below) instead of with
1420 Configure/Make.
1421
1422 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1423 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1424 exported in the Windows port". It has not yet been confirmed that the patch
1425 solves the problem, but it does no harm.
1426
1427 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1428 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1429 with CMake, and also correct the comment about stack recursion.
1430
1431 7. Remove the automatic building of pcredemo from the ./configure system and
1432 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1433 of a program that users should build themselves after PCRE is installed, so
1434 building it automatically is not really right. What is more, it gave
1435 trouble in some build environments.
1436
1437 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1438
1439
1440 Version 7.5 10-Jan-08
1441 ---------------------
1442
1443 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1444 values in parens when parsing an RE using the C++ wrapper."
1445
1446 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1447 Characters greater than 255 were excluded from the class instead of being
1448 included.
1449
1450 3. The same bug as (2) above applied to negated POSIX classes such as
1451 [:^space:].
1452
1453 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1454 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1455 I have changed it.
1456
1457 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1458 first named subpattern) and a construct such as (?&a) would reference the
1459 first named subpattern whose name started with "a" (in other words, the
1460 length check was missing). Both these problems are fixed. "Subpattern name
1461 expected" is now given for (?&) (a zero-length name), and this patch also
1462 makes it give the same error for \k'' (previously it complained that that
1463 was a reference to a non-existent subpattern).
1464
1465 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1466 this is right because (?- can be followed by option settings as well as by
1467 digits. I have, however, made the messages clearer.
1468
1469 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1470 than the number used in the conditional) now cause a compile-time error.
1471 This is actually not compatible with Perl, which accepts such patterns, but
1472 treats the conditional as always being FALSE (as PCRE used to), but it
1473 seems to me that giving a diagnostic is better.
1474
1475 8. Change "alphameric" to the more common word "alphanumeric" in comments
1476 and messages.
1477
1478 9. Fix two occurrences of "backslash" in comments that should have been
1479 "backspace".
1480
1481 10. Remove two redundant lines of code that can never be obeyed (their function
1482 was moved elsewhere).
1483
1484 11. The program that makes PCRE's Unicode character property table had a bug
1485 which caused it to generate incorrect table entries for sequences of
1486 characters that have the same character type, but are in different scripts.
1487 It amalgamated them into a single range, with the script of the first of
1488 them. In other words, some characters were in the wrong script. There were
1489 thirteen such cases, affecting characters in the following ranges:
1490
1491 U+002b0 - U+002c1
1492 U+0060c - U+0060d
1493 U+0061e - U+00612
1494 U+0064b - U+0065e
1495 U+0074d - U+0076d
1496 U+01800 - U+01805
1497 U+01d00 - U+01d77
1498 U+01d9b - U+01dbf
1499 U+0200b - U+0200f
1500 U+030fc - U+030fe
1501 U+03260 - U+0327f
1502 U+0fb46 - U+0fbb1
1503 U+10450 - U+1049d
1504
1505 12. The -o option (show only the matching part of a line) for pcregrep was not
1506 compatible with GNU grep in that, if there was more than one match in a
1507 line, it showed only the first of them. It now behaves in the same way as
1508 GNU grep.
1509
1510 13. If the -o and -v options were combined for pcregrep, it printed a blank
1511 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1512 does the same. The return code can be used to tell if there were any
1513 non-matching lines.
1514
1515 14. Added --file-offsets and --line-offsets to pcregrep.
1516
1517 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1518 infinitely looping recursion. The bug was that positive lookaheads were not
1519 being skipped when checking for a possible empty match (negative lookaheads
1520 and both kinds of lookbehind were skipped).
1521
1522 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1523 inclusion of <windows.h> to before rather than after the definition of
1524 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1525
1526 17. Specifying a possessive quantifier with a specific limit for a Unicode
1527 character property caused pcre_compile() to compile bad code, which led at
1528 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1529 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1530 caused the error; without that there was no problem.
1531
1532 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1533
1534 19. Added --enable-pcretest-libreadline.
1535
1536 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1537 RE::GlobalReplace(). As a result, the number of replacements returned was
1538 double what it should be. I removed one of the increments, but Craig sent a
1539 later patch that removed the other one (the right fix) and added unit tests
1540 that check the return values (which was not done before).
1541
1542 21. Several CMake things:
1543
1544 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1545 the names libpcre and libpcreposix, not just pcre and pcreposix.
1546
1547 (2) The above change means that pcretest and pcregrep are now correctly
1548 linked with the newly-built libraries, not previously installed ones.
1549
1550 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1551
1552 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1553 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1554 UTF-8 newline character). The key issue is that the pattern starts .*;
1555 this means that the match must be either at the beginning, or after a
1556 newline. The bug was in the code for advancing after a failed match and
1557 checking that the new position followed a newline. It was not taking
1558 account of UTF-8 characters correctly.
1559
1560 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1561 character classes. PCRE was not treating the sequence [:...:] as a
1562 character class unless the ... were all letters. Perl, however, seems to
1563 allow any characters between [: and :], though of course it rejects as
1564 unknown any "names" that contain non-letters, because all the known class
1565 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1566 for example, whereas PCRE did not - it did not recognize a POSIX character
1567 class. This seemed a bit dangerous, so the code has been changed to be
1568 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1569 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1570 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1571 Perl does, and where it didn't before.
1572
1573 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1574 Windows environments %n is disabled by default.
1575
1576
1577 Version 7.4 21-Sep-07
1578 ---------------------
1579
1580 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1581 means that a class such as [\s] counted as "explicit reference to CR or
1582 LF". That isn't really right - the whole point of the change was to try to
1583 help when there was an actual mention of one of the two characters. So now
1584 the change happens only if \r or \n (or a literal CR or LF) character is
1585 encountered.
1586
1587 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1588 of both had grown to the point where there were only 3 bits left.
1589 Fortunately, there was spare space in the data structure, and so I have
1590 moved the internal flags into a new 16-bit field to free up more option
1591 bits.
1592
1593 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1594 but did not set the internal JCHANGED flag - either of these is enough to
1595 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1596 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1597 start sets both bits.
1598
1599 4. Added options (at build time, compile time, exec time) to change \R from
1600 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1601
1602 5. doc/pcresyntax.html was missing from the distribution.
1603
1604 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1605 compatibility, even though it is no longer used.
1606
1607 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1608 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1609 windows.h file is present (where different names are used). [This was
1610 reversed later after testing - see 16 below.]
1611
1612 8. Changed all #include <config.h> to #include "config.h". There were also
1613 some further <pcre.h> cases that I changed to "pcre.h".
1614
1615 9. When pcregrep was used with the --colour option, it missed the line ending
1616 sequence off the lines that it output.
1617
1618 10. It was pointed out to me that arrays of string pointers cause lots of
1619 relocations when a shared library is dynamically loaded. A technique of
1620 using a single long string with a table of offsets can drastically reduce
1621 these. I have refactored PCRE in four places to do this. The result is
1622 dramatic:
1623
1624 Originally: 290
1625 After changing UCP table: 187
1626 After changing error message table: 43
1627 After changing table of "verbs" 36
1628 After changing table of Posix names 22
1629
1630 Thanks to the folks working on Gregex for glib for this insight.
1631
1632 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1633 unicode-properties was also set.
1634
1635 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1636
1637 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1638 checked only for CRLF.
1639
1640 14. Added casts to pcretest.c to avoid compiler warnings.
1641
1642 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1643
1644 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1645 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1646 entirely. This removes changes made in 7 above.
1647
1648 17. The CMake files have been updated, and there is now more information about
1649 building with CMake in the NON-UNIX-USE document.
1650
1651
1652 Version 7.3 28-Aug-07
1653 ---------------------
1654
1655 1. In the rejigging of the build system that eventually resulted in 7.1, the
1656 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1657 brackets there is not right, since it causes compilers to look for an
1658 installed pcre.h, not the version that is in the source that is being
1659 compiled (which of course may be different). I have changed it back to:
1660
1661 #include "pcre.h"
1662
1663 I have a vague recollection that the change was concerned with compiling in
1664 different directories, but in the new build system, that is taken care of
1665 by the VPATH setting the Makefile.
1666
1667 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1668 when the subject happened to end in the byte 0x85 (e.g. if the last
1669 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1670 characters but of course it shouldn't be taken as a newline when it is part
1671 of another character. The bug was that, for an unlimited repeat of . in
1672 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1673 characters when looking for a newline.
1674
1675 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1676
1677 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1678 in debug output.
1679
1680 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1681 long printing in the pcrecpp unittest when running under MinGW.
1682
1683 6. ESC_K was left out of the EBCDIC table.
1684
1685 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1686 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1687 limit also applies to "virtual nesting" when a pattern is recursive, and in
1688 this case 1000 isn't so big. I have been able to remove this limit at the
1689 expense of backing off one optimization in certain circumstances. Normally,
1690 when pcre_exec() would call its internal match() function recursively and
1691 immediately return the result unconditionally, it uses a "tail recursion"
1692 feature to save stack. However, when a subpattern that can match an empty
1693 string has an unlimited repetition quantifier, it no longer makes this
1694 optimization. That gives it a stack frame in which to save the data for
1695 checking that an empty string has been matched. Previously this was taken
1696 from the 1000-entry workspace that had been reserved. So now there is no
1697 explicit limit, but more stack is used.
1698
1699 8. Applied Daniel's patches to solve problems with the import/export magic
1700 syntax that is required for Windows, and which was going wrong for the
1701 pcreposix and pcrecpp parts of the library. These were overlooked when this
1702 problem was solved for the main library.
1703
1704 9. There were some crude static tests to avoid integer overflow when computing
1705 the size of patterns that contain repeated groups with explicit upper
1706 limits. As the maximum quantifier is 65535, the maximum group length was
1707 set at 30,000 so that the product of these two numbers did not overflow a
1708 32-bit integer. However, it turns out that people want to use groups that
1709 are longer than 30,000 bytes (though not repeat them that many times).
1710 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1711 made it possible to implement the integer overflow checks in a much more
1712 dynamic way, which I have now done. The artificial limitation on group
1713 length has been removed - we now have only the limit on the total length of
1714 the compiled pattern, which depends on the LINK_SIZE setting.
1715
1716 10. Fixed a bug in the documentation for get/copy named substring when
1717 duplicate names are permitted. If none of the named substrings are set, the
1718 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1719 empty string.
1720
1721 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1722 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1723 because the ] is interpreted as the first data character and the
1724 terminating ] is not found. PCRE has been made compatible with Perl in this
1725 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1726 cause memory overwriting.
1727
1728 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1729 string has been matched (to stop an infinite loop). It was not recognizing
1730 a conditional subpattern that could match an empty string if that
1731 subpattern was within another subpattern. For example, it looped when
1732 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1733 condition was not nested. This bug has been fixed.
1734
1735 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1736 past the start of the subject in the presence of bytes with the top bit
1737 set, for example "\x8aBCD".
1738
1739 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1740 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1741
1742 14. Optimized (?!) to (*FAIL).
1743
1744 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1745 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1746 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1747 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1748 does: it's just the validity check that is more restrictive.
1749
1750 16. Inserted checks for integer overflows during escape sequence (backslash)
1751 processing, and also fixed erroneous offset values for syntax errors during
1752 backslash processing.
1753
1754 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1755 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1756
1757 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1758 caused an overrun.
1759
1760 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1761 something other than just ASCII characters) inside a group that had an
1762 unlimited repeat caused a loop at compile time (while checking to see
1763 whether the group could match an empty string).
1764
1765 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1766 [\P{Any}] did so. (Error in the code for printing property names.)
1767
1768 21. An orphan \E inside a character class could cause a crash.
1769
1770 22. A repeated capturing bracket such as (A)? could cause a wild memory
1771 reference during compilation.
1772
1773 23. There are several functions in pcre_compile() that scan along a compiled
1774 expression for various reasons (e.g. to see if it's fixed length for look
1775 behind). There were bugs in these functions when a repeated \p or \P was
1776 present in the pattern. These operators have additional parameters compared
1777 with \d, etc, and these were not being taken into account when moving along
1778 the compiled data. Specifically:
1779
1780 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1781 length.
1782
1783 (b) An item such as \pL+ within a repeated group could cause crashes or
1784 loops.
1785
1786 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1787 "reference to non-existent subpattern" error.
1788
1789 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1790
1791 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1792 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1793
1794 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1795
1796 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1797 character were causing crashes (broken optimization).
1798
1799 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1800 \p or \P) caused a compile-time loop.
1801
1802 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1803 break. For example, the unstudied pattern [\r\n]A does not match the string
1804 "\r\nA" because change 7.0/46 below moves the current point on by two
1805 characters after failing to match at the start. However, the pattern \nA
1806 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1807 the same is true. There doesn't seem any very clean way out of this, but
1808 what I have chosen to do makes the common cases work: PCRE now takes note
1809 of whether there can be an explicit match for \r or \n anywhere in the
1810 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1811 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1812 pattern has explicit CR or LF references.
1813
1814 29. Added (*CR) etc for changing newline setting at start of pattern.
1815
1816
1817 Version 7.2 19-Jun-07
1818 ---------------------
1819
1820 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1821 which is apparently normally available under Windows.
1822
1823 2. Re-jig the pcregrep tests with different newline settings in an attempt
1824 to make them independent of the local environment's newline setting.
1825
1826 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1827
1828 4. Some of the "internals" tests were previously cut out when the link size
1829 was not 2, because the output contained actual offsets. The recent new
1830 "Z" feature of pcretest means that these can be cut out, making the tests
1831 usable with all link sizes.
1832
1833 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1834 stack recursion. This gives a massive performance boost under BSD, but just
1835 a small improvement under Linux. However, it saves one field in the frame
1836 in all cases.
1837
1838 6. Added more features from the forthcoming Perl 5.10:
1839
1840 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1841 recursion call. It refers to the nth most recently opened parentheses.
1842
1843 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1844 to be opened parentheses.
1845
1846 (c) Conditions that refer to capturing parentheses can be specified
1847 relatively, for example, (?(-2)... or (?(+3)...
1848
1849 (d) \K resets the start of the current match so that everything before
1850 is not part of it.
1851
1852 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1853
1854 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1855 reference syntax.
1856
1857 (g) (?| introduces a group in which the numbering of parentheses in each
1858 alternative starts with the same number.
1859
1860 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1861
1862 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1863 PCRE_INFO_JCHANGED.
1864
1865 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1866 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1867 for detecting groups that can match an empty string.
1868
1869 9. A pattern with a very large number of alternatives (more than several
1870 hundred) was running out of internal workspace during the pre-compile
1871 phase, where pcre_compile() figures out how much memory will be needed. A
1872 bit of new cunning has reduced the workspace needed for groups with
1873 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1874 workspace instead of running out of the 4096 that are available.
1875
1876 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1877
1878 11. Applied patch from Google to remove an optimization that didn't quite work.
1879 The report of the bug said:
1880
1881 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1882 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1883 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1884
1885 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1886 it matched the wrong number of bytes.
1887
1888
1889 Version 7.1 24-Apr-07
1890 ---------------------
1891
1892 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1893 that is more "standard", making use of automake and other Autotools. There
1894 is some re-arrangement of the files and adjustment of comments consequent
1895 on this.
1896
1897 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1898 for recursive directory scanning broke on some systems because the files
1899 are not scanned in any specific order and on different systems the order
1900 was different. A call to "sort" has been inserted into RunGrepTest for the
1901 approprate test as a short-term fix. In the longer term there may be an
1902 alternative.
1903
1904 3. I had an email from Eric Raymond about problems translating some of PCRE's
1905 man pages to HTML (despite the fact that I distribute HTML pages, some
1906 people do their own conversions for various reasons). The problems
1907 concerned the use of low-level troff macros .br and .in. I have therefore
1908 removed all such uses from the man pages (some were redundant, some could
1909 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1910 HTML has been updated to handle .nf/.fi and to complain if it encounters
1911 .br or .in.
1912
1913 4. Updated comments in configure.ac that get placed in config.h.in and also
1914 arranged for config.h to be included in the distribution, with the name
1915 config.h.generic, for the benefit of those who have to compile without
1916 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1917
1918 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1919 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1920 makevp.bat for the current PCRE, using the additional files
1921 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1922
1923 6. A Windows user reported a minor discrepancy with test 2, which turned out
1924 to be caused by a trailing space on an input line that had got lost in his
1925 copy. The trailing space was an accident, so I've just removed it.
1926
1927 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1928 that is needed.
1929
1930 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1931 as "const" (a) because they are and (b) because it helps the PHP
1932 maintainers who have recently made a script to detect big data structures
1933 in the php code that should be moved to the .rodata section. I remembered
1934 to update Builducptable as well, so it won't revert if ucptable.h is ever
1935 re-created.
1936
1937 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1938 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1939 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1940 support is not required. This saves 1.5-2K of code, which is important in
1941 some applications.
1942
1943 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1944 so as not to refer to the tables, even though these functions will never be
1945 called when UTF-8 support is disabled. Otherwise there are problems with a
1946 shared library.
1947
1948 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1949
1950 (a) It was defining its arguments as char * instead of void *.
1951
1952 (b) It was assuming that all moves were upwards in memory; this was true
1953 a long time ago when I wrote it, but is no longer the case.
1954
1955 The emulated memove() is provided for those environments that have neither
1956 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1957 is clearly not the case, as these two bugs were recently reported.
1958
1959 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1960 and Detrail to create the HTML documentation, the .txt form of the man
1961 pages, and it removes trailing spaces from listed files. It also creates
1962 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1963 case, it wraps all the #defines with #ifndefs. This script should be run
1964 before "make dist".
1965
1966 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1967 with Unicode property support.
1968
1969 (a) For a maximizing quantifier, if the two different cases of the
1970 character were of different lengths in their UTF-8 codings (there are
1971 some cases like this - I found 11), and the matching function had to
1972 back up over a mixture of the two cases, it incorrectly assumed they
1973 were both the same length.
1974
1975 (b) When PCRE was configured to use the heap rather than the stack for
1976 recursion during matching, it was not correctly preserving the data for
1977 the other case of a UTF-8 character when checking ahead for a match
1978 while processing a minimizing repeat. If the check also involved
1979 matching a wide character, but failed, corruption could cause an
1980 erroneous result when trying to check for a repeat of the original
1981 character.
1982
1983 13. Some tidying changes to the testing mechanism:
1984
1985 (a) The RunTest script now detects the internal link size and whether there
1986 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1987 values substituted by "configure". (The RunGrepTest script already did
1988 this for UTF-8.) The configure.ac script no longer substitutes the
1989 relevant variables.
1990
1991 (b) The debugging options /B and /D in pcretest show the compiled bytecode
1992 with length and offset values. This means that the output is different
1993 for different internal link sizes. Test 2 is skipped for link sizes
1994 other than 2 because of this, bypassing the problem. Unfortunately,
1995 there was also a test in test 3 (the locale tests) that used /B and
1996 failed for link sizes other than 2. Rather than cut the whole test out,
1997 I have added a new /Z option to pcretest that replaces the length and
1998 offset values with spaces. This is now used to make test 3 independent
1999 of link size. (Test 2 will be tidied up later.)
2000
2001 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2002 segmentation fault instead of returning the appropriate error message.
2003
2004 15. In multiline mode when the newline sequence was set to "any", the pattern
2005 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2006 This doesn't seem right; it now treats the CRLF combination as the line
2007 ending, and so does not match in that case. It's only a pattern such as ^$
2008 that would hit this one: something like ^ABC$ would have failed after \r
2009 and then tried again after \r\n.
2010
2011 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2012 in an attempt to make files that differ only in their line terminators
2013 compare equal. This works on Linux.
2014
2015 17. Under certain error circumstances pcregrep might try to free random memory
2016 as it exited. This is now fixed, thanks to valgrind.
2017
2018 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2019 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2020 was because its rules for how to advance for /g after matching an empty
2021 string at the end of a line did not allow for this case. They now check for
2022 it specially.
2023
2024 20. pcretest is supposed to handle patterns and data of any length, by
2025 extending its buffers when necessary. It was getting this wrong when the
2026 buffer for a data line had to be extended.
2027
2028 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2029 CRLF as a newline sequence.
2030
2031 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2032 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2033 I have nevertheless tidied it up.
2034
2035 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2036
2037 24. Added a man page for pcre-config.
2038
2039
2040 Version 7.0 19-Dec-06
2041 ---------------------
2042
2043 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2044 moving to gcc 4.1.1.
2045
2046 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2047 sys/time.h, which is documented as needed for this function. It doesn't
2048 seem to matter on Linux, but it showed up on some releases of OS X.
2049
2050 3. It seems that there are systems where bytes whose values are greater than
2051 127 match isprint() in the "C" locale. The "C" locale should be the
2052 default when a C program starts up. In most systems, only ASCII printing
2053 characters match isprint(). This difference caused the output from pcretest
2054 to vary, making some of the tests fail. I have changed pcretest so that:
2055
2056 (a) When it is outputting text in the compiled version of a pattern, bytes
2057 other than 32-126 are always shown as hex escapes.
2058
2059 (b) When it is outputting text that is a matched part of a subject string,
2060 it does the same, unless a different locale has been set for the match
2061 (using the /L modifier). In this case, it uses isprint() to decide.
2062
2063 4. Fixed a major bug that caused incorrect computation of the amount of memory
2064 required for a compiled pattern when options that changed within the
2065 pattern affected the logic of the preliminary scan that determines the
2066 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2067 that the computed length was too small. The symptoms of this bug were
2068 either the PCRE error "internal error: code overflow" from pcre_compile(),
2069 or a glibc crash with a message such as "pcretest: free(): invalid next
2070 size (fast)". Examples of patterns that provoked this bug (shown in
2071 pcretest format) are:
2072
2073 /(?-x: )/x
2074 /(?x)(?-x: \s*#\s*)/
2075 /((?i)[\x{c0}])/8
2076 /(?i:[\x{c0}])/8
2077
2078 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2079 is now done differently.
2080
2081 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2082 wrapper classes; (b) implement a new function in the C++ scanner that is
2083 more efficient than the old way of doing things because it avoids levels of
2084 recursion in the regex matching; (c) add a paragraph to the documentation
2085 for the FullMatch() function.
2086
2087 6. The escape sequence \n was being treated as whatever was defined as
2088 "newline". Not only was this contrary to the documentation, which states
2089 that \n is character 10 (hex 0A), but it also went horribly wrong when
2090 "newline" was defined as CRLF. This has been fixed.
2091
2092 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2093 was being set to -1 for the "end of line" case (supposedly a value that no
2094 character can have). Though this value is never used (the check for end of
2095 line is "zero bytes in current character"), it caused compiler complaints.
2096 I've changed it to 0xffffffff.
2097
2098 8. In pcre_version.c, the version string was being built by a sequence of
2099 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2100 string (as it is for production releases) called a macro with an empty
2101 argument. The C standard says the result of this is undefined. The gcc
2102 compiler treats it as an empty string (which was what was wanted) but it is
2103 reported that Visual C gives an error. The source has been hacked around to
2104 avoid this problem.
2105
2106 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2107 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2108 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2109 of them did).
2110
2111 10. Originally, pcretest opened its input and output without "b"; then I was
2112 told that "b" was needed in some environments, so it was added for release
2113 5.0 to both the input and output. (It makes no difference on Unix-like
2114 systems.) Later I was told that it is wrong for the input on Windows. I've
2115 now abstracted the modes into two macros, to make it easier to fiddle with
2116 them, and removed "b" from the input mode under Windows.
2117
2118 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2119
2120 12. Added -help and --help to pcretest as an official way of being reminded
2121 of the options.
2122
2123 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2124 and pcrecpp.cc because they annoy compilers at high warning levels.
2125
2126 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2127
2128 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2129 scripts are not C programs :-) and which was not noticed because it works
2130 on Linux.
2131
2132 16. pcretest is supposed to handle any length of pattern and data line (as one
2133 line or as a continued sequence of lines) by extending its input buffer if
2134 necessary. This feature was broken for very long pattern lines, leading to
2135 a string of junk being passed to pcre_compile() if the pattern was longer
2136 than about 50K.
2137
2138 17. I have done a major re-factoring of the way pcre_compile() computes the
2139 amount of memory needed for a compiled pattern. Previously, there was code
2140 that made a preliminary scan of the pattern in order to do this. That was
2141 OK when PCRE was new, but as the facilities have expanded, it has become
2142 harder and harder to keep it in step with the real compile phase, and there
2143 have been a number of bugs (see for example, 4 above). I have now found a
2144 cunning way of running the real compile function in a "fake" mode that
2145 enables it to compute how much memory it would need, while actually only
2146 ever using a few hundred bytes of working memory and without too many
2147 tests of the mode. This should make future maintenance and development
2148 easier. A side effect of this work is that the limit of 200 on the nesting
2149 depth of parentheses has been removed (though this was never a serious
2150 limitation, I suspect). However, there is a downside: pcre_compile() now
2151 runs more slowly than before (30% or more, depending on the pattern). I
2152 hope this isn't a big issue. There is no effect on runtime performance.
2153
2154 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2155 newline (only possible for the last line of a file) and it was a
2156 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2157
2158 19. Added additional timing features to pcretest. (1) The -tm option now times
2159 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2160 separate command line item, by a number that specifies the number of
2161 repeats to use when timing. The default is 50000; this gives better
2162 precision, but takes uncomfortably long for very large patterns.
2163
2164 20. Extended pcre_study() to be more clever in cases where a branch of a
2165 subpattern has no definite first character. For example, (a*|b*)[cd] would
2166 previously give no result from pcre_study(). Now it recognizes that the
2167 first character must be a, b, c, or d.
2168
2169 21. There was an incorrect error "recursive call could loop indefinitely" if
2170 a subpattern (or the entire pattern) that was being tested for matching an
2171 empty string contained only one non-empty item after a nested subpattern.
2172 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2173 incorrectly, because the \d was being skipped in the check.
2174
2175 22. The pcretest program now has a new pattern option /B and a command line
2176 option -b, which is equivalent to adding /B to every pattern. This causes
2177 it to show the compiled bytecode, without the additional information that
2178 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2179 is the same as /B/I).
2180
2181 23. A new optimization is now able automatically to treat some sequences such
2182 as a*b as a*+b. More specifically, if something simple (such as a character
2183 or a simple class like \d) has an unlimited quantifier, and is followed by
2184 something that cannot possibly match the quantified thing, the quantifier
2185 is automatically "possessified".
2186
2187 24. A recursive reference to a subpattern whose number was greater than 39
2188 went wrong under certain circumstances in UTF-8 mode. This bug could also
2189 have affected the operation of pcre_study().
2190
2191 25. Realized that a little bit of performance could be had by replacing
2192 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2193
2194 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2195
2196 27. Possessive quantifiers such as a++ were previously implemented by turning
2197 them into atomic groups such as ($>a+). Now they have their own opcodes,
2198 which improves performance. This includes the automatically created ones
2199 from 23 above.
2200
2201 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2202 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2203 the first matched character to be a colon. This applied both to named and
2204 numbered groups.
2205
2206 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2207
2208 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2209 building PCRE on Windows easier, so I have included it in the distribution.
2210
2211 31. There is now a check in pcretest against a ridiculously large number being
2212 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2213 loop, the loop is abandoned.
2214
2215 32. Forward references to subpatterns in conditions such as (?(2)...) where
2216 subpattern 2 is defined later cause pcre_compile() to search forwards in
2217 the pattern for the relevant set of parentheses. This search went wrong
2218 when there were unescaped parentheses in a character class, parentheses
2219 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2220
2221 33. "Subroutine" calls and backreferences were previously restricted to
2222 referencing subpatterns earlier in the regex. This restriction has now
2223 been removed.
2224
2225 34. Added a number of extra features that are going to be in Perl 5.10. On the
2226 whole, these are just syntactic alternatives for features that PCRE had
2227 previously implemented using the Python syntax or my own invention. The
2228 other formats are all retained for compatibility.
2229
2230 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2231 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2232 also .NET compatible.
2233
2234 (b) A recursion or subroutine call to a named group can now be defined as
2235 (?&name) as well as (?P>name).
2236
2237 (c) A backreference to a named group can now be defined as \k<name> or
2238 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2239 5.10, are also .NET compatible.
2240
2241 (d) A conditional reference to a named group can now use the syntax
2242 (?(<name>) or (?('name') as well as (?(name).
2243
2244 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2245 groups (named and numbered) that are never evaluated inline, but can be
2246 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2247 is always false. There may be only one alternative in such a group.
2248
2249 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2250 as the simple (?(R). The condition is true only if the most recent
2251 recursion is that of the given number or name. It does not search out
2252 through the entire recursion stack.
2253
2254 (g) The escape \gN or \g{N} has been added, where N is a positive or
2255 negative number, specifying an absolute or relative reference.
2256
2257 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2258 some "unreachable code" warnings.
2259
2260 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2261 things, this adds five new scripts.
2262
2263 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2264 There were also incompatibilities regarding the handling of \Q..\E inside
2265 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2266 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2267
2268 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2269 matches an empty string, and forcibly breaks the loop. There were bugs in
2270 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2271 against aaaa the result was just "a" rather than "aaaa", for example. Two
2272 separate and independent bugs (that affected different cases) have been
2273 fixed.
2274
2275 39. Refactored the code to abolish the use of different opcodes for small
2276 capturing bracket numbers. This is a tidy that I avoided doing when I
2277 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2278 The new approach is not only tidier, it makes it possible to reduce the
2279 memory needed to fix the previous bug (38).
2280
2281 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2282 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2283 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2284 mode.
2285
2286 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2287 report.
2288
2289 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2290 copy construction and assignment in the C++ wrapper.
2291
2292 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2293 couple of bugs that could have given wrong results in the "--newline=crlf"
2294 case.
2295
2296 44. Added a number of casts and did some reorganization of signed/unsigned int
2297 variables following suggestions from Dair Grant. Also renamed the variable
2298 "this" as "item" because it is a C++ keyword.
2299
2300 45. Arranged for dftables to add
2301
2302 #include "pcre_internal.h"
2303
2304 to pcre_chartables.c because without it, gcc 4.x may remove the array
2305 definition from the final binary if PCRE is built into a static library and
2306 dead code stripping is activated.
2307
2308 46. For an unanchored pattern, if a match attempt fails at the start of a
2309 newline sequence, and the newline setting is CRLF or ANY, and the next two
2310 characters are CRLF, advance by two characters instead of one.
2311
2312
2313 Version 6.7 04-Jul-06
2314 ---------------------
2315
2316 1. In order to handle tests when input lines are enormously long, pcretest has
2317 been re-factored so that it automatically extends its buffers when
2318 necessary. The code is crude, but this _is_ just a test program. The
2319 default size has been increased from 32K to 50K.
2320
2321 2. The code in pcre_study() was using the value of the re argument before
2322 testing it for NULL. (Of course, in any sensible call of the function, it
2323 won't be NULL.)
2324
2325 3. The memmove() emulation function in pcre_internal.h, which is used on
2326 systems that lack both memmove() and bcopy() - that is, hardly ever -
2327 was missing a "static" storage class specifier.
2328
2329 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2330 containing an extended class (one that cannot be represented by a bitmap
2331 because it contains high-valued characters or Unicode property items, e.g.
2332 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2333 pattern, but PCRE should not loop if you do not (it no longer does).
2334 [Detail: two cases were found: (a) a repeated subpattern containing an
2335 extended class; (b) a recursive reference to a subpattern that followed a
2336 previous extended class. It wasn't skipping over the extended class
2337 correctly when UTF-8 mode was not set.]
2338
2339 5. A negated single-character class was not being recognized as fixed-length
2340 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2341 compile error "lookbehind assertion is not fixed length".
2342
2343 6. The RunPerlTest auxiliary script was showing an unexpected difference
2344 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2345 write a Perl script that can interpret lines of an input file either as
2346 byte characters or as UTF-8, which is what "perltest" was being required to
2347 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2348 can't do is switch easily at run time between having the "use utf8;" pragma
2349 or not. In the end, I fudged it by using the RunPerlTest script to insert
2350 "use utf8;" explicitly for the UTF-8 tests.
2351
2352 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2353 the end of the subject string, contrary to the documentation and to what
2354 Perl does. This was true of both matching functions. Now it matches only at
2355 the start of the subject and immediately after *internal* newlines.
2356
2357 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2358 a pointer to an int instead of a pointer to an unsigned long int. This
2359 caused problems on 64-bit systems.
2360
2361 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2362 instance of the 'standard' template library not being so standard".
2363
2364 10. There was no check on the number of named subpatterns nor the maximum
2365 length of a subpattern name. The product of these values is used to compute
2366 the size of the memory block for a compiled pattern. By supplying a very
2367 long subpattern name and a large number of named subpatterns, the size
2368 computation could be caused to overflow. This is now prevented by limiting
2369 the length of names to 32 characters, and the number of named subpatterns
2370 to 10,000.
2371
2372 11. Subpatterns that are repeated with specific counts have to be replicated in
2373 the compiled pattern. The size of memory for this was computed from the
2374 length of the subpattern and the repeat count. The latter is limited to
2375 65535, but there was no limit on the former, meaning that integer overflow
2376 could in principle occur. The compiled length of a repeated subpattern is
2377 now limited to 30,000 bytes in order to prevent this.
2378
2379 12. Added the optional facility to have named substrings with the same name.
2380
2381 13. Added the ability to use a named substring as a condition, using the
2382 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2383 are numbers (not recommended). Forward references are permitted.
2384
2385 14. Added forward references in named backreferences (if you see what I mean).
2386
2387 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2388 pattern could run off the end of the subject. For example, the pattern
2389 "(?s)(.{1,5})"8 did this with the subject "ab".
2390
2391 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2392 PCRE_CASELESS was set when matching characters that were quantified with ?
2393 or *.
2394
2395 17. A character class other than a single negated character that had a minimum
2396 but no maximum quantifier - for example [ab]{6,} - was not handled
2397 correctly by pce_dfa_exec(). It would match only one character.
2398
2399 18. A valid (though odd) pattern that looked like a POSIX character
2400 class but used an invalid character after [ (for example [[,abc,]]) caused
2401 pcre_compile() to give the error "Failed: internal error: code overflow" or
2402 in some cases to crash with a glibc free() error. This could even happen if
2403 the pattern terminated after [[ but there just happened to be a sequence of
2404 letters, a binary zero, and a closing ] in the memory that followed.
2405
2406 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2407 over the years. Originally (before any Unicode support), just the bottom 8
2408 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2409 output from "man perlunicode" includes this:
2410
2411 The regular expression compiler produces polymorphic opcodes. That
2412 is, the pattern adapts to the data and automatically switches to
2413 the Unicode character scheme when presented with Unicode data--or
2414 instead uses a traditional byte scheme when presented with byte
2415 data.
2416
2417 Sadly, a wide octal escape does not cause a switch, and in a string with
2418 no other multibyte characters, these octal escapes are treated as before.
2419 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2420 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2421 Unicode string.
2422
2423 I have not perpetrated such confusion in PCRE. Up till now, it took just
2424 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2425 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2426 translate to the appropriate multibyte character.
2427
2428 29. Applied some refactoring to reduce the number of warnings from Microsoft
2429 and Borland compilers. This has included removing the fudge introduced
2430 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2431 a warning about an unused variable.
2432
2433 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2434 characters since release 4.0, because Perl (from release 5.004) does not.
2435 [Or at least, is documented not to: some releases seem to be in conflict
2436 with the documentation.] However, when a pattern was studied with
2437 pcre_study() and all its branches started with \s, PCRE still included VT
2438 as a possible starting character. Of course, this did no harm; it just
2439 caused an unnecessary match attempt.
2440
2441 22. Removed a now-redundant internal flag bit that recorded the fact that case
2442 dependency changed within the pattern. This was once needed for "required
2443 byte" processing, but is no longer used. This recovers a now-scarce options
2444 bit. Also moved the least significant internal flag bit to the most-
2445 significant bit of the word, which was not previously used (hangover from
2446 the days when it was an int rather than a uint) to free up another bit for
2447 the future.
2448
2449 23. Added support for CRLF line endings as well as CR and LF. As well as the
2450 default being selectable at build time, it can now be changed at runtime
2451 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2452 specify that it is scanning data with non-default line endings.
2453
2454 24. Changed the definition of CXXLINK to make it agree with the definition of
2455 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2456
2457 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2458 recursions. This makes a big different to stack usage for some patterns.
2459
2460 26. If a subpattern containing a named recursion or subroutine reference such
2461 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2462 the space required for the compiled pattern went wrong and gave too small a
2463 value. Depending on the environment, this could lead to "Failed: internal
2464 error: code overflow at offset 49" or "glibc detected double free or
2465 corruption" errors.
2466
2467 27. Applied patches from Google (a) to support the new newline modes and (b) to
2468 advance over multibyte UTF-8 characters in GlobalReplace.
2469
2470 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2471 difference for some implementation of PCRE in some Windows version.
2472
2473 29. Added some extra testing facilities to pcretest:
2474
2475 \q<number> in a data line sets the "match limit" value
2476 \Q<number> in a data line sets the "match recursion limt" value
2477 -S <number> sets the stack size, where <number> is in megabytes
2478
2479 The -S option isn't available for Windows.
2480
2481
2482 Version 6.6 06-Feb-06
2483 ---------------------
2484
2485 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2486 in pcreposix.h. I have copied the definition from pcre.h.
2487
2488 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2489 because pcre.h is no longer a built file.
2490
2491 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2492 not normally included in the compiled code.
2493
2494
2495 Version 6.5 01-Feb-06
2496 ---------------------
2497
2498 1. When using the partial match feature with pcre_dfa_exec(), it was not
2499 anchoring the second and subsequent partial matches at the new starting
2500 point. This could lead to incorrect results. For example, with the pattern
2501 /1234/, partially matching against "123" and then "a4" gave a match.
2502
2503 2. Changes to pcregrep:
2504
2505 (a) All non-match returns from pcre_exec() were being treated as failures
2506 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2507 error message is output. Some extra information is given for the
2508 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2509 probably the only errors that are likely to be caused by users (by
2510 specifying a regex that has nested indefinite repeats, for instance).
2511 If there are more than 20 of these errors, pcregrep is abandoned.
2512
2513 (b) A binary zero was treated as data while matching, but terminated the
2514 output line if it was written out. This has been fixed: binary zeroes
2515 are now no different to any other data bytes.
2516
2517 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2518 used to set a locale for matching. The --locale=xxxx long option has
2519 been added (no short equivalent) to specify a locale explicitly on the
2520 pcregrep command, overriding the environment variables.
2521
2522 (d) When -B was used with -n, some line numbers in the output were one less
2523 than they should have been.
2524
2525 (e) Added the -o (--only-matching) option.
2526
2527 (f) If -A or -C was used with -c (count only), some lines of context were
2528 accidentally printed for the final match.
2529
2530 (g) Added the -H (--with-filename) option.
2531
2532 (h) The combination of options -rh failed to suppress file names for files
2533 that were found from directory arguments.
2534
2535 (i) Added the -D (--devices) and -d (--directories) options.
2536
2537 (j) Added the -F (--fixed-strings) option.
2538
2539 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2540
2541 (l) Added the --colo(u)r option.
2542
2543 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2544 is not present by default.
2545
2546 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2547 items such as (?R) or (?1), when the recursion could match a number of
2548 alternatives. If it matched one of the alternatives, but subsequently,
2549 outside the recursion, there was a failure, the code tried to back up into
2550 the recursion. However, because of the way PCRE is implemented, this is not
2551 possible, and the result was an incorrect result from the match.
2552
2553 In order to prevent this happening, the specification of recursion has
2554 been changed so that all such subpatterns are automatically treated as
2555 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2556
2557 4. I had overlooked the fact that, in some locales, there are characters for
2558 which isalpha() is true but neither isupper() nor islower() are true. In
2559 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2560 and ordfeminine) are like this. This affected the treatment of \w and \W
2561 when they appeared in character classes, but not when they appeared outside
2562 a character class. The bit map for "word" characters is now created
2563 separately from the results of isalnum() instead of just taking it from the
2564 upper, lower, and digit maps. (Plus the underscore character, of course.)
2565
2566 5. The above bug also affected the handling of POSIX character classes such as
2567 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2568 permanent tables. Instead, the bit maps for such a class were previously
2569 created as the appropriate unions of the upper, lower, and digit bitmaps.
2570 Now they are created by subtraction from the [[:word:]] class, which has
2571 its own bitmap.
2572
2573 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2574 It is created by subtracting the vertical space characters (\x09, \x0a,
2575 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2576 subtraction was done in the overall bitmap for a character class, meaning
2577 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2578 be recognized. This bug has been fixed.
2579
2580 7. Patches from the folks at Google:
2581
2582 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2583 real life, but is still worth protecting against".
2584
2585 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2586 regular expressions".
2587
2588 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2589 have it.
2590
2591 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2592 "configure" and the latter not, in order to fix a problem somebody had
2593 with compiling the Arg class on HP-UX.
2594
2595 (e) Improve the error-handling of the C++ wrapper a little bit.
2596
2597 (f) New tests for checking recursion limiting.
2598
2599 8. The pcre_memmove() function, which is used only if the environment does not
2600 have a standard memmove() function (and is therefore rarely compiled),
2601 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2602 returning a result (though PCRE never actually uses the result).
2603
2604 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2605 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2606 returned instead of calling malloc() with an overflowing number that would
2607 most likely cause subsequent chaos.
2608
2609 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2610
2611 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2612 with this option is matched, the nmatch and pmatch options of regexec() are
2613 ignored.
2614
2615 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2616 provided in case anyone wants to the the POSIX interface with UTF-8
2617 strings.
2618
2619 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2620 C++ linking (needed for some HP-UX environments).
2621
2622 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2623 (unused parameter) and in the pcre_printint() function (omitted "default"
2624 switch label when the default is to do nothing).
2625
2626 15. Added some code to make it possible, when PCRE is compiled as a C++
2627 library, to replace subject pointers for pcre_exec() with a smart pointer
2628 class, thus making it possible to process discontinuous strings.
2629
2630 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2631 much the same function. They were added by different people who were trying
2632 to make PCRE easy to compile on non-Unix systems. It has been suggested
2633 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2634 for compiling on Windows systems. I have therefore replaced it with
2635 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2636 defaults to "extern" for C or "extern C" for C++, which works fine on
2637 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2638 SCOPE with something explicit in config.h. In addition:
2639
2640 (a) pcreposix.h still had just "extern" instead of either of these macros;
2641 I have replaced it with PCRE_DATA_SCOPE.
2642
2643 (b) Functions such as _pcre_xclass(), which are internal to the library,
2644 but external in the C sense, all had PCRE_EXPORT in their definitions.
2645 This is apparently wrong for the Windows case, so I have removed it.
2646 (It makes no difference on Unix-like systems.)
2647
2648 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2649 of recursive calls to match(). This is different to MATCH_LIMIT because
2650 that limits the total number of calls to match(), not all of which increase
2651 the depth of recursion. Limiting the recursion depth limits the amount of
2652 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2653 when PCRE is compiled, and changed at run time. A patch from Google adds
2654 this functionality to the C++ interface.
2655
2656 18. Changes to the handling of Unicode character properties:
2657
2658 (a) Updated the table to Unicode 4.1.0.
2659
2660 (b) Recognize characters that are not in the table as "Cn" (undefined).
2661
2662 (c) I revised the way the table is implemented to a much improved format
2663 which includes recognition of ranges. It now supports the ranges that
2664 are defined in UnicodeData.txt, and it also amalgamates other
2665 characters into ranges. This has reduced the number of entries in the
2666 table from around 16,000 to around 3,000, thus reducing its size
2667 considerably. I realized I did not need to use a tree structure after
2668 all - a binary chop search is just as efficient. Having reduced the
2669 number of entries, I extended their size from 6 bytes to 8 bytes to
2670 allow for more data.
2671
2672 (d) Added support for Unicode script names via properties such as \p{Han}.
2673
2674 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2675 matching that character.
2676
2677 20. When matching a repeated Unicode property with a minimum greater than zero,
2678 (for example \pL{2,}), PCRE could look past the end of the subject if it
2679 reached it while seeking the minimum number of characters. This could
2680 happen only if some of the characters were more than one byte long, because
2681 there is a check for at least the minimum number of bytes.
2682
2683 21. Refactored the implementation of \p and \P so as to be more general, to
2684 allow for more different types of property in future. This has changed the
2685 compiled form incompatibly. Anybody with saved compiled patterns that use
2686 \p or \P will have to recompile them.
2687
2688 22. Added "Any" and "L&" to the supported property types.
2689
2690 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2691 but give a compile time error if the value is greater than 0xff.
2692
2693 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2694 accidentally not being installed or uninstalled.
2695
2696 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2697 made were to insert the current release number. This seemed silly, because
2698 it made things harder for people building PCRE on systems that don't run
2699 "configure". I have turned pcre.h into a distributed file, no longer built
2700 by "configure", with the version identification directly included. There is
2701 no longer a pcre.h.in file.
2702
2703 However, this change necessitated a change to the pcre-config script as
2704 well. It is built from pcre-config.in, and one of the substitutions was the
2705 release number. I have updated configure.ac so that ./configure now finds
2706 the release number by grepping pcre.h.
2707
2708 26. Added the ability to run the tests under valgrind.
2709
2710
2711 Version 6.4 05-Sep-05
2712 ---------------------
2713
2714 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2715 "--" to be printed when multiple files were scanned, even when none of the
2716 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2717 consider it to be a bug, and have restored the previous behaviour.
2718
2719 2. A couple of code tidies to get rid of compiler warnings.
2720
2721 3. The pcretest program used to cheat by referring to symbols in the library
2722 whose names begin with _pcre_. These are internal symbols that are not
2723 really supposed to be visible externally, and in some environments it is
2724 possible to suppress them. The cheating is now confined to including
2725 certain files from the library's source, which is a bit cleaner.
2726
2727 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2728 file's purpose clearer.
2729
2730 5. Reorganized pcre_ucp_findchar().
2731
2732
2733 Version 6.3 15-Aug-05
2734 ---------------------
2735
2736 1. The file libpcre.pc.in did not have general read permission in the tarball.
2737
2738 2. There were some problems when building without C++ support:
2739
2740 (a) If C++ support was not built, "make install" and "make test" still
2741 tried to test it.
2742
2743 (b) There were problems when the value of CXX was explicitly set. Some
2744 changes have been made to try to fix these, and ...
2745
2746 (c) --disable-cpp can now be used to explicitly disable C++ support.
2747
2748 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2749 backslash in a target when C++ was disabled. This confuses some
2750 versions of "make", apparently. Using an intermediate variable solves
2751 this. (Same for CPP_LOBJ.)
2752
2753 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2754 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2755 necessary on certain architectures.
2756
2757 4. Added a setting of -export-symbols-regex to the link command to remove
2758 those symbols that are exported in the C sense, but actually are local
2759 within the library, and not documented. Their names all begin with
2760 "_pcre_". This is not a perfect job, because (a) we have to except some
2761 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2762 available (and never for static libraries). I have made a note to try to
2763 find a way round (a) in the future.
2764
2765
2766 Version 6.2 01-Aug-05
2767 ---------------------
2768
2769 1. There was no test for integer overflow of quantifier values. A construction
2770 such as {1111111111111111} would give undefined results. What is worse, if
2771 a minimum quantifier for a parenthesized subpattern overflowed and became
2772 negative, the calculation of the memory size went wrong. This could have
2773 led to memory overwriting.
2774
2775 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2776
2777 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2778 operating environments where this matters.
2779
2780 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2781 PCRE options from within the C++ wrapper.
2782
2783 5. Named capturing subpatterns were not being correctly counted when a pattern
2784 was compiled. This caused two problems: (a) If there were more than 100
2785 such subpatterns, the calculation of the memory needed for the whole
2786 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2787 back references of the form \12, where the number was greater than 9, were
2788 not recognized as back references, even though there were sufficient
2789 previous subpatterns.
2790
2791 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2792 versions of gcc, e.g. 2.95.4.
2793
2794
2795 Version 6.1 21-Jun-05
2796 ---------------------
2797
2798 1. There was one reference to the variable "posix" in pcretest.c that was not
2799 surrounded by "#if !defined NOPOSIX".
2800
2801 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2802 the cross-check on the old pcre_info() function, for the benefit of the
2803 cut-down version of PCRE that is currently imported into Exim.
2804
2805 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2806 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2807 patterns that start (?i) but I don't think that matters. The use of (?i) is
2808 just an example; this all applies to the other options as well.
2809
2810 4. Since libtool seems to echo the compile commands it is issuing, the output
2811 from "make" can be reduced a bit by putting "@" in front of each libtool
2812 compile command.
2813
2814 5. Patch from the folks at Google for configure.in to be a bit more thorough
2815 in checking for a suitable C++ installation before trying to compile the
2816 C++ stuff. This should fix a reported problem when a compiler was present,
2817 but no suitable headers.
2818
2819 6. The man pages all had just "PCRE" as their title. I have changed them to
2820 be the relevant file name. I have also arranged that these names are
2821 retained in the file doc/pcre.txt, which is a concatenation in text format
2822 of all the man pages except the little individual ones for each function.
2823
2824 7. The NON-UNIX-USE file had not been updated for the different set of source
2825 files that come with release 6. I also added a few comments about the C++
2826 wrapper.
2827
2828
2829 Version 6.0 07-Jun-05
2830 ---------------------
2831
2832 1. Some minor internal re-organization to help with my DFA experiments.
2833
2834 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2835 didn't matter for the library itself when fully configured, but did matter
2836 when compiling without UCP support, or within Exim, where the ucp files are
2837 not imported.
2838
2839 3. Refactoring of the library code to split up the various functions into
2840 different source modules. The addition of the new DFA matching code (see
2841 below) to a single monolithic source would have made it really too
2842 unwieldy, quite apart from causing all the code to be include in a
2843 statically linked application, when only some functions are used. This is
2844 relevant even without the DFA addition now that patterns can be compiled in
2845 one application and matched in another.
2846
2847 The downside of splitting up is that there have to be some external
2848 functions and data tables that are used internally in different modules of
2849 the library but which are not part of the API. These have all had their
2850 names changed to start with "_pcre_" so that they are unlikely to clash
2851 with other external names.
2852
2853 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2854 a different (DFA) algorithm. Although it is slower than the original
2855 function, it does have some advantages for certain types of matching
2856 problem.
2857
2858 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2859 including restarting after a partial match.
2860
2861 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2862 defined when compiling for Windows was sent to me. I have put it into the
2863 code, though I have no means of testing or verifying it.
2864
2865 7. Added the pcre_refcount() auxiliary function.
2866
2867 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2868 match before or at the first newline in the subject string. In pcretest,
2869 the /f option on a pattern can be used to set this.
2870
2871 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2872 would behave wrongly. This has been present in PCRE since release 4.0.
2873
2874 10. A number of changes to the pcregrep command:
2875
2876 (a) Refactored how -x works; insert ^(...)$ instead of setting
2877 PCRE_ANCHORED and checking the length, in preparation for adding
2878 something similar for -w.
2879
2880 (b) Added the -w (match as a word) option.
2881
2882 (c) Refactored the way lines are read and buffered so as to have more
2883 than one at a time available.
2884
2885 (d) Implemented a pcregrep test script.
2886
2887 (e) Added the -M (multiline match) option. This allows patterns to match
2888 over several lines of the subject. The buffering ensures that at least
2889 8K, or the rest of the document (whichever is the shorter) is available
2890 for matching (and similarly the previous 8K for lookbehind assertions).
2891
2892 (f) Changed the --help output so that it now says
2893
2894 -w, --word-regex(p)
2895
2896 instead of two lines, one with "regex" and the other with "regexp"
2897 because that confused at least one person since the short forms are the
2898 same. (This required a bit of code, as the output is generated
2899 automatically from a table. It wasn't just a text change.)
2900
2901 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2902 option but starts with a hyphen. Could be a pattern or a path name
2903 starting with a hyphen, for instance.
2904
2905 (h) "-" can be given as a file name to represent stdin.
2906
2907 (i) When file names are being printed, "(standard input)" is used for
2908 the standard input, for compatibility with GNU grep. Previously
2909 "<stdin>" was used.
2910
2911 (j) The option --label=xxx can be used to supply a name to be used for
2912 stdin when file names are being printed. There is no short form.
2913
2914 (k) Re-factored the options decoding logic because we are going to add
2915 two more options that take data. Such options can now be given in four
2916 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2917
2918 (l) Added the -A, -B, and -C options for requesting that lines of context
2919 around matches be printed.
2920
2921 (m) Added the -L option to print the names of files that do not contain
2922 any matching lines, that is, the complement of -l.
2923
2924 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2925 continue to scan other files.
2926
2927 (o) The -s option was incorrectly implemented. For compatibility with other
2928 greps, it now suppresses the error message for a non-existent or non-
2929 accessible file (but not the return code). There is a new option called
2930 -q that suppresses the output of matching lines, which was what -s was
2931 previously doing.
2932
2933 (p) Added --include and --exclude options to specify files for inclusion
2934 and exclusion when recursing.
2935
2936 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2937 Hopefully, it now does.
2938
2939 12. Missing cast in pcre_study().
2940
2941 13. Added an "uninstall" target to the makefile.
2942
2943 14. Replaced "extern" in the function prototypes in Makefile.in with
2944 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2945 world, but is set differently for Windows.
2946
2947 15. Added a second compiling function called pcre_compile2(). The only
2948 difference is that it has an extra argument, which is a pointer to an
2949 integer error code. When there is a compile-time failure, this is set
2950 non-zero, in addition to the error test pointer being set to point to an
2951 error message. The new argument may be NULL if no error number is required
2952 (but then you may as well call pcre_compile(), which is now just a
2953 wrapper). This facility is provided because some applications need a
2954 numeric error indication, but it has also enabled me to tidy up the way
2955 compile-time errors are handled in the POSIX wrapper.
2956
2957 16. Added VPATH=.libs to the makefile; this should help when building with one
2958 prefix path and installing with another. (Or so I'm told by someone who
2959 knows more about this stuff than I do.)
2960
2961 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2962 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2963 match everything, including newlines. This is not POSIX-compatible, but
2964 somebody wanted the feature. From pcretest it can be activated by using
2965 both the P and the s flags.
2966
2967 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2968
2969 19. libpcre.pc was being incorrectly installed as executable.
2970
2971 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2972 it now also looks for '\r' so that it will work unmodified on Windows.
2973
2974 21. Added Google's contributed C++ wrapper to the distribution.
2975
2976 22. Added some untidy missing memory free() calls in pcretest, to keep
2977 Electric Fence happy when testing.
2978
2979
2980
2981 Version 5.0 13-Sep-04
2982 ---------------------
2983
2984 1. Internal change: literal characters are no longer packed up into items
2985 containing multiple characters in a single byte-string. Each character
2986 is now matched using a separate opcode. However, there may be more than one
2987 byte in the character in UTF-8 mode.
2988
2989 2. The pcre_callout_block structure has two new fields: pattern_position and
2990 next_item_length. These contain the offset in the pattern to the next match
2991 item, and its length, respectively.
2992
2993 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
2994 insertion of callouts before each pattern item. Added the /C option to
2995 pcretest to make use of this.
2996
2997 4. On the advice of a Windows user, the lines
2998
2999 #if defined(_WIN32) || defined(WIN32)
3000 _setmode( _fileno( stdout ), 0x8000 );
3001 #endif /* defined(_WIN32) || defined(WIN32) */
3002
3003 have been added to the source of pcretest. This apparently does useful
3004 magic in relation to line terminators.
3005
3006 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3007 for the benefit of those environments where the "b" makes a difference.
3008
3009 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3010 to know about it. I have put a hack into configure.in that adds in code
3011 to set GCC=yes if CC=icc. This seems to end up at a point in the
3012 generated configure script that is early enough to affect the setting of
3013 compiler options, which is what is needed, but I have no means of testing
3014 whether it really works. (The user who reported this had patched the
3015 generated configure script, which of course I cannot do.)
3016
3017 LATER: After change 22 below (new libtool files), the configure script
3018 seems to know about icc (and also ecc). Therefore, I have commented out
3019 this hack in configure.in.
3020
3021 7. Added support for pkg-config (2 patches were sent in).
3022
3023 8. Negated POSIX character classes that used a combination of internal tables
3024 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3025 [[:^ascii]]. Typically, they would match almost any characters. The other
3026 POSIX classes were not broken in this way.
3027
3028 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3029 to find the match, as PCRE was deluded into thinking that the match had to
3030 start at the start point or following a newline. The same bug applied to
3031 patterns with negative forward assertions or any backward assertions
3032 preceding ".*" at the start, unless the pattern required a fixed first
3033 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3034
3035 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3036 starting at the last subject character, bytes beyond the end of the subject
3037 string were read.
3038
3039 11. Renamed the variable "class" as "classbits" to make life easier for C++
3040 users. (Previously there was a macro definition, but it apparently wasn't
3041 enough.)
3042
3043 12. Added the new field "tables" to the extra data so that tables can be passed
3044 in at exec time, or the internal tables can be re-selected. This allows
3045 a compiled regex to be saved and re-used at a later time by a different
3046 program that might have everything at different addresses.
3047
3048 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3049 -R library as well as a -L library.
3050
3051 14. The debugging options of pcretest (-d on the command line or D on a
3052 pattern) showed incorrect output for anything following an extended class
3053 that contained multibyte characters and which was followed by a quantifier.
3054
3055 15. Added optional support for general category Unicode character properties
3056 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3057 support. It adds about 90K to the size of the library. The meanings of the
3058 inbuilt class escapes such as \d and \s have NOT been changed.
3059
3060 16. Updated pcredemo.c to include calls to free() to release the memory for the
3061 compiled pattern.
3062
3063 17. The generated file chartables.c was being created in the source directory
3064 instead of in the building directory. This caused the build to fail if the
3065 source directory was different from the building directory, and was
3066 read-only.
3067
3068 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3069 file. No doubt somebody will tell me if they don't make sense... Also added
3070 Dan Mooney's comments about building on OpenVMS.
3071
3072 19. Added support for partial matching via the PCRE_PARTIAL option for
3073 pcre_exec() and the \P data escape in pcretest.
3074
3075 20. Extended pcretest with 3 new pattern features:
3076
3077 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3078 write the compiled pattern to the file whose name is "rest-of-line".
3079 This is a straight binary dump of the data, with the saved pointer to
3080 the character tables forced to be NULL. The study data, if any, is
3081 written too. After writing, pcretest reads a new pattern.
3082
3083 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3084 compiled pattern from the given file. There must not be any
3085 occurrences of "<" in the file name (pretty unlikely); if there are,
3086 pcretest will instead treat the initial "<" as a pattern delimiter.
3087 After reading in the pattern, pcretest goes on to read data lines as
3088 usual.
3089
3090 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3091 and 16-bit fields in a compiled pattern, to simulate a pattern that
3092 was compiled on a host of opposite endianness.
3093
3094 21. The pcre-exec() function can now cope with patterns that were compiled on
3095 hosts of opposite endianness, with this restriction:
3096
3097 As for any compiled expression that is saved and used later, the tables
3098 pointer field cannot be preserved; the extra_data field in the arguments
3099 to pcre_exec() should be used to pass in a tables address if a value
3100 other than the default internal tables were used at compile time.
3101
3102 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3103 now diagnosed as an error. Previously, most of the time, a negative number
3104 would have been treated as zero, but if in addition "ovector" was passed as
3105 NULL, a crash could occur.
3106
3107 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3108 new versions from the libtool 1.5 distribution (the last one is a copy of
3109 a file called libtool.m4). This seems to have fixed the need to patch
3110 "configure" to support Darwin 1.3 (which I used to do). However, I still
3111 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3112 workstation).
3113
3114 24. Changed the PCRE licence to be the more standard "BSD" licence.
3115
3116
3117 Version 4.5 01-Dec-03
3118 ---------------------
3119
3120 1. There has been some re-arrangement of the code for the match() function so
3121 that it can be compiled in a version that does not call itself recursively.
3122 Instead, it keeps those local variables that need separate instances for
3123 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3124 needs to "recurse". Keeping track of where control must go is done by means
3125 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3126 hide most of the details from the main code, and operates only if
3127 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3128 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3129 operating.
3130
3131 To make it easier for callers to provide specially tailored get/free
3132 functions for this usage, two new functions, pcre_stack_malloc, and
3133 pcre_stack_free, are used. They are always called in strict stacking order,
3134 and the size of block requested is always the same.
3135
3136 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3137 PCRE has been compiled to use the stack or the heap for recursion. The
3138 -C option of pcretest uses this to show which version is compiled.
3139
3140 A new data escape \S, is added to pcretest; it causes the amounts of store
3141 obtained and freed by both kinds of malloc/free at match time to be added
3142 to the output.
3143
3144 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3145 what's available on my current Linux desktop machine.
3146
3147 3. When matching a UTF-8 string, the test for a valid string at the start has
3148 been extended. If start_offset is not zero, PCRE now checks that it points
3149 to a byte that is the start of a UTF-8 character. If not, it returns
3150 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3151 this is necessary because there may be backward assertions in the pattern.
3152 When matching the same subject several times, it may save resources to use
3153 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3154
3155 4. The code for checking the validity of UTF-8 strings has been tightened so
3156 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3157 containing "overlong sequences".
3158
3159 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3160 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3161 should have been "&", but it just so happened that all the cases this let
3162 through by mistake were picked up later in the function.
3163
3164 6. I had used a variable called "isblank" - this is a C99 function, causing
3165 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3166
3167 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3168 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3169 so the version is shown; (c) stop "make test" echoing "./RunTest".
3170
3171 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3172
3173 9. The prototype for memmove() for systems that don't have it was using
3174 size_t, but the inclusion of the header that defines size_t was later. I've
3175 moved the #includes for the C headers earlier to avoid this.
3176
3177 10. Added some adjustments to the code to make it easier to compiler on certain
3178 special systems:
3179
3180 (a) Some "const" qualifiers were missing.
3181 (b) Added the macro EXPORT before all exported functions; by default this
3182 is defined to be empty.
3183 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3184 that it reads its output file name as an argument instead of writing
3185 to the standard output and assuming this can be redirected.
3186
3187 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3188 class containing characters with values greater than 255, PCRE compilation
3189 went into a loop.
3190
3191 12. A recursive reference to a subpattern that was within another subpattern
3192 that had a minimum quantifier of zero caused PCRE to crash. For example,
3193 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3194 recursion. If the recursively-called subpattern itself had a zero repeat,
3195 that was OK.
3196
3197 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3198 buffer into which it was copied (for escape processing) was still set at
3199 1024, so long lines caused crashes.
3200
3201 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3202 "internal error: code overflow...". This applied to any character class
3203 that was followed by a possessive quantifier.
3204
3205 15. Modified the Makefile to add libpcre.la as a prerequisite for
3206 libpcreposix.la because I was told this is needed for a parallel build to
3207 work.
3208
3209 16. If a pattern that contained .* following optional items at the start was
3210 studied, the wrong optimizing data was generated, leading to matching
3211 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3212 matching string must start with a or b or c. The correct conclusion for
3213 this pattern is that a match can start with any character.
3214
3215
3216 Version 4.4 13-Aug-03
3217 ---------------------
3218
3219 1. In UTF-8 mode, a character class containing characters with values between
3220 127 and 255 was not handled correctly if the compiled pattern was studied.
3221 In fixing this, I have also improved the studying algorithm for such
3222 classes (slightly).
3223
3224 2. Three internal functions had redundant arguments passed to them. Removal
3225 might give a very teeny performance improvement.
3226
3227 3. Documentation bug: the value of the capture_top field in a callout is *one
3228 more than* the number of the hightest numbered captured substring.
3229
3230 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3231 in incorrectly linking with a previously installed version. They now link
3232 explicitly with libpcre.la.
3233
3234 5. configure.in no longer needs to recognize Cygwin specially.
3235
3236 6. A problem in pcre.in for Windows platforms is fixed.
3237
3238 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3239 pcretest, it used to include the size of the study block as part of its
3240 output. Unfortunately, the structure contains a field that has a different
3241 size on different hardware architectures. This meant that the tests that
3242 showed this size failed. As the block is currently always of a fixed size,
3243 this information isn't actually particularly useful in pcretest output, so
3244 I have just removed it.
3245
3246 8. Three pre-processor statements accidentally did not start in column 1.
3247 Sadly, there are *still* compilers around that complain, even though
3248 standard C has not required this for well over a decade. Sigh.
3249
3250 9. In pcretest, the code for checking callouts passed small integers in the
3251 callout_data field, which is a void * field. However, some picky compilers
3252 complained about the casts involved for this on 64-bit systems. Now
3253 pcretest passes the address of the small integer instead, which should get
3254 rid of the warnings.
3255
3256 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3257 both compile and run time, and gives an error if an invalid UTF-8 sequence
3258 is found. There is a option for disabling this check in cases where the
3259 string is known to be correct and/or the maximum performance is wanted.
3260
3261 11. In response to a bug report, I changed one line in Makefile.in from
3262
3263 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3264 to
3265 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3266
3267 to look similar to other lines, but I have no way of telling whether this
3268 is the right thing to do, as I do not use Windows. No doubt I'll get told
3269 if it's wrong...
3270
3271
3272 Version 4.3 21-May-03
3273 ---------------------
3274
3275 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3276 Makefile.
3277
3278 2. Some refactoring to improve the quality of the code:
3279
3280 (i) The utf8_table... variables are now declared "const".
3281
3282 (ii) The code for \cx, which used the "case flipping" table to upper case
3283 lower case letters, now just substracts 32. This is ASCII-specific,
3284 but the whole concept of \cx is ASCII-specific, so it seems
3285 reasonable.
3286
3287 (iii) PCRE was using its character types table to recognize decimal and
3288 hexadecimal digits in the pattern. This is silly, because it handles
3289 only 0-9, a-f, and A-F, but the character types table is locale-
3290 specific, which means strange things might happen. A private
3291 table is now used for this - though it costs 256 bytes, a table is
3292 much faster than multiple explicit tests. Of course, the standard
3293 character types table is still used for matching digits in subject
3294 strings against \d.
3295
3296 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3297 ending in _t are). So I've renamed it as ESC_tee.
3298
3299 3. The first argument for regexec() in the POSIX wrapper should have been
3300 defined as "const".
3301
3302 4. Changed pcretest to use malloc() for its buffers so that they can be
3303 Electric Fenced for debugging.
3304
3305 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3306 to read one or more bytes before the start of the subject string. Often this
3307 had no effect on PCRE's behaviour, but in some circumstances it could
3308 provoke a segmentation fault.
3309
3310 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3311 to try to read one or more bytes before the start of the subject string.
3312
3313 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3314 UTF-8 support could misbehave in various ways if the subject string
3315 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3316 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3317 back over UTF-8 characters.)
3318
3319
3320 Version 4.2 14-Apr-03
3321 ---------------------
3322
3323 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3324
3325 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3326 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3327 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3328 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3329 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3330 and BUILD_EXEEXT
3331 Note: automatic setting of the BUILD variables is not yet working
3332 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3333 compile-time but not at link-time
3334 [LINK]: use for linking executables only
3335 make different versions for Windows and non-Windows
3336 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3337 libraries
3338 [LINK_FOR_BUILD]: new variable
3339 [OBJEXT]: use throughout
3340 [EXEEXT]: use throughout
3341 <winshared>: new target
3342 <wininstall>: new target
3343 <dftables.o>: use native compiler
3344 <dftables>: use native linker
3345 <install>: handle Windows platform correctly
3346 <clean>: ditto
3347 <check>: ditto
3348 copy DLL to top builddir before testing
3349
3350 As part of these changes, -no-undefined was removed again. This was reported
3351 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3352 in any case.
3353
3354 3. Some tidies to get rid of compiler warnings:
3355
3356 . In the match_data structure, match_limit was an unsigned long int, whereas
3357 match_call_count was an int. I've made them both unsigned long ints.
3358
3359 . In pcretest the fact that a const uschar * doesn't automatically cast to
3360 a void * provoked a warning.
3361
3362 . Turning on some more compiler warnings threw up some "shadow" variables
3363 and a few more missing casts.
3364
3365 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3366 option, a class that contained a single character with a value between 128
3367 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3368
3369 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3370 option, a class that contained several characters, but with at least one
3371 whose value was between 128 and 255 caused PCRE to crash.
3372
3373
3374 Version 4.1 12-Mar-03
3375 ---------------------
3376
3377 1. Compiling with gcc -pedantic found a couple of places where casts were
3378 needed, and a string in dftables.c that was longer than standard compilers are
3379 required to support.
3380
3381 2. Compiling with Sun's compiler found a few more places where the code could
3382 be tidied up in order to avoid warnings.
3383
3384 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3385 first of these names is deprecated in the latest Autoconf in favour of the name
3386 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3387 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3388 analogy I have changed it to CFLAGS_FOR_BUILD.
3389
3390 4. Added -no-undefined to the linking command in the Makefile, because this is
3391 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3392 linking step for the pcreposix library.
3393
3394 5. PCRE was failing to diagnose the case of two named groups with the same
3395 name.
3396
3397 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3398 literal character that is needed in the subject for a match, and scans along to
3399 ensure that it is present before embarking on the full matching process. This
3400 saves time in cases of nested unlimited repeats that are never going to match.
3401 Problem: the scan can take a lot of time if the subject is very long (e.g.
3402 megabytes), thus penalizing straightforward matches. It is now done only if the
3403 amount of subject to be scanned is less than 1000 bytes.
3404
3405 7. A lesser problem with the same optimization is that it was recording the
3406 first character of an anchored pattern as "needed", thus provoking a search
3407 right along the subject, even when the first match of the pattern was going to
3408 fail. The "needed" character is now not set for anchored patterns, unless it
3409 follows something in the pattern that is of non-fixed length. Thus, it still
3410 fulfils its original purpose of finding quick non-matches in cases of nested
3411 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3412
3413
3414 Version 4.0 17-Feb-03
3415 ---------------------
3416
3417 1. If a comment in an extended regex that started immediately after a meta-item
3418 extended to the end of string, PCRE compiled incorrect data. This could lead to
3419 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3420
3421 2. Moved to autoconf 2.53 and libtool 1.4.2.
3422
3423 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3424 the special perltest8 script is no longer needed - all the tests can be run
3425 from a single perltest script.
3426
3427 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3428 by \s. It has now been removed in PCRE. This means it isn't recognized as
3429 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3430 class [:space:] *does* include VT, thereby creating a mess.
3431
3432 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3433 space and tab.
3434
3435 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3436 its new features into the main test script, reducing the number of scripts.
3437
3438 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3439 were backward compatible, and made the (?i) apply to the whole pattern, as if
3440 /i were given. Now it behaves more logically, and applies the option setting
3441 only to what follows. PCRE has been changed to follow suit. However, if it
3442 finds options settings right at the start of the pattern, it extracts them into
3443 the global options, as before. Thus, they show up in the info data.
3444
3445 8. Added support for the \Q...\E escape sequence. Characters in between are
3446 treated as literals. This is slightly different from Perl in that $ and @ are
3447 also handled as literals inside the quotes. In Perl, they will cause variable
3448 interpolation. Note the following examples:
3449
3450 Pattern PCRE matches Perl matches
3451
3452 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3453 \Qabc\$xyz\E abc\$xyz abc\$xyz
3454 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3455
3456 For compatibility with Perl, \Q...\E sequences are recognized inside character
3457 classes as well as outside them.
3458
3459 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3460 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3461 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3462 signed/unsigned warnings.
3463
3464 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3465 option for pcretest, so I've replaced it by a simple function that does just
3466 that job.
3467
3468 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3469 "pcregrep -".
3470
3471 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3472 Java package. This provides some syntactic sugar for simple cases of what my
3473 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3474 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3475 item, you can use this simplified notation. Note that only makes sense with
3476 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3477 greediness, whatever the setting of the PCRE_UNGREEDY option.
3478
3479 13. A change of greediness default within a pattern was not taking effect at
3480 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3481 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3482 was abstracted outside.
3483
3484 14. PCRE now supports the \G assertion. It is true when the current matching
3485 position is at the start point of the match. This differs from \A when the
3486 starting offset is non-zero. Used with the /g option of pcretest (or similar
3487 code), it works in the same way as it does for Perl's /g option. If all
3488 alternatives of a regex begin with \G, the expression is anchored to the start
3489 match position, and the "anchored" flag is set in the compiled expression.
3490
3491 15. Some bugs concerning the handling of certain option changes within patterns
3492 have been fixed. These applied to options other than (?ims). For example,
3493 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3494 been the other way round. Some of this was related to change 7 above.
3495
3496 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3497 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3498 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3499 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3500
3501 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3502 mode. Unlike ".", it always matches newline, whatever the setting of
3503 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3504 assertions. Perl allows it, but it doesn't (in general) work because it can't
3505 calculate the length of the lookbehind. At least, that's the case for Perl
3506 5.8.0 - I've been told they are going to document that it doesn't work in
3507 future.
3508
3509 18. Added an error diagnosis for escapes that PCRE does not support: these are
3510 \L, \l, \N, \P, \p, \U, \u, and \X.
3511
3512 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3513 reading past the end of the pattern in cases such as /[abcd/.
3514
3515 20. PCRE was getting more memory than necessary for patterns with classes that
3516 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3517
3518 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3519 compiling PCRE for use with Virtual Pascal.
3520
3521 22. Small fix to the Makefile to make it work properly if the build is done
3522 outside the source tree.
3523
3524 23. Added a new extension: a condition to go with recursion. If a conditional
3525 subpattern starts with (?(R) the "true" branch is used if recursion has
3526 happened, whereas the "false" branch is used only at the top level.
3527
3528 24. When there was a very long string of literal characters (over 255 bytes
3529 without UTF support, over 250 bytes with UTF support), the computation of how
3530 much memory was required could be incorrect, leading to segfaults or other
3531 strange effects.
3532
3533 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3534 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3535 there was a subsequent back reference to those brackets. This meant that, for
3536 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3537 possible to check for precisely this case. All we can do is abandon the
3538 optimization if .* occurs inside capturing brackets when there are any back
3539 references whatsoever. (See below for a better fix that came later.)
3540
3541 26. The handling of the optimization for finding the first character of a
3542 non-anchored pattern, and for finding a character that is required later in the
3543 match were failing in some cases. This didn't break the matching; it just
3544 failed to optimize when it could. The way this is done has been re-implemented.
3545
3546 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3547
3548 28. Added a new feature that provides some of the functionality that Perl
3549 provides with (?{...}). The facility is termed a "callout". The way it is done
3550 in PCRE is for the caller to provide an optional function, by setting
3551 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3552 global variable. By default it is unset, which disables all calling out. To get
3553 the function called, the regex must include (?C) at appropriate points. This
3554 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3555 This provides a means of identifying different callout points. When PCRE
3556 reaches such a point in the regex, if pcre_callout has been set, the external
3557 function is called. It is provided with data in a structure called
3558 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3559 matching continues; if it returns a non-zero value, the match at the current
3560 point fails. However, backtracking will occur if possible. [This was changed
3561 later and other features added - see item 49 below.]
3562
3563 29. pcretest is upgraded to test the callout functionality. It provides a
3564 callout function that displays information. By default, it shows the start of
3565 the match and the current position in the text. There are some new data escapes
3566 to vary what happens:
3567
3568 \C+ in addition, show current contents of captured substrings
3569 \C- do not supply a callout function
3570 \C!n return 1 when callout number n is reached
3571 \C!n!m return 1 when callout number n is reached for the mth time
3572
3573 30. If pcregrep was called with the -l option and just a single file name, it
3574 output "<stdin>" if a match was found, instead of the file name.
3575
3576 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3577 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3578 pcre_exec(). This saves a malloc/free per call. The default value of
3579 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3580 when configuring.
3581
3582 32. The default maximum size of a compiled pattern is 64K. There have been a
3583 few cases of people hitting this limit. The code now uses macros to handle the
3584 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3585 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3586 configuring. Tests 2 and 5 work only with 2-byte links because they output
3587 debugging information about compiled patterns.
3588
3589 33. Internal code re-arrangements:
3590
3591 (a) Moved the debugging function for printing out a compiled regex into
3592 its own source file (printint.c) and used #include to pull it into
3593 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3594 separate copies.
3595
3596 (b) Defined the list of op-code names for debugging as a macro in
3597 internal.h so that it is next to the definition of the opcodes.
3598
3599 (c) Defined a table of op-code lengths for simpler skipping along compiled
3600 code. This is again a macro in internal.h so that it is next to the
3601 definition of the opcodes.
3602
3603 34. Added support for recursive calls to individual subpatterns, along the
3604 lines of Robin Houston's patch (but implemented somewhat differently).
3605
3606 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3607 allow it to read and process whole directories in Win32. This code was
3608 contributed by Lionel Fourquaux; it has not been tested by me.
3609
3610 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3611 used to name a group. Names consist of alphanumerics and underscores, and must
3612 be unique. Back references use the syntax (?P=name) and recursive calls use
3613 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3614 numbers. The function pcre_fullinfo() can be used after compilation to extract
3615 a name/number map. There are three relevant calls:
3616
3617 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3618 PCRE_INFO_NAMECOUNT yields the number of entries
3619 PCRE_INFO_NAMETABLE yields a pointer to the map.
3620
3621 The map is a vector of fixed-size entries. The size of each entry depends on
3622 the length of the longest name used. The first two bytes of each entry are the
3623 group number, most significant byte first. There follows the corresponding
3624 name, zero terminated. The names are in alphabetical order.
3625
3626 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3627 case instead of 255. Making it the same both with and without UTF-8 support
3628 means that the same test output works with both.
3629
3630 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3631 calling malloc() with a zero argument.
3632
3633 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3634 optimization. I've improved things by keeping a bitmap of backreferences with
3635 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3636 fact referenced, the optimization can be applied. It is unlikely that a
3637 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3638 the match to follow \n) will appear inside brackets with a number greater than
3639 31, but if it does, any back reference > 31 suppresses the optimization.
3640
3641 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3642 of disabling numbered capturing parentheses. Any opening parenthesis that is
3643 not followed by ? behaves as if it were followed by ?: but named parentheses
3644 can still be used for capturing (and they will acquire numbers in the usual
3645 way).
3646
3647 41. Redesigned the return codes from the match() function into yes/no/error so
3648 that errors can be passed back from deep inside the nested calls. A malloc
3649 failure while inside a recursive subpattern call now causes the
3650 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3651
3652 42. It is now possible to set a limit on the number of times the match()
3653 function is called in a call to pcre_exec(). This facility makes it possible to
3654 limit the amount of recursion and backtracking, though not in a directly
3655 obvious way, because the match() function is used in a number of different
3656 circumstances. The count starts from zero for each position in the subject
3657 string (for non-anchored patterns). The default limit is, for compatibility, a
3658 large number, namely 10 000 000. You can change this in two ways:
3659
3660 (a) When configuring PCRE before making, you can use --with-match-limit=n
3661 to set a default value for the compiled library.
3662
3663 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3664 a different value is set. See 45 below.
3665
3666 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3667
3668 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3669 of things that can be changed at compile time. The first argument specifies
3670 what is wanted and the second points to where the information is to be placed.
3671 The current list of available information is:
3672
3673 PCRE_CONFIG_UTF8
3674
3675 The output is an integer that is set to one if UTF-8 support is available;
3676 otherwise it is set to zero.
3677
3678 PCRE_CONFIG_NEWLINE
3679
3680 The output is an integer that it set to the value of the code that is used for
3681 newline. It is either LF (10) or CR (13).
3682
3683 PCRE_CONFIG_LINK_SIZE
3684
3685 The output is an integer that contains the number of bytes used for internal
3686 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3687
3688 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3689
3690 The output is an integer that contains the threshold above which the POSIX
3691 interface uses malloc() for output vectors. See item 31 above.
3692
3693 PCRE_CONFIG_MATCH_LIMIT
3694
3695 The output is an unsigned integer that contains the default limit of the number
3696 of match() calls in a pcre_exec() execution. See 42 above.
3697
3698 44. pcretest has been upgraded by the addition of the -C option. This causes it
3699 to extract all the available output from the new pcre_config() function, and to
3700 output it. The program then exits immediately.
3701
3702 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3703 order to support additional features. One way would have been to define
3704 pcre_exec2() (for example) with extra arguments, but this would not have been
3705 extensible, and would also have required all calls to the original function to
3706 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3707 is used for passing in "extra" data from pcre_study().
3708
3709 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3710 contains the following fields:
3711
3712 flags a bitmap indicating which of the following fields are set
3713 study_data opaque data from pcre_study()
3714 match_limit a way of specifying a limit on match() calls for a specific
3715 call to pcre_exec()
3716 callout_data data for callouts (see 49 below)
3717
3718 The flag bits are also defined in pcre.h, and are
3719
3720 PCRE_EXTRA_STUDY_DATA
3721 PCRE_EXTRA_MATCH_LIMIT
3722 PCRE_EXTRA_CALLOUT_DATA
3723
3724 The pcre_study() function now returns one of these new pcre_extra blocks, with
3725 the actual study data pointed to by the study_data field, and the
3726 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3727 before. That is, this change is entirely upwards-compatible and requires no
3728 change to existing code.
3729
3730 If you want to pass in additional data to pcre_exec(), you can either place it
3731 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3732 block.
3733
3734 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3735 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3736 times with different match limits, until it finds the minimum value needed for
3737 pcre_exec() to complete. The value is then output. This can be instructive; for
3738 most simple matches the number is quite small, but for pathological cases it
3739 gets very large very quickly.
3740
3741 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3742 returns the size of the data block pointed to by the study_data field in a
3743 pcre_extra block, that is, the value that was passed as the argument to
3744 pcre_malloc() when PCRE was getting memory in which to place the information
3745 created by pcre_study(). The fourth argument should point to a size_t variable.
3746 pcretest has been extended so that this information is shown after a successful
3747 pcre_study() call when information about the compiled regex is being displayed.
3748
3749 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3750 because what follows is always an absolute path. (Later: it turns out that this
3751 is more than cosmetic for MinGW, because it doesn't like empty path
3752 components.)
3753
3754 49. Some changes have been made to the callout feature (see 28 above):
3755
3756 (i) A callout function now has three choices for what it returns:
3757
3758 0 => success, carry on matching
3759 > 0 => failure at this point, but backtrack if possible
3760 < 0 => serious error, return this value from pcre_exec()
3761
3762 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3763 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3764 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3765 use by callout functions. It will never be used by PCRE itself.
3766
3767 (ii) The pcre_extra structure (see 45 above) has a void * field called
3768 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3769 pcre_callout_block structure has a field of the same name. The contents of
3770 the field passed in the pcre_extra structure are passed to the callout
3771 function in the corresponding field in the callout block. This makes it
3772 easier to use the same callout-containing regex from multiple threads. For
3773 testing, the pcretest program has a new data escape
3774
3775 \C*n pass the number n (may be negative) as callout_data
3776
3777 If the callout function in pcretest receives a non-zero value as
3778 callout_data, it returns that value.
3779
3780 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3781 there were some redundant $(CFLAGS) in commands that are now specified as
3782 $(LINK), which already includes $(CFLAGS).
3783
3784 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3785 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3786 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3787 one-byte characters throughout. Note that case-insensitive matching applies
3788 only to characters whose values are less than 256. PCRE doesn't support the
3789 notion of cases for higher-valued characters.
3790
3791 (i) A character class whose characters are all within 0-255 is handled as
3792 a bit map, and the map is inverted for negative classes. Previously, a
3793 character > 255 always failed to match such a class; however it should
3794 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3795
3796 (ii) A negated character class with a single character < 255 is coded as
3797 "not this character" (OP_NOT). This wasn't working properly when the test
3798 character was multibyte, either singly or repeated.
3799
3800 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3801 mode, for example: \x{100}{2,3}.
3802
3803 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3804 singly or repeated) now correctly test multibyte characters. However,
3805 PCRE doesn't recognize any characters with values greater than 255 as
3806 digits, spaces, or word characters. Such characters always match \D, \S,
3807 and \W, and never match \d, \s, or \w.
3808
3809 (v) Classes may now contain characters and character ranges with values
3810 greater than 255. For example: [ab\x{100}-\x{400}].
3811
3812 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3813 PCRE in UTF-8 mode.
3814
3815 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3816 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3817 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3818 value.)
3819
3820 53. The single man page has become too large. I have therefore split it up into
3821 a number of separate man pages. These also give rise to individual HTML pages;
3822 these are now put in a separate directory, and there is an index.html page that
3823 lists them all. Some hyperlinking between the pages has been installed.
3824
3825 54. Added convenience functions for handling named capturing parentheses.
3826
3827 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3828 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3829 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3830 are faulted.
3831
3832 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3833 calling configure. These values are used when compiling the dftables.c program
3834 which is run to generate the source of the default character tables. They
3835 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3836 you will need to set these values.
3837
3838 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3839
3840
3841 Version 3.9 02-Jan-02
3842 ---------------------
3843
3844 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3845
3846 2. If --disable-static was given, the building process failed when trying to
3847 build pcretest and pcregrep. (For some reason it was using libtool to compile
3848 them, which is not right, as they aren't part of the library.)
3849
3850
3851 Version 3.8 18-Dec-01
3852 ---------------------
3853
3854 1. The experimental UTF-8 code was completely screwed up. It was packing the
3855 bytes in the wrong order. How dumb can you get?
3856
3857
3858 Version 3.7 29-Oct-01
3859 ---------------------
3860
3861 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3862 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3863 this didn't happen under Solaris 8, where I normally test things.
3864
3865 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3866 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3867 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3868 this fix an hour or so after the initial 3.7 release.)
3869
3870
3871 Version 3.6 23-Oct-01
3872 ---------------------
3873
3874 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3875 offsets passed as NULL with zero offset count.
3876
3877 2. The config.guess and config.sub files had not been updated when I moved to
3878 the latest autoconf.
3879
3880
3881 Version 3.5 15-Aug-01
3882 ---------------------
3883
3884 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3885 had been forgotten.
3886
3887 2. By using declared but undefined structures, we can avoid using "void"
3888 definitions in pcre.h while keeping the internal definitions of the structures
3889 private.
3890
3891 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3892 user point of view, this means that both static and shared libraries are built
3893 by default, but this can be individually controlled. More of the work of
3894 handling this static/shared cases is now inside libtool instead of PCRE's make
3895 file.
3896
3897 4. The pcretest utility is now installed along with pcregrep because it is
3898 useful for users (to test regexs) and by doing this, it automatically gets
3899 relinked by libtool. The documentation has been turned into a man page, so
3900 there are now .1, .txt, and .html versions in /doc.
3901
3902 5. Upgrades to pcregrep:
3903 (i) Added long-form option names like gnu grep.
3904 (ii) Added --help to list all options with an explanatory phrase.
3905 (iii) Added -r, --recursive to recurse into sub-directories.
3906 (iv) Added -f, --file to read patterns from a file.
3907
3908 6. pcre_exec() was referring to its "code" argument before testing that
3909 argument for NULL (and giving an error if it was NULL).
3910
3911 7. Upgraded Makefile.in to allow for compiling in a different directory from
3912 the source directory.
3913
3914 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3915 options bits, the pointer it was passed was to an int instead of to an unsigned
3916 long int. This mattered only on 64-bit systems.
3917
3918 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3919 generated) instead of pcre.in, which it its source. Also made the same change
3920 in several of the .c files.
3921
3922 10. A new release of gcc defines printf() as a macro, which broke pcretest
3923 because it had an ifdef in the middle of a string argument for printf(). Fixed
3924 by using separate calls to printf().
3925
3926 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3927 script, to force use of CR or LF instead of \n in the source. On non-Unix
3928 systems, the value can be set in config.h.
3929
3930 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3931 absolute limit. Changed the text of the error message to make this clear, and
3932 likewise updated the man page.
3933
3934 13. The limit of 99 on the number of capturing subpatterns has been removed.
3935 The new limit is 65535, which I hope will not be a "real" limit.
3936
3937
3938 Version 3.4 22-Aug-00
3939 ---------------------
3940
3941 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3942
3943 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3944
3945
3946 Version 3.3 01-Aug-00
3947 ---------------------
3948
3949 1. If an octal character was given, but the value was greater than \377, it
3950 was not getting masked to the least significant bits, as documented. This could
3951 lead to crashes in some systems.
3952
3953 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3954 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3955
3956 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3957 These just pass their arguments on to (pcre_free)(), but they are provided
3958 because some uses of PCRE bind it to non-C systems that can call its functions,
3959 but cannot call free() or pcre_free() directly.
3960
3961 4. Add "make test" as a synonym for "make check". Corrected some comments in
3962 the Makefile.
3963
3964 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3965 Makefile.
3966
3967 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3968 command called pgrep for grepping around the active processes.
3969
3970 7. Added the beginnings of support for UTF-8 character strings.
3971
3972 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3973 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3974 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3975 out for the ar command.)
3976
3977
3978 Version 3.2 12-May-00
3979 ---------------------
3980
3981 This is purely a bug fixing release.
3982
3983 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3984 of ZA. This was just one example of several cases that could provoke this bug,
3985 which was introduced by change 9 of version 2.00. The code for breaking
3986 infinite loops after an iteration that matches an empty string was't working
3987 correctly.
3988
3989 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3990 when matched against abbab (for example). After matching an empty string, it
3991 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
3992 caused it to match further down the string than it should.
3993
3994 3. The code contained an inclusion of sys/types.h. It isn't clear why this
3995 was there because it doesn't seem to be needed, and it causes trouble on some
3996 systems, as it is not a Standard C header. It has been removed.
3997
3998 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
3999 were reported on the Macintosh. The changes were from
4000
4001 while ((c = *(++ptr)) != 0 && c != '\n');
4002 to
4003 while ((c = *(++ptr)) != 0 && c != '\n') ;
4004
4005 Totally extraordinary, but if that's what it takes...
4006
4007 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4008 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4009 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4010 assumes the way PCRE uses memmove() (always moving upwards).
4011
4012 6. PCRE is being used in one environment where strchr() is not available. There
4013 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4014 faster code anyway.
4015
4016
4017 Version 3.1 09-Feb-00
4018 ---------------------
4019
4020 The only change in this release is the fixing of some bugs in Makefile.in for
4021 the "install" target:
4022
4023 (1) It was failing to install pcreposix.h.
4024
4025 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4026
4027
4028 Version 3.0 01-Feb-00
4029 ---------------------
4030
4031 1. Add support for the /+ modifier to perltest (to output $` like it does in
4032 pcretest).
4033
4034 2. Add support for the /g modifier to perltest.
4035
4036 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4037 matches null strings.
4038
4039 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4040 pattern. Perl treats empty patterns specially - it reuses the most recent
4041 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4042 effect.
4043
4044 5. The POSIX interface was broken in that it was just handing over the POSIX
4045 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4046 required a bigger vector, with some working space on the end. This means that
4047 the POSIX wrapper now has to get and free some memory, and copy the results.
4048
4049 6. Added some simple autoconf support, placing the test data and the
4050 documentation in separate directories, re-organizing some of the
4051 information files, and making it build pcre-config (a GNU standard). Also added
4052 libtool support for building PCRE as a shared library, which is now the
4053 default.
4054
4055 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4056 09 are not valid octal constants. Single digits will be used for minor values
4057 less than 10.
4058
4059 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4060 existing programs that set these in the POSIX interface can use PCRE without
4061 modification.
4062
4063 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4064 return all that pcre_info() returns, plus additional data. The pcre_info()
4065 function is retained for compatibility, but is considered to be obsolete.
4066
4067 10. Added experimental recursion feature (?R) to handle one common case that
4068 Perl 5.6 will be able to do with (?p{...}).
4069
4070 11. Added support for POSIX character classes like [:alpha:], which Perl is
4071 adopting.
4072
4073
4074 Version 2.08 31-Aug-99
4075 ----------------------
4076
4077 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4078 trying to match at the startoffset position, but instead was moving forward to
4079 the next newline as if a previous match had failed.
4080
4081 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4082 and could get into a loop if a null string was matched other than at the start
4083 of the subject.
4084
4085 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4086 be distinguished at compile time, and for completeness also added PCRE_DATE.
4087
4088 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4089 in GnuWin32 environments.
4090
4091
4092 Version 2.07 29-Jul-99
4093 ----------------------
4094
4095 1. The documentation is now supplied in plain text form and HTML as well as in
4096 the form of man page sources.
4097
4098 2. C++ compilers don't like assigning (void *) values to other pointer types.
4099 In particular this affects malloc(). Although there is no problem in Standard
4100 C, I've put in casts to keep C++ compilers happy.
4101
4102 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4103 should be (const char *).
4104
4105 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4106 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4107 However, I haven't made this a standard facility. The documentation doesn't
4108 mention it, and the Makefile doesn't support it.
4109
4110 5. The Makefile now contains an "install" target, with editable destinations at
4111 the top of the file. The pcretest program is not installed.
4112
4113 6. pgrep -V now gives the PCRE version number and date.
4114
4115 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4116 causing the entire string to be ignored, instead of just the last character.
4117
4118 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4119 non-matching string, it can take a very, very long time, even for strings of
4120 quite modest length, because of the nested recursion. PCRE now does better in
4121 some of these cases. It does this by remembering the last required literal
4122 character in the pattern, and pre-searching the subject to ensure it is present
4123 before running the real match. In other words, it applies a heuristic to detect
4124 some types of certain failure quickly, and in the above example, if presented
4125 with a string that has no trailing " it gives "no match" very quickly.
4126
4127 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4128 other alternatives are tried instead.
4129
4130
4131 Version 2.06 09-Jun-99
4132 ----------------------
4133
4134 1. Change pcretest's output for amount of store used to show just the code
4135 space, because the remainder (the data block) varies in size between 32-bit and
4136 64-bit systems.
4137
4138 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4139 start matching at. This allows lookbehinds to work when searching for multiple
4140 occurrences in a string.
4141
4142 3. Added additional options to pcretest for testing multiple occurrences:
4143
4144 /+ outputs the rest of the string that follows a match
4145 /g loops for multiple occurrences, using the new startoffset argument
4146 /G loops for multiple occurrences by passing an incremented pointer
4147
4148 4. PCRE wasn't doing the "first character" optimization for patterns starting
4149 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4150 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4151 the letter 'x'. On long subject strings, this gives a significant speed-up.
4152
4153
4154 Version 2.05 21-Apr-99
4155 ----------------------
4156
4157 1. Changed the type of magic_number from int to long int so that it works
4158 properly on 16-bit systems.
4159
4160 2. Fixed a bug which caused patterns starting with .* not to work correctly
4161 when the subject string contained newline characters. PCRE was assuming
4162 anchoring for such patterns in all cases, which is not correct because .* will
4163 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4164 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4165 must be retried after every newline in the subject.
4166
4167
4168 Version 2.04 18-Feb-99
4169 ----------------------
4170
4171 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4172 computation of the store needed to hold the pattern was incorrect (too large).
4173 If such patterns were nested a few deep, this could multiply and become a real
4174 problem.
4175
4176 2. Added /M option to pcretest to show the memory requirement of a specific
4177 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4178
4179 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4180 compiled in such a way that the backtracking after subsequent failure was
4181 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4182 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4183
4184
4185 Version 2.03 02-Feb-99
4186 ----------------------
4187
4188 1. Fixed typo and small mistake in man page.
4189
4190 2. Added 4th condition (GPL supersedes if conflict) and created separate
4191 LICENCE file containing the conditions.
4192
4193 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4194 Perl, that is the internal \ allows the delimiter to be included in the
4195 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4196 the final delimiter, add \ to the end of the pattern (to test the error).
4197
4198 4. Added the convenience functions for extracting substrings after a successful
4199 match. Updated pcretest to make it able to test these functions.
4200
4201
4202 Version 2.02 14-Jan-99
4203 ----------------------
4204
4205 1. Initialized the working variables associated with each extraction so that
4206 their saving and restoring doesn't refer to uninitialized store.
4207
4208 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4209 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4210 fix the problem.
4211
4212 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4213 calls, and wasn't printing the correct value for compiling calls. Increased the
4214 default value of LOOPREPEAT, and the number of significant figures in the
4215 times.
4216
4217 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4218
4219 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4220 a building problem on Windows NT with a FAT file system.
4221
4222
4223 Version 2.01 21-Oct-98
4224 ----------------------
4225
4226 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4227 to character tables built by pcre_maketables() in the current locale. If NULL
4228 is passed, the default tables are used.
4229
4230
4231 Version 2.00 24-Sep-98
4232 ----------------------
4233
4234 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4235 it any more.
4236
4237 2. Allow quantification of (?>) groups, and make it work correctly.
4238
4239 3. The first character computation wasn't working for (?>) groups.
4240
4241 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4242 end of the subject) and add 5.005's \z, which really does match only at the
4243 very end of the subject.
4244
4245 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4246
4247 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4248 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4249 localized options. All options to pcre_study() were also removed.
4250
4251 7. Add other new features from 5.005:
4252
4253 $(?<= positive lookbehind
4254 $(?<! negative lookbehind
4255 (?imsx-imsx) added the unsetting capability
4256 such a setting is global if at outer level; local otherwise
4257 (?imsx-imsx:) non-capturing groups with option setting
4258 (?(cond)re|re) conditional pattern matching
4259
4260 A backreference to itself in a repeated group matches the previous
4261 captured string.
4262
4263 8. General tidying up of studying (both automatic and via "study")
4264 consequential on the addition of new assertions.
4265
4266 9. As in 5.005, unlimited repeated groups that could match an empty substring
4267 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4268 runtime if any iteration does actually match an empty substring.
4269
4270 10. Include the RunTest script in the distribution.
4271
4272 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4273 discrepancies, some of which were old and were also with respect to 5.004. They
4274 have now been fixed.
4275
4276
4277 Version 1.09 28-Apr-98
4278 ----------------------
4279
4280 1. A negated single character class followed by a quantifier with a minimum
4281 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4282 program crashes, or just wrong answers. This did not apply to negated classes
4283 containing more than one character, or to minima other than one.
4284
4285
4286 Version 1.08 27-Mar-98
4287 ----------------------
4288
4289 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4290
4291 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4292 latter must appear before anything that relies on it in the pattern.
4293
4294
4295 Version 1.07 16-Feb-98
4296 ----------------------
4297
4298 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4299 repeat of a potentially empty string).
4300
4301
4302 Version 1.06 23-Jan-98
4303 ----------------------
4304
4305 1. Added Markus Oberhumer's little patches for C++.
4306
4307 2. Literal strings longer than 255 characters were broken.
4308
4309
4310 Version 1.05 23-Dec-97
4311 ----------------------
4312
4313 1. Negated character classes containing more than one character were failing if
4314 PCRE_CASELESS was set at run time.
4315
4316
4317 Version 1.04 19-Dec-97
4318 ----------------------
4319
4320 1. Corrected the man page, where some "const" qualifiers had been omitted.
4321
4322 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4323 input syntax.
4324
4325 3. Fixed memory leak which occurred when a regex with back references was
4326 matched with an offsets vector that wasn't big enough. The temporary memory
4327 that is used in this case wasn't being freed if the match failed.
4328
4329 4. Tidied pcretest to ensure it frees memory that it gets.
4330
4331 5. Temporary memory was being obtained in the case where the passed offsets
4332 vector was exactly big enough.
4333
4334 6. Corrected definition of offsetof() from change 5 below.
4335
4336 7. I had screwed up change 6 below and broken the rules for the use of
4337 setjmp(). Now fixed.
4338
4339
4340 Version 1.03 18-Dec-97
4341 ----------------------
4342
4343 1. A erroneous regex with a missing opening parenthesis was correctly
4344 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4345 on some systems.
4346
4347 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4348 it was reported that one broken compiler failed on the former because "code" is
4349 also an independent variable.
4350
4351 3. The erroneous regex a[]b caused an array overrun reference.
4352
4353 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4354 fail on data ending with that character. (It was going on too far, and checking
4355 the next character, typically a binary zero.) This was specific to the
4356 optimized code for single-character negative classes.
4357
4358 5. Added a contributed patch from the TIN world which does the following:
4359
4360 + Add an undef for memmove, in case the the system defines a macro for it.
4361
4362 + Add a definition of offsetof(), in case there isn't one. (I don't know
4363 the reason behind this - offsetof() is part of the ANSI standard - but
4364 it does no harm).
4365
4366 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4367 most of the places where whitespace preceded '#'. I have given up and
4368 allowed the remaining 2 cases to be at the margin.
4369
4370 + Rename some variables in pcre to eliminate shadowing. This seems very
4371 pedantic, but does no harm, of course.
4372
4373 6. Moved the call to setjmp() into its own function, to get rid of warnings
4374 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4375
4376 7. Constructs such as \d{8,} were compiling into the equivalent of
4377 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4378 outcome, but in this particular case used more store than had been allocated,
4379 which caused the bug to be discovered because it threw up an internal error.
4380
4381 8. The debugging code in both pcre and pcretest for outputting the compiled
4382 form of a regex was going wrong in the case of back references followed by
4383 curly-bracketed repeats.
4384
4385
4386 Version 1.02 12-Dec-97
4387 ----------------------
4388
4389 1. Typos in pcre.3 and comments in the source fixed.
4390
4391 2. Applied a contributed patch to get rid of places where it used to remove
4392 'const' from variables, and fixed some signed/unsigned and uninitialized
4393 variable warnings.
4394
4395 3. Added the "runtest" target to Makefile.
4396
4397 4. Set default compiler flag to -O2 rather than just -O.
4398
4399
4400 Version 1.01 19-Nov-97
4401 ----------------------
4402
4403 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4404 like /([ab]*)*/, that is, for classes with more than one character in them.
4405
4406 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4407 as /((?>a*))*/ (a PCRE_EXTRA facility).
4408
4409
4410 Version 1.00 18-Nov-97
4411 ----------------------
4412
4413 1. Added compile-time macros to support systems such as SunOS4 which don't have
4414 memmove() or strerror() but have other things that can be used instead.
4415
4416 2. Arranged that "make clean" removes the executables.
4417
4418
4419 Version 0.99 27-Oct-97
4420 ----------------------
4421
4422 1. Fixed bug in code for optimizing classes with only one character. It was
4423 initializing a 32-byte map regardless, which could cause it to run off the end
4424 of the memory it had got.
4425
4426 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4427
4428
4429 Version 0.98 22-Oct-97
4430 ----------------------
4431
4432 1. Fixed bug in code for handling temporary memory usage when there are more
4433 back references than supplied space in the ovector. This could cause segfaults.
4434
4435
4436 Version 0.97 21-Oct-97
4437 ----------------------
4438
4439 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4440
4441 2. Optimized negated single characters not to use a bit map.
4442
4443 3. Brought error texts together as macro definitions; clarified some of them;
4444 fixed one that was wrong - it said "range out of order" when it meant "invalid
4445 escape sequence".
4446
4447 4. Changed some char * arguments to const char *.
4448
4449 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4450
4451 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4452 pcretest.
4453
4454
4455 Version 0.96 16-Oct-97
4456 ----------------------
4457
4458 1. Added a simple "pgrep" utility to the distribution.
4459
4460 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4461 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4462 where "ddd" means "one or more decimal digits".
4463
4464 3. Fixed serious bug. If a pattern had a back reference, but the call to
4465 pcre_exec() didn't supply a large enough ovector to record the related
4466 identifying subpattern, the match always failed. PCRE now remembers the number
4467 of the largest back reference, and gets some temporary memory in which to save
4468 the offsets during matching if necessary, in order to ensure that
4469 backreferences always work.
4470
4471 4. Increased the compatibility with Perl in a number of ways:
4472
4473 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4474 to request this handling. The option can be set at compile or exec time.
4475
4476 (b) $ matches before a terminating newline by default; an option
4477 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4478 mode). The option can be set at compile or exec time.
4479
4480 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4481 the same as Perl's. If the decimal number it represents is less than 10
4482 or there aren't that many previous left capturing parentheses, an octal
4483 escape is read. Inside a character class, it's always an octal escape,
4484 even if it is a single digit.
4485
4486 (d) An escaped but undefined alphabetic character is taken as a literal,
4487 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4488 escapes.
4489
4490 (e) {0} is now permitted. (The previous item is removed from the compiled
4491 pattern).
4492
4493 5. Changed all the names of code files so that the basic parts are no longer
4494 than 10 characters, and abolished the teeny "globals.c" file.
4495
4496 6. Changed the handling of character classes; they are now done with a 32-byte
4497 bit map always.
4498
4499 7. Added the -d and /D options to pcretest to make it possible to look at the
4500 internals of compilation without having to recompile pcre.
4501
4502
4503 Version 0.95 23-Sep-97
4504 ----------------------
4505
4506 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4507 \x20 at the start of a run of normal characters. These were being treated as
4508 real characters, instead of the source characters being re-checked.
4509
4510
4511 Version 0.94 18-Sep-97
4512 ----------------------
4513
4514 1. The functions are now thread-safe, with the caveat that the global variables
4515 containing pointers to malloc() and free() or alternative functions are the
4516 same for all threads.
4517
4518 2. Get pcre_study() to generate a bitmap of initial characters for non-
4519 anchored patterns when this is possible, and use it if passed to pcre_exec().
4520
4521
4522 Version 0.93 15-Sep-97
4523 ----------------------
4524
4525 1. /(b)|(:+)/ was computing an incorrect first character.
4526
4527 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4528 but not actually doing anything yet.
4529
4530 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4531 as Perl does (e.g. [-az] or [az-]).
4532
4533 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4534 all possible positions.
4535
4536 5. Split up into different modules to avoid including unneeded functions in a
4537 compiled binary. However, compile and exec are still in one module. The "study"
4538 function is split off.
4539
4540 6. The character tables are now in a separate module whose source is generated
4541 by an auxiliary program - but can then be edited by hand if required. There are
4542 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4543 toupper() in the code.
4544
4545 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4546 make them global. Abolish the function for setting them, as the caller can now
4547 set them directly.
4548
4549
4550 Version 0.92 11-Sep-97
4551 ----------------------
4552
4553 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4554 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4555
4556 2. Caseless matching was not working in character classes if the characters in
4557 the pattern were in upper case.
4558
4559 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4560
4561 4. Make PCRE_ANCHORED public and accept as a compile option.
4562
4563 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4564 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4565 pass them.
4566
4567 6. Give an error if bad option bits passed at compile or run time.
4568
4569 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4570 pcretest to cause it to pass that flag.
4571
4572 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4573 options, and the first character, if set.
4574
4575 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4576
4577
4578 Version 0.91 10-Sep-97
4579 ----------------------
4580
4581 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4582 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4583
4584 2. PCRE was looping on encountering an indefinitely repeated back reference to
4585 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4586 Perl does - treats the match as successful.
4587
4588 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12