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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 791 - (show annotations)
Wed Dec 7 16:38:03 2011 UTC (3 years, 8 months ago) by ph10
File size: 223097 byte(s)
Error occurred while calculating annotation data.
Updated RunTest.bat (don't use JIT for tests 7-10).
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.21 05-Dec-2011
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 25. The CheckMan script was not being included in the distribution. Also, added
104 an explicit "perl" to run Perl scripts from the PrepareRelease script
105 because this is reportedly needed in Windows.
106
107 26. If study data was being save in a file and studying had not found a set of
108 "starts with" bytes for the pattern, the data written to the file (though
109 never used) was taken from uninitialized memory and so caused valgrind to
110 complain.
111
112 27. Updated RunTest.bat as provided by Sheri Pierce.
113
114
115 Version 8.20 21-Oct-2011
116 ------------------------
117
118 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
119 a POSIX class. After further experiments with Perl, which convinced me that
120 Perl has bugs and confusions, a closing square bracket is no longer allowed
121 in a POSIX name. This bug also affected patterns with classes that started
122 with full stops.
123
124 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
125 captured substring, but while checking the failing first alternative,
126 substring 1 is temporarily captured. If the output vector supplied to
127 pcre_exec() was not big enough for this capture, the yield of the function
128 was still zero ("insufficient space for captured substrings"). This cannot
129 be totally fixed without adding another stack variable, which seems a lot
130 of expense for a edge case. However, I have improved the situation in cases
131 such as /(a)(b)x|abc/ matched against "abc", where the return code
132 indicates that fewer than the maximum number of slots in the ovector have
133 been set.
134
135 3. Related to (2) above: when there are more back references in a pattern than
136 slots in the output vector, pcre_exec() uses temporary memory during
137 matching, and copies in the captures as far as possible afterwards. It was
138 using the entire output vector, but this conflicts with the specification
139 that only 2/3 is used for passing back captured substrings. Now it uses
140 only the first 2/3, for compatibility. This is, of course, another edge
141 case.
142
143 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
144 main code base, and can be used by building with --enable-jit. When this is
145 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
146 runtime --no-jit option is given.
147
148 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
149 ovector, the return from the function was zero, implying that there were
150 other matches that did not fit. The correct "exactly full" value is now
151 returned.
152
153 6. If a subpattern that was called recursively or as a subroutine contained
154 (*PRUNE) or any other control that caused it to give a non-standard return,
155 invalid errors such as "Error -26 (nested recursion at the same subject
156 position)" or even infinite loops could occur.
157
158 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
159 computing the minimum length on reaching *ACCEPT, and so ended up with the
160 wrong value of 1 rather than 0. Further investigation indicates that
161 computing a minimum subject length in the presence of *ACCEPT is difficult
162 (think back references, subroutine calls), and so I have changed the code
163 so that no minimum is registered for a pattern that contains *ACCEPT.
164
165 8. If (*THEN) was present in the first (true) branch of a conditional group,
166 it was not handled as intended. [But see 16 below.]
167
168 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
169 Sheri Pierce.
170
171 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
172 the first byte in a match must be "a".
173
174 11. Change 17 for 8.13 increased the recursion depth for patterns like
175 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
176 pattern contains any instances of (*THEN). If it does not, the old
177 optimizations are restored. It would be nice to do this on a per-group
178 basis, but at the moment that is not feasible.
179
180 12. In some environments, the output of pcretest -C is CRLF terminated. This
181 broke RunTest's code that checks for the link size. A single white space
182 character after the value is now allowed for.
183
184 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
185 For "fr", it uses the Windows-specific input and output files.
186
187 14. If (*THEN) appeared in a group that was called recursively or as a
188 subroutine, it did not work as intended. [But see next item.]
189
190 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
191 pattern fragments (but not containing any | characters). If A and B are
192 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
193 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
194 D. In other words, Perl considers parentheses that do not contain any |
195 characters to be part of a surrounding alternative, whereas PCRE was
196 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
197 differently. PCRE now behaves in the same way as Perl, except in the case
198 of subroutine/recursion calls such as (?1) which have in any case always
199 been different (but PCRE had them first :-).
200
201 16. Related to 15 above: Perl does not treat the | in a conditional group as
202 creating alternatives. Such a group is treated in the same way as an
203 ordinary group without any | characters when processing (*THEN). PCRE has
204 been changed to match Perl's behaviour.
205
206 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
207 RunGrepTest script failed.
208
209 18. Change 22 for version 13 caused atomic groups to use more stack. This is
210 inevitable for groups that contain captures, but it can lead to a lot of
211 stack use in large patterns. The old behaviour has been restored for atomic
212 groups that do not contain any capturing parentheses.
213
214 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
215 suppress the check for a minimum subject length at run time. (If it was
216 given to pcre_exec() or pcre_dfa_exec() it did work.)
217
218 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
219 fail to work when decoding hex characters in data strings in EBCDIC
220 environments.
221
222 21. It appears that in at least one Mac OS environment, the isxdigit() function
223 is implemented as a macro that evaluates to its argument more than once,
224 contravening the C 90 Standard (I haven't checked a later standard). There
225 was an instance in pcretest which caused it to go wrong when processing
226 \x{...} escapes in subject strings. The has been rewritten to avoid using
227 things like p++ in the argument of isxdigit().
228
229
230 Version 8.13 16-Aug-2011
231 ------------------------
232
233 1. The Unicode data tables have been updated to Unicode 6.0.0.
234
235 2. Two minor typos in pcre_internal.h have been fixed.
236
237 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
238 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
239 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
240
241 4. There were a number of related bugs in the code for matching backrefences
242 caselessly in UTF-8 mode when codes for the characters concerned were
243 different numbers of bytes. For example, U+023A and U+2C65 are an upper
244 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
245 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
246 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
247 2-byte code at the end of the subject (it thought there wasn't enough data
248 left).
249
250 5. Comprehensive information about what went wrong is now returned by
251 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
252 as the output vector has at least 2 elements. The offset of the start of
253 the failing character and a reason code are placed in the vector.
254
255 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
256 now returned is for the first byte of the failing character, instead of the
257 last byte inspected. This is an incompatible change, but I hope it is small
258 enough not to be a problem. It makes the returned offset consistent with
259 pcre_exec() and pcre_dfa_exec().
260
261 7. pcretest now gives a text phrase as well as the error number when
262 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
263 failure, the offset and reason code are output.
264
265 8. When \R was used with a maximizing quantifier it failed to skip backwards
266 over a \r\n pair if the subsequent match failed. Instead, it just skipped
267 back over a single character (\n). This seems wrong (because it treated the
268 two characters as a single entity when going forwards), conflicts with the
269 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
270 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
271 has been changed.
272
273 9. Some internal refactoring has changed the processing so that the handling
274 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
275 time (the PCRE_DOTALL option was changed this way some time ago: version
276 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
277 which was always a bit of a fudge. It also means that there is one less
278 argument for the match() function, which reduces its stack requirements
279 slightly. This change also fixes an incompatibility with Perl: the pattern
280 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
281
282 10. More internal refactoring has drastically reduced the number of recursive
283 calls to match() for possessively repeated groups such as (abc)++ when
284 using pcre_exec().
285
286 11. While implementing 10, a number of bugs in the handling of groups were
287 discovered and fixed:
288
289 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
290 (a|)*(?1) gave a compile-time internal error.
291 ((a|)+)+ did not notice that the outer group could match an empty string.
292 (^a|^)+ was not marked as anchored.
293 (.*a|.*)+ was not marked as matching at start or after a newline.
294
295 12. Yet more internal refactoring has removed another argument from the match()
296 function. Special calls to this function are now indicated by setting a
297 value in a variable in the "match data" data block.
298
299 13. Be more explicit in pcre_study() instead of relying on "default" for
300 opcodes that mean there is no starting character; this means that when new
301 ones are added and accidentally left out of pcre_study(), testing should
302 pick them up.
303
304 14. The -s option of pcretest has been documented for ages as being an old
305 synonym of -m (show memory usage). I have changed it to mean "force study
306 for every regex", that is, assume /S for every regex. This is similar to -i
307 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
308 using it. It makes it easier to run collections of tests with and without
309 study enabled, and thereby test pcre_study() more easily. All the standard
310 tests are now run with and without -s (but some patterns can be marked as
311 "never study" - see 20 below).
312
313 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
314 restoration of the capturing data to the outer values was not happening
315 correctly.
316
317 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
318 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
319 pattern had matched an empty string, and so incorrectly returned a no
320 match.
321
322 17. There was optimizing code for the last branch of non-capturing parentheses,
323 and also for the obeyed branch of a conditional subexpression, which used
324 tail recursion to cut down on stack usage. Unfortunately, now that there is
325 the possibility of (*THEN) occurring in these branches, tail recursion is
326 no longer possible because the return has to be checked for (*THEN). These
327 two optimizations have therefore been removed. [But see 8.20/11 above.]
328
329 18. If a pattern containing \R was studied, it was assumed that \R always
330 matched two bytes, thus causing the minimum subject length to be
331 incorrectly computed because \R can also match just one byte.
332
333 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
334 was incorrectly computed.
335
336 20. If /S is present twice on a test pattern in pcretest input, it now
337 *disables* studying, thereby overriding the use of -s on the command line
338 (see 14 above). This is necessary for one or two tests to keep the output
339 identical in both cases.
340
341 21. When (*ACCEPT) was used in an assertion that matched an empty string and
342 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
343
344 22. When an atomic group that contained a capturing parenthesis was
345 successfully matched, but the branch in which it appeared failed, the
346 capturing was not being forgotten if a higher numbered group was later
347 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
348 group 1 to "a", when in fact it should be unset. This applied to multi-
349 branched capturing and non-capturing groups, repeated or not, and also to
350 positive assertions (capturing in negative assertions does not happen
351 in PCRE) and also to nested atomic groups.
352
353 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
354 subject after a captured substring, to make it easier to tell which of a
355 number of identical substrings has been captured.
356
357 24. The way atomic groups are processed by pcre_exec() has been changed so that
358 if they are repeated, backtracking one repetition now resets captured
359 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
360 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
361 "aaa". Previously, it would have been "a". As part of this code
362 refactoring, the way recursive calls are handled has also been changed.
363
364 25. If an assertion condition captured any substrings, they were not passed
365 back unless some other capturing happened later. For example, if
366 (?(?=(a))a) was matched against "a", no capturing was returned.
367
368 26. When studying a pattern that contained subroutine calls or assertions,
369 the code for finding the minimum length of a possible match was handling
370 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
371 group 1 called group 2 while simultaneously a separate group 2 called group
372 1). A stack overflow occurred in this case. I have fixed this by limiting
373 the recursion depth to 10.
374
375 27. Updated RunTest.bat in the distribution to the version supplied by Tom
376 Fortmann. This supports explicit test numbers on the command line, and has
377 argument validation and error reporting.
378
379 28. An instance of \X with an unlimited repeat could fail if at any point the
380 first character it looked at was a mark character.
381
382 29. Some minor code refactoring concerning Unicode properties and scripts
383 should reduce the stack requirement of match() slightly.
384
385 30. Added the '=' option to pcretest to check the setting of unused capturing
386 slots at the end of the pattern, which are documented as being -1, but are
387 not included in the return count.
388
389 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
390 compiled something random. Now it gives a compile-time error (as does
391 Perl).
392
393 32. A *MARK encountered during the processing of a positive assertion is now
394 recorded and passed back (compatible with Perl).
395
396 33. If --only-matching or --colour was set on a pcregrep call whose pattern
397 had alternative anchored branches, the search for a second match in a line
398 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
399 matched the line "0102" twice. The same bug affected patterns that started
400 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
401 twice.
402
403 34. Previously, PCRE did not allow quantification of assertions. However, Perl
404 does, and because of capturing effects, quantifying parenthesized
405 assertions may at times be useful. Quantifiers are now allowed for
406 parenthesized assertions.
407
408 35. A minor code tidy in pcre_compile() when checking options for \R usage.
409
410 36. \g was being checked for fancy things in a character class, when it should
411 just be a literal "g".
412
413 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
414 appearance of a nested POSIX class supersedes an apparent external class.
415 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
416 unescaped square brackets may also appear as part of class names. For
417 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
418 more like Perl. (But see 8.20/1 above.)
419
420 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
421 was because it thought it was \N{name}, which is not supported).
422
423 39. Add minix to OS list not supporting the -S option in pcretest.
424
425 40. PCRE tries to detect cases of infinite recursion at compile time, but it
426 cannot analyze patterns in sufficient detail to catch mutual recursions
427 such as ((?1))((?2)). There is now a runtime test that gives an error if a
428 subgroup is called recursively as a subpattern for a second time at the
429 same position in the subject string. In previous releases this might have
430 been caught by the recursion limit, or it might have run out of stack.
431
432 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
433 happen only once. PCRE was, however incorrectly giving a compile time error
434 "recursive call could loop indefinitely" because it cannot analyze the
435 pattern in sufficient detail. The compile time test no longer happens when
436 PCRE is compiling a conditional subpattern, but actual runaway loops are
437 now caught at runtime (see 40 above).
438
439 42. It seems that Perl allows any characters other than a closing parenthesis
440 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
441 has been changed to be the same.
442
443 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
444 as not to get warnings when autogen.sh is called. Also changed
445 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
446
447 44. To help people who use pcregrep to scan files containing exceedingly long
448 lines, the following changes have been made:
449
450 (a) The default value of the buffer size parameter has been increased from
451 8K to 20K. (The actual buffer used is three times this size.)
452
453 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
454 PCRE is built.
455
456 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
457 to be set at run time.
458
459 (d) Numerical values in pcregrep options can be followed by K or M, for
460 example --buffer-size=50K.
461
462 (e) If a line being scanned overflows pcregrep's buffer, an error is now
463 given and the return code is set to 2.
464
465 45. Add a pointer to the latest mark to the callout data block.
466
467 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
468 partial match of an empty string instead of no match. This was specific to
469 the use of ".".
470
471 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
472 complete match instead of a partial match. This bug was dependent on both
473 the PCRE_UTF8 and PCRE_DOTALL options being set.
474
475 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
476 starting byte set, because \b was not being ignored.
477
478
479 Version 8.12 15-Jan-2011
480 ------------------------
481
482 1. Fixed some typos in the markup of the man pages, and wrote a script that
483 checks for such things as part of the documentation building process.
484
485 2. On a big-endian 64-bit system, pcregrep did not correctly process the
486 --match-limit and --recursion-limit options (added for 8.11). In
487 particular, this made one of the standard tests fail. (The integer value
488 went into the wrong half of a long int.)
489
490 3. If the --colour option was given to pcregrep with -v (invert match), it
491 did strange things, either producing crazy output, or crashing. It should,
492 of course, ignore a request for colour when reporting lines that do not
493 match.
494
495 4. Another pcregrep bug caused similar problems if --colour was specified with
496 -M (multiline) and the pattern match finished with a line ending.
497
498 5. In pcregrep, when a pattern that ended with a literal newline sequence was
499 matched in multiline mode, the following line was shown as part of the
500 match. This seems wrong, so I have changed it.
501
502 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
503 the check for further matches in the same line (so they could be coloured)
504 to overrun the end of the current line. If another match was found, it was
505 incorrectly shown (and then shown again when found in the next line).
506
507 7. If pcregrep was compiled under Windows, there was a reference to the
508 function pcregrep_exit() before it was defined. I am assuming this was
509 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
510 reported by a user. I've moved the definition above the reference.
511
512
513 Version 8.11 10-Dec-2010
514 ------------------------
515
516 1. (*THEN) was not working properly if there were untried alternatives prior
517 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
518 backtracked to try for "b" instead of moving to the next alternative branch
519 at the same level (in this case, to look for "c"). The Perl documentation
520 is clear that when (*THEN) is backtracked onto, it goes to the "next
521 alternative in the innermost enclosing group".
522
523 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
524 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
525 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
526 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
527 (*THEN).
528
529 3. If \s appeared in a character class, it removed the VT character from
530 the class, even if it had been included by some previous item, for example
531 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
532 of \s, but is part of the POSIX "space" class.)
533
534 4. A partial match never returns an empty string (because you can always
535 match an empty string at the end of the subject); however the checking for
536 an empty string was starting at the "start of match" point. This has been
537 changed to the "earliest inspected character" point, because the returned
538 data for a partial match starts at this character. This means that, for
539 example, /(?<=abc)def/ gives a partial match for the subject "abc"
540 (previously it gave "no match").
541
542 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
543 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
544 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
545 has an implication that the given string is incomplete (because a partial
546 match is preferred over a full match). For this reason, these items now
547 give a partial match in this situation. [Aside: previously, the one case
548 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
549 match rather than a full match, which was wrong by the old rules, but is
550 now correct.]
551
552 6. There was a bug in the handling of #-introduced comments, recognized when
553 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
554 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
555 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
556 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
557 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
558 places in pcre_compile().
559
560 7. Related to (6) above, when pcre_compile() was skipping #-introduced
561 comments when looking ahead for named forward references to subpatterns,
562 the only newline sequence it recognized was NL. It now handles newlines
563 according to the set newline convention.
564
565 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
566 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
567 cater for a lack of strerror(). These oversights have been fixed.
568
569 9. Added --match-limit and --recursion-limit to pcregrep.
570
571 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
572
573 11. When the -o option was used, pcregrep was setting a return code of 1, even
574 when matches were found, and --line-buffered was not being honoured.
575
576 12. Added an optional parentheses number to the -o and --only-matching options
577 of pcregrep.
578
579 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
580 can match an empty string. The code to do it in pcretest and pcredemo
581 needed fixing:
582
583 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
584 only one byte after an empty string match just before CRLF (this case
585 just got forgotten; "any" and "anycrlf" were OK).
586
587 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
588 mode when an empty string match preceded an ASCII character followed by
589 a non-ASCII character. (The code for advancing by one character rather
590 than one byte was nonsense.)
591
592 (c) The pcredemo.c sample program did not have any code at all to handle
593 the cases when CRLF is a valid newline sequence.
594
595 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
596 as a starting offset was within the subject string. There is now a new
597 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
598 negative or greater than the length of the string. In order to test this,
599 pcretest is extended to allow the setting of negative starting offsets.
600
601 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
602 starting offset points to the beginning of a UTF-8 character was
603 unnecessarily clumsy. I tidied it up.
604
605 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
606 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
607
608 17. Nobody had reported that the --include_dir option, which was added in
609 release 7.7 should have been called --include-dir (hyphen, not underscore)
610 for compatibility with GNU grep. I have changed it to --include-dir, but
611 left --include_dir as an undocumented synonym, and the same for
612 --exclude-dir, though that is not available in GNU grep, at least as of
613 release 2.5.4.
614
615 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
616 characters from a string of bytes) have been redefined so as not to use
617 loops, in order to improve performance in some environments. At the same
618 time, I abstracted some of the common code into auxiliary macros to save
619 repetition (this should not affect the compiled code).
620
621 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
622 compile-time error is now given if \c is not followed by an ASCII
623 character, that is, a byte less than 128. (In EBCDIC mode, the code is
624 different, and any byte value is allowed.)
625
626 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
627 START_OPTIMIZE option, which is now allowed at compile time - but just
628 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
629 to pcregrep and other applications that have no direct access to PCRE
630 options. The new /Y option in pcretest sets this option when calling
631 pcre_compile().
632
633 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
634 back references. Groups containing recursive back references were forced to
635 be atomic by that change, but in the case of named groups, the amount of
636 memory required was incorrectly computed, leading to "Failed: internal
637 error: code overflow". This has been fixed.
638
639 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
640 pcretest.c, to avoid build problems in some Borland environments.
641
642
643 Version 8.10 25-Jun-2010
644 ------------------------
645
646 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
647 THEN.
648
649 2. (*ACCEPT) was not working when inside an atomic group.
650
651 3. Inside a character class, \B is treated as a literal by default, but
652 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
653 causes the error). The code is unchanged, but I tidied the documentation.
654
655 4. Inside a character class, PCRE always treated \R and \X as literals,
656 whereas Perl faults them if its -w option is set. I have changed PCRE so
657 that it faults them when PCRE_EXTRA is set.
658
659 5. Added support for \N, which always matches any character other than
660 newline. (It is the same as "." when PCRE_DOTALL is not set.)
661
662 6. When compiling pcregrep with newer versions of gcc which may have
663 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
664 declared with attribute warn_unused_result" were given. Just casting the
665 result to (void) does not stop the warnings; a more elaborate fudge is
666 needed. I've used a macro to implement this.
667
668 7. Minor change to pcretest.c to avoid a compiler warning.
669
670 8. Added four artifical Unicode properties to help with an option to make
671 \s etc use properties (see next item). The new properties are: Xan
672 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
673
674 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
675 use Unicode properties. (*UCP) at the start of a pattern can be used to set
676 this option. Modified pcretest to add /W to test this facility. Added
677 REG_UCP to make it available via the POSIX interface.
678
679 10. Added --line-buffered to pcregrep.
680
681 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
682 studied, and the match started with a letter with a code point greater than
683 127 whose first byte was different to the first byte of the other case of
684 the letter, the other case of this starting letter was not recognized
685 (#976).
686
687 12. If a pattern that was studied started with a repeated Unicode property
688 test, for example, \p{Nd}+, there was the theoretical possibility of
689 setting up an incorrect bitmap of starting bytes, but fortunately it could
690 not have actually happened in practice until change 8 above was made (it
691 added property types that matched character-matching opcodes).
692
693 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
694 possible starting bytes for non-anchored patterns.
695
696 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
697 \R, and also a number of cases that involve Unicode properties, both
698 explicit and implicit when PCRE_UCP is set.
699
700 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
701 input, it could crash or give wrong results if characters with values
702 greater than 0xc0 were present in the subject string. (Detail: it assumed
703 UTF-8 input when processing these items.)
704
705 16. Added a lot of (int) casts to avoid compiler warnings in systems where
706 size_t is 64-bit (#991).
707
708 17. Added a check for running out of memory when PCRE is compiled with
709 --disable-stack-for-recursion (#990).
710
711 18. If the last data line in a file for pcretest does not have a newline on
712 the end, a newline was missing in the output.
713
714 19. The default pcre_chartables.c file recognizes only ASCII characters (values
715 less than 128) in its various bitmaps. However, there is a facility for
716 generating tables according to the current locale when PCRE is compiled. It
717 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
718 characters, are recognized by isspace() and therefore were getting set in
719 these tables, and indeed these tables seem to approximate to ISO 8859. This
720 caused a problem in UTF-8 mode when pcre_study() was used to create a list
721 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
722 which of course cannot start UTF-8 characters. I have changed the code so
723 that only real ASCII characters (less than 128) and the correct starting
724 bytes for UTF-8 encodings are set for characters greater than 127 when in
725 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
726 altogether.)
727
728 20. Added the /T option to pcretest so as to be able to run tests with non-
729 standard character tables, thus making it possible to include the tests
730 used for 19 above in the standard set of tests.
731
732 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
733 reference to a subpattern the other side of a comment that contains an
734 opening parenthesis caused either an internal compiling error, or a
735 reference to the wrong subpattern.
736
737
738 Version 8.02 19-Mar-2010
739 ------------------------
740
741 1. The Unicode data tables have been updated to Unicode 5.2.0.
742
743 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
744 configured.
745
746 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
747 original author of that file, following a query about its status.
748
749 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
750 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
751
752 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
753 quantifier applied to a forward-referencing subroutine call, could compile
754 incorrect code or give the error "internal error: previously-checked
755 referenced subpattern not found".
756
757 6. Both MS Visual Studio and Symbian OS have problems with initializing
758 variables to point to external functions. For these systems, therefore,
759 pcre_malloc etc. are now initialized to local functions that call the
760 relevant global functions.
761
762 7. There were two entries missing in the vectors called coptable and poptable
763 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
764 I've fixed the data, and added a kludgy way of testing at compile time that
765 the lengths are correct (equal to the number of opcodes).
766
767 8. Following on from 7, I added a similar kludge to check the length of the
768 eint vector in pcreposix.c.
769
770 9. Error texts for pcre_compile() are held as one long string to avoid too
771 much relocation at load time. To find a text, the string is searched,
772 counting zeros. There was no check for running off the end of the string,
773 which could happen if a new error number was added without updating the
774 string.
775
776 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
777
778 11. \K was not working if it appeared in an atomic group or in a group that
779 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
780 \K is "not well defined" if used in an assertion. PCRE now accepts it if
781 the assertion is positive, but not if it is negative.
782
783 12. Change 11 fortuitously reduced the size of the stack frame used in the
784 "match()" function of pcre_exec.c by one pointer. Forthcoming
785 implementation of support for (*MARK) will need an extra pointer on the
786 stack; I have reserved it now, so that the stack frame size does not
787 decrease.
788
789 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
790 item in branch that calls a recursion is a subroutine call - as in the
791 second branch in the above example - was incorrectly given the compile-
792 time error "recursive call could loop indefinitely" because pcre_compile()
793 was not correctly checking the subroutine for matching a non-empty string.
794
795 14. The checks for overrunning compiling workspace could trigger after an
796 overrun had occurred. This is a "should never occur" error, but it can be
797 triggered by pathological patterns such as hundreds of nested parentheses.
798 The checks now trigger 100 bytes before the end of the workspace.
799
800 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
801
802
803 Version 8.01 19-Jan-2010
804 ------------------------
805
806 1. If a pattern contained a conditional subpattern with only one branch (in
807 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
808 computed the wrong minimum data length (which is of course zero for such
809 subpatterns). This could cause incorrect "no match" results.
810
811 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
812 the pattern is reset in the first branch, pcre_compile() failed with
813 "internal error: code overflow at offset...". This happened only when
814 the reset was to the original external option setting. (An optimization
815 abstracts leading options settings into an external setting, which was the
816 cause of this.)
817
818 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
819 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
820 assertion pattern did not match (meaning that the assertion was true), it
821 was incorrectly treated as false if the SKIP had been reached during the
822 matching. This also applied to assertions used as conditions.
823
824 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
825 assertion subpattern, including such a pattern used as a condition,
826 unpredictable results occurred, instead of the error return
827 PCRE_ERROR_DFA_UITEM.
828
829 5. The C++ GlobalReplace function was not working like Perl for the special
830 situation when an empty string is matched. It now does the fancy magic
831 stuff that is necessary.
832
833 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
834 removed. (These were left over from very, very early versions of PCRE.)
835
836 7. Some cosmetic changes to the code to make life easier when compiling it
837 as part of something else:
838
839 (a) Change DEBUG to PCRE_DEBUG.
840
841 (b) In pcre_compile(), rename the member of the "branch_chain" structure
842 called "current" as "current_branch", to prevent a collision with the
843 Linux macro when compiled as a kernel module.
844
845 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
846 prevent a collision with the Linux macro when compiled as a kernel
847 module.
848
849 8. In pcre_compile() there are some checks for integer overflows that used to
850 cast potentially large values to (double). This has been changed to that
851 when building, a check for int64_t is made, and if it is found, it is used
852 instead, thus avoiding the use of floating point arithmetic. (There is no
853 other use of FP in PCRE.) If int64_t is not found, the fallback is to
854 double.
855
856 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
857 2005 (difference between two addresses compared to an unsigned value).
858
859 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
860 custom one, because of the following reported problem in Windows:
861
862 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
863 under Win32.
864 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
865 therefore missing the function definition.
866 - The compiler thus generates a "C" signature for the test function.
867 - The linker fails to find the "C" function.
868 - PCRE fails to configure if asked to do so against libbz2.
869
870 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
871 messages were output:
872
873 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
874 rerunning libtoolize, to keep the correct libtool macros in-tree.
875 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
876
877 I have done both of these things.
878
879 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
880 most of the time, it *can* run out if it is given a pattern that contains a
881 runaway infinite recursion. I updated the discussion in the pcrestack man
882 page.
883
884 13. Now that we have gone to the x.xx style of version numbers, the minor
885 version may start with zero. Using 08 or 09 is a bad idea because users
886 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
887 interpreted as invalid octal numbers. I've updated the previous comment in
888 configure.ac, and also added a check that gives an error if 08 or 09 are
889 used.
890
891 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
892 causing partial matching to fail when the end of the subject matched \W
893 in a UTF-8 pattern where \W was quantified with a minimum of 3.
894
895 15. There were some discrepancies between the declarations in pcre_internal.h
896 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
897 their definitions. The declarations used "const uschar *" and the
898 definitions used USPTR. Even though USPTR is normally defined as "const
899 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
900 reported that: "This difference in casting confuses some C++ compilers, for
901 example, SunCC recognizes above declarations as different functions and
902 generates broken code for hbpcre." I have changed the declarations to use
903 USPTR.
904
905 16. GNU libtool is named differently on some systems. The autogen.sh script now
906 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
907 (FreeBSD).
908
909 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
910 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
911 comment: "Figure out how to create a longlong from a string: strtoll and
912 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
913 instance, but it only takes 2 args instead of 3!"
914
915 18. A subtle bug concerned with back references has been fixed by a change of
916 specification, with a corresponding code fix. A pattern such as
917 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
918 refers, was giving matches when it shouldn't. For example, xa=xaaa would
919 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
920 same bug. Such groups have to be quantified to be useful, or contained
921 inside another quantified group. (If there's no repetition, the reference
922 can never match.) The problem arises because, having left the group and
923 moved on to the rest of the pattern, a later failure that backtracks into
924 the group uses the captured value from the final iteration of the group
925 rather than the correct earlier one. I have fixed this in PCRE by forcing
926 any group that contains a reference to itself to be an atomic group; that
927 is, there cannot be any backtracking into it once it has completed. This is
928 similar to recursive and subroutine calls.
929
930
931 Version 8.00 19-Oct-09
932 ----------------------
933
934 1. The table for translating pcre_compile() error codes into POSIX error codes
935 was out-of-date, and there was no check on the pcre_compile() error code
936 being within the table. This could lead to an OK return being given in
937 error.
938
939 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
940 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
941 in a Windows environment.
942
943 3. The pcregrep --count option prints the count for each file even when it is
944 zero, as does GNU grep. However, pcregrep was also printing all files when
945 --files-with-matches was added. Now, when both options are given, it prints
946 counts only for those files that have at least one match. (GNU grep just
947 prints the file name in this circumstance, but including the count seems
948 more useful - otherwise, why use --count?) Also ensured that the
949 combination -clh just lists non-zero counts, with no names.
950
951 4. The long form of the pcregrep -F option was incorrectly implemented as
952 --fixed_strings instead of --fixed-strings. This is an incompatible change,
953 but it seems right to fix it, and I didn't think it was worth preserving
954 the old behaviour.
955
956 5. The command line items --regex=pattern and --regexp=pattern were not
957 recognized by pcregrep, which required --regex pattern or --regexp pattern
958 (with a space rather than an '='). The man page documented the '=' forms,
959 which are compatible with GNU grep; these now work.
960
961 6. No libpcreposix.pc file was created for pkg-config; there was just
962 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
963
964 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
965 when UCP support is not needed, by modifying the Python script that
966 generates it from Unicode data files. This should not matter if the module
967 is correctly used as a library, but I received one complaint about 50K of
968 unwanted data. My guess is that the person linked everything into his
969 program rather than using a library. Anyway, it does no harm.
970
971 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
972 was a minimum greater than 1 for a wide character in a possessive
973 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
974 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
975 character. Chaos in the form of incorrect output or a compiling loop could
976 result.
977
978 9. The restrictions on what a pattern can contain when partial matching is
979 requested for pcre_exec() have been removed. All patterns can now be
980 partially matched by this function. In addition, if there are at least two
981 slots in the offset vector, the offset of the earliest inspected character
982 for the match and the offset of the end of the subject are set in them when
983 PCRE_ERROR_PARTIAL is returned.
984
985 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
986 synonymous with PCRE_PARTIAL, for backwards compatibility, and
987 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
988 and may be more useful for multi-segment matching.
989
990 11. Partial matching with pcre_exec() is now more intuitive. A partial match
991 used to be given if ever the end of the subject was reached; now it is
992 given only if matching could not proceed because another character was
993 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
994 string "Z", which now yields "no match" instead of "partial match". In the
995 case of pcre_dfa_exec(), "no match" is given if every matching path for the
996 final character ended with (*FAIL).
997
998 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
999 if the pattern had a "must contain" character that was already found in the
1000 earlier partial match, unless partial matching was again requested. For
1001 example, with the pattern /dog.(body)?/, the "must contain" character is
1002 "g". If the first part-match was for the string "dog", restarting with
1003 "sbody" failed. This bug has been fixed.
1004
1005 13. The string returned by pcre_dfa_exec() after a partial match has been
1006 changed so that it starts at the first inspected character rather than the
1007 first character of the match. This makes a difference only if the pattern
1008 starts with a lookbehind assertion or \b or \B (\K is not supported by
1009 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1010 matching functions compatible, and I think it's the right thing to do.
1011
1012 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1013 so that the demonstration program is easily available in environments where
1014 PCRE has not been installed from source.
1015
1016 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1017 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1018 library.
1019
1020 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1021 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1022 is not the first non-POSIX option to be added. Clearly some people find
1023 these options useful.
1024
1025 17. If a caller to the POSIX matching function regexec() passes a non-zero
1026 value for nmatch with a NULL value for pmatch, the value of
1027 nmatch is forced to zero.
1028
1029 18. RunGrepTest did not have a test for the availability of the -u option of
1030 the diff command, as RunTest does. It now checks in the same way as
1031 RunTest, and also checks for the -b option.
1032
1033 19. If an odd number of negated classes containing just a single character
1034 interposed, within parentheses, between a forward reference to a named
1035 subpattern and the definition of the subpattern, compilation crashed with
1036 an internal error, complaining that it could not find the referenced
1037 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1038 [The bug was that it was starting one character too far in when skipping
1039 over the character class, thus treating the ] as data rather than
1040 terminating the class. This meant it could skip too much.]
1041
1042 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1043 /g option in pcretest when the pattern contains \K, which makes it possible
1044 to have an empty string match not at the start, even when the pattern is
1045 anchored. Updated pcretest and pcredemo to use this option.
1046
1047 21. If the maximum number of capturing subpatterns in a recursion was greater
1048 than the maximum at the outer level, the higher number was returned, but
1049 with unset values at the outer level. The correct (outer level) value is
1050 now given.
1051
1052 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1053 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1054 make it do so. The string so far is captured, making this feature
1055 compatible with Perl.
1056
1057 23. The tests have been re-organized, adding tests 11 and 12, to make it
1058 possible to check the Perl 5.10 features against Perl 5.10.
1059
1060 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1061 pattern matches a fixed length string. PCRE did not allow this; now it
1062 does. Neither allows recursion.
1063
1064 25. I finally figured out how to implement a request to provide the minimum
1065 length of subject string that was needed in order to match a given pattern.
1066 (It was back references and recursion that I had previously got hung up
1067 on.) This code has now been added to pcre_study(); it finds a lower bound
1068 to the length of subject needed. It is not necessarily the greatest lower
1069 bound, but using it to avoid searching strings that are too short does give
1070 some useful speed-ups. The value is available to calling programs via
1071 pcre_fullinfo().
1072
1073 26. While implementing 25, I discovered to my embarrassment that pcretest had
1074 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1075 study optimizations had never been tested with that matching function.
1076 Oops. What is worse, even when it was passed study data, there was a bug in
1077 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1078 were also very few tests of studied patterns with pcre_dfa_exec().
1079
1080 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1081 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1082 on the other side of the coin, they are no longer allowed to have different
1083 names, because these cannot be distinguished in PCRE, and this has caused
1084 confusion. (This is a difference from Perl.)
1085
1086 28. When duplicate subpattern names are present (necessarily with different
1087 numbers, as required by 27 above), and a test is made by name in a
1088 conditional pattern, either for a subpattern having been matched, or for
1089 recursion in such a pattern, all the associated numbered subpatterns are
1090 tested, and the overall condition is true if the condition is true for any
1091 one of them. This is the way Perl works, and is also more like the way
1092 testing by number works.
1093
1094
1095 Version 7.9 11-Apr-09
1096 ---------------------
1097
1098 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1099 (pcretest), all targets were linked against these libraries. This included
1100 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1101 libraries. This caused unwanted dependencies to be created. This problem
1102 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1103 pcretest is linked with readline.
1104
1105 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1106 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1107 moved outside it again, because FALSE and TRUE are already defined in AIX,
1108 but BOOL is not.
1109
1110 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1111 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1112
1113 4. The pcregrep documentation said spaces were inserted as well as colons (or
1114 hyphens) following file names and line numbers when outputting matching
1115 lines. This is not true; no spaces are inserted. I have also clarified the
1116 wording for the --colour (or --color) option.
1117
1118 5. In pcregrep, when --colour was used with -o, the list of matching strings
1119 was not coloured; this is different to GNU grep, so I have changed it to be
1120 the same.
1121
1122 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1123 each matching line was coloured. Now it goes on to look for further matches
1124 of any of the test patterns, which is the same behaviour as GNU grep.
1125
1126 7. A pattern that could match an empty string could cause pcregrep to loop; it
1127 doesn't make sense to accept an empty string match in pcregrep, so I have
1128 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1129 seems to be how GNU grep behaves.
1130
1131 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1132 start or after a newline", because the conditional assertion was not being
1133 correctly handled. The rule now is that both the assertion and what follows
1134 in the first alternative must satisfy the test.
1135
1136 9. If auto-callout was enabled in a pattern with a conditional group whose
1137 condition was an assertion, PCRE could crash during matching, both with
1138 pcre_exec() and pcre_dfa_exec().
1139
1140 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1141 used for matching.
1142
1143 11. Unicode property support in character classes was not working for
1144 characters (bytes) greater than 127 when not in UTF-8 mode.
1145
1146 12. Added the -M command line option to pcretest.
1147
1148 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1149
1150 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1151
1152 16. Added comments and documentation about mis-use of no_arg in the C++
1153 wrapper.
1154
1155 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1156 from Martin Jerabek that uses macro names for all relevant character and
1157 string constants.
1158
1159 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1160 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1161 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1162 these, but not everybody uses configure.
1163
1164 19. A conditional group that had only one branch was not being correctly
1165 recognized as an item that could match an empty string. This meant that an
1166 enclosing group might also not be so recognized, causing infinite looping
1167 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1168 with the subject "ab", where knowledge that the repeated group can match
1169 nothing is needed in order to break the loop.
1170
1171 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1172 exec(), but without supplying a callout function, matching went wrong.
1173
1174 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1175 leak if the size of the offset vector was greater than 30. When the vector
1176 is smaller, the saved offsets during recursion go onto a local stack
1177 vector, but for larger vectors malloc() is used. It was failing to free
1178 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1179 error, in fact).
1180
1181 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1182 heapframe that is used only when UTF-8 support is enabled. This caused no
1183 problem, but was untidy.
1184
1185 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1186 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1187 included within another project.
1188
1189 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1190 slightly modified by me:
1191
1192 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1193 not building pcregrep.
1194
1195 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1196 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1197
1198 25. Forward references, both numeric and by name, in patterns that made use of
1199 duplicate group numbers, could behave incorrectly or give incorrect errors,
1200 because when scanning forward to find the reference group, PCRE was not
1201 taking into account the duplicate group numbers. A pattern such as
1202 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1203
1204 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1205 the feature of a custom pointer more persuasive (as requested by a user).
1206
1207 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1208 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1209 pre-defined.
1210
1211 28. Added support for (*UTF8) at the start of a pattern.
1212
1213 29. Arrange for flags added by the "release type" setting in CMake to be shown
1214 in the configuration summary.
1215
1216
1217 Version 7.8 05-Sep-08
1218 ---------------------
1219
1220 1. Replaced UCP searching code with optimized version as implemented for Ad
1221 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1222 stage table and inline lookup instead of a function, giving speed ups of 2
1223 to 5 times on some simple patterns that I tested. Permission was given to
1224 distribute the MultiStage2.py script that generates the tables (it's not in
1225 the tarball, but is in the Subversion repository).
1226
1227 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1228 scripts.
1229
1230 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1231 a group with a zero qualifier. The result of the study could be incorrect,
1232 or the function might crash, depending on the pattern.
1233
1234 4. Caseless matching was not working for non-ASCII characters in back
1235 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1236 It now works when Unicode Property Support is available.
1237
1238 5. In pcretest, an escape such as \x{de} in the data was always generating
1239 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1240 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1241 truncation.
1242
1243 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1244
1245 7. Added two (int) casts to pcregrep when printing the difference of two
1246 pointers, in case they are 64-bit values.
1247
1248 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1249 test 2 if it fails.
1250
1251 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1252 and a #define of that name to empty if it is not externally set. This is to
1253 allow users of MSVC to set it if necessary.
1254
1255 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1256 the convenience functions in the pcre_get.c source file.
1257
1258 11. An option change at the start of a pattern that had top-level alternatives
1259 could cause overwriting and/or a crash. This command provoked a crash in
1260 some environments:
1261
1262 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1263
1264 This potential security problem was recorded as CVE-2008-2371.
1265
1266 12. For a pattern where the match had to start at the beginning or immediately
1267 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1268 pcre_dfa_exec() could read past the end of the passed subject if there was
1269 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1270 pcretest so that it places the subject at the end of its malloc-ed buffer.
1271
1272 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1273 exec() might read past the end of the data buffer in UTF-8 mode.
1274
1275 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1276 the data contained the byte 0x85 as part of a UTF-8 character within its
1277 first line. This applied both to normal and DFA matching.
1278
1279 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1280 /^[^d]*?$/8 failed to match "abc".
1281
1282 16. Added a missing copyright notice to pcrecpp_internal.h.
1283
1284 17. Make it more clear in the documentation that values returned from
1285 pcre_exec() in ovector are byte offsets, not character counts.
1286
1287 18. Tidied a few places to stop certain compilers from issuing warnings.
1288
1289 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1290 supplied by Stefan Weber. I made a further small update for 7.8 because
1291 there is a change of source arrangements: the pcre_searchfuncs.c module is
1292 replaced by pcre_ucd.c.
1293
1294
1295 Version 7.7 07-May-08
1296 ---------------------
1297
1298 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1299 a string to a long long, pretend we don't even have a long long." This is
1300 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1301
1302 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1303 pre-7.6 versions, which defined a global no_arg variable instead of putting
1304 it in the RE class. (See also #8 below.)
1305
1306 3. Remove a line of dead code, identified by coverity and reported by Nuno
1307 Lopes.
1308
1309 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1310
1311 (1) The include/exclude patterns were being applied to the whole pathnames
1312 of files, instead of just to the final components.
1313
1314 (2) If there was more than one level of directory, the subdirectories were
1315 skipped unless they satisfied the include/exclude conditions. This is
1316 inconsistent with GNU grep (and could even be seen as contrary to the
1317 pcregrep specification - which I improved to make it absolutely clear).
1318 The action now is always to scan all levels of directory, and just
1319 apply the include/exclude patterns to regular files.
1320
1321 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1322 --exclude_dir in the tests to avoid scanning .svn directories.
1323
1324 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1325 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1326 doesn't support NULs in patterns.
1327
1328 7. Added some missing "const"s to declarations of static tables in
1329 pcre_compile.c and pcre_dfa_exec.c.
1330
1331 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1332 caused by fix #2 above. (Subsequently also a second patch to fix the
1333 first patch. And a third patch - this was a messy problem.)
1334
1335 9. Applied Craig's patch to remove the use of push_back().
1336
1337 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1338 matching function regexec().
1339
1340 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1341 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1342 references. PCRE supports relative numbers with this syntax (I don't think
1343 Oniguruma does).
1344
1345 12. Previously, a group with a zero repeat such as (...){0} was completely
1346 omitted from the compiled regex. However, this means that if the group
1347 was called as a subroutine from elsewhere in the pattern, things went wrong
1348 (an internal error was given). Such groups are now left in the compiled
1349 pattern, with a new opcode that causes them to be skipped at execution
1350 time.
1351
1352 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1353 to the way PCRE behaves:
1354
1355 (a) A lone ] character is dis-allowed (Perl treats it as data).
1356
1357 (b) A back reference to an unmatched subpattern matches an empty string
1358 (Perl fails the current match path).
1359
1360 (c) A data ] in a character class must be notated as \] because if the
1361 first data character in a class is ], it defines an empty class. (In
1362 Perl it is not possible to have an empty class.) The empty class []
1363 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1364 The negative empty class [^] matches any one character, independently
1365 of the DOTALL setting.
1366
1367 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1368 non-existent subpattern following a character class starting with ']' and
1369 containing () gave an internal compiling error instead of "reference to
1370 non-existent subpattern". Fortunately, when the pattern did exist, the
1371 compiled code was correct. (When scanning forwards to check for the
1372 existencd of the subpattern, it was treating the data ']' as terminating
1373 the class, so got the count wrong. When actually compiling, the reference
1374 was subsequently set up correctly.)
1375
1376 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1377 it was being rejected as not supported by pcre_dfa_exec(), even though
1378 other assertions are supported. I have made pcre_dfa_exec() support
1379 (*FAIL).
1380
1381 16. The implementation of 13c above involved the invention of a new opcode,
1382 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1383 cannot be changed at match time, I realized I could make a small
1384 improvement to matching performance by compiling OP_ALLANY instead of
1385 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1386 on the OP_ANY path.
1387
1388 17. Compiling pcretest on Windows with readline support failed without the
1389 following two fixes: (1) Make the unistd.h include conditional on
1390 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1391
1392 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1393 ncurses library to be included for pcretest when ReadLine support is
1394 requested, but also to allow for it to be overridden. This patch came from
1395 Daniel Bergström.
1396
1397 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1398 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1399 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1400 spotting this.
1401
1402
1403 Version 7.6 28-Jan-08
1404 ---------------------
1405
1406 1. A character class containing a very large number of characters with
1407 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1408 overflow.
1409
1410 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1411 HAVE_LONG_LONG is not defined.
1412
1413 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1414 bring them up to date and include new features. This patch includes:
1415
1416 - Fixed PH's badly added libz and libbz2 support.
1417 - Fixed a problem with static linking.
1418 - Added pcredemo. [But later removed - see 7 below.]
1419 - Fixed dftables problem and added an option.
1420 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1421 HAVE_LONG_LONG.
1422 - Added readline support for pcretest.
1423 - Added an listing of the option settings after cmake has run.
1424
1425 4. A user submitted a patch to Makefile that makes it easy to create
1426 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1427 Makefile.am that cause it to include this special target, without
1428 affecting anything else. Note that the same mingw target plus all
1429 the other distribution libraries and programs are now supported
1430 when configuring with CMake (see 6 below) instead of with
1431 Configure/Make.
1432
1433 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1434 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1435 exported in the Windows port". It has not yet been confirmed that the patch
1436 solves the problem, but it does no harm.
1437
1438 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1439 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1440 with CMake, and also correct the comment about stack recursion.
1441
1442 7. Remove the automatic building of pcredemo from the ./configure system and
1443 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1444 of a program that users should build themselves after PCRE is installed, so
1445 building it automatically is not really right. What is more, it gave
1446 trouble in some build environments.
1447
1448 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1449
1450
1451 Version 7.5 10-Jan-08
1452 ---------------------
1453
1454 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1455 values in parens when parsing an RE using the C++ wrapper."
1456
1457 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1458 Characters greater than 255 were excluded from the class instead of being
1459 included.
1460
1461 3. The same bug as (2) above applied to negated POSIX classes such as
1462 [:^space:].
1463
1464 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1465 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1466 I have changed it.
1467
1468 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1469 first named subpattern) and a construct such as (?&a) would reference the
1470 first named subpattern whose name started with "a" (in other words, the
1471 length check was missing). Both these problems are fixed. "Subpattern name
1472 expected" is now given for (?&) (a zero-length name), and this patch also
1473 makes it give the same error for \k'' (previously it complained that that
1474 was a reference to a non-existent subpattern).
1475
1476 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1477 this is right because (?- can be followed by option settings as well as by
1478 digits. I have, however, made the messages clearer.
1479
1480 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1481 than the number used in the conditional) now cause a compile-time error.
1482 This is actually not compatible with Perl, which accepts such patterns, but
1483 treats the conditional as always being FALSE (as PCRE used to), but it
1484 seems to me that giving a diagnostic is better.
1485
1486 8. Change "alphameric" to the more common word "alphanumeric" in comments
1487 and messages.
1488
1489 9. Fix two occurrences of "backslash" in comments that should have been
1490 "backspace".
1491
1492 10. Remove two redundant lines of code that can never be obeyed (their function
1493 was moved elsewhere).
1494
1495 11. The program that makes PCRE's Unicode character property table had a bug
1496 which caused it to generate incorrect table entries for sequences of
1497 characters that have the same character type, but are in different scripts.
1498 It amalgamated them into a single range, with the script of the first of
1499 them. In other words, some characters were in the wrong script. There were
1500 thirteen such cases, affecting characters in the following ranges:
1501
1502 U+002b0 - U+002c1
1503 U+0060c - U+0060d
1504 U+0061e - U+00612
1505 U+0064b - U+0065e
1506 U+0074d - U+0076d
1507 U+01800 - U+01805
1508 U+01d00 - U+01d77
1509 U+01d9b - U+01dbf
1510 U+0200b - U+0200f
1511 U+030fc - U+030fe
1512 U+03260 - U+0327f
1513 U+0fb46 - U+0fbb1
1514 U+10450 - U+1049d
1515
1516 12. The -o option (show only the matching part of a line) for pcregrep was not
1517 compatible with GNU grep in that, if there was more than one match in a
1518 line, it showed only the first of them. It now behaves in the same way as
1519 GNU grep.
1520
1521 13. If the -o and -v options were combined for pcregrep, it printed a blank
1522 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1523 does the same. The return code can be used to tell if there were any
1524 non-matching lines.
1525
1526 14. Added --file-offsets and --line-offsets to pcregrep.
1527
1528 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1529 infinitely looping recursion. The bug was that positive lookaheads were not
1530 being skipped when checking for a possible empty match (negative lookaheads
1531 and both kinds of lookbehind were skipped).
1532
1533 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1534 inclusion of <windows.h> to before rather than after the definition of
1535 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1536
1537 17. Specifying a possessive quantifier with a specific limit for a Unicode
1538 character property caused pcre_compile() to compile bad code, which led at
1539 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1540 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1541 caused the error; without that there was no problem.
1542
1543 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1544
1545 19. Added --enable-pcretest-libreadline.
1546
1547 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1548 RE::GlobalReplace(). As a result, the number of replacements returned was
1549 double what it should be. I removed one of the increments, but Craig sent a
1550 later patch that removed the other one (the right fix) and added unit tests
1551 that check the return values (which was not done before).
1552
1553 21. Several CMake things:
1554
1555 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1556 the names libpcre and libpcreposix, not just pcre and pcreposix.
1557
1558 (2) The above change means that pcretest and pcregrep are now correctly
1559 linked with the newly-built libraries, not previously installed ones.
1560
1561 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1562
1563 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1564 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1565 UTF-8 newline character). The key issue is that the pattern starts .*;
1566 this means that the match must be either at the beginning, or after a
1567 newline. The bug was in the code for advancing after a failed match and
1568 checking that the new position followed a newline. It was not taking
1569 account of UTF-8 characters correctly.
1570
1571 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1572 character classes. PCRE was not treating the sequence [:...:] as a
1573 character class unless the ... were all letters. Perl, however, seems to
1574 allow any characters between [: and :], though of course it rejects as
1575 unknown any "names" that contain non-letters, because all the known class
1576 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1577 for example, whereas PCRE did not - it did not recognize a POSIX character
1578 class. This seemed a bit dangerous, so the code has been changed to be
1579 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1580 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1581 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1582 Perl does, and where it didn't before.
1583
1584 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1585 Windows environments %n is disabled by default.
1586
1587
1588 Version 7.4 21-Sep-07
1589 ---------------------
1590
1591 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1592 means that a class such as [\s] counted as "explicit reference to CR or
1593 LF". That isn't really right - the whole point of the change was to try to
1594 help when there was an actual mention of one of the two characters. So now
1595 the change happens only if \r or \n (or a literal CR or LF) character is
1596 encountered.
1597
1598 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1599 of both had grown to the point where there were only 3 bits left.
1600 Fortunately, there was spare space in the data structure, and so I have
1601 moved the internal flags into a new 16-bit field to free up more option
1602 bits.
1603
1604 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1605 but did not set the internal JCHANGED flag - either of these is enough to
1606 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1607 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1608 start sets both bits.
1609
1610 4. Added options (at build time, compile time, exec time) to change \R from
1611 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1612
1613 5. doc/pcresyntax.html was missing from the distribution.
1614
1615 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1616 compatibility, even though it is no longer used.
1617
1618 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1619 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1620 windows.h file is present (where different names are used). [This was
1621 reversed later after testing - see 16 below.]
1622
1623 8. Changed all #include <config.h> to #include "config.h". There were also
1624 some further <pcre.h> cases that I changed to "pcre.h".
1625
1626 9. When pcregrep was used with the --colour option, it missed the line ending
1627 sequence off the lines that it output.
1628
1629 10. It was pointed out to me that arrays of string pointers cause lots of
1630 relocations when a shared library is dynamically loaded. A technique of
1631 using a single long string with a table of offsets can drastically reduce
1632 these. I have refactored PCRE in four places to do this. The result is
1633 dramatic:
1634
1635 Originally: 290
1636 After changing UCP table: 187
1637 After changing error message table: 43
1638 After changing table of "verbs" 36
1639 After changing table of Posix names 22
1640
1641 Thanks to the folks working on Gregex for glib for this insight.
1642
1643 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1644 unicode-properties was also set.
1645
1646 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1647
1648 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1649 checked only for CRLF.
1650
1651 14. Added casts to pcretest.c to avoid compiler warnings.
1652
1653 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1654
1655 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1656 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1657 entirely. This removes changes made in 7 above.
1658
1659 17. The CMake files have been updated, and there is now more information about
1660 building with CMake in the NON-UNIX-USE document.
1661
1662
1663 Version 7.3 28-Aug-07
1664 ---------------------
1665
1666 1. In the rejigging of the build system that eventually resulted in 7.1, the
1667 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1668 brackets there is not right, since it causes compilers to look for an
1669 installed pcre.h, not the version that is in the source that is being
1670 compiled (which of course may be different). I have changed it back to:
1671
1672 #include "pcre.h"
1673
1674 I have a vague recollection that the change was concerned with compiling in
1675 different directories, but in the new build system, that is taken care of
1676 by the VPATH setting the Makefile.
1677
1678 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1679 when the subject happened to end in the byte 0x85 (e.g. if the last
1680 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1681 characters but of course it shouldn't be taken as a newline when it is part
1682 of another character. The bug was that, for an unlimited repeat of . in
1683 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1684 characters when looking for a newline.
1685
1686 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1687
1688 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1689 in debug output.
1690
1691 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1692 long printing in the pcrecpp unittest when running under MinGW.
1693
1694 6. ESC_K was left out of the EBCDIC table.
1695
1696 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1697 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1698 limit also applies to "virtual nesting" when a pattern is recursive, and in
1699 this case 1000 isn't so big. I have been able to remove this limit at the
1700 expense of backing off one optimization in certain circumstances. Normally,
1701 when pcre_exec() would call its internal match() function recursively and
1702 immediately return the result unconditionally, it uses a "tail recursion"
1703 feature to save stack. However, when a subpattern that can match an empty
1704 string has an unlimited repetition quantifier, it no longer makes this
1705 optimization. That gives it a stack frame in which to save the data for
1706 checking that an empty string has been matched. Previously this was taken
1707 from the 1000-entry workspace that had been reserved. So now there is no
1708 explicit limit, but more stack is used.
1709
1710 8. Applied Daniel's patches to solve problems with the import/export magic
1711 syntax that is required for Windows, and which was going wrong for the
1712 pcreposix and pcrecpp parts of the library. These were overlooked when this
1713 problem was solved for the main library.
1714
1715 9. There were some crude static tests to avoid integer overflow when computing
1716 the size of patterns that contain repeated groups with explicit upper
1717 limits. As the maximum quantifier is 65535, the maximum group length was
1718 set at 30,000 so that the product of these two numbers did not overflow a
1719 32-bit integer. However, it turns out that people want to use groups that
1720 are longer than 30,000 bytes (though not repeat them that many times).
1721 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1722 made it possible to implement the integer overflow checks in a much more
1723 dynamic way, which I have now done. The artificial limitation on group
1724 length has been removed - we now have only the limit on the total length of
1725 the compiled pattern, which depends on the LINK_SIZE setting.
1726
1727 10. Fixed a bug in the documentation for get/copy named substring when
1728 duplicate names are permitted. If none of the named substrings are set, the
1729 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1730 empty string.
1731
1732 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1733 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1734 because the ] is interpreted as the first data character and the
1735 terminating ] is not found. PCRE has been made compatible with Perl in this
1736 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1737 cause memory overwriting.
1738
1739 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1740 string has been matched (to stop an infinite loop). It was not recognizing
1741 a conditional subpattern that could match an empty string if that
1742 subpattern was within another subpattern. For example, it looped when
1743 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1744 condition was not nested. This bug has been fixed.
1745
1746 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1747 past the start of the subject in the presence of bytes with the top bit
1748 set, for example "\x8aBCD".
1749
1750 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1751 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1752
1753 14. Optimized (?!) to (*FAIL).
1754
1755 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1756 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1757 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1758 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1759 does: it's just the validity check that is more restrictive.
1760
1761 16. Inserted checks for integer overflows during escape sequence (backslash)
1762 processing, and also fixed erroneous offset values for syntax errors during
1763 backslash processing.
1764
1765 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1766 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1767
1768 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1769 caused an overrun.
1770
1771 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1772 something other than just ASCII characters) inside a group that had an
1773 unlimited repeat caused a loop at compile time (while checking to see
1774 whether the group could match an empty string).
1775
1776 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1777 [\P{Any}] did so. (Error in the code for printing property names.)
1778
1779 21. An orphan \E inside a character class could cause a crash.
1780
1781 22. A repeated capturing bracket such as (A)? could cause a wild memory
1782 reference during compilation.
1783
1784 23. There are several functions in pcre_compile() that scan along a compiled
1785 expression for various reasons (e.g. to see if it's fixed length for look
1786 behind). There were bugs in these functions when a repeated \p or \P was
1787 present in the pattern. These operators have additional parameters compared
1788 with \d, etc, and these were not being taken into account when moving along
1789 the compiled data. Specifically:
1790
1791 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1792 length.
1793
1794 (b) An item such as \pL+ within a repeated group could cause crashes or
1795 loops.
1796
1797 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1798 "reference to non-existent subpattern" error.
1799
1800 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1801
1802 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1803 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1804
1805 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1806
1807 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1808 character were causing crashes (broken optimization).
1809
1810 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1811 \p or \P) caused a compile-time loop.
1812
1813 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1814 break. For example, the unstudied pattern [\r\n]A does not match the string
1815 "\r\nA" because change 7.0/46 below moves the current point on by two
1816 characters after failing to match at the start. However, the pattern \nA
1817 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1818 the same is true. There doesn't seem any very clean way out of this, but
1819 what I have chosen to do makes the common cases work: PCRE now takes note
1820 of whether there can be an explicit match for \r or \n anywhere in the
1821 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1822 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1823 pattern has explicit CR or LF references.
1824
1825 29. Added (*CR) etc for changing newline setting at start of pattern.
1826
1827
1828 Version 7.2 19-Jun-07
1829 ---------------------
1830
1831 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1832 which is apparently normally available under Windows.
1833
1834 2. Re-jig the pcregrep tests with different newline settings in an attempt
1835 to make them independent of the local environment's newline setting.
1836
1837 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1838
1839 4. Some of the "internals" tests were previously cut out when the link size
1840 was not 2, because the output contained actual offsets. The recent new
1841 "Z" feature of pcretest means that these can be cut out, making the tests
1842 usable with all link sizes.
1843
1844 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1845 stack recursion. This gives a massive performance boost under BSD, but just
1846 a small improvement under Linux. However, it saves one field in the frame
1847 in all cases.
1848
1849 6. Added more features from the forthcoming Perl 5.10:
1850
1851 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1852 recursion call. It refers to the nth most recently opened parentheses.
1853
1854 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1855 to be opened parentheses.
1856
1857 (c) Conditions that refer to capturing parentheses can be specified
1858 relatively, for example, (?(-2)... or (?(+3)...
1859
1860 (d) \K resets the start of the current match so that everything before
1861 is not part of it.
1862
1863 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1864
1865 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1866 reference syntax.
1867
1868 (g) (?| introduces a group in which the numbering of parentheses in each
1869 alternative starts with the same number.
1870
1871 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1872
1873 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1874 PCRE_INFO_JCHANGED.
1875
1876 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1877 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1878 for detecting groups that can match an empty string.
1879
1880 9. A pattern with a very large number of alternatives (more than several
1881 hundred) was running out of internal workspace during the pre-compile
1882 phase, where pcre_compile() figures out how much memory will be needed. A
1883 bit of new cunning has reduced the workspace needed for groups with
1884 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1885 workspace instead of running out of the 4096 that are available.
1886
1887 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1888
1889 11. Applied patch from Google to remove an optimization that didn't quite work.
1890 The report of the bug said:
1891
1892 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1893 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1894 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1895
1896 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1897 it matched the wrong number of bytes.
1898
1899
1900 Version 7.1 24-Apr-07
1901 ---------------------
1902
1903 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1904 that is more "standard", making use of automake and other Autotools. There
1905 is some re-arrangement of the files and adjustment of comments consequent
1906 on this.
1907
1908 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1909 for recursive directory scanning broke on some systems because the files
1910 are not scanned in any specific order and on different systems the order
1911 was different. A call to "sort" has been inserted into RunGrepTest for the
1912 approprate test as a short-term fix. In the longer term there may be an
1913 alternative.
1914
1915 3. I had an email from Eric Raymond about problems translating some of PCRE's
1916 man pages to HTML (despite the fact that I distribute HTML pages, some
1917 people do their own conversions for various reasons). The problems
1918 concerned the use of low-level troff macros .br and .in. I have therefore
1919 removed all such uses from the man pages (some were redundant, some could
1920 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1921 HTML has been updated to handle .nf/.fi and to complain if it encounters
1922 .br or .in.
1923
1924 4. Updated comments in configure.ac that get placed in config.h.in and also
1925 arranged for config.h to be included in the distribution, with the name
1926 config.h.generic, for the benefit of those who have to compile without
1927 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1928
1929 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1930 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1931 makevp.bat for the current PCRE, using the additional files
1932 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1933
1934 6. A Windows user reported a minor discrepancy with test 2, which turned out
1935 to be caused by a trailing space on an input line that had got lost in his
1936 copy. The trailing space was an accident, so I've just removed it.
1937
1938 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1939 that is needed.
1940
1941 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1942 as "const" (a) because they are and (b) because it helps the PHP
1943 maintainers who have recently made a script to detect big data structures
1944 in the php code that should be moved to the .rodata section. I remembered
1945 to update Builducptable as well, so it won't revert if ucptable.h is ever
1946 re-created.
1947
1948 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1949 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1950 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1951 support is not required. This saves 1.5-2K of code, which is important in
1952 some applications.
1953
1954 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1955 so as not to refer to the tables, even though these functions will never be
1956 called when UTF-8 support is disabled. Otherwise there are problems with a
1957 shared library.
1958
1959 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1960
1961 (a) It was defining its arguments as char * instead of void *.
1962
1963 (b) It was assuming that all moves were upwards in memory; this was true
1964 a long time ago when I wrote it, but is no longer the case.
1965
1966 The emulated memove() is provided for those environments that have neither
1967 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1968 is clearly not the case, as these two bugs were recently reported.
1969
1970 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1971 and Detrail to create the HTML documentation, the .txt form of the man
1972 pages, and it removes trailing spaces from listed files. It also creates
1973 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1974 case, it wraps all the #defines with #ifndefs. This script should be run
1975 before "make dist".
1976
1977 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1978 with Unicode property support.
1979
1980 (a) For a maximizing quantifier, if the two different cases of the
1981 character were of different lengths in their UTF-8 codings (there are
1982 some cases like this - I found 11), and the matching function had to
1983 back up over a mixture of the two cases, it incorrectly assumed they
1984 were both the same length.
1985
1986 (b) When PCRE was configured to use the heap rather than the stack for
1987 recursion during matching, it was not correctly preserving the data for
1988 the other case of a UTF-8 character when checking ahead for a match
1989 while processing a minimizing repeat. If the check also involved
1990 matching a wide character, but failed, corruption could cause an
1991 erroneous result when trying to check for a repeat of the original
1992 character.
1993
1994 13. Some tidying changes to the testing mechanism:
1995
1996 (a) The RunTest script now detects the internal link size and whether there
1997 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1998 values substituted by "configure". (The RunGrepTest script already did
1999 this for UTF-8.) The configure.ac script no longer substitutes the
2000 relevant variables.
2001
2002 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2003 with length and offset values. This means that the output is different
2004 for different internal link sizes. Test 2 is skipped for link sizes
2005 other than 2 because of this, bypassing the problem. Unfortunately,
2006 there was also a test in test 3 (the locale tests) that used /B and
2007 failed for link sizes other than 2. Rather than cut the whole test out,
2008 I have added a new /Z option to pcretest that replaces the length and
2009 offset values with spaces. This is now used to make test 3 independent
2010 of link size. (Test 2 will be tidied up later.)
2011
2012 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2013 segmentation fault instead of returning the appropriate error message.
2014
2015 15. In multiline mode when the newline sequence was set to "any", the pattern
2016 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2017 This doesn't seem right; it now treats the CRLF combination as the line
2018 ending, and so does not match in that case. It's only a pattern such as ^$
2019 that would hit this one: something like ^ABC$ would have failed after \r
2020 and then tried again after \r\n.
2021
2022 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2023 in an attempt to make files that differ only in their line terminators
2024 compare equal. This works on Linux.
2025
2026 17. Under certain error circumstances pcregrep might try to free random memory
2027 as it exited. This is now fixed, thanks to valgrind.
2028
2029 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2030 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2031 was because its rules for how to advance for /g after matching an empty
2032 string at the end of a line did not allow for this case. They now check for
2033 it specially.
2034
2035 20. pcretest is supposed to handle patterns and data of any length, by
2036 extending its buffers when necessary. It was getting this wrong when the
2037 buffer for a data line had to be extended.
2038
2039 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2040 CRLF as a newline sequence.
2041
2042 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2043 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2044 I have nevertheless tidied it up.
2045
2046 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2047
2048 24. Added a man page for pcre-config.
2049
2050
2051 Version 7.0 19-Dec-06
2052 ---------------------
2053
2054 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2055 moving to gcc 4.1.1.
2056
2057 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2058 sys/time.h, which is documented as needed for this function. It doesn't
2059 seem to matter on Linux, but it showed up on some releases of OS X.
2060
2061 3. It seems that there are systems where bytes whose values are greater than
2062 127 match isprint() in the "C" locale. The "C" locale should be the
2063 default when a C program starts up. In most systems, only ASCII printing
2064 characters match isprint(). This difference caused the output from pcretest
2065 to vary, making some of the tests fail. I have changed pcretest so that:
2066
2067 (a) When it is outputting text in the compiled version of a pattern, bytes
2068 other than 32-126 are always shown as hex escapes.
2069
2070 (b) When it is outputting text that is a matched part of a subject string,
2071 it does the same, unless a different locale has been set for the match
2072 (using the /L modifier). In this case, it uses isprint() to decide.
2073
2074 4. Fixed a major bug that caused incorrect computation of the amount of memory
2075 required for a compiled pattern when options that changed within the
2076 pattern affected the logic of the preliminary scan that determines the
2077 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2078 that the computed length was too small. The symptoms of this bug were
2079 either the PCRE error "internal error: code overflow" from pcre_compile(),
2080 or a glibc crash with a message such as "pcretest: free(): invalid next
2081 size (fast)". Examples of patterns that provoked this bug (shown in
2082 pcretest format) are:
2083
2084 /(?-x: )/x
2085 /(?x)(?-x: \s*#\s*)/
2086 /((?i)[\x{c0}])/8
2087 /(?i:[\x{c0}])/8
2088
2089 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2090 is now done differently.
2091
2092 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2093 wrapper classes; (b) implement a new function in the C++ scanner that is
2094 more efficient than the old way of doing things because it avoids levels of
2095 recursion in the regex matching; (c) add a paragraph to the documentation
2096 for the FullMatch() function.
2097
2098 6. The escape sequence \n was being treated as whatever was defined as
2099 "newline". Not only was this contrary to the documentation, which states
2100 that \n is character 10 (hex 0A), but it also went horribly wrong when
2101 "newline" was defined as CRLF. This has been fixed.
2102
2103 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2104 was being set to -1 for the "end of line" case (supposedly a value that no
2105 character can have). Though this value is never used (the check for end of
2106 line is "zero bytes in current character"), it caused compiler complaints.
2107 I've changed it to 0xffffffff.
2108
2109 8. In pcre_version.c, the version string was being built by a sequence of
2110 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2111 string (as it is for production releases) called a macro with an empty
2112 argument. The C standard says the result of this is undefined. The gcc
2113 compiler treats it as an empty string (which was what was wanted) but it is
2114 reported that Visual C gives an error. The source has been hacked around to
2115 avoid this problem.
2116
2117 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2118 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2119 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2120 of them did).
2121
2122 10. Originally, pcretest opened its input and output without "b"; then I was
2123 told that "b" was needed in some environments, so it was added for release
2124 5.0 to both the input and output. (It makes no difference on Unix-like
2125 systems.) Later I was told that it is wrong for the input on Windows. I've
2126 now abstracted the modes into two macros, to make it easier to fiddle with
2127 them, and removed "b" from the input mode under Windows.
2128
2129 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2130
2131 12. Added -help and --help to pcretest as an official way of being reminded
2132 of the options.
2133
2134 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2135 and pcrecpp.cc because they annoy compilers at high warning levels.
2136
2137 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2138
2139 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2140 scripts are not C programs :-) and which was not noticed because it works
2141 on Linux.
2142
2143 16. pcretest is supposed to handle any length of pattern and data line (as one
2144 line or as a continued sequence of lines) by extending its input buffer if
2145 necessary. This feature was broken for very long pattern lines, leading to
2146 a string of junk being passed to pcre_compile() if the pattern was longer
2147 than about 50K.
2148
2149 17. I have done a major re-factoring of the way pcre_compile() computes the
2150 amount of memory needed for a compiled pattern. Previously, there was code
2151 that made a preliminary scan of the pattern in order to do this. That was
2152 OK when PCRE was new, but as the facilities have expanded, it has become
2153 harder and harder to keep it in step with the real compile phase, and there
2154 have been a number of bugs (see for example, 4 above). I have now found a
2155 cunning way of running the real compile function in a "fake" mode that
2156 enables it to compute how much memory it would need, while actually only
2157 ever using a few hundred bytes of working memory and without too many
2158 tests of the mode. This should make future maintenance and development
2159 easier. A side effect of this work is that the limit of 200 on the nesting
2160 depth of parentheses has been removed (though this was never a serious
2161 limitation, I suspect). However, there is a downside: pcre_compile() now
2162 runs more slowly than before (30% or more, depending on the pattern). I
2163 hope this isn't a big issue. There is no effect on runtime performance.
2164
2165 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2166 newline (only possible for the last line of a file) and it was a
2167 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2168
2169 19. Added additional timing features to pcretest. (1) The -tm option now times
2170 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2171 separate command line item, by a number that specifies the number of
2172 repeats to use when timing. The default is 50000; this gives better
2173 precision, but takes uncomfortably long for very large patterns.
2174
2175 20. Extended pcre_study() to be more clever in cases where a branch of a
2176 subpattern has no definite first character. For example, (a*|b*)[cd] would
2177 previously give no result from pcre_study(). Now it recognizes that the
2178 first character must be a, b, c, or d.
2179
2180 21. There was an incorrect error "recursive call could loop indefinitely" if
2181 a subpattern (or the entire pattern) that was being tested for matching an
2182 empty string contained only one non-empty item after a nested subpattern.
2183 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2184 incorrectly, because the \d was being skipped in the check.
2185
2186 22. The pcretest program now has a new pattern option /B and a command line
2187 option -b, which is equivalent to adding /B to every pattern. This causes
2188 it to show the compiled bytecode, without the additional information that
2189 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2190 is the same as /B/I).
2191
2192 23. A new optimization is now able automatically to treat some sequences such
2193 as a*b as a*+b. More specifically, if something simple (such as a character
2194 or a simple class like \d) has an unlimited quantifier, and is followed by
2195 something that cannot possibly match the quantified thing, the quantifier
2196 is automatically "possessified".
2197
2198 24. A recursive reference to a subpattern whose number was greater than 39
2199 went wrong under certain circumstances in UTF-8 mode. This bug could also
2200 have affected the operation of pcre_study().
2201
2202 25. Realized that a little bit of performance could be had by replacing
2203 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2204
2205 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2206
2207 27. Possessive quantifiers such as a++ were previously implemented by turning
2208 them into atomic groups such as ($>a+). Now they have their own opcodes,
2209 which improves performance. This includes the automatically created ones
2210 from 23 above.
2211
2212 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2213 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2214 the first matched character to be a colon. This applied both to named and
2215 numbered groups.
2216
2217 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2218
2219 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2220 building PCRE on Windows easier, so I have included it in the distribution.
2221
2222 31. There is now a check in pcretest against a ridiculously large number being
2223 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2224 loop, the loop is abandoned.
2225
2226 32. Forward references to subpatterns in conditions such as (?(2)...) where
2227 subpattern 2 is defined later cause pcre_compile() to search forwards in
2228 the pattern for the relevant set of parentheses. This search went wrong
2229 when there were unescaped parentheses in a character class, parentheses
2230 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2231
2232 33. "Subroutine" calls and backreferences were previously restricted to
2233 referencing subpatterns earlier in the regex. This restriction has now
2234 been removed.
2235
2236 34. Added a number of extra features that are going to be in Perl 5.10. On the
2237 whole, these are just syntactic alternatives for features that PCRE had
2238 previously implemented using the Python syntax or my own invention. The
2239 other formats are all retained for compatibility.
2240
2241 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2242 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2243 also .NET compatible.
2244
2245 (b) A recursion or subroutine call to a named group can now be defined as
2246 (?&name) as well as (?P>name).
2247
2248 (c) A backreference to a named group can now be defined as \k<name> or
2249 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2250 5.10, are also .NET compatible.
2251
2252 (d) A conditional reference to a named group can now use the syntax
2253 (?(<name>) or (?('name') as well as (?(name).
2254
2255 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2256 groups (named and numbered) that are never evaluated inline, but can be
2257 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2258 is always false. There may be only one alternative in such a group.
2259
2260 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2261 as the simple (?(R). The condition is true only if the most recent
2262 recursion is that of the given number or name. It does not search out
2263 through the entire recursion stack.
2264
2265 (g) The escape \gN or \g{N} has been added, where N is a positive or
2266 negative number, specifying an absolute or relative reference.
2267
2268 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2269 some "unreachable code" warnings.
2270
2271 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2272 things, this adds five new scripts.
2273
2274 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2275 There were also incompatibilities regarding the handling of \Q..\E inside
2276 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2277 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2278
2279 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2280 matches an empty string, and forcibly breaks the loop. There were bugs in
2281 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2282 against aaaa the result was just "a" rather than "aaaa", for example. Two
2283 separate and independent bugs (that affected different cases) have been
2284 fixed.
2285
2286 39. Refactored the code to abolish the use of different opcodes for small
2287 capturing bracket numbers. This is a tidy that I avoided doing when I
2288 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2289 The new approach is not only tidier, it makes it possible to reduce the
2290 memory needed to fix the previous bug (38).
2291
2292 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2293 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2294 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2295 mode.
2296
2297 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2298 report.
2299
2300 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2301 copy construction and assignment in the C++ wrapper.
2302
2303 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2304 couple of bugs that could have given wrong results in the "--newline=crlf"
2305 case.
2306
2307 44. Added a number of casts and did some reorganization of signed/unsigned int
2308 variables following suggestions from Dair Grant. Also renamed the variable
2309 "this" as "item" because it is a C++ keyword.
2310
2311 45. Arranged for dftables to add
2312
2313 #include "pcre_internal.h"
2314
2315 to pcre_chartables.c because without it, gcc 4.x may remove the array
2316 definition from the final binary if PCRE is built into a static library and
2317 dead code stripping is activated.
2318
2319 46. For an unanchored pattern, if a match attempt fails at the start of a
2320 newline sequence, and the newline setting is CRLF or ANY, and the next two
2321 characters are CRLF, advance by two characters instead of one.
2322
2323
2324 Version 6.7 04-Jul-06
2325 ---------------------
2326
2327 1. In order to handle tests when input lines are enormously long, pcretest has
2328 been re-factored so that it automatically extends its buffers when
2329 necessary. The code is crude, but this _is_ just a test program. The
2330 default size has been increased from 32K to 50K.
2331
2332 2. The code in pcre_study() was using the value of the re argument before
2333 testing it for NULL. (Of course, in any sensible call of the function, it
2334 won't be NULL.)
2335
2336 3. The memmove() emulation function in pcre_internal.h, which is used on
2337 systems that lack both memmove() and bcopy() - that is, hardly ever -
2338 was missing a "static" storage class specifier.
2339
2340 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2341 containing an extended class (one that cannot be represented by a bitmap
2342 because it contains high-valued characters or Unicode property items, e.g.
2343 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2344 pattern, but PCRE should not loop if you do not (it no longer does).
2345 [Detail: two cases were found: (a) a repeated subpattern containing an
2346 extended class; (b) a recursive reference to a subpattern that followed a
2347 previous extended class. It wasn't skipping over the extended class
2348 correctly when UTF-8 mode was not set.]
2349
2350 5. A negated single-character class was not being recognized as fixed-length
2351 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2352 compile error "lookbehind assertion is not fixed length".
2353
2354 6. The RunPerlTest auxiliary script was showing an unexpected difference
2355 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2356 write a Perl script that can interpret lines of an input file either as
2357 byte characters or as UTF-8, which is what "perltest" was being required to
2358 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2359 can't do is switch easily at run time between having the "use utf8;" pragma
2360 or not. In the end, I fudged it by using the RunPerlTest script to insert
2361 "use utf8;" explicitly for the UTF-8 tests.
2362
2363 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2364 the end of the subject string, contrary to the documentation and to what
2365 Perl does. This was true of both matching functions. Now it matches only at
2366 the start of the subject and immediately after *internal* newlines.
2367
2368 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2369 a pointer to an int instead of a pointer to an unsigned long int. This
2370 caused problems on 64-bit systems.
2371
2372 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2373 instance of the 'standard' template library not being so standard".
2374
2375 10. There was no check on the number of named subpatterns nor the maximum
2376 length of a subpattern name. The product of these values is used to compute
2377 the size of the memory block for a compiled pattern. By supplying a very
2378 long subpattern name and a large number of named subpatterns, the size
2379 computation could be caused to overflow. This is now prevented by limiting
2380 the length of names to 32 characters, and the number of named subpatterns
2381 to 10,000.
2382
2383 11. Subpatterns that are repeated with specific counts have to be replicated in
2384 the compiled pattern. The size of memory for this was computed from the
2385 length of the subpattern and the repeat count. The latter is limited to
2386 65535, but there was no limit on the former, meaning that integer overflow
2387 could in principle occur. The compiled length of a repeated subpattern is
2388 now limited to 30,000 bytes in order to prevent this.
2389
2390 12. Added the optional facility to have named substrings with the same name.
2391
2392 13. Added the ability to use a named substring as a condition, using the
2393 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2394 are numbers (not recommended). Forward references are permitted.
2395
2396 14. Added forward references in named backreferences (if you see what I mean).
2397
2398 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2399 pattern could run off the end of the subject. For example, the pattern
2400 "(?s)(.{1,5})"8 did this with the subject "ab".
2401
2402 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2403 PCRE_CASELESS was set when matching characters that were quantified with ?
2404 or *.
2405
2406 17. A character class other than a single negated character that had a minimum
2407 but no maximum quantifier - for example [ab]{6,} - was not handled
2408 correctly by pce_dfa_exec(). It would match only one character.
2409
2410 18. A valid (though odd) pattern that looked like a POSIX character
2411 class but used an invalid character after [ (for example [[,abc,]]) caused
2412 pcre_compile() to give the error "Failed: internal error: code overflow" or
2413 in some cases to crash with a glibc free() error. This could even happen if
2414 the pattern terminated after [[ but there just happened to be a sequence of
2415 letters, a binary zero, and a closing ] in the memory that followed.
2416
2417 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2418 over the years. Originally (before any Unicode support), just the bottom 8
2419 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2420 output from "man perlunicode" includes this:
2421
2422 The regular expression compiler produces polymorphic opcodes. That
2423 is, the pattern adapts to the data and automatically switches to
2424 the Unicode character scheme when presented with Unicode data--or
2425 instead uses a traditional byte scheme when presented with byte
2426 data.
2427
2428 Sadly, a wide octal escape does not cause a switch, and in a string with
2429 no other multibyte characters, these octal escapes are treated as before.
2430 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2431 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2432 Unicode string.
2433
2434 I have not perpetrated such confusion in PCRE. Up till now, it took just
2435 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2436 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2437 translate to the appropriate multibyte character.
2438
2439 29. Applied some refactoring to reduce the number of warnings from Microsoft
2440 and Borland compilers. This has included removing the fudge introduced
2441 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2442 a warning about an unused variable.
2443
2444 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2445 characters since release 4.0, because Perl (from release 5.004) does not.
2446 [Or at least, is documented not to: some releases seem to be in conflict
2447 with the documentation.] However, when a pattern was studied with
2448 pcre_study() and all its branches started with \s, PCRE still included VT
2449 as a possible starting character. Of course, this did no harm; it just
2450 caused an unnecessary match attempt.
2451
2452 22. Removed a now-redundant internal flag bit that recorded the fact that case
2453 dependency changed within the pattern. This was once needed for "required
2454 byte" processing, but is no longer used. This recovers a now-scarce options
2455 bit. Also moved the least significant internal flag bit to the most-
2456 significant bit of the word, which was not previously used (hangover from
2457 the days when it was an int rather than a uint) to free up another bit for
2458 the future.
2459
2460 23. Added support for CRLF line endings as well as CR and LF. As well as the
2461 default being selectable at build time, it can now be changed at runtime
2462 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2463 specify that it is scanning data with non-default line endings.
2464
2465 24. Changed the definition of CXXLINK to make it agree with the definition of
2466 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2467
2468 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2469 recursions. This makes a big different to stack usage for some patterns.
2470
2471 26. If a subpattern containing a named recursion or subroutine reference such
2472 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2473 the space required for the compiled pattern went wrong and gave too small a
2474 value. Depending on the environment, this could lead to "Failed: internal
2475 error: code overflow at offset 49" or "glibc detected double free or
2476 corruption" errors.
2477
2478 27. Applied patches from Google (a) to support the new newline modes and (b) to
2479 advance over multibyte UTF-8 characters in GlobalReplace.
2480
2481 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2482 difference for some implementation of PCRE in some Windows version.
2483
2484 29. Added some extra testing facilities to pcretest:
2485
2486 \q<number> in a data line sets the "match limit" value
2487 \Q<number> in a data line sets the "match recursion limt" value
2488 -S <number> sets the stack size, where <number> is in megabytes
2489
2490 The -S option isn't available for Windows.
2491
2492
2493 Version 6.6 06-Feb-06
2494 ---------------------
2495
2496 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2497 in pcreposix.h. I have copied the definition from pcre.h.
2498
2499 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2500 because pcre.h is no longer a built file.
2501
2502 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2503 not normally included in the compiled code.
2504
2505
2506 Version 6.5 01-Feb-06
2507 ---------------------
2508
2509 1. When using the partial match feature with pcre_dfa_exec(), it was not
2510 anchoring the second and subsequent partial matches at the new starting
2511 point. This could lead to incorrect results. For example, with the pattern
2512 /1234/, partially matching against "123" and then "a4" gave a match.
2513
2514 2. Changes to pcregrep:
2515
2516 (a) All non-match returns from pcre_exec() were being treated as failures
2517 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2518 error message is output. Some extra information is given for the
2519 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2520 probably the only errors that are likely to be caused by users (by
2521 specifying a regex that has nested indefinite repeats, for instance).
2522 If there are more than 20 of these errors, pcregrep is abandoned.
2523
2524 (b) A binary zero was treated as data while matching, but terminated the
2525 output line if it was written out. This has been fixed: binary zeroes
2526 are now no different to any other data bytes.
2527
2528 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2529 used to set a locale for matching. The --locale=xxxx long option has
2530 been added (no short equivalent) to specify a locale explicitly on the
2531 pcregrep command, overriding the environment variables.
2532
2533 (d) When -B was used with -n, some line numbers in the output were one less
2534 than they should have been.
2535
2536 (e) Added the -o (--only-matching) option.
2537
2538 (f) If -A or -C was used with -c (count only), some lines of context were
2539 accidentally printed for the final match.
2540
2541 (g) Added the -H (--with-filename) option.
2542
2543 (h) The combination of options -rh failed to suppress file names for files
2544 that were found from directory arguments.
2545
2546 (i) Added the -D (--devices) and -d (--directories) options.
2547
2548 (j) Added the -F (--fixed-strings) option.
2549
2550 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2551
2552 (l) Added the --colo(u)r option.
2553
2554 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2555 is not present by default.
2556
2557 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2558 items such as (?R) or (?1), when the recursion could match a number of
2559 alternatives. If it matched one of the alternatives, but subsequently,
2560 outside the recursion, there was a failure, the code tried to back up into
2561 the recursion. However, because of the way PCRE is implemented, this is not
2562 possible, and the result was an incorrect result from the match.
2563
2564 In order to prevent this happening, the specification of recursion has
2565 been changed so that all such subpatterns are automatically treated as
2566 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2567
2568 4. I had overlooked the fact that, in some locales, there are characters for
2569 which isalpha() is true but neither isupper() nor islower() are true. In
2570 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2571 and ordfeminine) are like this. This affected the treatment of \w and \W
2572 when they appeared in character classes, but not when they appeared outside
2573 a character class. The bit map for "word" characters is now created
2574 separately from the results of isalnum() instead of just taking it from the
2575 upper, lower, and digit maps. (Plus the underscore character, of course.)
2576
2577 5. The above bug also affected the handling of POSIX character classes such as
2578 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2579 permanent tables. Instead, the bit maps for such a class were previously
2580 created as the appropriate unions of the upper, lower, and digit bitmaps.
2581 Now they are created by subtraction from the [[:word:]] class, which has
2582 its own bitmap.
2583
2584 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2585 It is created by subtracting the vertical space characters (\x09, \x0a,
2586 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2587 subtraction was done in the overall bitmap for a character class, meaning
2588 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2589 be recognized. This bug has been fixed.
2590
2591 7. Patches from the folks at Google:
2592
2593 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2594 real life, but is still worth protecting against".
2595
2596 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2597 regular expressions".
2598
2599 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2600 have it.
2601
2602 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2603 "configure" and the latter not, in order to fix a problem somebody had
2604 with compiling the Arg class on HP-UX.
2605
2606 (e) Improve the error-handling of the C++ wrapper a little bit.
2607
2608 (f) New tests for checking recursion limiting.
2609
2610 8. The pcre_memmove() function, which is used only if the environment does not
2611 have a standard memmove() function (and is therefore rarely compiled),
2612 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2613 returning a result (though PCRE never actually uses the result).
2614
2615 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2616 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2617 returned instead of calling malloc() with an overflowing number that would
2618 most likely cause subsequent chaos.
2619
2620 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2621
2622 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2623 with this option is matched, the nmatch and pmatch options of regexec() are
2624 ignored.
2625
2626 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2627 provided in case anyone wants to the the POSIX interface with UTF-8
2628 strings.
2629
2630 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2631 C++ linking (needed for some HP-UX environments).
2632
2633 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2634 (unused parameter) and in the pcre_printint() function (omitted "default"
2635 switch label when the default is to do nothing).
2636
2637 15. Added some code to make it possible, when PCRE is compiled as a C++
2638 library, to replace subject pointers for pcre_exec() with a smart pointer
2639 class, thus making it possible to process discontinuous strings.
2640
2641 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2642 much the same function. They were added by different people who were trying
2643 to make PCRE easy to compile on non-Unix systems. It has been suggested
2644 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2645 for compiling on Windows systems. I have therefore replaced it with
2646 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2647 defaults to "extern" for C or "extern C" for C++, which works fine on
2648 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2649 SCOPE with something explicit in config.h. In addition:
2650
2651 (a) pcreposix.h still had just "extern" instead of either of these macros;
2652 I have replaced it with PCRE_DATA_SCOPE.
2653
2654 (b) Functions such as _pcre_xclass(), which are internal to the library,
2655 but external in the C sense, all had PCRE_EXPORT in their definitions.
2656 This is apparently wrong for the Windows case, so I have removed it.
2657 (It makes no difference on Unix-like systems.)
2658
2659 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2660 of recursive calls to match(). This is different to MATCH_LIMIT because
2661 that limits the total number of calls to match(), not all of which increase
2662 the depth of recursion. Limiting the recursion depth limits the amount of
2663 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2664 when PCRE is compiled, and changed at run time. A patch from Google adds
2665 this functionality to the C++ interface.
2666
2667 18. Changes to the handling of Unicode character properties:
2668
2669 (a) Updated the table to Unicode 4.1.0.
2670
2671 (b) Recognize characters that are not in the table as "Cn" (undefined).
2672
2673 (c) I revised the way the table is implemented to a much improved format
2674 which includes recognition of ranges. It now supports the ranges that
2675 are defined in UnicodeData.txt, and it also amalgamates other
2676 characters into ranges. This has reduced the number of entries in the
2677 table from around 16,000 to around 3,000, thus reducing its size
2678 considerably. I realized I did not need to use a tree structure after
2679 all - a binary chop search is just as efficient. Having reduced the
2680 number of entries, I extended their size from 6 bytes to 8 bytes to
2681 allow for more data.
2682
2683 (d) Added support for Unicode script names via properties such as \p{Han}.
2684
2685 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2686 matching that character.
2687
2688 20. When matching a repeated Unicode property with a minimum greater than zero,
2689 (for example \pL{2,}), PCRE could look past the end of the subject if it
2690 reached it while seeking the minimum number of characters. This could
2691 happen only if some of the characters were more than one byte long, because
2692 there is a check for at least the minimum number of bytes.
2693
2694 21. Refactored the implementation of \p and \P so as to be more general, to
2695 allow for more different types of property in future. This has changed the
2696 compiled form incompatibly. Anybody with saved compiled patterns that use
2697 \p or \P will have to recompile them.
2698
2699 22. Added "Any" and "L&" to the supported property types.
2700
2701 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2702 but give a compile time error if the value is greater than 0xff.
2703
2704 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2705 accidentally not being installed or uninstalled.
2706
2707 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2708 made were to insert the current release number. This seemed silly, because
2709 it made things harder for people building PCRE on systems that don't run
2710 "configure". I have turned pcre.h into a distributed file, no longer built
2711 by "configure", with the version identification directly included. There is
2712 no longer a pcre.h.in file.
2713
2714 However, this change necessitated a change to the pcre-config script as
2715 well. It is built from pcre-config.in, and one of the substitutions was the
2716 release number. I have updated configure.ac so that ./configure now finds
2717 the release number by grepping pcre.h.
2718
2719 26. Added the ability to run the tests under valgrind.
2720
2721
2722 Version 6.4 05-Sep-05
2723 ---------------------
2724
2725 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2726 "--" to be printed when multiple files were scanned, even when none of the
2727 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2728 consider it to be a bug, and have restored the previous behaviour.
2729
2730 2. A couple of code tidies to get rid of compiler warnings.
2731
2732 3. The pcretest program used to cheat by referring to symbols in the library
2733 whose names begin with _pcre_. These are internal symbols that are not
2734 really supposed to be visible externally, and in some environments it is
2735 possible to suppress them. The cheating is now confined to including
2736 certain files from the library's source, which is a bit cleaner.
2737
2738 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2739 file's purpose clearer.
2740
2741 5. Reorganized pcre_ucp_findchar().
2742
2743
2744 Version 6.3 15-Aug-05
2745 ---------------------
2746
2747 1. The file libpcre.pc.in did not have general read permission in the tarball.
2748
2749 2. There were some problems when building without C++ support:
2750
2751 (a) If C++ support was not built, "make install" and "make test" still
2752 tried to test it.
2753
2754 (b) There were problems when the value of CXX was explicitly set. Some
2755 changes have been made to try to fix these, and ...
2756
2757 (c) --disable-cpp can now be used to explicitly disable C++ support.
2758
2759 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2760 backslash in a target when C++ was disabled. This confuses some
2761 versions of "make", apparently. Using an intermediate variable solves
2762 this. (Same for CPP_LOBJ.)
2763
2764 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2765 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2766 necessary on certain architectures.
2767
2768 4. Added a setting of -export-symbols-regex to the link command to remove
2769 those symbols that are exported in the C sense, but actually are local
2770 within the library, and not documented. Their names all begin with
2771 "_pcre_". This is not a perfect job, because (a) we have to except some
2772 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2773 available (and never for static libraries). I have made a note to try to
2774 find a way round (a) in the future.
2775
2776
2777 Version 6.2 01-Aug-05
2778 ---------------------
2779
2780 1. There was no test for integer overflow of quantifier values. A construction
2781 such as {1111111111111111} would give undefined results. What is worse, if
2782 a minimum quantifier for a parenthesized subpattern overflowed and became
2783 negative, the calculation of the memory size went wrong. This could have
2784 led to memory overwriting.
2785
2786 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2787
2788 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2789 operating environments where this matters.
2790
2791 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2792 PCRE options from within the C++ wrapper.
2793
2794 5. Named capturing subpatterns were not being correctly counted when a pattern
2795 was compiled. This caused two problems: (a) If there were more than 100
2796 such subpatterns, the calculation of the memory needed for the whole
2797 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2798 back references of the form \12, where the number was greater than 9, were
2799 not recognized as back references, even though there were sufficient
2800 previous subpatterns.
2801
2802 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2803 versions of gcc, e.g. 2.95.4.
2804
2805
2806 Version 6.1 21-Jun-05
2807 ---------------------
2808
2809 1. There was one reference to the variable "posix" in pcretest.c that was not
2810 surrounded by "#if !defined NOPOSIX".
2811
2812 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2813 the cross-check on the old pcre_info() function, for the benefit of the
2814 cut-down version of PCRE that is currently imported into Exim.
2815
2816 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2817 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2818 patterns that start (?i) but I don't think that matters. The use of (?i) is
2819 just an example; this all applies to the other options as well.
2820
2821 4. Since libtool seems to echo the compile commands it is issuing, the output
2822 from "make" can be reduced a bit by putting "@" in front of each libtool
2823 compile command.
2824
2825 5. Patch from the folks at Google for configure.in to be a bit more thorough
2826 in checking for a suitable C++ installation before trying to compile the
2827 C++ stuff. This should fix a reported problem when a compiler was present,
2828 but no suitable headers.
2829
2830 6. The man pages all had just "PCRE" as their title. I have changed them to
2831 be the relevant file name. I have also arranged that these names are
2832 retained in the file doc/pcre.txt, which is a concatenation in text format
2833 of all the man pages except the little individual ones for each function.
2834
2835 7. The NON-UNIX-USE file had not been updated for the different set of source
2836 files that come with release 6. I also added a few comments about the C++
2837 wrapper.
2838
2839
2840 Version 6.0 07-Jun-05
2841 ---------------------
2842
2843 1. Some minor internal re-organization to help with my DFA experiments.
2844
2845 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2846 didn't matter for the library itself when fully configured, but did matter
2847 when compiling without UCP support, or within Exim, where the ucp files are
2848 not imported.
2849
2850 3. Refactoring of the library code to split up the various functions into
2851 different source modules. The addition of the new DFA matching code (see
2852 below) to a single monolithic source would have made it really too
2853 unwieldy, quite apart from causing all the code to be include in a
2854 statically linked application, when only some functions are used. This is
2855 relevant even without the DFA addition now that patterns can be compiled in
2856 one application and matched in another.
2857
2858 The downside of splitting up is that there have to be some external
2859 functions and data tables that are used internally in different modules of
2860 the library but which are not part of the API. These have all had their
2861 names changed to start with "_pcre_" so that they are unlikely to clash
2862 with other external names.
2863
2864 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2865 a different (DFA) algorithm. Although it is slower than the original
2866 function, it does have some advantages for certain types of matching
2867 problem.
2868
2869 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2870 including restarting after a partial match.
2871
2872 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2873 defined when compiling for Windows was sent to me. I have put it into the
2874 code, though I have no means of testing or verifying it.
2875
2876 7. Added the pcre_refcount() auxiliary function.
2877
2878 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2879 match before or at the first newline in the subject string. In pcretest,
2880 the /f option on a pattern can be used to set this.
2881
2882 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2883 would behave wrongly. This has been present in PCRE since release 4.0.
2884
2885 10. A number of changes to the pcregrep command:
2886
2887 (a) Refactored how -x works; insert ^(...)$ instead of setting
2888 PCRE_ANCHORED and checking the length, in preparation for adding
2889 something similar for -w.
2890
2891 (b) Added the -w (match as a word) option.
2892
2893 (c) Refactored the way lines are read and buffered so as to have more
2894 than one at a time available.
2895
2896 (d) Implemented a pcregrep test script.
2897
2898 (e) Added the -M (multiline match) option. This allows patterns to match
2899 over several lines of the subject. The buffering ensures that at least
2900 8K, or the rest of the document (whichever is the shorter) is available
2901 for matching (and similarly the previous 8K for lookbehind assertions).
2902
2903 (f) Changed the --help output so that it now says
2904
2905 -w, --word-regex(p)
2906
2907 instead of two lines, one with "regex" and the other with "regexp"
2908 because that confused at least one person since the short forms are the
2909 same. (This required a bit of code, as the output is generated
2910 automatically from a table. It wasn't just a text change.)
2911
2912 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2913 option but starts with a hyphen. Could be a pattern or a path name
2914 starting with a hyphen, for instance.
2915
2916 (h) "-" can be given as a file name to represent stdin.
2917
2918 (i) When file names are being printed, "(standard input)" is used for
2919 the standard input, for compatibility with GNU grep. Previously
2920 "<stdin>" was used.
2921
2922 (j) The option --label=xxx can be used to supply a name to be used for
2923 stdin when file names are being printed. There is no short form.
2924
2925 (k) Re-factored the options decoding logic because we are going to add
2926 two more options that take data. Such options can now be given in four
2927 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2928
2929 (l) Added the -A, -B, and -C options for requesting that lines of context
2930 around matches be printed.
2931
2932 (m) Added the -L option to print the names of files that do not contain
2933 any matching lines, that is, the complement of -l.
2934
2935 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2936 continue to scan other files.
2937
2938 (o) The -s option was incorrectly implemented. For compatibility with other
2939 greps, it now suppresses the error message for a non-existent or non-
2940 accessible file (but not the return code). There is a new option called
2941 -q that suppresses the output of matching lines, which was what -s was
2942 previously doing.
2943
2944 (p) Added --include and --exclude options to specify files for inclusion
2945 and exclusion when recursing.
2946
2947 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2948 Hopefully, it now does.
2949
2950 12. Missing cast in pcre_study().
2951
2952 13. Added an "uninstall" target to the makefile.
2953
2954 14. Replaced "extern" in the function prototypes in Makefile.in with
2955 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2956 world, but is set differently for Windows.
2957
2958 15. Added a second compiling function called pcre_compile2(). The only
2959 difference is that it has an extra argument, which is a pointer to an
2960 integer error code. When there is a compile-time failure, this is set
2961 non-zero, in addition to the error test pointer being set to point to an
2962 error message. The new argument may be NULL if no error number is required
2963 (but then you may as well call pcre_compile(), which is now just a
2964 wrapper). This facility is provided because some applications need a
2965 numeric error indication, but it has also enabled me to tidy up the way
2966 compile-time errors are handled in the POSIX wrapper.
2967
2968 16. Added VPATH=.libs to the makefile; this should help when building with one
2969 prefix path and installing with another. (Or so I'm told by someone who
2970 knows more about this stuff than I do.)
2971
2972 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2973 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2974 match everything, including newlines. This is not POSIX-compatible, but
2975 somebody wanted the feature. From pcretest it can be activated by using
2976 both the P and the s flags.
2977
2978 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2979
2980 19. libpcre.pc was being incorrectly installed as executable.
2981
2982 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2983 it now also looks for '\r' so that it will work unmodified on Windows.
2984
2985 21. Added Google's contributed C++ wrapper to the distribution.
2986
2987 22. Added some untidy missing memory free() calls in pcretest, to keep
2988 Electric Fence happy when testing.
2989
2990
2991
2992 Version 5.0 13-Sep-04
2993 ---------------------
2994
2995 1. Internal change: literal characters are no longer packed up into items
2996 containing multiple characters in a single byte-string. Each character
2997 is now matched using a separate opcode. However, there may be more than one
2998 byte in the character in UTF-8 mode.
2999
3000 2. The pcre_callout_block structure has two new fields: pattern_position and
3001 next_item_length. These contain the offset in the pattern to the next match
3002 item, and its length, respectively.
3003
3004 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3005 insertion of callouts before each pattern item. Added the /C option to
3006 pcretest to make use of this.
3007
3008 4. On the advice of a Windows user, the lines
3009
3010 #if defined(_WIN32) || defined(WIN32)
3011 _setmode( _fileno( stdout ), 0x8000 );
3012 #endif /* defined(_WIN32) || defined(WIN32) */
3013
3014 have been added to the source of pcretest. This apparently does useful
3015 magic in relation to line terminators.
3016
3017 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3018 for the benefit of those environments where the "b" makes a difference.
3019
3020 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3021 to know about it. I have put a hack into configure.in that adds in code
3022 to set GCC=yes if CC=icc. This seems to end up at a point in the
3023 generated configure script that is early enough to affect the setting of
3024 compiler options, which is what is needed, but I have no means of testing
3025 whether it really works. (The user who reported this had patched the
3026 generated configure script, which of course I cannot do.)
3027
3028 LATER: After change 22 below (new libtool files), the configure script
3029 seems to know about icc (and also ecc). Therefore, I have commented out
3030 this hack in configure.in.
3031
3032 7. Added support for pkg-config (2 patches were sent in).
3033
3034 8. Negated POSIX character classes that used a combination of internal tables
3035 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3036 [[:^ascii]]. Typically, they would match almost any characters. The other
3037 POSIX classes were not broken in this way.
3038
3039 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3040 to find the match, as PCRE was deluded into thinking that the match had to
3041 start at the start point or following a newline. The same bug applied to
3042 patterns with negative forward assertions or any backward assertions
3043 preceding ".*" at the start, unless the pattern required a fixed first
3044 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3045
3046 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3047 starting at the last subject character, bytes beyond the end of the subject
3048 string were read.
3049
3050 11. Renamed the variable "class" as "classbits" to make life easier for C++
3051 users. (Previously there was a macro definition, but it apparently wasn't
3052 enough.)
3053
3054 12. Added the new field "tables" to the extra data so that tables can be passed
3055 in at exec time, or the internal tables can be re-selected. This allows
3056 a compiled regex to be saved and re-used at a later time by a different
3057 program that might have everything at different addresses.
3058
3059 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3060 -R library as well as a -L library.
3061
3062 14. The debugging options of pcretest (-d on the command line or D on a
3063 pattern) showed incorrect output for anything following an extended class
3064 that contained multibyte characters and which was followed by a quantifier.
3065
3066 15. Added optional support for general category Unicode character properties
3067 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3068 support. It adds about 90K to the size of the library. The meanings of the
3069 inbuilt class escapes such as \d and \s have NOT been changed.
3070
3071 16. Updated pcredemo.c to include calls to free() to release the memory for the
3072 compiled pattern.
3073
3074 17. The generated file chartables.c was being created in the source directory
3075 instead of in the building directory. This caused the build to fail if the
3076 source directory was different from the building directory, and was
3077 read-only.
3078
3079 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3080 file. No doubt somebody will tell me if they don't make sense... Also added
3081 Dan Mooney's comments about building on OpenVMS.
3082
3083 19. Added support for partial matching via the PCRE_PARTIAL option for
3084 pcre_exec() and the \P data escape in pcretest.
3085
3086 20. Extended pcretest with 3 new pattern features:
3087
3088 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3089 write the compiled pattern to the file whose name is "rest-of-line".
3090 This is a straight binary dump of the data, with the saved pointer to
3091 the character tables forced to be NULL. The study data, if any, is
3092 written too. After writing, pcretest reads a new pattern.
3093
3094 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3095 compiled pattern from the given file. There must not be any
3096 occurrences of "<" in the file name (pretty unlikely); if there are,
3097 pcretest will instead treat the initial "<" as a pattern delimiter.
3098 After reading in the pattern, pcretest goes on to read data lines as
3099 usual.
3100
3101 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3102 and 16-bit fields in a compiled pattern, to simulate a pattern that
3103 was compiled on a host of opposite endianness.
3104
3105 21. The pcre-exec() function can now cope with patterns that were compiled on
3106 hosts of opposite endianness, with this restriction:
3107
3108 As for any compiled expression that is saved and used later, the tables
3109 pointer field cannot be preserved; the extra_data field in the arguments
3110 to pcre_exec() should be used to pass in a tables address if a value
3111 other than the default internal tables were used at compile time.
3112
3113 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3114 now diagnosed as an error. Previously, most of the time, a negative number
3115 would have been treated as zero, but if in addition "ovector" was passed as
3116 NULL, a crash could occur.
3117
3118 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3119 new versions from the libtool 1.5 distribution (the last one is a copy of
3120 a file called libtool.m4). This seems to have fixed the need to patch
3121 "configure" to support Darwin 1.3 (which I used to do). However, I still
3122 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3123 workstation).
3124
3125 24. Changed the PCRE licence to be the more standard "BSD" licence.
3126
3127
3128 Version 4.5 01-Dec-03
3129 ---------------------
3130
3131 1. There has been some re-arrangement of the code for the match() function so
3132 that it can be compiled in a version that does not call itself recursively.
3133 Instead, it keeps those local variables that need separate instances for
3134 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3135 needs to "recurse". Keeping track of where control must go is done by means
3136 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3137 hide most of the details from the main code, and operates only if
3138 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3139 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3140 operating.
3141
3142 To make it easier for callers to provide specially tailored get/free
3143 functions for this usage, two new functions, pcre_stack_malloc, and
3144 pcre_stack_free, are used. They are always called in strict stacking order,
3145 and the size of block requested is always the same.
3146
3147 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3148 PCRE has been compiled to use the stack or the heap for recursion. The
3149 -C option of pcretest uses this to show which version is compiled.
3150
3151 A new data escape \S, is added to pcretest; it causes the amounts of store
3152 obtained and freed by both kinds of malloc/free at match time to be added
3153 to the output.
3154
3155 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3156 what's available on my current Linux desktop machine.
3157
3158 3. When matching a UTF-8 string, the test for a valid string at the start has
3159 been extended. If start_offset is not zero, PCRE now checks that it points
3160 to a byte that is the start of a UTF-8 character. If not, it returns
3161 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3162 this is necessary because there may be backward assertions in the pattern.
3163 When matching the same subject several times, it may save resources to use
3164 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3165
3166 4. The code for checking the validity of UTF-8 strings has been tightened so
3167 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3168 containing "overlong sequences".
3169
3170 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3171 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3172 should have been "&", but it just so happened that all the cases this let
3173 through by mistake were picked up later in the function.
3174
3175 6. I had used a variable called "isblank" - this is a C99 function, causing
3176 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3177
3178 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3179 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3180 so the version is shown; (c) stop "make test" echoing "./RunTest".
3181
3182 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3183
3184 9. The prototype for memmove() for systems that don't have it was using
3185 size_t, but the inclusion of the header that defines size_t was later. I've
3186 moved the #includes for the C headers earlier to avoid this.
3187
3188 10. Added some adjustments to the code to make it easier to compiler on certain
3189 special systems:
3190
3191 (a) Some "const" qualifiers were missing.
3192 (b) Added the macro EXPORT before all exported functions; by default this
3193 is defined to be empty.
3194 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3195 that it reads its output file name as an argument instead of writing
3196 to the standard output and assuming this can be redirected.
3197
3198 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3199 class containing characters with values greater than 255, PCRE compilation
3200 went into a loop.
3201
3202 12. A recursive reference to a subpattern that was within another subpattern
3203 that had a minimum quantifier of zero caused PCRE to crash. For example,
3204 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3205 recursion. If the recursively-called subpattern itself had a zero repeat,
3206 that was OK.
3207
3208 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3209 buffer into which it was copied (for escape processing) was still set at
3210 1024, so long lines caused crashes.
3211
3212 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3213 "internal error: code overflow...". This applied to any character class
3214 that was followed by a possessive quantifier.
3215
3216 15. Modified the Makefile to add libpcre.la as a prerequisite for
3217 libpcreposix.la because I was told this is needed for a parallel build to
3218 work.
3219
3220 16. If a pattern that contained .* following optional items at the start was
3221 studied, the wrong optimizing data was generated, leading to matching
3222 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3223 matching string must start with a or b or c. The correct conclusion for
3224 this pattern is that a match can start with any character.
3225
3226
3227 Version 4.4 13-Aug-03
3228 ---------------------
3229
3230 1. In UTF-8 mode, a character class containing characters with values between
3231 127 and 255 was not handled correctly if the compiled pattern was studied.
3232 In fixing this, I have also improved the studying algorithm for such
3233 classes (slightly).
3234
3235 2. Three internal functions had redundant arguments passed to them. Removal
3236 might give a very teeny performance improvement.
3237
3238 3. Documentation bug: the value of the capture_top field in a callout is *one
3239 more than* the number of the hightest numbered captured substring.
3240
3241 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3242 in incorrectly linking with a previously installed version. They now link
3243 explicitly with libpcre.la.
3244
3245 5. configure.in no longer needs to recognize Cygwin specially.
3246
3247 6. A problem in pcre.in for Windows platforms is fixed.
3248
3249 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3250 pcretest, it used to include the size of the study block as part of its
3251 output. Unfortunately, the structure contains a field that has a different
3252 size on different hardware architectures. This meant that the tests that
3253 showed this size failed. As the block is currently always of a fixed size,
3254 this information isn't actually particularly useful in pcretest output, so
3255 I have just removed it.
3256
3257 8. Three pre-processor statements accidentally did not start in column 1.
3258 Sadly, there are *still* compilers around that complain, even though
3259 standard C has not required this for well over a decade. Sigh.
3260
3261 9. In pcretest, the code for checking callouts passed small integers in the
3262 callout_data field, which is a void * field. However, some picky compilers
3263 complained about the casts involved for this on 64-bit systems. Now
3264 pcretest passes the address of the small integer instead, which should get
3265 rid of the warnings.
3266
3267 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3268 both compile and run time, and gives an error if an invalid UTF-8 sequence
3269 is found. There is a option for disabling this check in cases where the
3270 string is known to be correct and/or the maximum performance is wanted.
3271
3272 11. In response to a bug report, I changed one line in Makefile.in from
3273
3274 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3275 to
3276 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3277
3278 to look similar to other lines, but I have no way of telling whether this
3279 is the right thing to do, as I do not use Windows. No doubt I'll get told
3280 if it's wrong...
3281
3282
3283 Version 4.3 21-May-03
3284 ---------------------
3285
3286 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3287 Makefile.
3288
3289 2. Some refactoring to improve the quality of the code:
3290
3291 (i) The utf8_table... variables are now declared "const".
3292
3293 (ii) The code for \cx, which used the "case flipping" table to upper case
3294 lower case letters, now just substracts 32. This is ASCII-specific,
3295 but the whole concept of \cx is ASCII-specific, so it seems
3296 reasonable.
3297
3298 (iii) PCRE was using its character types table to recognize decimal and
3299 hexadecimal digits in the pattern. This is silly, because it handles
3300 only 0-9, a-f, and A-F, but the character types table is locale-
3301 specific, which means strange things might happen. A private
3302 table is now used for this - though it costs 256 bytes, a table is
3303 much faster than multiple explicit tests. Of course, the standard
3304 character types table is still used for matching digits in subject
3305 strings against \d.
3306
3307 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3308 ending in _t are). So I've renamed it as ESC_tee.
3309
3310 3. The first argument for regexec() in the POSIX wrapper should have been
3311 defined as "const".
3312
3313 4. Changed pcretest to use malloc() for its buffers so that they can be
3314 Electric Fenced for debugging.
3315
3316 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3317 to read one or more bytes before the start of the subject string. Often this
3318 had no effect on PCRE's behaviour, but in some circumstances it could
3319 provoke a segmentation fault.
3320
3321 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3322 to try to read one or more bytes before the start of the subject string.
3323
3324 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3325 UTF-8 support could misbehave in various ways if the subject string
3326 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3327 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3328 back over UTF-8 characters.)
3329
3330
3331 Version 4.2 14-Apr-03
3332 ---------------------
3333
3334 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3335
3336 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3337 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3338 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3339 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3340 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3341 and BUILD_EXEEXT
3342 Note: automatic setting of the BUILD variables is not yet working
3343 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3344 compile-time but not at link-time
3345 [LINK]: use for linking executables only
3346 make different versions for Windows and non-Windows
3347 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3348 libraries
3349 [LINK_FOR_BUILD]: new variable
3350 [OBJEXT]: use throughout
3351 [EXEEXT]: use throughout
3352 <winshared>: new target
3353 <wininstall>: new target
3354 <dftables.o>: use native compiler
3355 <dftables>: use native linker
3356 <install>: handle Windows platform correctly
3357 <clean>: ditto
3358 <check>: ditto
3359 copy DLL to top builddir before testing
3360
3361 As part of these changes, -no-undefined was removed again. This was reported
3362 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3363 in any case.
3364
3365 3. Some tidies to get rid of compiler warnings:
3366
3367 . In the match_data structure, match_limit was an unsigned long int, whereas
3368 match_call_count was an int. I've made them both unsigned long ints.
3369
3370 . In pcretest the fact that a const uschar * doesn't automatically cast to
3371 a void * provoked a warning.
3372
3373 . Turning on some more compiler warnings threw up some "shadow" variables
3374 and a few more missing casts.
3375
3376 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3377 option, a class that contained a single character with a value between 128
3378 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3379
3380 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3381 option, a class that contained several characters, but with at least one
3382 whose value was between 128 and 255 caused PCRE to crash.
3383
3384
3385 Version 4.1 12-Mar-03
3386 ---------------------
3387
3388 1. Compiling with gcc -pedantic found a couple of places where casts were
3389 needed, and a string in dftables.c that was longer than standard compilers are
3390 required to support.
3391
3392 2. Compiling with Sun's compiler found a few more places where the code could
3393 be tidied up in order to avoid warnings.
3394
3395 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3396 first of these names is deprecated in the latest Autoconf in favour of the name
3397 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3398 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3399 analogy I have changed it to CFLAGS_FOR_BUILD.
3400
3401 4. Added -no-undefined to the linking command in the Makefile, because this is
3402 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3403 linking step for the pcreposix library.
3404
3405 5. PCRE was failing to diagnose the case of two named groups with the same
3406 name.
3407
3408 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3409 literal character that is needed in the subject for a match, and scans along to
3410 ensure that it is present before embarking on the full matching process. This
3411 saves time in cases of nested unlimited repeats that are never going to match.
3412 Problem: the scan can take a lot of time if the subject is very long (e.g.
3413 megabytes), thus penalizing straightforward matches. It is now done only if the
3414 amount of subject to be scanned is less than 1000 bytes.
3415
3416 7. A lesser problem with the same optimization is that it was recording the
3417 first character of an anchored pattern as "needed", thus provoking a search
3418 right along the subject, even when the first match of the pattern was going to
3419 fail. The "needed" character is now not set for anchored patterns, unless it
3420 follows something in the pattern that is of non-fixed length. Thus, it still
3421 fulfils its original purpose of finding quick non-matches in cases of nested
3422 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3423
3424
3425 Version 4.0 17-Feb-03
3426 ---------------------
3427
3428 1. If a comment in an extended regex that started immediately after a meta-item
3429 extended to the end of string, PCRE compiled incorrect data. This could lead to
3430 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3431
3432 2. Moved to autoconf 2.53 and libtool 1.4.2.
3433
3434 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3435 the special perltest8 script is no longer needed - all the tests can be run
3436 from a single perltest script.
3437
3438 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3439 by \s. It has now been removed in PCRE. This means it isn't recognized as
3440 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3441 class [:space:] *does* include VT, thereby creating a mess.
3442
3443 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3444 space and tab.
3445
3446 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3447 its new features into the main test script, reducing the number of scripts.
3448
3449 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3450 were backward compatible, and made the (?i) apply to the whole pattern, as if
3451 /i were given. Now it behaves more logically, and applies the option setting
3452 only to what follows. PCRE has been changed to follow suit. However, if it
3453 finds options settings right at the start of the pattern, it extracts them into
3454 the global options, as before. Thus, they show up in the info data.
3455
3456 8. Added support for the \Q...\E escape sequence. Characters in between are
3457 treated as literals. This is slightly different from Perl in that $ and @ are
3458 also handled as literals inside the quotes. In Perl, they will cause variable
3459 interpolation. Note the following examples:
3460
3461 Pattern PCRE matches Perl matches
3462
3463 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3464 \Qabc\$xyz\E abc\$xyz abc\$xyz
3465 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3466
3467 For compatibility with Perl, \Q...\E sequences are recognized inside character
3468 classes as well as outside them.
3469
3470 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3471 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3472 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3473 signed/unsigned warnings.
3474
3475 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3476 option for pcretest, so I've replaced it by a simple function that does just
3477 that job.
3478
3479 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3480 "pcregrep -".
3481
3482 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3483 Java package. This provides some syntactic sugar for simple cases of what my
3484 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3485 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3486 item, you can use this simplified notation. Note that only makes sense with
3487 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3488 greediness, whatever the setting of the PCRE_UNGREEDY option.
3489
3490 13. A change of greediness default within a pattern was not taking effect at
3491 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3492 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3493 was abstracted outside.
3494
3495 14. PCRE now supports the \G assertion. It is true when the current matching
3496 position is at the start point of the match. This differs from \A when the
3497 starting offset is non-zero. Used with the /g option of pcretest (or similar
3498 code), it works in the same way as it does for Perl's /g option. If all
3499 alternatives of a regex begin with \G, the expression is anchored to the start
3500 match position, and the "anchored" flag is set in the compiled expression.
3501
3502 15. Some bugs concerning the handling of certain option changes within patterns
3503 have been fixed. These applied to options other than (?ims). For example,
3504 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3505 been the other way round. Some of this was related to change 7 above.
3506
3507 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3508 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3509 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3510 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3511
3512 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3513 mode. Unlike ".", it always matches newline, whatever the setting of
3514 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3515 assertions. Perl allows it, but it doesn't (in general) work because it can't
3516 calculate the length of the lookbehind. At least, that's the case for Perl
3517 5.8.0 - I've been told they are going to document that it doesn't work in
3518 future.
3519
3520 18. Added an error diagnosis for escapes that PCRE does not support: these are
3521 \L, \l, \N, \P, \p, \U, \u, and \X.
3522
3523 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3524 reading past the end of the pattern in cases such as /[abcd/.
3525
3526 20. PCRE was getting more memory than necessary for patterns with classes that
3527 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3528
3529 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3530 compiling PCRE for use with Virtual Pascal.
3531
3532 22. Small fix to the Makefile to make it work properly if the build is done
3533 outside the source tree.
3534
3535 23. Added a new extension: a condition to go with recursion. If a conditional
3536 subpattern starts with (?(R) the "true" branch is used if recursion has
3537 happened, whereas the "false" branch is used only at the top level.
3538
3539 24. When there was a very long string of literal characters (over 255 bytes
3540 without UTF support, over 250 bytes with UTF support), the computation of how
3541 much memory was required could be incorrect, leading to segfaults or other
3542 strange effects.
3543
3544 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3545 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3546 there was a subsequent back reference to those brackets. This meant that, for
3547 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3548 possible to check for precisely this case. All we can do is abandon the
3549 optimization if .* occurs inside capturing brackets when there are any back
3550 references whatsoever. (See below for a better fix that came later.)
3551
3552 26. The handling of the optimization for finding the first character of a
3553 non-anchored pattern, and for finding a character that is required later in the
3554 match were failing in some cases. This didn't break the matching; it just
3555 failed to optimize when it could. The way this is done has been re-implemented.
3556
3557 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3558
3559 28. Added a new feature that provides some of the functionality that Perl
3560 provides with (?{...}). The facility is termed a "callout". The way it is done
3561 in PCRE is for the caller to provide an optional function, by setting
3562 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3563 global variable. By default it is unset, which disables all calling out. To get
3564 the function called, the regex must include (?C) at appropriate points. This
3565 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3566 This provides a means of identifying different callout points. When PCRE
3567 reaches such a point in the regex, if pcre_callout has been set, the external
3568 function is called. It is provided with data in a structure called
3569 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3570 matching continues; if it returns a non-zero value, the match at the current
3571 point fails. However, backtracking will occur if possible. [This was changed
3572 later and other features added - see item 49 below.]
3573
3574 29. pcretest is upgraded to test the callout functionality. It provides a
3575 callout function that displays information. By default, it shows the start of
3576 the match and the current position in the text. There are some new data escapes
3577 to vary what happens:
3578
3579 \C+ in addition, show current contents of captured substrings
3580 \C- do not supply a callout function
3581 \C!n return 1 when callout number n is reached
3582 \C!n!m return 1 when callout number n is reached for the mth time
3583
3584 30. If pcregrep was called with the -l option and just a single file name, it
3585 output "<stdin>" if a match was found, instead of the file name.
3586
3587 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3588 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3589 pcre_exec(). This saves a malloc/free per call. The default value of
3590 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3591 when configuring.
3592
3593 32. The default maximum size of a compiled pattern is 64K. There have been a
3594 few cases of people hitting this limit. The code now uses macros to handle the
3595 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3596 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3597 configuring. Tests 2 and 5 work only with 2-byte links because they output
3598 debugging information about compiled patterns.
3599
3600 33. Internal code re-arrangements:
3601
3602 (a) Moved the debugging function for printing out a compiled regex into
3603 its own source file (printint.c) and used #include to pull it into
3604 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3605 separate copies.
3606
3607 (b) Defined the list of op-code names for debugging as a macro in
3608 internal.h so that it is next to the definition of the opcodes.
3609
3610 (c) Defined a table of op-code lengths for simpler skipping along compiled
3611 code. This is again a macro in internal.h so that it is next to the
3612 definition of the opcodes.
3613
3614 34. Added support for recursive calls to individual subpatterns, along the
3615 lines of Robin Houston's patch (but implemented somewhat differently).
3616
3617 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3618 allow it to read and process whole directories in Win32. This code was
3619 contributed by Lionel Fourquaux; it has not been tested by me.
3620
3621 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3622 used to name a group. Names consist of alphanumerics and underscores, and must
3623 be unique. Back references use the syntax (?P=name) and recursive calls use
3624 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3625 numbers. The function pcre_fullinfo() can be used after compilation to extract
3626 a name/number map. There are three relevant calls:
3627
3628 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3629 PCRE_INFO_NAMECOUNT yields the number of entries
3630 PCRE_INFO_NAMETABLE yields a pointer to the map.
3631
3632 The map is a vector of fixed-size entries. The size of each entry depends on
3633 the length of the longest name used. The first two bytes of each entry are the
3634 group number, most significant byte first. There follows the corresponding
3635 name, zero terminated. The names are in alphabetical order.
3636
3637 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3638 case instead of 255. Making it the same both with and without UTF-8 support
3639 means that the same test output works with both.
3640
3641 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3642 calling malloc() with a zero argument.
3643
3644 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3645 optimization. I've improved things by keeping a bitmap of backreferences with
3646 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3647 fact referenced, the optimization can be applied. It is unlikely that a
3648 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3649 the match to follow \n) will appear inside brackets with a number greater than
3650 31, but if it does, any back reference > 31 suppresses the optimization.
3651
3652 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3653 of disabling numbered capturing parentheses. Any opening parenthesis that is
3654 not followed by ? behaves as if it were followed by ?: but named parentheses
3655 can still be used for capturing (and they will acquire numbers in the usual
3656 way).
3657
3658 41. Redesigned the return codes from the match() function into yes/no/error so
3659 that errors can be passed back from deep inside the nested calls. A malloc
3660 failure while inside a recursive subpattern call now causes the
3661 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3662
3663 42. It is now possible to set a limit on the number of times the match()
3664 function is called in a call to pcre_exec(). This facility makes it possible to
3665 limit the amount of recursion and backtracking, though not in a directly
3666 obvious way, because the match() function is used in a number of different
3667 circumstances. The count starts from zero for each position in the subject
3668 string (for non-anchored patterns). The default limit is, for compatibility, a
3669 large number, namely 10 000 000. You can change this in two ways:
3670
3671 (a) When configuring PCRE before making, you can use --with-match-limit=n
3672 to set a default value for the compiled library.
3673
3674 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3675 a different value is set. See 45 below.
3676
3677 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3678
3679 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3680 of things that can be changed at compile time. The first argument specifies
3681 what is wanted and the second points to where the information is to be placed.
3682 The current list of available information is:
3683
3684 PCRE_CONFIG_UTF8
3685
3686 The output is an integer that is set to one if UTF-8 support is available;
3687 otherwise it is set to zero.
3688
3689 PCRE_CONFIG_NEWLINE
3690
3691 The output is an integer that it set to the value of the code that is used for
3692 newline. It is either LF (10) or CR (13).
3693
3694 PCRE_CONFIG_LINK_SIZE
3695
3696 The output is an integer that contains the number of bytes used for internal
3697 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3698
3699 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3700
3701 The output is an integer that contains the threshold above which the POSIX
3702 interface uses malloc() for output vectors. See item 31 above.
3703
3704 PCRE_CONFIG_MATCH_LIMIT
3705
3706 The output is an unsigned integer that contains the default limit of the number
3707 of match() calls in a pcre_exec() execution. See 42 above.
3708
3709 44. pcretest has been upgraded by the addition of the -C option. This causes it
3710 to extract all the available output from the new pcre_config() function, and to
3711 output it. The program then exits immediately.
3712
3713 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3714 order to support additional features. One way would have been to define
3715 pcre_exec2() (for example) with extra arguments, but this would not have been
3716 extensible, and would also have required all calls to the original function to
3717 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3718 is used for passing in "extra" data from pcre_study().
3719
3720 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3721 contains the following fields:
3722
3723 flags a bitmap indicating which of the following fields are set
3724 study_data opaque data from pcre_study()
3725 match_limit a way of specifying a limit on match() calls for a specific
3726 call to pcre_exec()
3727 callout_data data for callouts (see 49 below)
3728
3729 The flag bits are also defined in pcre.h, and are
3730
3731 PCRE_EXTRA_STUDY_DATA
3732 PCRE_EXTRA_MATCH_LIMIT
3733 PCRE_EXTRA_CALLOUT_DATA
3734
3735 The pcre_study() function now returns one of these new pcre_extra blocks, with
3736 the actual study data pointed to by the study_data field, and the
3737 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3738 before. That is, this change is entirely upwards-compatible and requires no
3739 change to existing code.
3740
3741 If you want to pass in additional data to pcre_exec(), you can either place it
3742 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3743 block.
3744
3745 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3746 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3747 times with different match limits, until it finds the minimum value needed for
3748 pcre_exec() to complete. The value is then output. This can be instructive; for
3749 most simple matches the number is quite small, but for pathological cases it
3750 gets very large very quickly.
3751
3752 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3753 returns the size of the data block pointed to by the study_data field in a
3754 pcre_extra block, that is, the value that was passed as the argument to
3755 pcre_malloc() when PCRE was getting memory in which to place the information
3756 created by pcre_study(). The fourth argument should point to a size_t variable.
3757 pcretest has been extended so that this information is shown after a successful
3758 pcre_study() call when information about the compiled regex is being displayed.
3759
3760 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3761 because what follows is always an absolute path. (Later: it turns out that this
3762 is more than cosmetic for MinGW, because it doesn't like empty path
3763 components.)
3764
3765 49. Some changes have been made to the callout feature (see 28 above):
3766
3767 (i) A callout function now has three choices for what it returns:
3768
3769 0 => success, carry on matching
3770 > 0 => failure at this point, but backtrack if possible
3771 < 0 => serious error, return this value from pcre_exec()
3772
3773 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3774 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3775 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3776 use by callout functions. It will never be used by PCRE itself.
3777
3778 (ii) The pcre_extra structure (see 45 above) has a void * field called
3779 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3780 pcre_callout_block structure has a field of the same name. The contents of
3781 the field passed in the pcre_extra structure are passed to the callout
3782 function in the corresponding field in the callout block. This makes it
3783 easier to use the same callout-containing regex from multiple threads. For
3784 testing, the pcretest program has a new data escape
3785
3786 \C*n pass the number n (may be negative) as callout_data
3787
3788 If the callout function in pcretest receives a non-zero value as
3789 callout_data, it returns that value.
3790
3791 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3792 there were some redundant $(CFLAGS) in commands that are now specified as
3793 $(LINK), which already includes $(CFLAGS).
3794
3795 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3796 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3797 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3798 one-byte characters throughout. Note that case-insensitive matching applies
3799 only to characters whose values are less than 256. PCRE doesn't support the
3800 notion of cases for higher-valued characters.
3801
3802 (i) A character class whose characters are all within 0-255 is handled as
3803 a bit map, and the map is inverted for negative classes. Previously, a
3804 character > 255 always failed to match such a class; however it should
3805 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3806
3807 (ii) A negated character class with a single character < 255 is coded as
3808 "not this character" (OP_NOT). This wasn't working properly when the test
3809 character was multibyte, either singly or repeated.
3810
3811 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3812 mode, for example: \x{100}{2,3}.
3813
3814 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3815 singly or repeated) now correctly test multibyte characters. However,
3816 PCRE doesn't recognize any characters with values greater than 255 as
3817 digits, spaces, or word characters. Such characters always match \D, \S,
3818 and \W, and never match \d, \s, or \w.
3819
3820 (v) Classes may now contain characters and character ranges with values
3821 greater than 255. For example: [ab\x{100}-\x{400}].
3822
3823 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3824 PCRE in UTF-8 mode.
3825
3826 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3827 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3828 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3829 value.)
3830
3831 53. The single man page has become too large. I have therefore split it up into
3832 a number of separate man pages. These also give rise to individual HTML pages;
3833 these are now put in a separate directory, and there is an index.html page that
3834 lists them all. Some hyperlinking between the pages has been installed.
3835
3836 54. Added convenience functions for handling named capturing parentheses.
3837
3838 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3839 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3840 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3841 are faulted.
3842
3843 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3844 calling configure. These values are used when compiling the dftables.c program
3845 which is run to generate the source of the default character tables. They
3846 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3847 you will need to set these values.
3848
3849 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3850
3851
3852 Version 3.9 02-Jan-02
3853 ---------------------
3854
3855 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3856
3857 2. If --disable-static was given, the building process failed when trying to
3858 build pcretest and pcregrep. (For some reason it was using libtool to compile
3859 them, which is not right, as they aren't part of the library.)
3860
3861
3862 Version 3.8 18-Dec-01
3863 ---------------------
3864
3865 1. The experimental UTF-8 code was completely screwed up. It was packing the
3866 bytes in the wrong order. How dumb can you get?
3867
3868
3869 Version 3.7 29-Oct-01
3870 ---------------------
3871
3872 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3873 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3874 this didn't happen under Solaris 8, where I normally test things.
3875
3876 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3877 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3878 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3879 this fix an hour or so after the initial 3.7 release.)
3880
3881
3882 Version 3.6 23-Oct-01
3883 ---------------------
3884
3885 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3886 offsets passed as NULL with zero offset count.
3887
3888 2. The config.guess and config.sub files had not been updated when I moved to
3889 the latest autoconf.
3890
3891
3892 Version 3.5 15-Aug-01
3893 ---------------------
3894
3895 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3896 had been forgotten.
3897
3898 2. By using declared but undefined structures, we can avoid using "void"
3899 definitions in pcre.h while keeping the internal definitions of the structures
3900 private.
3901
3902 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3903 user point of view, this means that both static and shared libraries are built
3904 by default, but this can be individually controlled. More of the work of
3905 handling this static/shared cases is now inside libtool instead of PCRE's make
3906 file.
3907
3908 4. The pcretest utility is now installed along with pcregrep because it is
3909 useful for users (to test regexs) and by doing this, it automatically gets
3910 relinked by libtool. The documentation has been turned into a man page, so
3911 there are now .1, .txt, and .html versions in /doc.
3912
3913 5. Upgrades to pcregrep:
3914 (i) Added long-form option names like gnu grep.
3915 (ii) Added --help to list all options with an explanatory phrase.
3916 (iii) Added -r, --recursive to recurse into sub-directories.
3917 (iv) Added -f, --file to read patterns from a file.
3918
3919 6. pcre_exec() was referring to its "code" argument before testing that
3920 argument for NULL (and giving an error if it was NULL).
3921
3922 7. Upgraded Makefile.in to allow for compiling in a different directory from
3923 the source directory.
3924
3925 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3926 options bits, the pointer it was passed was to an int instead of to an unsigned
3927 long int. This mattered only on 64-bit systems.
3928
3929 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3930 generated) instead of pcre.in, which it its source. Also made the same change
3931 in several of the .c files.
3932
3933 10. A new release of gcc defines printf() as a macro, which broke pcretest
3934 because it had an ifdef in the middle of a string argument for printf(). Fixed
3935 by using separate calls to printf().
3936
3937 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3938 script, to force use of CR or LF instead of \n in the source. On non-Unix
3939 systems, the value can be set in config.h.
3940
3941 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3942 absolute limit. Changed the text of the error message to make this clear, and
3943 likewise updated the man page.
3944
3945 13. The limit of 99 on the number of capturing subpatterns has been removed.
3946 The new limit is 65535, which I hope will not be a "real" limit.
3947
3948
3949 Version 3.4 22-Aug-00
3950 ---------------------
3951
3952 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3953
3954 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3955
3956
3957 Version 3.3 01-Aug-00
3958 ---------------------
3959
3960 1. If an octal character was given, but the value was greater than \377, it
3961 was not getting masked to the least significant bits, as documented. This could
3962 lead to crashes in some systems.
3963
3964 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3965 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3966
3967 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3968 These just pass their arguments on to (pcre_free)(), but they are provided
3969 because some uses of PCRE bind it to non-C systems that can call its functions,
3970 but cannot call free() or pcre_free() directly.
3971
3972 4. Add "make test" as a synonym for "make check". Corrected some comments in
3973 the Makefile.
3974
3975 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3976 Makefile.
3977
3978 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3979 command called pgrep for grepping around the active processes.
3980
3981 7. Added the beginnings of support for UTF-8 character strings.
3982
3983 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3984 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3985 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3986 out for the ar command.)
3987
3988
3989 Version 3.2 12-May-00
3990 ---------------------
3991
3992 This is purely a bug fixing release.
3993
3994 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3995 of ZA. This was just one example of several cases that could provoke this bug,
3996 which was introduced by change 9 of version 2.00. The code for breaking
3997 infinite loops after an iteration that matches an empty string was't working
3998 correctly.
3999
4000 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4001 when matched against abbab (for example). After matching an empty string, it
4002 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4003 caused it to match further down the string than it should.
4004
4005 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4006 was there because it doesn't seem to be needed, and it causes trouble on some
4007 systems, as it is not a Standard C header. It has been removed.
4008
4009 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4010 were reported on the Macintosh. The changes were from
4011
4012 while ((c = *(++ptr)) != 0 && c != '\n');
4013 to
4014 while ((c = *(++ptr)) != 0 && c != '\n') ;
4015
4016 Totally extraordinary, but if that's what it takes...
4017
4018 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4019 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4020 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4021 assumes the way PCRE uses memmove() (always moving upwards).
4022
4023 6. PCRE is being used in one environment where strchr() is not available. There
4024 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4025 faster code anyway.
4026
4027
4028 Version 3.1 09-Feb-00
4029 ---------------------
4030
4031 The only change in this release is the fixing of some bugs in Makefile.in for
4032 the "install" target:
4033
4034 (1) It was failing to install pcreposix.h.
4035
4036 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4037
4038
4039 Version 3.0 01-Feb-00
4040 ---------------------
4041
4042 1. Add support for the /+ modifier to perltest (to output $` like it does in
4043 pcretest).
4044
4045 2. Add support for the /g modifier to perltest.
4046
4047 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4048 matches null strings.
4049
4050 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4051 pattern. Perl treats empty patterns specially - it reuses the most recent
4052 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4053 effect.
4054
4055 5. The POSIX interface was broken in that it was just handing over the POSIX
4056 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4057 required a bigger vector, with some working space on the end. This means that
4058 the POSIX wrapper now has to get and free some memory, and copy the results.
4059
4060 6. Added some simple autoconf support, placing the test data and the
4061 documentation in separate directories, re-organizing some of the
4062 information files, and making it build pcre-config (a GNU standard). Also added
4063 libtool support for building PCRE as a shared library, which is now the
4064 default.
4065
4066 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4067 09 are not valid octal constants. Single digits will be used for minor values
4068 less than 10.
4069
4070 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4071 existing programs that set these in the POSIX interface can use PCRE without
4072 modification.
4073
4074 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4075 return all that pcre_info() returns, plus additional data. The pcre_info()
4076 function is retained for compatibility, but is considered to be obsolete.
4077
4078 10. Added experimental recursion feature (?R) to handle one common case that
4079 Perl 5.6 will be able to do with (?p{...}).
4080
4081 11. Added support for POSIX character classes like [:alpha:], which Perl is
4082 adopting.
4083
4084
4085 Version 2.08 31-Aug-99
4086 ----------------------
4087
4088 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4089 trying to match at the startoffset position, but instead was moving forward to
4090 the next newline as if a previous match had failed.
4091
4092 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4093 and could get into a loop if a null string was matched other than at the start
4094 of the subject.
4095
4096 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4097 be distinguished at compile time, and for completeness also added PCRE_DATE.
4098
4099 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4100 in GnuWin32 environments.
4101
4102
4103 Version 2.07 29-Jul-99
4104 ----------------------
4105
4106 1. The documentation is now supplied in plain text form and HTML as well as in
4107 the form of man page sources.
4108
4109 2. C++ compilers don't like assigning (void *) values to other pointer types.
4110 In particular this affects malloc(). Although there is no problem in Standard
4111 C, I've put in casts to keep C++ compilers happy.
4112
4113 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4114 should be (const char *).
4115
4116 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4117 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4118 However, I haven't made this a standard facility. The documentation doesn't
4119 mention it, and the Makefile doesn't support it.
4120
4121 5. The Makefile now contains an "install" target, with editable destinations at
4122 the top of the file. The pcretest program is not installed.
4123
4124 6. pgrep -V now gives the PCRE version number and date.
4125
4126 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4127 causing the entire string to be ignored, instead of just the last character.
4128
4129 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4130 non-matching string, it can take a very, very long time, even for strings of
4131 quite modest length, because of the nested recursion. PCRE now does better in
4132 some of these cases. It does this by remembering the last required literal
4133 character in the pattern, and pre-searching the subject to ensure it is present
4134 before running the real match. In other words, it applies a heuristic to detect
4135 some types of certain failure quickly, and in the above example, if presented
4136 with a string that has no trailing " it gives "no match" very quickly.
4137
4138 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4139 other alternatives are tried instead.
4140
4141
4142 Version 2.06 09-Jun-99
4143 ----------------------
4144
4145 1. Change pcretest's output for amount of store used to show just the code
4146 space, because the remainder (the data block) varies in size between 32-bit and
4147 64-bit systems.
4148
4149 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4150 start matching at. This allows lookbehinds to work when searching for multiple
4151 occurrences in a string.
4152
4153 3. Added additional options to pcretest for testing multiple occurrences:
4154
4155 /+ outputs the rest of the string that follows a match
4156 /g loops for multiple occurrences, using the new startoffset argument
4157 /G loops for multiple occurrences by passing an incremented pointer
4158
4159 4. PCRE wasn't doing the "first character" optimization for patterns starting
4160 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4161 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4162 the letter 'x'. On long subject strings, this gives a significant speed-up.
4163
4164
4165 Version 2.05 21-Apr-99
4166 ----------------------
4167
4168 1. Changed the type of magic_number from int to long int so that it works
4169 properly on 16-bit systems.
4170
4171 2. Fixed a bug which caused patterns starting with .* not to work correctly
4172 when the subject string contained newline characters. PCRE was assuming
4173 anchoring for such patterns in all cases, which is not correct because .* will
4174 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4175 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4176 must be retried after every newline in the subject.
4177
4178
4179 Version 2.04 18-Feb-99
4180 ----------------------
4181
4182 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4183 computation of the store needed to hold the pattern was incorrect (too large).
4184 If such patterns were nested a few deep, this could multiply and become a real
4185 problem.
4186
4187 2. Added /M option to pcretest to show the memory requirement of a specific
4188 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4189
4190 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4191 compiled in such a way that the backtracking after subsequent failure was
4192 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4193 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4194
4195
4196 Version 2.03 02-Feb-99
4197 ----------------------
4198
4199 1. Fixed typo and small mistake in man page.
4200
4201 2. Added 4th condition (GPL supersedes if conflict) and created separate
4202 LICENCE file containing the conditions.
4203
4204 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4205 Perl, that is the internal \ allows the delimiter to be included in the
4206 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4207 the final delimiter, add \ to the end of the pattern (to test the error).
4208
4209 4. Added the convenience functions for extracting substrings after a successful
4210 match. Updated pcretest to make it able to test these functions.
4211
4212
4213 Version 2.02 14-Jan-99
4214 ----------------------
4215
4216 1. Initialized the working variables associated with each extraction so that
4217 their saving and restoring doesn't refer to uninitialized store.
4218
4219 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4220 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4221 fix the problem.
4222
4223 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4224 calls, and wasn't printing the correct value for compiling calls. Increased the
4225 default value of LOOPREPEAT, and the number of significant figures in the
4226 times.
4227
4228 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4229
4230 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4231 a building problem on Windows NT with a FAT file system.
4232
4233
4234 Version 2.01 21-Oct-98
4235 ----------------------
4236
4237 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4238 to character tables built by pcre_maketables() in the current locale. If NULL
4239 is passed, the default tables are used.
4240
4241
4242 Version 2.00 24-Sep-98
4243 ----------------------
4244
4245 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4246 it any more.
4247
4248 2. Allow quantification of (?>) groups, and make it work correctly.
4249
4250 3. The first character computation wasn't working for (?>) groups.
4251
4252 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4253 end of the subject) and add 5.005's \z, which really does match only at the
4254 very end of the subject.
4255
4256 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4257
4258 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4259 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4260 localized options. All options to pcre_study() were also removed.
4261
4262 7. Add other new features from 5.005:
4263
4264 $(?<= positive lookbehind
4265 $(?<! negative lookbehind
4266 (?imsx-imsx) added the unsetting capability
4267 such a setting is global if at outer level; local otherwise
4268 (?imsx-imsx:) non-capturing groups with option setting
4269 (?(cond)re|re) conditional pattern matching
4270
4271 A backreference to itself in a repeated group matches the previous
4272 captured string.
4273
4274 8. General tidying up of studying (both automatic and via "study")
4275 consequential on the addition of new assertions.
4276
4277 9. As in 5.005, unlimited repeated groups that could match an empty substring
4278 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4279 runtime if any iteration does actually match an empty substring.
4280
4281 10. Include the RunTest script in the distribution.
4282
4283 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4284 discrepancies, some of which were old and were also with respect to 5.004. They
4285 have now been fixed.
4286
4287
4288 Version 1.09 28-Apr-98
4289 ----------------------
4290
4291 1. A negated single character class followed by a quantifier with a minimum
4292 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4293 program crashes, or just wrong answers. This did not apply to negated classes
4294 containing more than one character, or to minima other than one.
4295
4296
4297 Version 1.08 27-Mar-98
4298 ----------------------
4299
4300 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4301
4302 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4303 latter must appear before anything that relies on it in the pattern.
4304
4305
4306 Version 1.07 16-Feb-98
4307 ----------------------
4308
4309 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4310 repeat of a potentially empty string).
4311
4312
4313 Version 1.06 23-Jan-98
4314 ----------------------
4315
4316 1. Added Markus Oberhumer's little patches for C++.
4317
4318 2. Literal strings longer than 255 characters were broken.
4319
4320
4321 Version 1.05 23-Dec-97
4322 ----------------------
4323
4324 1. Negated character classes containing more than one character were failing if
4325 PCRE_CASELESS was set at run time.
4326
4327
4328 Version 1.04 19-Dec-97
4329 ----------------------
4330
4331 1. Corrected the man page, where some "const" qualifiers had been omitted.
4332
4333 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4334 input syntax.
4335
4336 3. Fixed memory leak which occurred when a regex with back references was
4337 matched with an offsets vector that wasn't big enough. The temporary memory
4338 that is used in this case wasn't being freed if the match failed.
4339
4340 4. Tidied pcretest to ensure it frees memory that it gets.
4341
4342 5. Temporary memory was being obtained in the case where the passed offsets
4343 vector was exactly big enough.
4344
4345 6. Corrected definition of offsetof() from change 5 below.
4346
4347 7. I had screwed up change 6 below and broken the rules for the use of
4348 setjmp(). Now fixed.
4349
4350
4351 Version 1.03 18-Dec-97
4352 ----------------------
4353
4354 1. A erroneous regex with a missing opening parenthesis was correctly
4355 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4356 on some systems.
4357
4358 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4359 it was reported that one broken compiler failed on the former because "code" is
4360 also an independent variable.
4361
4362 3. The erroneous regex a[]b caused an array overrun reference.
4363
4364 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4365 fail on data ending with that character. (It was going on too far, and checking
4366 the next character, typically a binary zero.) This was specific to the
4367 optimized code for single-character negative classes.
4368
4369 5. Added a contributed patch from the TIN world which does the following:
4370
4371 + Add an undef for memmove, in case the the system defines a macro for it.
4372
4373 + Add a definition of offsetof(), in case there isn't one. (I don't know
4374 the reason behind this - offsetof() is part of the ANSI standard - but
4375 it does no harm).
4376
4377 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4378 most of the places where whitespace preceded '#'. I have given up and
4379 allowed the remaining 2 cases to be at the margin.
4380
4381 + Rename some variables in pcre to eliminate shadowing. This seems very
4382 pedantic, but does no harm, of course.
4383
4384 6. Moved the call to setjmp() into its own function, to get rid of warnings
4385 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4386
4387 7. Constructs such as \d{8,} were compiling into the equivalent of
4388 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4389 outcome, but in this particular case used more store than had been allocated,
4390 which caused the bug to be discovered because it threw up an internal error.
4391
4392 8. The debugging code in both pcre and pcretest for outputting the compiled
4393 form of a regex was going wrong in the case of back references followed by
4394 curly-bracketed repeats.
4395
4396
4397 Version 1.02 12-Dec-97
4398 ----------------------
4399
4400 1. Typos in pcre.3 and comments in the source fixed.
4401
4402 2. Applied a contributed patch to get rid of places where it used to remove
4403 'const' from variables, and fixed some signed/unsigned and uninitialized
4404 variable warnings.
4405
4406 3. Added the "runtest" target to Makefile.
4407
4408 4. Set default compiler flag to -O2 rather than just -O.
4409
4410
4411 Version 1.01 19-Nov-97
4412 ----------------------
4413
4414 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4415 like /([ab]*)*/, that is, for classes with more than one character in them.
4416
4417 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4418 as /((?>a*))*/ (a PCRE_EXTRA facility).
4419
4420
4421 Version 1.00 18-Nov-97
4422 ----------------------
4423
4424 1. Added compile-time macros to support systems such as SunOS4 which don't have
4425 memmove() or strerror() but have other things that can be used instead.
4426
4427 2. Arranged that "make clean" removes the executables.
4428
4429
4430 Version 0.99 27-Oct-97
4431 ----------------------
4432
4433 1. Fixed bug in code for optimizing classes with only one character. It was
4434 initializing a 32-byte map regardless, which could cause it to run off the end
4435 of the memory it had got.
4436
4437 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4438
4439
4440 Version 0.98 22-Oct-97
4441 ----------------------
4442
4443 1. Fixed bug in code for handling temporary memory usage when there are more
4444 back references than supplied space in the ovector. This could cause segfaults.
4445
4446
4447 Version 0.97 21-Oct-97
4448 ----------------------
4449
4450 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4451
4452 2. Optimized negated single characters not to use a bit map.
4453
4454 3. Brought error texts together as macro definitions; clarified some of them;
4455 fixed one that was wrong - it said "range out of order" when it meant "invalid
4456 escape sequence".
4457
4458 4. Changed some char * arguments to const char *.
4459
4460 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4461
4462 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4463 pcretest.
4464
4465
4466 Version 0.96 16-Oct-97
4467 ----------------------
4468
4469 1. Added a simple "pgrep" utility to the distribution.
4470
4471 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4472 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4473 where "ddd" means "one or more decimal digits".
4474
4475 3. Fixed serious bug. If a pattern had a back reference, but the call to
4476 pcre_exec() didn't supply a large enough ovector to record the related
4477 identifying subpattern, the match always failed. PCRE now remembers the number
4478 of the largest back reference, and gets some temporary memory in which to save
4479 the offsets during matching if necessary, in order to ensure that
4480 backreferences always work.
4481
4482 4. Increased the compatibility with Perl in a number of ways:
4483
4484 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4485 to request this handling. The option can be set at compile or exec time.
4486
4487 (b) $ matches before a terminating newline by default; an option
4488 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4489 mode). The option can be set at compile or exec time.
4490
4491 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4492 the same as Perl's. If the decimal number it represents is less than 10
4493 or there aren't that many previous left capturing parentheses, an octal
4494 escape is read. Inside a character class, it's always an octal escape,
4495 even if it is a single digit.
4496
4497 (d) An escaped but undefined alphabetic character is taken as a literal,
4498 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4499 escapes.
4500
4501 (e) {0} is now permitted. (The previous item is removed from the compiled
4502 pattern).
4503
4504 5. Changed all the names of code files so that the basic parts are no longer
4505 than 10 characters, and abolished the teeny "globals.c" file.
4506
4507 6. Changed the handling of character classes; they are now done with a 32-byte
4508 bit map always.
4509
4510 7. Added the -d and /D options to pcretest to make it possible to look at the
4511 internals of compilation without having to recompile pcre.
4512
4513
4514 Version 0.95 23-Sep-97
4515 ----------------------
4516
4517 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4518 \x20 at the start of a run of normal characters. These were being treated as
4519 real characters, instead of the source characters being re-checked.
4520
4521
4522 Version 0.94 18-Sep-97
4523 ----------------------
4524
4525 1. The functions are now thread-safe, with the caveat that the global variables
4526 containing pointers to malloc() and free() or alternative functions are the
4527 same for all threads.
4528
4529 2. Get pcre_study() to generate a bitmap of initial characters for non-
4530 anchored patterns when this is possible, and use it if passed to pcre_exec().
4531
4532
4533 Version 0.93 15-Sep-97
4534 ----------------------
4535
4536 1. /(b)|(:+)/ was computing an incorrect first character.
4537
4538 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4539 but not actually doing anything yet.
4540
4541 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4542 as Perl does (e.g. [-az] or [az-]).
4543
4544 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4545 all possible positions.
4546
4547 5. Split up into different modules to avoid including unneeded functions in a
4548 compiled binary. However, compile and exec are still in one module. The "study"
4549 function is split off.
4550
4551 6. The character tables are now in a separate module whose source is generated
4552 by an auxiliary program - but can then be edited by hand if required. There are
4553 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4554 toupper() in the code.
4555
4556 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4557 make them global. Abolish the function for setting them, as the caller can now
4558 set them directly.
4559
4560
4561 Version 0.92 11-Sep-97
4562 ----------------------
4563
4564 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4565 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4566
4567 2. Caseless matching was not working in character classes if the characters in
4568 the pattern were in upper case.
4569
4570 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4571
4572 4. Make PCRE_ANCHORED public and accept as a compile option.
4573
4574 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4575 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4576 pass them.
4577
4578 6. Give an error if bad option bits passed at compile or run time.
4579
4580 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4581 pcretest to cause it to pass that flag.
4582
4583 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4584 options, and the first character, if set.
4585
4586 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4587
4588
4589 Version 0.91 10-Sep-97
4590 ----------------------
4591
4592 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4593 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4594
4595 2. PCRE was looping on encountering an indefinitely repeated back reference to
4596 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4597 Perl does - treats the match as successful.
4598
4599 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5