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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 873 - (show annotations) (download)
Sat Jan 14 16:45:24 2012 UTC (2 years, 7 months ago) by ph10
File size: 224764 byte(s)
Documentation minor edits.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.30
5 ------------
6
7 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
8 name is defined in ctype.h.
9
10 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
11 only in quite long subpatterns.
12
13 3. Removed the function pcre_info(), which has been obsolete and deprecated
14 since it was replaced by pcre_fullinfo() in February 2000.
15
16 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
17 match a (*MARK), and the match failed at the start of the subject, a
18 reference to memory before the start of the subject could occur. This bug
19 was introduced by fix 17 of release 8.21.
20
21 5. A reference to an unset group with zero minimum repetition was giving
22 totally wrong answers (in non-JavaScript-compatibility mode). For example,
23 /(another)?(\1?)test/ matched against "hello world test". This bug was
24 introduced in release 8.13.
25
26 6. Add support for 16-bit character strings (a large amount of work involving
27 many changes and refactorings).
28
29 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31 from a file.
32
33 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34 rounding is not applied in this particular case).
35
36 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
37 if they appear, or are escaped, in patterns.
38
39
40 Version 8.21 12-Dec-2011
41 ------------------------
42
43 1. Updating the JIT compiler.
44
45 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
46 are added as well.
47
48 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
49 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
50 calling _pcre_jit_exec. Some extra comments are added.
51
52 4. (*MARK) settings inside atomic groups that do not contain any capturing
53 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
54 was introduced by change 18 for 8.20.
55
56 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
57 ECMA-262 standard.
58
59 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
60 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
61 This bug was probably introduced by change 9 of 8.13.
62
63 7. While fixing 6 above, I noticed that a number of other items were being
64 incorrectly rejected as "not fixed length". This arose partly because newer
65 opcodes had not been added to the fixed-length checking code. I have (a)
66 corrected the bug and added tests for these items, and (b) arranged for an
67 error to occur if an unknown opcode is encountered while checking for fixed
68 length instead of just assuming "not fixed length". The items that were
69 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
70 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
71 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
72
73 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
74 being incorrectly compiled and would have given unpredicatble results.
75
76 9. A possessively repeated subpattern with minimum repeat count greater than
77 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
78 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
79 the first (A) could occur when it should not.
80
81 10. Add a cast and remove a redundant test from the code.
82
83 11. JIT should use pcre_malloc/pcre_free for allocation.
84
85 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
86 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
87 is anything other than /usr, -L is still shown).
88
89 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
90
91 14. Perl does not support \N without a following name in a [] class; PCRE now
92 also gives an error.
93
94 15. If a forward reference was repeated with an upper limit of around 2000,
95 it caused the error "internal error: overran compiling workspace". The
96 maximum number of forward references (including repeats) was limited by the
97 internal workspace, and dependent on the LINK_SIZE. The code has been
98 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
99 the default depends on LINK_SIZE. There is a new upper limit (for safety)
100 of around 200,000 forward references. While doing this, I also speeded up
101 the filling in of repeated forward references.
102
103 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
104 incorrectly expecting the subject to contain another "a" after the start.
105
106 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
107 in the match, the SKIP should be ignored. This was not happening; instead
108 the SKIP was being treated as NOMATCH. For patterns such as
109 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
110 tested.
111
112 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
113 now much more compatible with Perl, in particular in cases where the result
114 is a non-match for a non-anchored pattern. For example, if
115 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
116 "m", where previously it did not return a name. A side effect of this
117 change is that for partial matches, the last encountered mark name is
118 returned, as for non matches. A number of tests that were previously not
119 Perl-compatible have been moved into the Perl-compatible test files. The
120 refactoring has had the pleasing side effect of removing one argument from
121 the match() function, thus reducing its stack requirements.
122
123 19. If the /S+ option was used in pcretest to study a pattern using JIT,
124 subsequent uses of /S (without +) incorrectly behaved like /S+.
125
126 21. Retrieve executable code size support for the JIT compiler and fixing
127 some warnings.
128
129 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
130 not work when the shorter character appeared right at the end of the
131 subject string.
132
133 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
134 systems.
135
136 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
137 output it when the /M option is used in pcretest.
138
139 25. The CheckMan script was not being included in the distribution. Also, added
140 an explicit "perl" to run Perl scripts from the PrepareRelease script
141 because this is reportedly needed in Windows.
142
143 26. If study data was being save in a file and studying had not found a set of
144 "starts with" bytes for the pattern, the data written to the file (though
145 never used) was taken from uninitialized memory and so caused valgrind to
146 complain.
147
148 27. Updated RunTest.bat as provided by Sheri Pierce.
149
150 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
151
152 29. Computation of memory usage for the table of capturing group names was
153 giving an unnecessarily large value.
154
155
156 Version 8.20 21-Oct-2011
157 ------------------------
158
159 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
160 a POSIX class. After further experiments with Perl, which convinced me that
161 Perl has bugs and confusions, a closing square bracket is no longer allowed
162 in a POSIX name. This bug also affected patterns with classes that started
163 with full stops.
164
165 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
166 captured substring, but while checking the failing first alternative,
167 substring 1 is temporarily captured. If the output vector supplied to
168 pcre_exec() was not big enough for this capture, the yield of the function
169 was still zero ("insufficient space for captured substrings"). This cannot
170 be totally fixed without adding another stack variable, which seems a lot
171 of expense for a edge case. However, I have improved the situation in cases
172 such as /(a)(b)x|abc/ matched against "abc", where the return code
173 indicates that fewer than the maximum number of slots in the ovector have
174 been set.
175
176 3. Related to (2) above: when there are more back references in a pattern than
177 slots in the output vector, pcre_exec() uses temporary memory during
178 matching, and copies in the captures as far as possible afterwards. It was
179 using the entire output vector, but this conflicts with the specification
180 that only 2/3 is used for passing back captured substrings. Now it uses
181 only the first 2/3, for compatibility. This is, of course, another edge
182 case.
183
184 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
185 main code base, and can be used by building with --enable-jit. When this is
186 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
187 runtime --no-jit option is given.
188
189 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
190 ovector, the return from the function was zero, implying that there were
191 other matches that did not fit. The correct "exactly full" value is now
192 returned.
193
194 6. If a subpattern that was called recursively or as a subroutine contained
195 (*PRUNE) or any other control that caused it to give a non-standard return,
196 invalid errors such as "Error -26 (nested recursion at the same subject
197 position)" or even infinite loops could occur.
198
199 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
200 computing the minimum length on reaching *ACCEPT, and so ended up with the
201 wrong value of 1 rather than 0. Further investigation indicates that
202 computing a minimum subject length in the presence of *ACCEPT is difficult
203 (think back references, subroutine calls), and so I have changed the code
204 so that no minimum is registered for a pattern that contains *ACCEPT.
205
206 8. If (*THEN) was present in the first (true) branch of a conditional group,
207 it was not handled as intended. [But see 16 below.]
208
209 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
210 Sheri Pierce.
211
212 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
213 the first byte in a match must be "a".
214
215 11. Change 17 for 8.13 increased the recursion depth for patterns like
216 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
217 pattern contains any instances of (*THEN). If it does not, the old
218 optimizations are restored. It would be nice to do this on a per-group
219 basis, but at the moment that is not feasible.
220
221 12. In some environments, the output of pcretest -C is CRLF terminated. This
222 broke RunTest's code that checks for the link size. A single white space
223 character after the value is now allowed for.
224
225 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
226 For "fr", it uses the Windows-specific input and output files.
227
228 14. If (*THEN) appeared in a group that was called recursively or as a
229 subroutine, it did not work as intended. [But see next item.]
230
231 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
232 pattern fragments (but not containing any | characters). If A and B are
233 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
234 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
235 D. In other words, Perl considers parentheses that do not contain any |
236 characters to be part of a surrounding alternative, whereas PCRE was
237 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
238 differently. PCRE now behaves in the same way as Perl, except in the case
239 of subroutine/recursion calls such as (?1) which have in any case always
240 been different (but PCRE had them first :-).
241
242 16. Related to 15 above: Perl does not treat the | in a conditional group as
243 creating alternatives. Such a group is treated in the same way as an
244 ordinary group without any | characters when processing (*THEN). PCRE has
245 been changed to match Perl's behaviour.
246
247 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
248 RunGrepTest script failed.
249
250 18. Change 22 for version 13 caused atomic groups to use more stack. This is
251 inevitable for groups that contain captures, but it can lead to a lot of
252 stack use in large patterns. The old behaviour has been restored for atomic
253 groups that do not contain any capturing parentheses.
254
255 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
256 suppress the check for a minimum subject length at run time. (If it was
257 given to pcre_exec() or pcre_dfa_exec() it did work.)
258
259 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
260 fail to work when decoding hex characters in data strings in EBCDIC
261 environments.
262
263 21. It appears that in at least one Mac OS environment, the isxdigit() function
264 is implemented as a macro that evaluates to its argument more than once,
265 contravening the C 90 Standard (I haven't checked a later standard). There
266 was an instance in pcretest which caused it to go wrong when processing
267 \x{...} escapes in subject strings. The has been rewritten to avoid using
268 things like p++ in the argument of isxdigit().
269
270
271 Version 8.13 16-Aug-2011
272 ------------------------
273
274 1. The Unicode data tables have been updated to Unicode 6.0.0.
275
276 2. Two minor typos in pcre_internal.h have been fixed.
277
278 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
279 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
280 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
281
282 4. There were a number of related bugs in the code for matching backrefences
283 caselessly in UTF-8 mode when codes for the characters concerned were
284 different numbers of bytes. For example, U+023A and U+2C65 are an upper
285 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
286 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
287 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
288 2-byte code at the end of the subject (it thought there wasn't enough data
289 left).
290
291 5. Comprehensive information about what went wrong is now returned by
292 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
293 as the output vector has at least 2 elements. The offset of the start of
294 the failing character and a reason code are placed in the vector.
295
296 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
297 now returned is for the first byte of the failing character, instead of the
298 last byte inspected. This is an incompatible change, but I hope it is small
299 enough not to be a problem. It makes the returned offset consistent with
300 pcre_exec() and pcre_dfa_exec().
301
302 7. pcretest now gives a text phrase as well as the error number when
303 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
304 failure, the offset and reason code are output.
305
306 8. When \R was used with a maximizing quantifier it failed to skip backwards
307 over a \r\n pair if the subsequent match failed. Instead, it just skipped
308 back over a single character (\n). This seems wrong (because it treated the
309 two characters as a single entity when going forwards), conflicts with the
310 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
311 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
312 has been changed.
313
314 9. Some internal refactoring has changed the processing so that the handling
315 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
316 time (the PCRE_DOTALL option was changed this way some time ago: version
317 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
318 which was always a bit of a fudge. It also means that there is one less
319 argument for the match() function, which reduces its stack requirements
320 slightly. This change also fixes an incompatibility with Perl: the pattern
321 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
322
323 10. More internal refactoring has drastically reduced the number of recursive
324 calls to match() for possessively repeated groups such as (abc)++ when
325 using pcre_exec().
326
327 11. While implementing 10, a number of bugs in the handling of groups were
328 discovered and fixed:
329
330 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
331 (a|)*(?1) gave a compile-time internal error.
332 ((a|)+)+ did not notice that the outer group could match an empty string.
333 (^a|^)+ was not marked as anchored.
334 (.*a|.*)+ was not marked as matching at start or after a newline.
335
336 12. Yet more internal refactoring has removed another argument from the match()
337 function. Special calls to this function are now indicated by setting a
338 value in a variable in the "match data" data block.
339
340 13. Be more explicit in pcre_study() instead of relying on "default" for
341 opcodes that mean there is no starting character; this means that when new
342 ones are added and accidentally left out of pcre_study(), testing should
343 pick them up.
344
345 14. The -s option of pcretest has been documented for ages as being an old
346 synonym of -m (show memory usage). I have changed it to mean "force study
347 for every regex", that is, assume /S for every regex. This is similar to -i
348 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
349 using it. It makes it easier to run collections of tests with and without
350 study enabled, and thereby test pcre_study() more easily. All the standard
351 tests are now run with and without -s (but some patterns can be marked as
352 "never study" - see 20 below).
353
354 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
355 restoration of the capturing data to the outer values was not happening
356 correctly.
357
358 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
359 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
360 pattern had matched an empty string, and so incorrectly returned a no
361 match.
362
363 17. There was optimizing code for the last branch of non-capturing parentheses,
364 and also for the obeyed branch of a conditional subexpression, which used
365 tail recursion to cut down on stack usage. Unfortunately, now that there is
366 the possibility of (*THEN) occurring in these branches, tail recursion is
367 no longer possible because the return has to be checked for (*THEN). These
368 two optimizations have therefore been removed. [But see 8.20/11 above.]
369
370 18. If a pattern containing \R was studied, it was assumed that \R always
371 matched two bytes, thus causing the minimum subject length to be
372 incorrectly computed because \R can also match just one byte.
373
374 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
375 was incorrectly computed.
376
377 20. If /S is present twice on a test pattern in pcretest input, it now
378 *disables* studying, thereby overriding the use of -s on the command line
379 (see 14 above). This is necessary for one or two tests to keep the output
380 identical in both cases.
381
382 21. When (*ACCEPT) was used in an assertion that matched an empty string and
383 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
384
385 22. When an atomic group that contained a capturing parenthesis was
386 successfully matched, but the branch in which it appeared failed, the
387 capturing was not being forgotten if a higher numbered group was later
388 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
389 group 1 to "a", when in fact it should be unset. This applied to multi-
390 branched capturing and non-capturing groups, repeated or not, and also to
391 positive assertions (capturing in negative assertions does not happen
392 in PCRE) and also to nested atomic groups.
393
394 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
395 subject after a captured substring, to make it easier to tell which of a
396 number of identical substrings has been captured.
397
398 24. The way atomic groups are processed by pcre_exec() has been changed so that
399 if they are repeated, backtracking one repetition now resets captured
400 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
401 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
402 "aaa". Previously, it would have been "a". As part of this code
403 refactoring, the way recursive calls are handled has also been changed.
404
405 25. If an assertion condition captured any substrings, they were not passed
406 back unless some other capturing happened later. For example, if
407 (?(?=(a))a) was matched against "a", no capturing was returned.
408
409 26. When studying a pattern that contained subroutine calls or assertions,
410 the code for finding the minimum length of a possible match was handling
411 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
412 group 1 called group 2 while simultaneously a separate group 2 called group
413 1). A stack overflow occurred in this case. I have fixed this by limiting
414 the recursion depth to 10.
415
416 27. Updated RunTest.bat in the distribution to the version supplied by Tom
417 Fortmann. This supports explicit test numbers on the command line, and has
418 argument validation and error reporting.
419
420 28. An instance of \X with an unlimited repeat could fail if at any point the
421 first character it looked at was a mark character.
422
423 29. Some minor code refactoring concerning Unicode properties and scripts
424 should reduce the stack requirement of match() slightly.
425
426 30. Added the '=' option to pcretest to check the setting of unused capturing
427 slots at the end of the pattern, which are documented as being -1, but are
428 not included in the return count.
429
430 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
431 compiled something random. Now it gives a compile-time error (as does
432 Perl).
433
434 32. A *MARK encountered during the processing of a positive assertion is now
435 recorded and passed back (compatible with Perl).
436
437 33. If --only-matching or --colour was set on a pcregrep call whose pattern
438 had alternative anchored branches, the search for a second match in a line
439 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
440 matched the line "0102" twice. The same bug affected patterns that started
441 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
442 twice.
443
444 34. Previously, PCRE did not allow quantification of assertions. However, Perl
445 does, and because of capturing effects, quantifying parenthesized
446 assertions may at times be useful. Quantifiers are now allowed for
447 parenthesized assertions.
448
449 35. A minor code tidy in pcre_compile() when checking options for \R usage.
450
451 36. \g was being checked for fancy things in a character class, when it should
452 just be a literal "g".
453
454 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
455 appearance of a nested POSIX class supersedes an apparent external class.
456 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
457 unescaped square brackets may also appear as part of class names. For
458 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
459 more like Perl. (But see 8.20/1 above.)
460
461 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
462 was because it thought it was \N{name}, which is not supported).
463
464 39. Add minix to OS list not supporting the -S option in pcretest.
465
466 40. PCRE tries to detect cases of infinite recursion at compile time, but it
467 cannot analyze patterns in sufficient detail to catch mutual recursions
468 such as ((?1))((?2)). There is now a runtime test that gives an error if a
469 subgroup is called recursively as a subpattern for a second time at the
470 same position in the subject string. In previous releases this might have
471 been caught by the recursion limit, or it might have run out of stack.
472
473 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
474 happen only once. PCRE was, however incorrectly giving a compile time error
475 "recursive call could loop indefinitely" because it cannot analyze the
476 pattern in sufficient detail. The compile time test no longer happens when
477 PCRE is compiling a conditional subpattern, but actual runaway loops are
478 now caught at runtime (see 40 above).
479
480 42. It seems that Perl allows any characters other than a closing parenthesis
481 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
482 has been changed to be the same.
483
484 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
485 as not to get warnings when autogen.sh is called. Also changed
486 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
487
488 44. To help people who use pcregrep to scan files containing exceedingly long
489 lines, the following changes have been made:
490
491 (a) The default value of the buffer size parameter has been increased from
492 8K to 20K. (The actual buffer used is three times this size.)
493
494 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
495 PCRE is built.
496
497 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
498 to be set at run time.
499
500 (d) Numerical values in pcregrep options can be followed by K or M, for
501 example --buffer-size=50K.
502
503 (e) If a line being scanned overflows pcregrep's buffer, an error is now
504 given and the return code is set to 2.
505
506 45. Add a pointer to the latest mark to the callout data block.
507
508 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
509 partial match of an empty string instead of no match. This was specific to
510 the use of ".".
511
512 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
513 complete match instead of a partial match. This bug was dependent on both
514 the PCRE_UTF8 and PCRE_DOTALL options being set.
515
516 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
517 starting byte set, because \b was not being ignored.
518
519
520 Version 8.12 15-Jan-2011
521 ------------------------
522
523 1. Fixed some typos in the markup of the man pages, and wrote a script that
524 checks for such things as part of the documentation building process.
525
526 2. On a big-endian 64-bit system, pcregrep did not correctly process the
527 --match-limit and --recursion-limit options (added for 8.11). In
528 particular, this made one of the standard tests fail. (The integer value
529 went into the wrong half of a long int.)
530
531 3. If the --colour option was given to pcregrep with -v (invert match), it
532 did strange things, either producing crazy output, or crashing. It should,
533 of course, ignore a request for colour when reporting lines that do not
534 match.
535
536 4. Another pcregrep bug caused similar problems if --colour was specified with
537 -M (multiline) and the pattern match finished with a line ending.
538
539 5. In pcregrep, when a pattern that ended with a literal newline sequence was
540 matched in multiline mode, the following line was shown as part of the
541 match. This seems wrong, so I have changed it.
542
543 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
544 the check for further matches in the same line (so they could be coloured)
545 to overrun the end of the current line. If another match was found, it was
546 incorrectly shown (and then shown again when found in the next line).
547
548 7. If pcregrep was compiled under Windows, there was a reference to the
549 function pcregrep_exit() before it was defined. I am assuming this was
550 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
551 reported by a user. I've moved the definition above the reference.
552
553
554 Version 8.11 10-Dec-2010
555 ------------------------
556
557 1. (*THEN) was not working properly if there were untried alternatives prior
558 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
559 backtracked to try for "b" instead of moving to the next alternative branch
560 at the same level (in this case, to look for "c"). The Perl documentation
561 is clear that when (*THEN) is backtracked onto, it goes to the "next
562 alternative in the innermost enclosing group".
563
564 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
565 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
566 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
567 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
568 (*THEN).
569
570 3. If \s appeared in a character class, it removed the VT character from
571 the class, even if it had been included by some previous item, for example
572 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
573 of \s, but is part of the POSIX "space" class.)
574
575 4. A partial match never returns an empty string (because you can always
576 match an empty string at the end of the subject); however the checking for
577 an empty string was starting at the "start of match" point. This has been
578 changed to the "earliest inspected character" point, because the returned
579 data for a partial match starts at this character. This means that, for
580 example, /(?<=abc)def/ gives a partial match for the subject "abc"
581 (previously it gave "no match").
582
583 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
584 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
585 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
586 has an implication that the given string is incomplete (because a partial
587 match is preferred over a full match). For this reason, these items now
588 give a partial match in this situation. [Aside: previously, the one case
589 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
590 match rather than a full match, which was wrong by the old rules, but is
591 now correct.]
592
593 6. There was a bug in the handling of #-introduced comments, recognized when
594 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
595 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
596 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
597 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
598 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
599 places in pcre_compile().
600
601 7. Related to (6) above, when pcre_compile() was skipping #-introduced
602 comments when looking ahead for named forward references to subpatterns,
603 the only newline sequence it recognized was NL. It now handles newlines
604 according to the set newline convention.
605
606 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
607 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
608 cater for a lack of strerror(). These oversights have been fixed.
609
610 9. Added --match-limit and --recursion-limit to pcregrep.
611
612 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
613
614 11. When the -o option was used, pcregrep was setting a return code of 1, even
615 when matches were found, and --line-buffered was not being honoured.
616
617 12. Added an optional parentheses number to the -o and --only-matching options
618 of pcregrep.
619
620 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
621 can match an empty string. The code to do it in pcretest and pcredemo
622 needed fixing:
623
624 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
625 only one byte after an empty string match just before CRLF (this case
626 just got forgotten; "any" and "anycrlf" were OK).
627
628 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
629 mode when an empty string match preceded an ASCII character followed by
630 a non-ASCII character. (The code for advancing by one character rather
631 than one byte was nonsense.)
632
633 (c) The pcredemo.c sample program did not have any code at all to handle
634 the cases when CRLF is a valid newline sequence.
635
636 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
637 as a starting offset was within the subject string. There is now a new
638 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
639 negative or greater than the length of the string. In order to test this,
640 pcretest is extended to allow the setting of negative starting offsets.
641
642 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
643 starting offset points to the beginning of a UTF-8 character was
644 unnecessarily clumsy. I tidied it up.
645
646 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
647 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
648
649 17. Nobody had reported that the --include_dir option, which was added in
650 release 7.7 should have been called --include-dir (hyphen, not underscore)
651 for compatibility with GNU grep. I have changed it to --include-dir, but
652 left --include_dir as an undocumented synonym, and the same for
653 --exclude-dir, though that is not available in GNU grep, at least as of
654 release 2.5.4.
655
656 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
657 characters from a string of bytes) have been redefined so as not to use
658 loops, in order to improve performance in some environments. At the same
659 time, I abstracted some of the common code into auxiliary macros to save
660 repetition (this should not affect the compiled code).
661
662 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
663 compile-time error is now given if \c is not followed by an ASCII
664 character, that is, a byte less than 128. (In EBCDIC mode, the code is
665 different, and any byte value is allowed.)
666
667 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
668 START_OPTIMIZE option, which is now allowed at compile time - but just
669 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
670 to pcregrep and other applications that have no direct access to PCRE
671 options. The new /Y option in pcretest sets this option when calling
672 pcre_compile().
673
674 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
675 back references. Groups containing recursive back references were forced to
676 be atomic by that change, but in the case of named groups, the amount of
677 memory required was incorrectly computed, leading to "Failed: internal
678 error: code overflow". This has been fixed.
679
680 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
681 pcretest.c, to avoid build problems in some Borland environments.
682
683
684 Version 8.10 25-Jun-2010
685 ------------------------
686
687 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
688 THEN.
689
690 2. (*ACCEPT) was not working when inside an atomic group.
691
692 3. Inside a character class, \B is treated as a literal by default, but
693 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
694 causes the error). The code is unchanged, but I tidied the documentation.
695
696 4. Inside a character class, PCRE always treated \R and \X as literals,
697 whereas Perl faults them if its -w option is set. I have changed PCRE so
698 that it faults them when PCRE_EXTRA is set.
699
700 5. Added support for \N, which always matches any character other than
701 newline. (It is the same as "." when PCRE_DOTALL is not set.)
702
703 6. When compiling pcregrep with newer versions of gcc which may have
704 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
705 declared with attribute warn_unused_result" were given. Just casting the
706 result to (void) does not stop the warnings; a more elaborate fudge is
707 needed. I've used a macro to implement this.
708
709 7. Minor change to pcretest.c to avoid a compiler warning.
710
711 8. Added four artifical Unicode properties to help with an option to make
712 \s etc use properties (see next item). The new properties are: Xan
713 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
714
715 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
716 use Unicode properties. (*UCP) at the start of a pattern can be used to set
717 this option. Modified pcretest to add /W to test this facility. Added
718 REG_UCP to make it available via the POSIX interface.
719
720 10. Added --line-buffered to pcregrep.
721
722 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
723 studied, and the match started with a letter with a code point greater than
724 127 whose first byte was different to the first byte of the other case of
725 the letter, the other case of this starting letter was not recognized
726 (#976).
727
728 12. If a pattern that was studied started with a repeated Unicode property
729 test, for example, \p{Nd}+, there was the theoretical possibility of
730 setting up an incorrect bitmap of starting bytes, but fortunately it could
731 not have actually happened in practice until change 8 above was made (it
732 added property types that matched character-matching opcodes).
733
734 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
735 possible starting bytes for non-anchored patterns.
736
737 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
738 \R, and also a number of cases that involve Unicode properties, both
739 explicit and implicit when PCRE_UCP is set.
740
741 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
742 input, it could crash or give wrong results if characters with values
743 greater than 0xc0 were present in the subject string. (Detail: it assumed
744 UTF-8 input when processing these items.)
745
746 16. Added a lot of (int) casts to avoid compiler warnings in systems where
747 size_t is 64-bit (#991).
748
749 17. Added a check for running out of memory when PCRE is compiled with
750 --disable-stack-for-recursion (#990).
751
752 18. If the last data line in a file for pcretest does not have a newline on
753 the end, a newline was missing in the output.
754
755 19. The default pcre_chartables.c file recognizes only ASCII characters (values
756 less than 128) in its various bitmaps. However, there is a facility for
757 generating tables according to the current locale when PCRE is compiled. It
758 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
759 characters, are recognized by isspace() and therefore were getting set in
760 these tables, and indeed these tables seem to approximate to ISO 8859. This
761 caused a problem in UTF-8 mode when pcre_study() was used to create a list
762 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
763 which of course cannot start UTF-8 characters. I have changed the code so
764 that only real ASCII characters (less than 128) and the correct starting
765 bytes for UTF-8 encodings are set for characters greater than 127 when in
766 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
767 altogether.)
768
769 20. Added the /T option to pcretest so as to be able to run tests with non-
770 standard character tables, thus making it possible to include the tests
771 used for 19 above in the standard set of tests.
772
773 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
774 reference to a subpattern the other side of a comment that contains an
775 opening parenthesis caused either an internal compiling error, or a
776 reference to the wrong subpattern.
777
778
779 Version 8.02 19-Mar-2010
780 ------------------------
781
782 1. The Unicode data tables have been updated to Unicode 5.2.0.
783
784 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
785 configured.
786
787 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
788 original author of that file, following a query about its status.
789
790 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
791 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
792
793 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
794 quantifier applied to a forward-referencing subroutine call, could compile
795 incorrect code or give the error "internal error: previously-checked
796 referenced subpattern not found".
797
798 6. Both MS Visual Studio and Symbian OS have problems with initializing
799 variables to point to external functions. For these systems, therefore,
800 pcre_malloc etc. are now initialized to local functions that call the
801 relevant global functions.
802
803 7. There were two entries missing in the vectors called coptable and poptable
804 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
805 I've fixed the data, and added a kludgy way of testing at compile time that
806 the lengths are correct (equal to the number of opcodes).
807
808 8. Following on from 7, I added a similar kludge to check the length of the
809 eint vector in pcreposix.c.
810
811 9. Error texts for pcre_compile() are held as one long string to avoid too
812 much relocation at load time. To find a text, the string is searched,
813 counting zeros. There was no check for running off the end of the string,
814 which could happen if a new error number was added without updating the
815 string.
816
817 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
818
819 11. \K was not working if it appeared in an atomic group or in a group that
820 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
821 \K is "not well defined" if used in an assertion. PCRE now accepts it if
822 the assertion is positive, but not if it is negative.
823
824 12. Change 11 fortuitously reduced the size of the stack frame used in the
825 "match()" function of pcre_exec.c by one pointer. Forthcoming
826 implementation of support for (*MARK) will need an extra pointer on the
827 stack; I have reserved it now, so that the stack frame size does not
828 decrease.
829
830 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
831 item in branch that calls a recursion is a subroutine call - as in the
832 second branch in the above example - was incorrectly given the compile-
833 time error "recursive call could loop indefinitely" because pcre_compile()
834 was not correctly checking the subroutine for matching a non-empty string.
835
836 14. The checks for overrunning compiling workspace could trigger after an
837 overrun had occurred. This is a "should never occur" error, but it can be
838 triggered by pathological patterns such as hundreds of nested parentheses.
839 The checks now trigger 100 bytes before the end of the workspace.
840
841 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
842
843
844 Version 8.01 19-Jan-2010
845 ------------------------
846
847 1. If a pattern contained a conditional subpattern with only one branch (in
848 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
849 computed the wrong minimum data length (which is of course zero for such
850 subpatterns). This could cause incorrect "no match" results.
851
852 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
853 the pattern is reset in the first branch, pcre_compile() failed with
854 "internal error: code overflow at offset...". This happened only when
855 the reset was to the original external option setting. (An optimization
856 abstracts leading options settings into an external setting, which was the
857 cause of this.)
858
859 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
860 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
861 assertion pattern did not match (meaning that the assertion was true), it
862 was incorrectly treated as false if the SKIP had been reached during the
863 matching. This also applied to assertions used as conditions.
864
865 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
866 assertion subpattern, including such a pattern used as a condition,
867 unpredictable results occurred, instead of the error return
868 PCRE_ERROR_DFA_UITEM.
869
870 5. The C++ GlobalReplace function was not working like Perl for the special
871 situation when an empty string is matched. It now does the fancy magic
872 stuff that is necessary.
873
874 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
875 removed. (These were left over from very, very early versions of PCRE.)
876
877 7. Some cosmetic changes to the code to make life easier when compiling it
878 as part of something else:
879
880 (a) Change DEBUG to PCRE_DEBUG.
881
882 (b) In pcre_compile(), rename the member of the "branch_chain" structure
883 called "current" as "current_branch", to prevent a collision with the
884 Linux macro when compiled as a kernel module.
885
886 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
887 prevent a collision with the Linux macro when compiled as a kernel
888 module.
889
890 8. In pcre_compile() there are some checks for integer overflows that used to
891 cast potentially large values to (double). This has been changed to that
892 when building, a check for int64_t is made, and if it is found, it is used
893 instead, thus avoiding the use of floating point arithmetic. (There is no
894 other use of FP in PCRE.) If int64_t is not found, the fallback is to
895 double.
896
897 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
898 2005 (difference between two addresses compared to an unsigned value).
899
900 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
901 custom one, because of the following reported problem in Windows:
902
903 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
904 under Win32.
905 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
906 therefore missing the function definition.
907 - The compiler thus generates a "C" signature for the test function.
908 - The linker fails to find the "C" function.
909 - PCRE fails to configure if asked to do so against libbz2.
910
911 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
912 messages were output:
913
914 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
915 rerunning libtoolize, to keep the correct libtool macros in-tree.
916 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
917
918 I have done both of these things.
919
920 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
921 most of the time, it *can* run out if it is given a pattern that contains a
922 runaway infinite recursion. I updated the discussion in the pcrestack man
923 page.
924
925 13. Now that we have gone to the x.xx style of version numbers, the minor
926 version may start with zero. Using 08 or 09 is a bad idea because users
927 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
928 interpreted as invalid octal numbers. I've updated the previous comment in
929 configure.ac, and also added a check that gives an error if 08 or 09 are
930 used.
931
932 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
933 causing partial matching to fail when the end of the subject matched \W
934 in a UTF-8 pattern where \W was quantified with a minimum of 3.
935
936 15. There were some discrepancies between the declarations in pcre_internal.h
937 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
938 their definitions. The declarations used "const uschar *" and the
939 definitions used USPTR. Even though USPTR is normally defined as "const
940 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
941 reported that: "This difference in casting confuses some C++ compilers, for
942 example, SunCC recognizes above declarations as different functions and
943 generates broken code for hbpcre." I have changed the declarations to use
944 USPTR.
945
946 16. GNU libtool is named differently on some systems. The autogen.sh script now
947 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
948 (FreeBSD).
949
950 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
951 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
952 comment: "Figure out how to create a longlong from a string: strtoll and
953 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
954 instance, but it only takes 2 args instead of 3!"
955
956 18. A subtle bug concerned with back references has been fixed by a change of
957 specification, with a corresponding code fix. A pattern such as
958 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
959 refers, was giving matches when it shouldn't. For example, xa=xaaa would
960 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
961 same bug. Such groups have to be quantified to be useful, or contained
962 inside another quantified group. (If there's no repetition, the reference
963 can never match.) The problem arises because, having left the group and
964 moved on to the rest of the pattern, a later failure that backtracks into
965 the group uses the captured value from the final iteration of the group
966 rather than the correct earlier one. I have fixed this in PCRE by forcing
967 any group that contains a reference to itself to be an atomic group; that
968 is, there cannot be any backtracking into it once it has completed. This is
969 similar to recursive and subroutine calls.
970
971
972 Version 8.00 19-Oct-09
973 ----------------------
974
975 1. The table for translating pcre_compile() error codes into POSIX error codes
976 was out-of-date, and there was no check on the pcre_compile() error code
977 being within the table. This could lead to an OK return being given in
978 error.
979
980 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
981 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
982 in a Windows environment.
983
984 3. The pcregrep --count option prints the count for each file even when it is
985 zero, as does GNU grep. However, pcregrep was also printing all files when
986 --files-with-matches was added. Now, when both options are given, it prints
987 counts only for those files that have at least one match. (GNU grep just
988 prints the file name in this circumstance, but including the count seems
989 more useful - otherwise, why use --count?) Also ensured that the
990 combination -clh just lists non-zero counts, with no names.
991
992 4. The long form of the pcregrep -F option was incorrectly implemented as
993 --fixed_strings instead of --fixed-strings. This is an incompatible change,
994 but it seems right to fix it, and I didn't think it was worth preserving
995 the old behaviour.
996
997 5. The command line items --regex=pattern and --regexp=pattern were not
998 recognized by pcregrep, which required --regex pattern or --regexp pattern
999 (with a space rather than an '='). The man page documented the '=' forms,
1000 which are compatible with GNU grep; these now work.
1001
1002 6. No libpcreposix.pc file was created for pkg-config; there was just
1003 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1004
1005 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1006 when UCP support is not needed, by modifying the Python script that
1007 generates it from Unicode data files. This should not matter if the module
1008 is correctly used as a library, but I received one complaint about 50K of
1009 unwanted data. My guess is that the person linked everything into his
1010 program rather than using a library. Anyway, it does no harm.
1011
1012 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1013 was a minimum greater than 1 for a wide character in a possessive
1014 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1015 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1016 character. Chaos in the form of incorrect output or a compiling loop could
1017 result.
1018
1019 9. The restrictions on what a pattern can contain when partial matching is
1020 requested for pcre_exec() have been removed. All patterns can now be
1021 partially matched by this function. In addition, if there are at least two
1022 slots in the offset vector, the offset of the earliest inspected character
1023 for the match and the offset of the end of the subject are set in them when
1024 PCRE_ERROR_PARTIAL is returned.
1025
1026 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1027 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1028 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1029 and may be more useful for multi-segment matching.
1030
1031 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1032 used to be given if ever the end of the subject was reached; now it is
1033 given only if matching could not proceed because another character was
1034 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1035 string "Z", which now yields "no match" instead of "partial match". In the
1036 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1037 final character ended with (*FAIL).
1038
1039 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1040 if the pattern had a "must contain" character that was already found in the
1041 earlier partial match, unless partial matching was again requested. For
1042 example, with the pattern /dog.(body)?/, the "must contain" character is
1043 "g". If the first part-match was for the string "dog", restarting with
1044 "sbody" failed. This bug has been fixed.
1045
1046 13. The string returned by pcre_dfa_exec() after a partial match has been
1047 changed so that it starts at the first inspected character rather than the
1048 first character of the match. This makes a difference only if the pattern
1049 starts with a lookbehind assertion or \b or \B (\K is not supported by
1050 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1051 matching functions compatible, and I think it's the right thing to do.
1052
1053 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1054 so that the demonstration program is easily available in environments where
1055 PCRE has not been installed from source.
1056
1057 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1058 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1059 library.
1060
1061 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1062 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1063 is not the first non-POSIX option to be added. Clearly some people find
1064 these options useful.
1065
1066 17. If a caller to the POSIX matching function regexec() passes a non-zero
1067 value for nmatch with a NULL value for pmatch, the value of
1068 nmatch is forced to zero.
1069
1070 18. RunGrepTest did not have a test for the availability of the -u option of
1071 the diff command, as RunTest does. It now checks in the same way as
1072 RunTest, and also checks for the -b option.
1073
1074 19. If an odd number of negated classes containing just a single character
1075 interposed, within parentheses, between a forward reference to a named
1076 subpattern and the definition of the subpattern, compilation crashed with
1077 an internal error, complaining that it could not find the referenced
1078 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1079 [The bug was that it was starting one character too far in when skipping
1080 over the character class, thus treating the ] as data rather than
1081 terminating the class. This meant it could skip too much.]
1082
1083 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1084 /g option in pcretest when the pattern contains \K, which makes it possible
1085 to have an empty string match not at the start, even when the pattern is
1086 anchored. Updated pcretest and pcredemo to use this option.
1087
1088 21. If the maximum number of capturing subpatterns in a recursion was greater
1089 than the maximum at the outer level, the higher number was returned, but
1090 with unset values at the outer level. The correct (outer level) value is
1091 now given.
1092
1093 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1094 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1095 make it do so. The string so far is captured, making this feature
1096 compatible with Perl.
1097
1098 23. The tests have been re-organized, adding tests 11 and 12, to make it
1099 possible to check the Perl 5.10 features against Perl 5.10.
1100
1101 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1102 pattern matches a fixed length string. PCRE did not allow this; now it
1103 does. Neither allows recursion.
1104
1105 25. I finally figured out how to implement a request to provide the minimum
1106 length of subject string that was needed in order to match a given pattern.
1107 (It was back references and recursion that I had previously got hung up
1108 on.) This code has now been added to pcre_study(); it finds a lower bound
1109 to the length of subject needed. It is not necessarily the greatest lower
1110 bound, but using it to avoid searching strings that are too short does give
1111 some useful speed-ups. The value is available to calling programs via
1112 pcre_fullinfo().
1113
1114 26. While implementing 25, I discovered to my embarrassment that pcretest had
1115 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1116 study optimizations had never been tested with that matching function.
1117 Oops. What is worse, even when it was passed study data, there was a bug in
1118 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1119 were also very few tests of studied patterns with pcre_dfa_exec().
1120
1121 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1122 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1123 on the other side of the coin, they are no longer allowed to have different
1124 names, because these cannot be distinguished in PCRE, and this has caused
1125 confusion. (This is a difference from Perl.)
1126
1127 28. When duplicate subpattern names are present (necessarily with different
1128 numbers, as required by 27 above), and a test is made by name in a
1129 conditional pattern, either for a subpattern having been matched, or for
1130 recursion in such a pattern, all the associated numbered subpatterns are
1131 tested, and the overall condition is true if the condition is true for any
1132 one of them. This is the way Perl works, and is also more like the way
1133 testing by number works.
1134
1135
1136 Version 7.9 11-Apr-09
1137 ---------------------
1138
1139 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1140 (pcretest), all targets were linked against these libraries. This included
1141 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1142 libraries. This caused unwanted dependencies to be created. This problem
1143 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1144 pcretest is linked with readline.
1145
1146 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1147 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1148 moved outside it again, because FALSE and TRUE are already defined in AIX,
1149 but BOOL is not.
1150
1151 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1152 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1153
1154 4. The pcregrep documentation said spaces were inserted as well as colons (or
1155 hyphens) following file names and line numbers when outputting matching
1156 lines. This is not true; no spaces are inserted. I have also clarified the
1157 wording for the --colour (or --color) option.
1158
1159 5. In pcregrep, when --colour was used with -o, the list of matching strings
1160 was not coloured; this is different to GNU grep, so I have changed it to be
1161 the same.
1162
1163 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1164 each matching line was coloured. Now it goes on to look for further matches
1165 of any of the test patterns, which is the same behaviour as GNU grep.
1166
1167 7. A pattern that could match an empty string could cause pcregrep to loop; it
1168 doesn't make sense to accept an empty string match in pcregrep, so I have
1169 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1170 seems to be how GNU grep behaves.
1171
1172 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1173 start or after a newline", because the conditional assertion was not being
1174 correctly handled. The rule now is that both the assertion and what follows
1175 in the first alternative must satisfy the test.
1176
1177 9. If auto-callout was enabled in a pattern with a conditional group whose
1178 condition was an assertion, PCRE could crash during matching, both with
1179 pcre_exec() and pcre_dfa_exec().
1180
1181 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1182 used for matching.
1183
1184 11. Unicode property support in character classes was not working for
1185 characters (bytes) greater than 127 when not in UTF-8 mode.
1186
1187 12. Added the -M command line option to pcretest.
1188
1189 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1190
1191 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1192
1193 16. Added comments and documentation about mis-use of no_arg in the C++
1194 wrapper.
1195
1196 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1197 from Martin Jerabek that uses macro names for all relevant character and
1198 string constants.
1199
1200 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1201 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1202 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1203 these, but not everybody uses configure.
1204
1205 19. A conditional group that had only one branch was not being correctly
1206 recognized as an item that could match an empty string. This meant that an
1207 enclosing group might also not be so recognized, causing infinite looping
1208 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1209 with the subject "ab", where knowledge that the repeated group can match
1210 nothing is needed in order to break the loop.
1211
1212 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1213 exec(), but without supplying a callout function, matching went wrong.
1214
1215 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1216 leak if the size of the offset vector was greater than 30. When the vector
1217 is smaller, the saved offsets during recursion go onto a local stack
1218 vector, but for larger vectors malloc() is used. It was failing to free
1219 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1220 error, in fact).
1221
1222 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1223 heapframe that is used only when UTF-8 support is enabled. This caused no
1224 problem, but was untidy.
1225
1226 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1227 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1228 included within another project.
1229
1230 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1231 slightly modified by me:
1232
1233 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1234 not building pcregrep.
1235
1236 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1237 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1238
1239 25. Forward references, both numeric and by name, in patterns that made use of
1240 duplicate group numbers, could behave incorrectly or give incorrect errors,
1241 because when scanning forward to find the reference group, PCRE was not
1242 taking into account the duplicate group numbers. A pattern such as
1243 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1244
1245 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1246 the feature of a custom pointer more persuasive (as requested by a user).
1247
1248 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1249 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1250 pre-defined.
1251
1252 28. Added support for (*UTF8) at the start of a pattern.
1253
1254 29. Arrange for flags added by the "release type" setting in CMake to be shown
1255 in the configuration summary.
1256
1257
1258 Version 7.8 05-Sep-08
1259 ---------------------
1260
1261 1. Replaced UCP searching code with optimized version as implemented for Ad
1262 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1263 stage table and inline lookup instead of a function, giving speed ups of 2
1264 to 5 times on some simple patterns that I tested. Permission was given to
1265 distribute the MultiStage2.py script that generates the tables (it's not in
1266 the tarball, but is in the Subversion repository).
1267
1268 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1269 scripts.
1270
1271 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1272 a group with a zero qualifier. The result of the study could be incorrect,
1273 or the function might crash, depending on the pattern.
1274
1275 4. Caseless matching was not working for non-ASCII characters in back
1276 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1277 It now works when Unicode Property Support is available.
1278
1279 5. In pcretest, an escape such as \x{de} in the data was always generating
1280 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1281 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1282 truncation.
1283
1284 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1285
1286 7. Added two (int) casts to pcregrep when printing the difference of two
1287 pointers, in case they are 64-bit values.
1288
1289 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1290 test 2 if it fails.
1291
1292 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1293 and a #define of that name to empty if it is not externally set. This is to
1294 allow users of MSVC to set it if necessary.
1295
1296 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1297 the convenience functions in the pcre_get.c source file.
1298
1299 11. An option change at the start of a pattern that had top-level alternatives
1300 could cause overwriting and/or a crash. This command provoked a crash in
1301 some environments:
1302
1303 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1304
1305 This potential security problem was recorded as CVE-2008-2371.
1306
1307 12. For a pattern where the match had to start at the beginning or immediately
1308 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1309 pcre_dfa_exec() could read past the end of the passed subject if there was
1310 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1311 pcretest so that it places the subject at the end of its malloc-ed buffer.
1312
1313 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1314 exec() might read past the end of the data buffer in UTF-8 mode.
1315
1316 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1317 the data contained the byte 0x85 as part of a UTF-8 character within its
1318 first line. This applied both to normal and DFA matching.
1319
1320 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1321 /^[^d]*?$/8 failed to match "abc".
1322
1323 16. Added a missing copyright notice to pcrecpp_internal.h.
1324
1325 17. Make it more clear in the documentation that values returned from
1326 pcre_exec() in ovector are byte offsets, not character counts.
1327
1328 18. Tidied a few places to stop certain compilers from issuing warnings.
1329
1330 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1331 supplied by Stefan Weber. I made a further small update for 7.8 because
1332 there is a change of source arrangements: the pcre_searchfuncs.c module is
1333 replaced by pcre_ucd.c.
1334
1335
1336 Version 7.7 07-May-08
1337 ---------------------
1338
1339 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1340 a string to a long long, pretend we don't even have a long long." This is
1341 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1342
1343 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1344 pre-7.6 versions, which defined a global no_arg variable instead of putting
1345 it in the RE class. (See also #8 below.)
1346
1347 3. Remove a line of dead code, identified by coverity and reported by Nuno
1348 Lopes.
1349
1350 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1351
1352 (1) The include/exclude patterns were being applied to the whole pathnames
1353 of files, instead of just to the final components.
1354
1355 (2) If there was more than one level of directory, the subdirectories were
1356 skipped unless they satisfied the include/exclude conditions. This is
1357 inconsistent with GNU grep (and could even be seen as contrary to the
1358 pcregrep specification - which I improved to make it absolutely clear).
1359 The action now is always to scan all levels of directory, and just
1360 apply the include/exclude patterns to regular files.
1361
1362 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1363 --exclude_dir in the tests to avoid scanning .svn directories.
1364
1365 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1366 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1367 doesn't support NULs in patterns.
1368
1369 7. Added some missing "const"s to declarations of static tables in
1370 pcre_compile.c and pcre_dfa_exec.c.
1371
1372 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1373 caused by fix #2 above. (Subsequently also a second patch to fix the
1374 first patch. And a third patch - this was a messy problem.)
1375
1376 9. Applied Craig's patch to remove the use of push_back().
1377
1378 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1379 matching function regexec().
1380
1381 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1382 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1383 references. PCRE supports relative numbers with this syntax (I don't think
1384 Oniguruma does).
1385
1386 12. Previously, a group with a zero repeat such as (...){0} was completely
1387 omitted from the compiled regex. However, this means that if the group
1388 was called as a subroutine from elsewhere in the pattern, things went wrong
1389 (an internal error was given). Such groups are now left in the compiled
1390 pattern, with a new opcode that causes them to be skipped at execution
1391 time.
1392
1393 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1394 to the way PCRE behaves:
1395
1396 (a) A lone ] character is dis-allowed (Perl treats it as data).
1397
1398 (b) A back reference to an unmatched subpattern matches an empty string
1399 (Perl fails the current match path).
1400
1401 (c) A data ] in a character class must be notated as \] because if the
1402 first data character in a class is ], it defines an empty class. (In
1403 Perl it is not possible to have an empty class.) The empty class []
1404 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1405 The negative empty class [^] matches any one character, independently
1406 of the DOTALL setting.
1407
1408 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1409 non-existent subpattern following a character class starting with ']' and
1410 containing () gave an internal compiling error instead of "reference to
1411 non-existent subpattern". Fortunately, when the pattern did exist, the
1412 compiled code was correct. (When scanning forwards to check for the
1413 existencd of the subpattern, it was treating the data ']' as terminating
1414 the class, so got the count wrong. When actually compiling, the reference
1415 was subsequently set up correctly.)
1416
1417 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1418 it was being rejected as not supported by pcre_dfa_exec(), even though
1419 other assertions are supported. I have made pcre_dfa_exec() support
1420 (*FAIL).
1421
1422 16. The implementation of 13c above involved the invention of a new opcode,
1423 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1424 cannot be changed at match time, I realized I could make a small
1425 improvement to matching performance by compiling OP_ALLANY instead of
1426 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1427 on the OP_ANY path.
1428
1429 17. Compiling pcretest on Windows with readline support failed without the
1430 following two fixes: (1) Make the unistd.h include conditional on
1431 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1432
1433 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1434 ncurses library to be included for pcretest when ReadLine support is
1435 requested, but also to allow for it to be overridden. This patch came from
1436 Daniel Bergström.
1437
1438 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1439 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1440 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1441 spotting this.
1442
1443
1444 Version 7.6 28-Jan-08
1445 ---------------------
1446
1447 1. A character class containing a very large number of characters with
1448 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1449 overflow.
1450
1451 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1452 HAVE_LONG_LONG is not defined.
1453
1454 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1455 bring them up to date and include new features. This patch includes:
1456
1457 - Fixed PH's badly added libz and libbz2 support.
1458 - Fixed a problem with static linking.
1459 - Added pcredemo. [But later removed - see 7 below.]
1460 - Fixed dftables problem and added an option.
1461 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1462 HAVE_LONG_LONG.
1463 - Added readline support for pcretest.
1464 - Added an listing of the option settings after cmake has run.
1465
1466 4. A user submitted a patch to Makefile that makes it easy to create
1467 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1468 Makefile.am that cause it to include this special target, without
1469 affecting anything else. Note that the same mingw target plus all
1470 the other distribution libraries and programs are now supported
1471 when configuring with CMake (see 6 below) instead of with
1472 Configure/Make.
1473
1474 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1475 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1476 exported in the Windows port". It has not yet been confirmed that the patch
1477 solves the problem, but it does no harm.
1478
1479 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1480 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1481 with CMake, and also correct the comment about stack recursion.
1482
1483 7. Remove the automatic building of pcredemo from the ./configure system and
1484 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1485 of a program that users should build themselves after PCRE is installed, so
1486 building it automatically is not really right. What is more, it gave
1487 trouble in some build environments.
1488
1489 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1490
1491
1492 Version 7.5 10-Jan-08
1493 ---------------------
1494
1495 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1496 values in parens when parsing an RE using the C++ wrapper."
1497
1498 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1499 Characters greater than 255 were excluded from the class instead of being
1500 included.
1501
1502 3. The same bug as (2) above applied to negated POSIX classes such as
1503 [:^space:].
1504
1505 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1506 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1507 I have changed it.
1508
1509 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1510 first named subpattern) and a construct such as (?&a) would reference the
1511 first named subpattern whose name started with "a" (in other words, the
1512 length check was missing). Both these problems are fixed. "Subpattern name
1513 expected" is now given for (?&) (a zero-length name), and this patch also
1514 makes it give the same error for \k'' (previously it complained that that
1515 was a reference to a non-existent subpattern).
1516
1517 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1518 this is right because (?- can be followed by option settings as well as by
1519 digits. I have, however, made the messages clearer.
1520
1521 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1522 than the number used in the conditional) now cause a compile-time error.
1523 This is actually not compatible with Perl, which accepts such patterns, but
1524 treats the conditional as always being FALSE (as PCRE used to), but it
1525 seems to me that giving a diagnostic is better.
1526
1527 8. Change "alphameric" to the more common word "alphanumeric" in comments
1528 and messages.
1529
1530 9. Fix two occurrences of "backslash" in comments that should have been
1531 "backspace".
1532
1533 10. Remove two redundant lines of code that can never be obeyed (their function
1534 was moved elsewhere).
1535
1536 11. The program that makes PCRE's Unicode character property table had a bug
1537 which caused it to generate incorrect table entries for sequences of
1538 characters that have the same character type, but are in different scripts.
1539 It amalgamated them into a single range, with the script of the first of
1540 them. In other words, some characters were in the wrong script. There were
1541 thirteen such cases, affecting characters in the following ranges:
1542
1543 U+002b0 - U+002c1
1544 U+0060c - U+0060d
1545 U+0061e - U+00612
1546 U+0064b - U+0065e
1547 U+0074d - U+0076d
1548 U+01800 - U+01805
1549 U+01d00 - U+01d77
1550 U+01d9b - U+01dbf
1551 U+0200b - U+0200f
1552 U+030fc - U+030fe
1553 U+03260 - U+0327f
1554 U+0fb46 - U+0fbb1
1555 U+10450 - U+1049d
1556
1557 12. The -o option (show only the matching part of a line) for pcregrep was not
1558 compatible with GNU grep in that, if there was more than one match in a
1559 line, it showed only the first of them. It now behaves in the same way as
1560 GNU grep.
1561
1562 13. If the -o and -v options were combined for pcregrep, it printed a blank
1563 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1564 does the same. The return code can be used to tell if there were any
1565 non-matching lines.
1566
1567 14. Added --file-offsets and --line-offsets to pcregrep.
1568
1569 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1570 infinitely looping recursion. The bug was that positive lookaheads were not
1571 being skipped when checking for a possible empty match (negative lookaheads
1572 and both kinds of lookbehind were skipped).
1573
1574 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1575 inclusion of <windows.h> to before rather than after the definition of
1576 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1577
1578 17. Specifying a possessive quantifier with a specific limit for a Unicode
1579 character property caused pcre_compile() to compile bad code, which led at
1580 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1581 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1582 caused the error; without that there was no problem.
1583
1584 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1585
1586 19. Added --enable-pcretest-libreadline.
1587
1588 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1589 RE::GlobalReplace(). As a result, the number of replacements returned was
1590 double what it should be. I removed one of the increments, but Craig sent a
1591 later patch that removed the other one (the right fix) and added unit tests
1592 that check the return values (which was not done before).
1593
1594 21. Several CMake things:
1595
1596 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1597 the names libpcre and libpcreposix, not just pcre and pcreposix.
1598
1599 (2) The above change means that pcretest and pcregrep are now correctly
1600 linked with the newly-built libraries, not previously installed ones.
1601
1602 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1603
1604 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1605 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1606 UTF-8 newline character). The key issue is that the pattern starts .*;
1607 this means that the match must be either at the beginning, or after a
1608 newline. The bug was in the code for advancing after a failed match and
1609 checking that the new position followed a newline. It was not taking
1610 account of UTF-8 characters correctly.
1611
1612 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1613 character classes. PCRE was not treating the sequence [:...:] as a
1614 character class unless the ... were all letters. Perl, however, seems to
1615 allow any characters between [: and :], though of course it rejects as
1616 unknown any "names" that contain non-letters, because all the known class
1617 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1618 for example, whereas PCRE did not - it did not recognize a POSIX character
1619 class. This seemed a bit dangerous, so the code has been changed to be
1620 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1621 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1622 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1623 Perl does, and where it didn't before.
1624
1625 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1626 Windows environments %n is disabled by default.
1627
1628
1629 Version 7.4 21-Sep-07
1630 ---------------------
1631
1632 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1633 means that a class such as [\s] counted as "explicit reference to CR or
1634 LF". That isn't really right - the whole point of the change was to try to
1635 help when there was an actual mention of one of the two characters. So now
1636 the change happens only if \r or \n (or a literal CR or LF) character is
1637 encountered.
1638
1639 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1640 of both had grown to the point where there were only 3 bits left.
1641 Fortunately, there was spare space in the data structure, and so I have
1642 moved the internal flags into a new 16-bit field to free up more option
1643 bits.
1644
1645 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1646 but did not set the internal JCHANGED flag - either of these is enough to
1647 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1648 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1649 start sets both bits.
1650
1651 4. Added options (at build time, compile time, exec time) to change \R from
1652 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1653
1654 5. doc/pcresyntax.html was missing from the distribution.
1655
1656 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1657 compatibility, even though it is no longer used.
1658
1659 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1660 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1661 windows.h file is present (where different names are used). [This was
1662 reversed later after testing - see 16 below.]
1663
1664 8. Changed all #include <config.h> to #include "config.h". There were also
1665 some further <pcre.h> cases that I changed to "pcre.h".
1666
1667 9. When pcregrep was used with the --colour option, it missed the line ending
1668 sequence off the lines that it output.
1669
1670 10. It was pointed out to me that arrays of string pointers cause lots of
1671 relocations when a shared library is dynamically loaded. A technique of
1672 using a single long string with a table of offsets can drastically reduce
1673 these. I have refactored PCRE in four places to do this. The result is
1674 dramatic:
1675
1676 Originally: 290
1677 After changing UCP table: 187
1678 After changing error message table: 43
1679 After changing table of "verbs" 36
1680 After changing table of Posix names 22
1681
1682 Thanks to the folks working on Gregex for glib for this insight.
1683
1684 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1685 unicode-properties was also set.
1686
1687 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1688
1689 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1690 checked only for CRLF.
1691
1692 14. Added casts to pcretest.c to avoid compiler warnings.
1693
1694 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1695
1696 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1697 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1698 entirely. This removes changes made in 7 above.
1699
1700 17. The CMake files have been updated, and there is now more information about
1701 building with CMake in the NON-UNIX-USE document.
1702
1703
1704 Version 7.3 28-Aug-07
1705 ---------------------
1706
1707 1. In the rejigging of the build system that eventually resulted in 7.1, the
1708 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1709 brackets there is not right, since it causes compilers to look for an
1710 installed pcre.h, not the version that is in the source that is being
1711 compiled (which of course may be different). I have changed it back to:
1712
1713 #include "pcre.h"
1714
1715 I have a vague recollection that the change was concerned with compiling in
1716 different directories, but in the new build system, that is taken care of
1717 by the VPATH setting the Makefile.
1718
1719 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1720 when the subject happened to end in the byte 0x85 (e.g. if the last
1721 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1722 characters but of course it shouldn't be taken as a newline when it is part
1723 of another character. The bug was that, for an unlimited repeat of . in
1724 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1725 characters when looking for a newline.
1726
1727 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1728
1729 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1730 in debug output.
1731
1732 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1733 long printing in the pcrecpp unittest when running under MinGW.
1734
1735 6. ESC_K was left out of the EBCDIC table.
1736
1737 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1738 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1739 limit also applies to "virtual nesting" when a pattern is recursive, and in
1740 this case 1000 isn't so big. I have been able to remove this limit at the
1741 expense of backing off one optimization in certain circumstances. Normally,
1742 when pcre_exec() would call its internal match() function recursively and
1743 immediately return the result unconditionally, it uses a "tail recursion"
1744 feature to save stack. However, when a subpattern that can match an empty
1745 string has an unlimited repetition quantifier, it no longer makes this
1746 optimization. That gives it a stack frame in which to save the data for
1747 checking that an empty string has been matched. Previously this was taken
1748 from the 1000-entry workspace that had been reserved. So now there is no
1749 explicit limit, but more stack is used.
1750
1751 8. Applied Daniel's patches to solve problems with the import/export magic
1752 syntax that is required for Windows, and which was going wrong for the
1753 pcreposix and pcrecpp parts of the library. These were overlooked when this
1754 problem was solved for the main library.
1755
1756 9. There were some crude static tests to avoid integer overflow when computing
1757 the size of patterns that contain repeated groups with explicit upper
1758 limits. As the maximum quantifier is 65535, the maximum group length was
1759 set at 30,000 so that the product of these two numbers did not overflow a
1760 32-bit integer. However, it turns out that people want to use groups that
1761 are longer than 30,000 bytes (though not repeat them that many times).
1762 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1763 made it possible to implement the integer overflow checks in a much more
1764 dynamic way, which I have now done. The artificial limitation on group
1765 length has been removed - we now have only the limit on the total length of
1766 the compiled pattern, which depends on the LINK_SIZE setting.
1767
1768 10. Fixed a bug in the documentation for get/copy named substring when
1769 duplicate names are permitted. If none of the named substrings are set, the
1770 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1771 empty string.
1772
1773 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1774 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1775 because the ] is interpreted as the first data character and the
1776 terminating ] is not found. PCRE has been made compatible with Perl in this
1777 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1778 cause memory overwriting.
1779
1780 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1781 string has been matched (to stop an infinite loop). It was not recognizing
1782 a conditional subpattern that could match an empty string if that
1783 subpattern was within another subpattern. For example, it looped when
1784 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1785 condition was not nested. This bug has been fixed.
1786
1787 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1788 past the start of the subject in the presence of bytes with the top bit
1789 set, for example "\x8aBCD".
1790
1791 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1792 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1793
1794 14. Optimized (?!) to (*FAIL).
1795
1796 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1797 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1798 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1799 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1800 does: it's just the validity check that is more restrictive.
1801
1802 16. Inserted checks for integer overflows during escape sequence (backslash)
1803 processing, and also fixed erroneous offset values for syntax errors during
1804 backslash processing.
1805
1806 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1807 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1808
1809 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1810 caused an overrun.
1811
1812 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1813 something other than just ASCII characters) inside a group that had an
1814 unlimited repeat caused a loop at compile time (while checking to see
1815 whether the group could match an empty string).
1816
1817 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1818 [\P{Any}] did so. (Error in the code for printing property names.)
1819
1820 21. An orphan \E inside a character class could cause a crash.
1821
1822 22. A repeated capturing bracket such as (A)? could cause a wild memory
1823 reference during compilation.
1824
1825 23. There are several functions in pcre_compile() that scan along a compiled
1826 expression for various reasons (e.g. to see if it's fixed length for look
1827 behind). There were bugs in these functions when a repeated \p or \P was
1828 present in the pattern. These operators have additional parameters compared
1829 with \d, etc, and these were not being taken into account when moving along
1830 the compiled data. Specifically:
1831
1832 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1833 length.
1834
1835 (b) An item such as \pL+ within a repeated group could cause crashes or
1836 loops.
1837
1838 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1839 "reference to non-existent subpattern" error.
1840
1841 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1842
1843 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1844 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1845
1846 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1847
1848 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1849 character were causing crashes (broken optimization).
1850
1851 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1852 \p or \P) caused a compile-time loop.
1853
1854 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1855 break. For example, the unstudied pattern [\r\n]A does not match the string
1856 "\r\nA" because change 7.0/46 below moves the current point on by two
1857 characters after failing to match at the start. However, the pattern \nA
1858 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1859 the same is true. There doesn't seem any very clean way out of this, but
1860 what I have chosen to do makes the common cases work: PCRE now takes note
1861 of whether there can be an explicit match for \r or \n anywhere in the
1862 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1863 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1864 pattern has explicit CR or LF references.
1865
1866 29. Added (*CR) etc for changing newline setting at start of pattern.
1867
1868
1869 Version 7.2 19-Jun-07
1870 ---------------------
1871
1872 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1873 which is apparently normally available under Windows.
1874
1875 2. Re-jig the pcregrep tests with different newline settings in an attempt
1876 to make them independent of the local environment's newline setting.
1877
1878 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1879
1880 4. Some of the "internals" tests were previously cut out when the link size
1881 was not 2, because the output contained actual offsets. The recent new
1882 "Z" feature of pcretest means that these can be cut out, making the tests
1883 usable with all link sizes.
1884
1885 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1886 stack recursion. This gives a massive performance boost under BSD, but just
1887 a small improvement under Linux. However, it saves one field in the frame
1888 in all cases.
1889
1890 6. Added more features from the forthcoming Perl 5.10:
1891
1892 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1893 recursion call. It refers to the nth most recently opened parentheses.
1894
1895 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1896 to be opened parentheses.
1897
1898 (c) Conditions that refer to capturing parentheses can be specified
1899 relatively, for example, (?(-2)... or (?(+3)...
1900
1901 (d) \K resets the start of the current match so that everything before
1902 is not part of it.
1903
1904 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1905
1906 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1907 reference syntax.
1908
1909 (g) (?| introduces a group in which the numbering of parentheses in each
1910 alternative starts with the same number.
1911
1912 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1913
1914 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1915 PCRE_INFO_JCHANGED.
1916
1917 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1918 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1919 for detecting groups that can match an empty string.
1920
1921 9. A pattern with a very large number of alternatives (more than several
1922 hundred) was running out of internal workspace during the pre-compile
1923 phase, where pcre_compile() figures out how much memory will be needed. A
1924 bit of new cunning has reduced the workspace needed for groups with
1925 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1926 workspace instead of running out of the 4096 that are available.
1927
1928 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1929
1930 11. Applied patch from Google to remove an optimization that didn't quite work.
1931 The report of the bug said:
1932
1933 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1934 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1935 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1936
1937 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1938 it matched the wrong number of bytes.
1939
1940
1941 Version 7.1 24-Apr-07
1942 ---------------------
1943
1944 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1945 that is more "standard", making use of automake and other Autotools. There
1946 is some re-arrangement of the files and adjustment of comments consequent
1947 on this.
1948
1949 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1950 for recursive directory scanning broke on some systems because the files
1951 are not scanned in any specific order and on different systems the order
1952 was different. A call to "sort" has been inserted into RunGrepTest for the
1953 approprate test as a short-term fix. In the longer term there may be an
1954 alternative.
1955
1956 3. I had an email from Eric Raymond about problems translating some of PCRE's
1957 man pages to HTML (despite the fact that I distribute HTML pages, some
1958 people do their own conversions for various reasons). The problems
1959 concerned the use of low-level troff macros .br and .in. I have therefore
1960 removed all such uses from the man pages (some were redundant, some could
1961 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1962 HTML has been updated to handle .nf/.fi and to complain if it encounters
1963 .br or .in.
1964
1965 4. Updated comments in configure.ac that get placed in config.h.in and also
1966 arranged for config.h to be included in the distribution, with the name
1967 config.h.generic, for the benefit of those who have to compile without
1968 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1969
1970 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1971 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1972 makevp.bat for the current PCRE, using the additional files
1973 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1974
1975 6. A Windows user reported a minor discrepancy with test 2, which turned out
1976 to be caused by a trailing space on an input line that had got lost in his
1977 copy. The trailing space was an accident, so I've just removed it.
1978
1979 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1980 that is needed.
1981
1982 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1983 as "const" (a) because they are and (b) because it helps the PHP
1984 maintainers who have recently made a script to detect big data structures
1985 in the php code that should be moved to the .rodata section. I remembered
1986 to update Builducptable as well, so it won't revert if ucptable.h is ever
1987 re-created.
1988
1989 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1990 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1991 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1992 support is not required. This saves 1.5-2K of code, which is important in
1993 some applications.
1994
1995 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1996 so as not to refer to the tables, even though these functions will never be
1997 called when UTF-8 support is disabled. Otherwise there are problems with a
1998 shared library.
1999
2000 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2001
2002 (a) It was defining its arguments as char * instead of void *.
2003
2004 (b) It was assuming that all moves were upwards in memory; this was true
2005 a long time ago when I wrote it, but is no longer the case.
2006
2007 The emulated memove() is provided for those environments that have neither
2008 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2009 is clearly not the case, as these two bugs were recently reported.
2010
2011 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2012 and Detrail to create the HTML documentation, the .txt form of the man
2013 pages, and it removes trailing spaces from listed files. It also creates
2014 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2015 case, it wraps all the #defines with #ifndefs. This script should be run
2016 before "make dist".
2017
2018 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2019 with Unicode property support.
2020
2021 (a) For a maximizing quantifier, if the two different cases of the
2022 character were of different lengths in their UTF-8 codings (there are
2023 some cases like this - I found 11), and the matching function had to
2024 back up over a mixture of the two cases, it incorrectly assumed they
2025 were both the same length.
2026
2027 (b) When PCRE was configured to use the heap rather than the stack for
2028 recursion during matching, it was not correctly preserving the data for
2029 the other case of a UTF-8 character when checking ahead for a match
2030 while processing a minimizing repeat. If the check also involved
2031 matching a wide character, but failed, corruption could cause an
2032 erroneous result when trying to check for a repeat of the original
2033 character.
2034
2035 13. Some tidying changes to the testing mechanism:
2036
2037 (a) The RunTest script now detects the internal link size and whether there
2038 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2039 values substituted by "configure". (The RunGrepTest script already did
2040 this for UTF-8.) The configure.ac script no longer substitutes the
2041 relevant variables.
2042
2043 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2044 with length and offset values. This means that the output is different
2045 for different internal link sizes. Test 2 is skipped for link sizes
2046 other than 2 because of this, bypassing the problem. Unfortunately,
2047 there was also a test in test 3 (the locale tests) that used /B and
2048 failed for link sizes other than 2. Rather than cut the whole test out,
2049 I have added a new /Z option to pcretest that replaces the length and
2050 offset values with spaces. This is now used to make test 3 independent
2051 of link size. (Test 2 will be tidied up later.)
2052
2053 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2054 segmentation fault instead of returning the appropriate error message.
2055
2056 15. In multiline mode when the newline sequence was set to "any", the pattern
2057 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2058 This doesn't seem right; it now treats the CRLF combination as the line
2059 ending, and so does not match in that case. It's only a pattern such as ^$
2060 that would hit this one: something like ^ABC$ would have failed after \r
2061 and then tried again after \r\n.
2062
2063 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2064 in an attempt to make files that differ only in their line terminators
2065 compare equal. This works on Linux.
2066
2067 17. Under certain error circumstances pcregrep might try to free random memory
2068 as it exited. This is now fixed, thanks to valgrind.
2069
2070 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2071 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2072 was because its rules for how to advance for /g after matching an empty
2073 string at the end of a line did not allow for this case. They now check for
2074 it specially.
2075
2076 20. pcretest is supposed to handle patterns and data of any length, by
2077 extending its buffers when necessary. It was getting this wrong when the
2078 buffer for a data line had to be extended.
2079
2080 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2081 CRLF as a newline sequence.
2082
2083 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2084 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2085 I have nevertheless tidied it up.
2086
2087 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2088
2089 24. Added a man page for pcre-config.
2090
2091
2092 Version 7.0 19-Dec-06
2093 ---------------------
2094
2095 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2096 moving to gcc 4.1.1.
2097
2098 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2099 sys/time.h, which is documented as needed for this function. It doesn't
2100 seem to matter on Linux, but it showed up on some releases of OS X.
2101
2102 3. It seems that there are systems where bytes whose values are greater than
2103 127 match isprint() in the "C" locale. The "C" locale should be the
2104 default when a C program starts up. In most systems, only ASCII printing
2105 characters match isprint(). This difference caused the output from pcretest
2106 to vary, making some of the tests fail. I have changed pcretest so that:
2107
2108 (a) When it is outputting text in the compiled version of a pattern, bytes
2109 other than 32-126 are always shown as hex escapes.
2110
2111 (b) When it is outputting text that is a matched part of a subject string,
2112 it does the same, unless a different locale has been set for the match
2113 (using the /L modifier). In this case, it uses isprint() to decide.
2114
2115 4. Fixed a major bug that caused incorrect computation of the amount of memory
2116 required for a compiled pattern when options that changed within the
2117 pattern affected the logic of the preliminary scan that determines the
2118 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2119 that the computed length was too small. The symptoms of this bug were
2120 either the PCRE error "internal error: code overflow" from pcre_compile(),
2121 or a glibc crash with a message such as "pcretest: free(): invalid next
2122 size (fast)". Examples of patterns that provoked this bug (shown in
2123 pcretest format) are:
2124
2125 /(?-x: )/x
2126 /(?x)(?-x: \s*#\s*)/
2127 /((?i)[\x{c0}])/8
2128 /(?i:[\x{c0}])/8
2129
2130 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2131 is now done differently.
2132
2133 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2134 wrapper classes; (b) implement a new function in the C++ scanner that is
2135 more efficient than the old way of doing things because it avoids levels of
2136 recursion in the regex matching; (c) add a paragraph to the documentation
2137 for the FullMatch() function.
2138
2139 6. The escape sequence \n was being treated as whatever was defined as
2140 "newline". Not only was this contrary to the documentation, which states
2141 that \n is character 10 (hex 0A), but it also went horribly wrong when
2142 "newline" was defined as CRLF. This has been fixed.
2143
2144 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2145 was being set to -1 for the "end of line" case (supposedly a value that no
2146 character can have). Though this value is never used (the check for end of
2147 line is "zero bytes in current character"), it caused compiler complaints.
2148 I've changed it to 0xffffffff.
2149
2150 8. In pcre_version.c, the version string was being built by a sequence of
2151 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2152 string (as it is for production releases) called a macro with an empty
2153 argument. The C standard says the result of this is undefined. The gcc
2154 compiler treats it as an empty string (which was what was wanted) but it is
2155 reported that Visual C gives an error. The source has been hacked around to
2156 avoid this problem.
2157
2158 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2159 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2160 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2161 of them did).
2162
2163 10. Originally, pcretest opened its input and output without "b"; then I was
2164 told that "b" was needed in some environments, so it was added for release
2165 5.0 to both the input and output. (It makes no difference on Unix-like
2166 systems.) Later I was told that it is wrong for the input on Windows. I've
2167 now abstracted the modes into two macros, to make it easier to fiddle with
2168 them, and removed "b" from the input mode under Windows.
2169
2170 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2171
2172 12. Added -help and --help to pcretest as an official way of being reminded
2173 of the options.
2174
2175 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2176 and pcrecpp.cc because they annoy compilers at high warning levels.
2177
2178 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2179
2180 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2181 scripts are not C programs :-) and which was not noticed because it works
2182 on Linux.
2183
2184 16. pcretest is supposed to handle any length of pattern and data line (as one
2185 line or as a continued sequence of lines) by extending its input buffer if
2186 necessary. This feature was broken for very long pattern lines, leading to
2187 a string of junk being passed to pcre_compile() if the pattern was longer
2188 than about 50K.
2189
2190 17. I have done a major re-factoring of the way pcre_compile() computes the
2191 amount of memory needed for a compiled pattern. Previously, there was code
2192 that made a preliminary scan of the pattern in order to do this. That was
2193 OK when PCRE was new, but as the facilities have expanded, it has become
2194 harder and harder to keep it in step with the real compile phase, and there
2195 have been a number of bugs (see for example, 4 above). I have now found a
2196 cunning way of running the real compile function in a "fake" mode that
2197 enables it to compute how much memory it would need, while actually only
2198 ever using a few hundred bytes of working memory and without too many
2199 tests of the mode. This should make future maintenance and development
2200 easier. A side effect of this work is that the limit of 200 on the nesting
2201 depth of parentheses has been removed (though this was never a serious
2202 limitation, I suspect). However, there is a downside: pcre_compile() now
2203 runs more slowly than before (30% or more, depending on the pattern). I
2204 hope this isn't a big issue. There is no effect on runtime performance.
2205
2206 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2207 newline (only possible for the last line of a file) and it was a
2208 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2209
2210 19. Added additional timing features to pcretest. (1) The -tm option now times
2211 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2212 separate command line item, by a number that specifies the number of
2213 repeats to use when timing. The default is 50000; this gives better
2214 precision, but takes uncomfortably long for very large patterns.
2215
2216 20. Extended pcre_study() to be more clever in cases where a branch of a
2217 subpattern has no definite first character. For example, (a*|b*)[cd] would
2218 previously give no result from pcre_study(). Now it recognizes that the
2219 first character must be a, b, c, or d.
2220
2221 21. There was an incorrect error "recursive call could loop indefinitely" if
2222 a subpattern (or the entire pattern) that was being tested for matching an
2223 empty string contained only one non-empty item after a nested subpattern.
2224 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2225 incorrectly, because the \d was being skipped in the check.
2226
2227 22. The pcretest program now has a new pattern option /B and a command line
2228 option -b, which is equivalent to adding /B to every pattern. This causes
2229 it to show the compiled bytecode, without the additional information that
2230 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2231 is the same as /B/I).
2232
2233 23. A new optimization is now able automatically to treat some sequences such
2234 as a*b as a*+b. More specifically, if something simple (such as a character
2235 or a simple class like \d) has an unlimited quantifier, and is followed by
2236 something that cannot possibly match the quantified thing, the quantifier
2237 is automatically "possessified".
2238
2239 24. A recursive reference to a subpattern whose number was greater than 39
2240 went wrong under certain circumstances in UTF-8 mode. This bug could also
2241 have affected the operation of pcre_study().
2242
2243 25. Realized that a little bit of performance could be had by replacing
2244 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2245
2246 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2247
2248 27. Possessive quantifiers such as a++ were previously implemented by turning
2249 them into atomic groups such as ($>a+). Now they have their own opcodes,
2250 which improves performance. This includes the automatically created ones
2251 from 23 above.
2252
2253 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2254 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2255 the first matched character to be a colon. This applied both to named and
2256 numbered groups.
2257
2258 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2259
2260 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2261 building PCRE on Windows easier, so I have included it in the distribution.
2262
2263 31. There is now a check in pcretest against a ridiculously large number being
2264 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2265 loop, the loop is abandoned.
2266
2267 32. Forward references to subpatterns in conditions such as (?(2)...) where
2268 subpattern 2 is defined later cause pcre_compile() to search forwards in
2269 the pattern for the relevant set of parentheses. This search went wrong
2270 when there were unescaped parentheses in a character class, parentheses
2271 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2272
2273 33. "Subroutine" calls and backreferences were previously restricted to
2274 referencing subpatterns earlier in the regex. This restriction has now
2275 been removed.
2276
2277 34. Added a number of extra features that are going to be in Perl 5.10. On the
2278 whole, these are just syntactic alternatives for features that PCRE had
2279 previously implemented using the Python syntax or my own invention. The
2280 other formats are all retained for compatibility.
2281
2282 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2283 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2284 also .NET compatible.
2285
2286 (b) A recursion or subroutine call to a named group can now be defined as
2287 (?&name) as well as (?P>name).
2288
2289 (c) A backreference to a named group can now be defined as \k<name> or
2290 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2291 5.10, are also .NET compatible.
2292
2293 (d) A conditional reference to a named group can now use the syntax
2294 (?(<name>) or (?('name') as well as (?(name).
2295
2296 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2297 groups (named and numbered) that are never evaluated inline, but can be
2298 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2299 is always false. There may be only one alternative in such a group.
2300
2301 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2302 as the simple (?(R). The condition is true only if the most recent
2303 recursion is that of the given number or name. It does not search out
2304 through the entire recursion stack.
2305
2306 (g) The escape \gN or \g{N} has been added, where N is a positive or
2307 negative number, specifying an absolute or relative reference.
2308
2309 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2310 some "unreachable code" warnings.
2311
2312 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2313 things, this adds five new scripts.
2314
2315 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2316 There were also incompatibilities regarding the handling of \Q..\E inside
2317 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2318 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2319
2320 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2321 matches an empty string, and forcibly breaks the loop. There were bugs in
2322 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2323 against aaaa the result was just "a" rather than "aaaa", for example. Two
2324 separate and independent bugs (that affected different cases) have been
2325 fixed.
2326
2327 39. Refactored the code to abolish the use of different opcodes for small
2328 capturing bracket numbers. This is a tidy that I avoided doing when I
2329 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2330 The new approach is not only tidier, it makes it possible to reduce the
2331 memory needed to fix the previous bug (38).
2332
2333 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2334 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2335 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2336 mode.
2337
2338 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2339 report.
2340
2341 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2342 copy construction and assignment in the C++ wrapper.
2343
2344 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2345 couple of bugs that could have given wrong results in the "--newline=crlf"
2346 case.
2347
2348 44. Added a number of casts and did some reorganization of signed/unsigned int
2349 variables following suggestions from Dair Grant. Also renamed the variable
2350 "this" as "item" because it is a C++ keyword.
2351
2352 45. Arranged for dftables to add
2353
2354 #include "pcre_internal.h"
2355
2356 to pcre_chartables.c because without it, gcc 4.x may remove the array
2357 definition from the final binary if PCRE is built into a static library and
2358 dead code stripping is activated.
2359
2360 46. For an unanchored pattern, if a match attempt fails at the start of a
2361 newline sequence, and the newline setting is CRLF or ANY, and the next two
2362 characters are CRLF, advance by two characters instead of one.
2363
2364
2365 Version 6.7 04-Jul-06
2366 ---------------------
2367
2368 1. In order to handle tests when input lines are enormously long, pcretest has
2369 been re-factored so that it automatically extends its buffers when
2370 necessary. The code is crude, but this _is_ just a test program. The
2371 default size has been increased from 32K to 50K.
2372
2373 2. The code in pcre_study() was using the value of the re argument before
2374 testing it for NULL. (Of course, in any sensible call of the function, it
2375 won't be NULL.)
2376
2377 3. The memmove() emulation function in pcre_internal.h, which is used on
2378 systems that lack both memmove() and bcopy() - that is, hardly ever -
2379 was missing a "static" storage class specifier.
2380
2381 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2382 containing an extended class (one that cannot be represented by a bitmap
2383 because it contains high-valued characters or Unicode property items, e.g.
2384 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2385 pattern, but PCRE should not loop if you do not (it no longer does).
2386 [Detail: two cases were found: (a) a repeated subpattern containing an
2387 extended class; (b) a recursive reference to a subpattern that followed a
2388 previous extended class. It wasn't skipping over the extended class
2389 correctly when UTF-8 mode was not set.]
2390
2391 5. A negated single-character class was not being recognized as fixed-length
2392 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2393 compile error "lookbehind assertion is not fixed length".
2394
2395 6. The RunPerlTest auxiliary script was showing an unexpected difference
2396 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2397 write a Perl script that can interpret lines of an input file either as
2398 byte characters or as UTF-8, which is what "perltest" was being required to
2399 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2400 can't do is switch easily at run time between having the "use utf8;" pragma
2401 or not. In the end, I fudged it by using the RunPerlTest script to insert
2402 "use utf8;" explicitly for the UTF-8 tests.
2403
2404 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2405 the end of the subject string, contrary to the documentation and to what
2406 Perl does. This was true of both matching functions. Now it matches only at
2407 the start of the subject and immediately after *internal* newlines.
2408
2409 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2410 a pointer to an int instead of a pointer to an unsigned long int. This
2411 caused problems on 64-bit systems.
2412
2413 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2414 instance of the 'standard' template library not being so standard".
2415
2416 10. There was no check on the number of named subpatterns nor the maximum
2417 length of a subpattern name. The product of these values is used to compute
2418 the size of the memory block for a compiled pattern. By supplying a very
2419 long subpattern name and a large number of named subpatterns, the size
2420 computation could be caused to overflow. This is now prevented by limiting
2421 the length of names to 32 characters, and the number of named subpatterns
2422 to 10,000.
2423
2424 11. Subpatterns that are repeated with specific counts have to be replicated in
2425 the compiled pattern. The size of memory for this was computed from the
2426 length of the subpattern and the repeat count. The latter is limited to
2427 65535, but there was no limit on the former, meaning that integer overflow
2428 could in principle occur. The compiled length of a repeated subpattern is
2429 now limited to 30,000 bytes in order to prevent this.
2430
2431 12. Added the optional facility to have named substrings with the same name.
2432
2433 13. Added the ability to use a named substring as a condition, using the
2434 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2435 are numbers (not recommended). Forward references are permitted.
2436
2437 14. Added forward references in named backreferences (if you see what I mean).
2438
2439 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2440 pattern could run off the end of the subject. For example, the pattern
2441 "(?s)(.{1,5})"8 did this with the subject "ab".
2442
2443 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2444 PCRE_CASELESS was set when matching characters that were quantified with ?
2445 or *.
2446
2447 17. A character class other than a single negated character that had a minimum
2448 but no maximum quantifier - for example [ab]{6,} - was not handled
2449 correctly by pce_dfa_exec(). It would match only one character.
2450
2451 18. A valid (though odd) pattern that looked like a POSIX character
2452 class but used an invalid character after [ (for example [[,abc,]]) caused
2453 pcre_compile() to give the error "Failed: internal error: code overflow" or
2454 in some cases to crash with a glibc free() error. This could even happen if
2455 the pattern terminated after [[ but there just happened to be a sequence of
2456 letters, a binary zero, and a closing ] in the memory that followed.
2457
2458 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2459 over the years. Originally (before any Unicode support), just the bottom 8
2460 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2461 output from "man perlunicode" includes this:
2462
2463 The regular expression compiler produces polymorphic opcodes. That
2464 is, the pattern adapts to the data and automatically switches to
2465 the Unicode character scheme when presented with Unicode data--or
2466 instead uses a traditional byte scheme when presented with byte
2467 data.
2468
2469 Sadly, a wide octal escape does not cause a switch, and in a string with
2470 no other multibyte characters, these octal escapes are treated as before.
2471 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2472 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2473 Unicode string.
2474
2475 I have not perpetrated such confusion in PCRE. Up till now, it took just
2476 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2477 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2478 translate to the appropriate multibyte character.
2479
2480 29. Applied some refactoring to reduce the number of warnings from Microsoft
2481 and Borland compilers. This has included removing the fudge introduced
2482 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2483 a warning about an unused variable.
2484
2485 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2486 characters since release 4.0, because Perl (from release 5.004) does not.
2487 [Or at least, is documented not to: some releases seem to be in conflict
2488 with the documentation.] However, when a pattern was studied with
2489 pcre_study() and all its branches started with \s, PCRE still included VT
2490 as a possible starting character. Of course, this did no harm; it just
2491 caused an unnecessary match attempt.
2492
2493 22. Removed a now-redundant internal flag bit that recorded the fact that case
2494 dependency changed within the pattern. This was once needed for "required
2495 byte" processing, but is no longer used. This recovers a now-scarce options
2496 bit. Also moved the least significant internal flag bit to the most-
2497 significant bit of the word, which was not previously used (hangover from
2498 the days when it was an int rather than a uint) to free up another bit for
2499 the future.
2500
2501 23. Added support for CRLF line endings as well as CR and LF. As well as the
2502 default being selectable at build time, it can now be changed at runtime
2503 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2504 specify that it is scanning data with non-default line endings.
2505
2506 24. Changed the definition of CXXLINK to make it agree with the definition of
2507 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2508
2509 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2510 recursions. This makes a big different to stack usage for some patterns.
2511
2512 26. If a subpattern containing a named recursion or subroutine reference such
2513 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2514 the space required for the compiled pattern went wrong and gave too small a
2515 value. Depending on the environment, this could lead to "Failed: internal
2516 error: code overflow at offset 49" or "glibc detected double free or
2517 corruption" errors.
2518
2519 27. Applied patches from Google (a) to support the new newline modes and (b) to
2520 advance over multibyte UTF-8 characters in GlobalReplace.
2521
2522 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2523 difference for some implementation of PCRE in some Windows version.
2524
2525 29. Added some extra testing facilities to pcretest:
2526
2527 \q<number> in a data line sets the "match limit" value
2528 \Q<number> in a data line sets the "match recursion limt" value
2529 -S <number> sets the stack size, where <number> is in megabytes
2530
2531 The -S option isn't available for Windows.
2532
2533
2534 Version 6.6 06-Feb-06
2535 ---------------------
2536
2537 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2538 in pcreposix.h. I have copied the definition from pcre.h.
2539
2540 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2541 because pcre.h is no longer a built file.
2542
2543 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2544 not normally included in the compiled code.
2545
2546
2547 Version 6.5 01-Feb-06
2548 ---------------------
2549
2550 1. When using the partial match feature with pcre_dfa_exec(), it was not
2551 anchoring the second and subsequent partial matches at the new starting
2552 point. This could lead to incorrect results. For example, with the pattern
2553 /1234/, partially matching against "123" and then "a4" gave a match.
2554
2555 2. Changes to pcregrep:
2556
2557 (a) All non-match returns from pcre_exec() were being treated as failures
2558 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2559 error message is output. Some extra information is given for the
2560 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2561 probably the only errors that are likely to be caused by users (by
2562 specifying a regex that has nested indefinite repeats, for instance).
2563 If there are more than 20 of these errors, pcregrep is abandoned.
2564
2565 (b) A binary zero was treated as data while matching, but terminated the
2566 output line if it was written out. This has been fixed: binary zeroes
2567 are now no different to any other data bytes.
2568
2569 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2570 used to set a locale for matching. The --locale=xxxx long option has
2571 been added (no short equivalent) to specify a locale explicitly on the
2572 pcregrep command, overriding the environment variables.
2573
2574 (d) When -B was used with -n, some line numbers in the output were one less
2575 than they should have been.
2576
2577 (e) Added the -o (--only-matching) option.
2578
2579 (f) If -A or -C was used with -c (count only), some lines of context were
2580 accidentally printed for the final match.
2581
2582 (g) Added the -H (--with-filename) option.
2583
2584 (h) The combination of options -rh failed to suppress file names for files
2585 that were found from directory arguments.
2586
2587 (i) Added the -D (--devices) and -d (--directories) options.
2588
2589 (j) Added the -F (--fixed-strings) option.
2590
2591 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2592
2593 (l) Added the --colo(u)r option.
2594
2595 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2596 is not present by default.
2597
2598 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2599 items such as (?R) or (?1), when the recursion could match a number of
2600 alternatives. If it matched one of the alternatives, but subsequently,
2601 outside the recursion, there was a failure, the code tried to back up into
2602 the recursion. However, because of the way PCRE is implemented, this is not
2603 possible, and the result was an incorrect result from the match.
2604
2605 In order to prevent this happening, the specification of recursion has
2606 been changed so that all such subpatterns are automatically treated as
2607 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2608
2609 4. I had overlooked the fact that, in some locales, there are characters for
2610 which isalpha() is true but neither isupper() nor islower() are true. In
2611 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2612 and ordfeminine) are like this. This affected the treatment of \w and \W
2613 when they appeared in character classes, but not when they appeared outside
2614 a character class. The bit map for "word" characters is now created
2615 separately from the results of isalnum() instead of just taking it from the
2616 upper, lower, and digit maps. (Plus the underscore character, of course.)
2617
2618 5. The above bug also affected the handling of POSIX character classes such as
2619 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2620 permanent tables. Instead, the bit maps for such a class were previously
2621 created as the appropriate unions of the upper, lower, and digit bitmaps.
2622 Now they are created by subtraction from the [[:word:]] class, which has
2623 its own bitmap.
2624
2625 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2626 It is created by subtracting the vertical space characters (\x09, \x0a,
2627 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2628 subtraction was done in the overall bitmap for a character class, meaning
2629 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2630 be recognized. This bug has been fixed.
2631
2632 7. Patches from the folks at Google:
2633
2634 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2635 real life, but is still worth protecting against".
2636
2637 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2638 regular expressions".
2639
2640 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2641 have it.
2642
2643 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2644 "configure" and the latter not, in order to fix a problem somebody had
2645 with compiling the Arg class on HP-UX.
2646
2647 (e) Improve the error-handling of the C++ wrapper a little bit.
2648
2649 (f) New tests for checking recursion limiting.
2650
2651 8. The pcre_memmove() function, which is used only if the environment does not
2652 have a standard memmove() function (and is therefore rarely compiled),
2653 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2654 returning a result (though PCRE never actually uses the result).
2655
2656 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2657 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2658 returned instead of calling malloc() with an overflowing number that would
2659 most likely cause subsequent chaos.
2660
2661 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2662
2663 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2664 with this option is matched, the nmatch and pmatch options of regexec() are
2665 ignored.
2666
2667 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2668 provided in case anyone wants to the the POSIX interface with UTF-8
2669 strings.
2670
2671 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2672 C++ linking (needed for some HP-UX environments).
2673
2674 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2675 (unused parameter) and in the pcre_printint() function (omitted "default"
2676 switch label when the default is to do nothing).
2677
2678 15. Added some code to make it possible, when PCRE is compiled as a C++
2679 library, to replace subject pointers for pcre_exec() with a smart pointer
2680 class, thus making it possible to process discontinuous strings.
2681
2682 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2683 much the same function. They were added by different people who were trying
2684 to make PCRE easy to compile on non-Unix systems. It has been suggested
2685 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2686 for compiling on Windows systems. I have therefore replaced it with
2687 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2688 defaults to "extern" for C or "extern C" for C++, which works fine on
2689 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2690 SCOPE with something explicit in config.h. In addition:
2691
2692 (a) pcreposix.h still had just "extern" instead of either of these macros;
2693 I have replaced it with PCRE_DATA_SCOPE.
2694
2695 (b) Functions such as _pcre_xclass(), which are internal to the library,
2696 but external in the C sense, all had PCRE_EXPORT in their definitions.
2697 This is apparently wrong for the Windows case, so I have removed it.
2698 (It makes no difference on Unix-like systems.)
2699
2700 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2701 of recursive calls to match(). This is different to MATCH_LIMIT because
2702 that limits the total number of calls to match(), not all of which increase
2703 the depth of recursion. Limiting the recursion depth limits the amount of
2704 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2705 when PCRE is compiled, and changed at run time. A patch from Google adds
2706 this functionality to the C++ interface.
2707
2708 18. Changes to the handling of Unicode character properties:
2709
2710 (a) Updated the table to Unicode 4.1.0.
2711
2712 (b) Recognize characters that are not in the table as "Cn" (undefined).
2713
2714 (c) I revised the way the table is implemented to a much improved format
2715 which includes recognition of ranges. It now supports the ranges that
2716 are defined in UnicodeData.txt, and it also amalgamates other
2717 characters into ranges. This has reduced the number of entries in the
2718 table from around 16,000 to around 3,000, thus reducing its size
2719 considerably. I realized I did not need to use a tree structure after
2720 all - a binary chop search is just as efficient. Having reduced the
2721 number of entries, I extended their size from 6 bytes to 8 bytes to
2722 allow for more data.
2723
2724 (d) Added support for Unicode script names via properties such as \p{Han}.
2725
2726 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2727 matching that character.
2728
2729 20. When matching a repeated Unicode property with a minimum greater than zero,
2730 (for example \pL{2,}), PCRE could look past the end of the subject if it
2731 reached it while seeking the minimum number of characters. This could
2732 happen only if some of the characters were more than one byte long, because
2733 there is a check for at least the minimum number of bytes.
2734
2735 21. Refactored the implementation of \p and \P so as to be more general, to
2736 allow for more different types of property in future. This has changed the
2737 compiled form incompatibly. Anybody with saved compiled patterns that use
2738 \p or \P will have to recompile them.
2739
2740 22. Added "Any" and "L&" to the supported property types.
2741
2742 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2743 but give a compile time error if the value is greater than 0xff.
2744
2745 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2746 accidentally not being installed or uninstalled.
2747
2748 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2749 made were to insert the current release number. This seemed silly, because
2750 it made things harder for people building PCRE on systems that don't run
2751 "configure". I have turned pcre.h into a distributed file, no longer built
2752 by "configure", with the version identification directly included. There is
2753 no longer a pcre.h.in file.
2754
2755 However, this change necessitated a change to the pcre-config script as
2756 well. It is built from pcre-config.in, and one of the substitutions was the
2757 release number. I have updated configure.ac so that ./configure now finds
2758 the release number by grepping pcre.h.
2759
2760 26. Added the ability to run the tests under valgrind.
2761
2762
2763 Version 6.4 05-Sep-05
2764 ---------------------
2765
2766 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2767 "--" to be printed when multiple files were scanned, even when none of the
2768 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2769 consider it to be a bug, and have restored the previous behaviour.
2770
2771 2. A couple of code tidies to get rid of compiler warnings.
2772
2773 3. The pcretest program used to cheat by referring to symbols in the library
2774 whose names begin with _pcre_. These are internal symbols that are not
2775 really supposed to be visible externally, and in some environments it is
2776 possible to suppress them. The cheating is now confined to including
2777 certain files from the library's source, which is a bit cleaner.
2778
2779 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2780 file's purpose clearer.
2781
2782 5. Reorganized pcre_ucp_findchar().
2783
2784
2785 Version 6.3 15-Aug-05
2786 ---------------------
2787
2788 1. The file libpcre.pc.in did not have general read permission in the tarball.
2789
2790 2. There were some problems when building without C++ support:
2791
2792 (a) If C++ support was not built, "make install" and "make test" still
2793 tried to test it.
2794
2795 (b) There were problems when the value of CXX was explicitly set. Some
2796 changes have been made to try to fix these, and ...
2797
2798 (c) --disable-cpp can now be used to explicitly disable C++ support.
2799
2800 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2801 backslash in a target when C++ was disabled. This confuses some
2802 versions of "make", apparently. Using an intermediate variable solves
2803 this. (Same for CPP_LOBJ.)
2804
2805 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2806 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2807 necessary on certain architectures.
2808
2809 4. Added a setting of -export-symbols-regex to the link command to remove
2810 those symbols that are exported in the C sense, but actually are local
2811 within the library, and not documented. Their names all begin with
2812 "_pcre_". This is not a perfect job, because (a) we have to except some
2813 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2814 available (and never for static libraries). I have made a note to try to
2815 find a way round (a) in the future.
2816
2817
2818 Version 6.2 01-Aug-05
2819 ---------------------
2820
2821 1. There was no test for integer overflow of quantifier values. A construction
2822 such as {1111111111111111} would give undefined results. What is worse, if
2823 a minimum quantifier for a parenthesized subpattern overflowed and became
2824 negative, the calculation of the memory size went wrong. This could have
2825 led to memory overwriting.
2826
2827 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2828
2829 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2830 operating environments where this matters.
2831
2832 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2833 PCRE options from within the C++ wrapper.
2834
2835 5. Named capturing subpatterns were not being correctly counted when a pattern
2836 was compiled. This caused two problems: (a) If there were more than 100
2837 such subpatterns, the calculation of the memory needed for the whole
2838 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2839 back references of the form \12, where the number was greater than 9, were
2840 not recognized as back references, even though there were sufficient
2841 previous subpatterns.
2842
2843 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2844 versions of gcc, e.g. 2.95.4.
2845
2846
2847 Version 6.1 21-Jun-05
2848 ---------------------
2849
2850 1. There was one reference to the variable "posix" in pcretest.c that was not
2851 surrounded by "#if !defined NOPOSIX".
2852
2853 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2854 the cross-check on the old pcre_info() function, for the benefit of the
2855 cut-down version of PCRE that is currently imported into Exim.
2856
2857 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2858 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2859 patterns that start (?i) but I don't think that matters. The use of (?i) is
2860 just an example; this all applies to the other options as well.
2861
2862 4. Since libtool seems to echo the compile commands it is issuing, the output
2863 from "make" can be reduced a bit by putting "@" in front of each libtool
2864 compile command.
2865
2866 5. Patch from the folks at Google for configure.in to be a bit more thorough
2867 in checking for a suitable C++ installation before trying to compile the
2868 C++ stuff. This should fix a reported problem when a compiler was present,
2869 but no suitable headers.
2870
2871 6. The man pages all had just "PCRE" as their title. I have changed them to
2872 be the relevant file name. I have also arranged that these names are
2873 retained in the file doc/pcre.txt, which is a concatenation in text format
2874 of all the man pages except the little individual ones for each function.
2875
2876 7. The NON-UNIX-USE file had not been updated for the different set of source
2877 files that come with release 6. I also added a few comments about the C++
2878 wrapper.
2879
2880
2881 Version 6.0 07-Jun-05
2882 ---------------------
2883
2884 1. Some minor internal re-organization to help with my DFA experiments.
2885
2886 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2887 didn't matter for the library itself when fully configured, but did matter
2888 when compiling without UCP support, or within Exim, where the ucp files are
2889 not imported.
2890
2891 3. Refactoring of the library code to split up the various functions into
2892 different source modules. The addition of the new DFA matching code (see
2893 below) to a single monolithic source would have made it really too
2894 unwieldy, quite apart from causing all the code to be include in a
2895 statically linked application, when only some functions are used. This is
2896 relevant even without the DFA addition now that patterns can be compiled in
2897 one application and matched in another.
2898
2899 The downside of splitting up is that there have to be some external
2900 functions and data tables that are used internally in different modules of
2901 the library but which are not part of the API. These have all had their
2902 names changed to start with "_pcre_" so that they are unlikely to clash
2903 with other external names.
2904
2905 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2906 a different (DFA) algorithm. Although it is slower than the original
2907 function, it does have some advantages for certain types of matching
2908 problem.
2909
2910 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2911 including restarting after a partial match.
2912
2913 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2914 defined when compiling for Windows was sent to me. I have put it into the
2915 code, though I have no means of testing or verifying it.
2916
2917 7. Added the pcre_refcount() auxiliary function.
2918
2919 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2920 match before or at the first newline in the subject string. In pcretest,
2921 the /f option on a pattern can be used to set this.
2922
2923 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2924 would behave wrongly. This has been present in PCRE since release 4.0.
2925
2926 10. A number of changes to the pcregrep command:
2927
2928 (a) Refactored how -x works; insert ^(...)$ instead of setting
2929 PCRE_ANCHORED and checking the length, in preparation for adding
2930 something similar for -w.
2931
2932 (b) Added the -w (match as a word) option.
2933
2934 (c) Refactored the way lines are read and buffered so as to have more
2935 than one at a time available.
2936
2937 (d) Implemented a pcregrep test script.
2938
2939 (e) Added the -M (multiline match) option. This allows patterns to match
2940 over several lines of the subject. The buffering ensures that at least
2941 8K, or the rest of the document (whichever is the shorter) is available
2942 for matching (and similarly the previous 8K for lookbehind assertions).
2943
2944 (f) Changed the --help output so that it now says
2945
2946 -w, --word-regex(p)
2947
2948 instead of two lines, one with "regex" and the other with "regexp"
2949 because that confused at least one person since the short forms are the
2950 same. (This required a bit of code, as the output is generated
2951 automatically from a table. It wasn't just a text change.)
2952
2953 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2954 option but starts with a hyphen. Could be a pattern or a path name
2955 starting with a hyphen, for instance.
2956
2957 (h) "-" can be given as a file name to represent stdin.
2958
2959 (i) When file names are being printed, "(standard input)" is used for
2960 the standard input, for compatibility with GNU grep. Previously
2961 "<stdin>" was used.
2962
2963 (j) The option --label=xxx can be used to supply a name to be used for
2964 stdin when file names are being printed. There is no short form.
2965
2966 (k) Re-factored the options decoding logic because we are going to add
2967 two more options that take data. Such options can now be given in four
2968 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2969
2970 (l) Added the -A, -B, and -C options for requesting that lines of context
2971 around matches be printed.
2972
2973 (m) Added the -L option to print the names of files that do not contain
2974 any matching lines, that is, the complement of -l.
2975
2976 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2977 continue to scan other files.
2978
2979 (o) The -s option was incorrectly implemented. For compatibility with other
2980 greps, it now suppresses the error message for a non-existent or non-
2981 accessible file (but not the return code). There is a new option called
2982 -q that suppresses the output of matching lines, which was what -s was
2983 previously doing.
2984
2985 (p) Added --include and --exclude options to specify files for inclusion
2986 and exclusion when recursing.
2987
2988 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2989 Hopefully, it now does.
2990
2991 12. Missing cast in pcre_study().
2992
2993 13. Added an "uninstall" target to the makefile.
2994
2995 14. Replaced "extern" in the function prototypes in Makefile.in with
2996 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2997 world, but is set differently for Windows.
2998
2999 15. Added a second compiling function called pcre_compile2(). The only
3000 difference is that it has an extra argument, which is a pointer to an
3001 integer error code. When there is a compile-time failure, this is set
3002 non-zero, in addition to the error test pointer being set to point to an
3003 error message. The new argument may be NULL if no error number is required
3004 (but then you may as well call pcre_compile(), which is now just a
3005 wrapper). This facility is provided because some applications need a
3006 numeric error indication, but it has also enabled me to tidy up the way
3007 compile-time errors are handled in the POSIX wrapper.
3008
3009 16. Added VPATH=.libs to the makefile; this should help when building with one
3010 prefix path and installing with another. (Or so I'm told by someone who
3011 knows more about this stuff than I do.)
3012
3013 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3014 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3015 match everything, including newlines. This is not POSIX-compatible, but
3016 somebody wanted the feature. From pcretest it can be activated by using
3017 both the P and the s flags.
3018
3019 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3020
3021 19. libpcre.pc was being incorrectly installed as executable.
3022
3023 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3024 it now also looks for '\r' so that it will work unmodified on Windows.
3025
3026 21. Added Google's contributed C++ wrapper to the distribution.
3027
3028 22. Added some untidy missing memory free() calls in pcretest, to keep
3029 Electric Fence happy when testing.
3030
3031
3032
3033 Version 5.0 13-Sep-04
3034 ---------------------
3035
3036 1. Internal change: literal characters are no longer packed up into items
3037 containing multiple characters in a single byte-string. Each character
3038 is now matched using a separate opcode. However, there may be more than one
3039 byte in the character in UTF-8 mode.
3040
3041 2. The pcre_callout_block structure has two new fields: pattern_position and
3042 next_item_length. These contain the offset in the pattern to the next match
3043 item, and its length, respectively.
3044
3045 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3046 insertion of callouts before each pattern item. Added the /C option to
3047 pcretest to make use of this.
3048
3049 4. On the advice of a Windows user, the lines
3050
3051 #if defined(_WIN32) || defined(WIN32)
3052 _setmode( _fileno( stdout ), 0x8000 );
3053 #endif /* defined(_WIN32) || defined(WIN32) */
3054
3055 have been added to the source of pcretest. This apparently does useful
3056 magic in relation to line terminators.
3057
3058 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3059 for the benefit of those environments where the "b" makes a difference.
3060
3061 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3062 to know about it. I have put a hack into configure.in that adds in code
3063 to set GCC=yes if CC=icc. This seems to end up at a point in the
3064 generated configure script that is early enough to affect the setting of
3065 compiler options, which is what is needed, but I have no means of testing
3066 whether it really works. (The user who reported this had patched the
3067 generated configure script, which of course I cannot do.)
3068
3069 LATER: After change 22 below (new libtool files), the configure script
3070 seems to know about icc (and also ecc). Therefore, I have commented out
3071 this hack in configure.in.
3072
3073 7. Added support for pkg-config (2 patches were sent in).
3074
3075 8. Negated POSIX character classes that used a combination of internal tables
3076 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3077 [[:^ascii]]. Typically, they would match almost any characters. The other
3078 POSIX classes were not broken in this way.
3079
3080 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3081 to find the match, as PCRE was deluded into thinking that the match had to
3082 start at the start point or following a newline. The same bug applied to
3083 patterns with negative forward assertions or any backward assertions
3084 preceding ".*" at the start, unless the pattern required a fixed first
3085 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3086
3087 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3088 starting at the last subject character, bytes beyond the end of the subject
3089 string were read.
3090
3091 11. Renamed the variable "class" as "classbits" to make life easier for C++
3092 users. (Previously there was a macro definition, but it apparently wasn't
3093 enough.)
3094
3095 12. Added the new field "tables" to the extra data so that tables can be passed
3096 in at exec time, or the internal tables can be re-selected. This allows
3097 a compiled regex to be saved and re-used at a later time by a different
3098 program that might have everything at different addresses.
3099
3100 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3101 -R library as well as a -L library.
3102
3103 14. The debugging options of pcretest (-d on the command line or D on a
3104 pattern) showed incorrect output for anything following an extended class
3105 that contained multibyte characters and which was followed by a quantifier.
3106
3107 15. Added optional support for general category Unicode character properties
3108 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3109 support. It adds about 90K to the size of the library. The meanings of the
3110 inbuilt class escapes such as \d and \s have NOT been changed.
3111
3112 16. Updated pcredemo.c to include calls to free() to release the memory for the
3113 compiled pattern.
3114
3115 17. The generated file chartables.c was being created in the source directory
3116 instead of in the building directory. This caused the build to fail if the
3117 source directory was different from the building directory, and was
3118 read-only.
3119
3120 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3121 file. No doubt somebody will tell me if they don't make sense... Also added
3122 Dan Mooney's comments about building on OpenVMS.
3123
3124 19. Added support for partial matching via the PCRE_PARTIAL option for
3125 pcre_exec() and the \P data escape in pcretest.
3126
3127 20. Extended pcretest with 3 new pattern features:
3128
3129 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3130 write the compiled pattern to the file whose name is "rest-of-line".
3131 This is a straight binary dump of the data, with the saved pointer to
3132 the character tables forced to be NULL. The study data, if any, is
3133 written too. After writing, pcretest reads a new pattern.
3134
3135 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3136 compiled pattern from the given file. There must not be any
3137 occurrences of "<" in the file name (pretty unlikely); if there are,
3138 pcretest will instead treat the initial "<" as a pattern delimiter.
3139 After reading in the pattern, pcretest goes on to read data lines as
3140 usual.
3141
3142 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3143 and 16-bit fields in a compiled pattern, to simulate a pattern that
3144 was compiled on a host of opposite endianness.
3145
3146 21. The pcre-exec() function can now cope with patterns that were compiled on
3147 hosts of opposite endianness, with this restriction:
3148
3149 As for any compiled expression that is saved and used later, the tables
3150 pointer field cannot be preserved; the extra_data field in the arguments
3151 to pcre_exec() should be used to pass in a tables address if a value
3152 other than the default internal tables were used at compile time.
3153
3154 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3155 now diagnosed as an error. Previously, most of the time, a negative number
3156 would have been treated as zero, but if in addition "ovector" was passed as
3157 NULL, a crash could occur.
3158
3159 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3160 new versions from the libtool 1.5 distribution (the last one is a copy of
3161 a file called libtool.m4). This seems to have fixed the need to patch
3162 "configure" to support Darwin 1.3 (which I used to do). However, I still
3163 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3164 workstation).
3165
3166 24. Changed the PCRE licence to be the more standard "BSD" licence.
3167
3168
3169 Version 4.5 01-Dec-03
3170 ---------------------
3171
3172 1. There has been some re-arrangement of the code for the match() function so
3173 that it can be compiled in a version that does not call itself recursively.
3174 Instead, it keeps those local variables that need separate instances for
3175 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3176 needs to "recurse". Keeping track of where control must go is done by means
3177 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3178 hide most of the details from the main code, and operates only if
3179 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3180 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3181 operating.
3182
3183 To make it easier for callers to provide specially tailored get/free
3184 functions for this usage, two new functions, pcre_stack_malloc, and
3185 pcre_stack_free, are used. They are always called in strict stacking order,
3186 and the size of block requested is always the same.
3187
3188 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3189 PCRE has been compiled to use the stack or the heap for recursion. The
3190 -C option of pcretest uses this to show which version is compiled.
3191
3192 A new data escape \S, is added to pcretest; it causes the amounts of store
3193 obtained and freed by both kinds of malloc/free at match time to be added
3194 to the output.
3195
3196 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3197 what's available on my current Linux desktop machine.
3198
3199 3. When matching a UTF-8 string, the test for a valid string at the start has
3200 been extended. If start_offset is not zero, PCRE now checks that it points
3201 to a byte that is the start of a UTF-8 character. If not, it returns
3202 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3203 this is necessary because there may be backward assertions in the pattern.
3204 When matching the same subject several times, it may save resources to use
3205 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3206
3207 4. The code for checking the validity of UTF-8 strings has been tightened so
3208 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3209 containing "overlong sequences".
3210
3211 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3212 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3213 should have been "&", but it just so happened that all the cases this let
3214 through by mistake were picked up later in the function.
3215
3216 6. I had used a variable called "isblank" - this is a C99 function, causing
3217 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3218
3219 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3220 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3221 so the version is shown; (c) stop "make test" echoing "./RunTest".
3222
3223 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3224
3225 9. The prototype for memmove() for systems that don't have it was using
3226 size_t, but the inclusion of the header that defines size_t was later. I've
3227 moved the #includes for the C headers earlier to avoid this.
3228
3229 10. Added some adjustments to the code to make it easier to compiler on certain
3230 special systems:
3231
3232 (a) Some "const" qualifiers were missing.
3233 (b) Added the macro EXPORT before all exported functions; by default this
3234 is defined to be empty.
3235 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3236 that it reads its output file name as an argument instead of writing
3237 to the standard output and assuming this can be redirected.
3238
3239 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3240 class containing characters with values greater than 255, PCRE compilation
3241 went into a loop.
3242
3243 12. A recursive reference to a subpattern that was within another subpattern
3244 that had a minimum quantifier of zero caused PCRE to crash. For example,
3245 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3246 recursion. If the recursively-called subpattern itself had a zero repeat,
3247 that was OK.
3248
3249 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3250 buffer into which it was copied (for escape processing) was still set at
3251 1024, so long lines caused crashes.
3252
3253 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3254 "internal error: code overflow...". This applied to any character class
3255 that was followed by a possessive quantifier.
3256
3257 15. Modified the Makefile to add libpcre.la as a prerequisite for
3258 libpcreposix.la because I was told this is needed for a parallel build to
3259 work.
3260
3261 16. If a pattern that contained .* following optional items at the start was
3262 studied, the wrong optimizing data was generated, leading to matching
3263 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3264 matching string must start with a or b or c. The correct conclusion for
3265 this pattern is that a match can start with any character.
3266
3267
3268 Version 4.4 13-Aug-03
3269 ---------------------
3270
3271 1. In UTF-8 mode, a character class containing characters with values between
3272 127 and 255 was not handled correctly if the compiled pattern was studied.
3273 In fixing this, I have also improved the studying algorithm for such
3274 classes (slightly).
3275
3276 2. Three internal functions had redundant arguments passed to them. Removal
3277 might give a very teeny performance improvement.
3278
3279 3. Documentation bug: the value of the capture_top field in a callout is *one
3280 more than* the number of the hightest numbered captured substring.
3281
3282 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3283 in incorrectly linking with a previously installed version. They now link
3284 explicitly with libpcre.la.
3285
3286 5. configure.in no longer needs to recognize Cygwin specially.
3287
3288 6. A problem in pcre.in for Windows platforms is fixed.
3289
3290 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3291 pcretest, it used to include the size of the study block as part of its
3292 output. Unfortunately, the structure contains a field that has a different
3293 size on different hardware architectures. This meant that the tests that
3294 showed this size failed. As the block is currently always of a fixed size,
3295 this information isn't actually particularly useful in pcretest output, so
3296 I have just removed it.
3297
3298 8. Three pre-processor statements accidentally did not start in column 1.
3299 Sadly, there are *still* compilers around that complain, even though
3300 standard C has not required this for well over a decade. Sigh.
3301
3302 9. In pcretest, the code for checking callouts passed small integers in the
3303 callout_data field, which is a void * field. However, some picky compilers
3304 complained about the casts involved for this on 64-bit systems. Now
3305 pcretest passes the address of the small integer instead, which should get
3306 rid of the warnings.
3307
3308 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3309 both compile and run time, and gives an error if an invalid UTF-8 sequence
3310 is found. There is a option for disabling this check in cases where the
3311 string is known to be correct and/or the maximum performance is wanted.
3312
3313 11. In response to a bug report, I changed one line in Makefile.in from
3314
3315 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3316 to
3317 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3318
3319 to look similar to other lines, but I have no way of telling whether this
3320 is the right thing to do, as I do not use Windows. No doubt I'll get told
3321 if it's wrong...
3322
3323
3324 Version 4.3 21-May-03
3325 ---------------------
3326
3327 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3328 Makefile.
3329
3330 2. Some refactoring to improve the quality of the code:
3331
3332 (i) The utf8_table... variables are now declared "const".
3333
3334 (ii) The code for \cx, which used the "case flipping" table to upper case
3335 lower case letters, now just substracts 32. This is ASCII-specific,
3336 but the whole concept of \cx is ASCII-specific, so it seems
3337 reasonable.
3338
3339 (iii) PCRE was using its character types table to recognize decimal and
3340 hexadecimal digits in the pattern. This is silly, because it handles
3341 only 0-9, a-f, and A-F, but the character types table is locale-
3342 specific, which means strange things might happen. A private
3343 table is now used for this - though it costs 256 bytes, a table is
3344 much faster than multiple explicit tests. Of course, the standard
3345 character types table is still used for matching digits in subject
3346 strings against \d.
3347
3348 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3349 ending in _t are). So I've renamed it as ESC_tee.
3350
3351 3. The first argument for regexec() in the POSIX wrapper should have been
3352 defined as "const".
3353
3354 4. Changed pcretest to use malloc() for its buffers so that they can be
3355 Electric Fenced for debugging.
3356
3357 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3358 to read one or more bytes before the start of the subject string. Often this
3359 had no effect on PCRE's behaviour, but in some circumstances it could
3360 provoke a segmentation fault.
3361
3362 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3363 to try to read one or more bytes before the start of the subject string.
3364
3365 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3366 UTF-8 support could misbehave in various ways if the subject string
3367 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3368 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3369 back over UTF-8 characters.)
3370
3371
3372 Version 4.2 14-Apr-03
3373 ---------------------
3374
3375 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3376
3377 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3378 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3379 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3380 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3381 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3382 and BUILD_EXEEXT
3383 Note: automatic setting of the BUILD variables is not yet working
3384 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3385 compile-time but not at link-time
3386 [LINK]: use for linking executables only
3387 make different versions for Windows and non-Windows
3388 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3389 libraries
3390 [LINK_FOR_BUILD]: new variable
3391 [OBJEXT]: use throughout
3392 [EXEEXT]: use throughout
3393 <winshared>: new target
3394 <wininstall>: new target
3395 <dftables.o>: use native compiler
3396 <dftables>: use native linker
3397 <install>: handle Windows platform correctly
3398 <clean>: ditto
3399 <check>: ditto
3400 copy DLL to top builddir before testing
3401
3402 As part of these changes, -no-undefined was removed again. This was reported
3403 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3404 in any case.
3405
3406 3. Some tidies to get rid of compiler warnings:
3407
3408 . In the match_data structure, match_limit was an unsigned long int, whereas
3409 match_call_count was an int. I've made them both unsigned long ints.
3410
3411 . In pcretest the fact that a const uschar * doesn't automatically cast to
3412 a void * provoked a warning.
3413
3414 . Turning on some more compiler warnings threw up some "shadow" variables
3415 and a few more missing casts.
3416
3417 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3418 option, a class that contained a single character with a value between 128
3419 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3420
3421 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3422 option, a class that contained several characters, but with at least one
3423 whose value was between 128 and 255 caused PCRE to crash.
3424
3425
3426 Version 4.1 12-Mar-03
3427 ---------------------
3428
3429 1. Compiling with gcc -pedantic found a couple of places where casts were
3430 needed, and a string in dftables.c that was longer than standard compilers are
3431 required to support.
3432
3433 2. Compiling with Sun's compiler found a few more places where the code could
3434 be tidied up in order to avoid warnings.
3435
3436 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3437 first of these names is deprecated in the latest Autoconf in favour of the name
3438 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3439 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3440 analogy I have changed it to CFLAGS_FOR_BUILD.
3441
3442 4. Added -no-undefined to the linking command in the Makefile, because this is
3443 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3444 linking step for the pcreposix library.
3445
3446 5. PCRE was failing to diagnose the case of two named groups with the same
3447 name.
3448
3449 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3450 literal character that is needed in the subject for a match, and scans along to
3451 ensure that it is present before embarking on the full matching process. This
3452 saves time in cases of nested unlimited repeats that are never going to match.
3453 Problem: the scan can take a lot of time if the subject is very long (e.g.
3454 megabytes), thus penalizing straightforward matches. It is now done only if the
3455 amount of subject to be scanned is less than 1000 bytes.
3456
3457 7. A lesser problem with the same optimization is that it was recording the
3458 first character of an anchored pattern as "needed", thus provoking a search
3459 right along the subject, even when the first match of the pattern was going to
3460 fail. The "needed" character is now not set for anchored patterns, unless it
3461 follows something in the pattern that is of non-fixed length. Thus, it still
3462 fulfils its original purpose of finding quick non-matches in cases of nested
3463 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3464
3465
3466 Version 4.0 17-Feb-03
3467 ---------------------
3468
3469 1. If a comment in an extended regex that started immediately after a meta-item
3470 extended to the end of string, PCRE compiled incorrect data. This could lead to
3471 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3472
3473 2. Moved to autoconf 2.53 and libtool 1.4.2.
3474
3475 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3476 the special perltest8 script is no longer needed - all the tests can be run
3477 from a single perltest script.
3478
3479 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3480 by \s. It has now been removed in PCRE. This means it isn't recognized as
3481 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3482 class [:space:] *does* include VT, thereby creating a mess.
3483
3484 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3485 space and tab.
3486
3487 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3488 its new features into the main test script, reducing the number of scripts.
3489
3490 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3491 were backward compatible, and made the (?i) apply to the whole pattern, as if
3492 /i were given. Now it behaves more logically, and applies the option setting
3493 only to what follows. PCRE has been changed to follow suit. However, if it
3494 finds options settings right at the start of the pattern, it extracts them into
3495 the global options, as before. Thus, they show up in the info data.
3496
3497 8. Added support for the \Q...\E escape sequence. Characters in between are
3498 treated as literals. This is slightly different from Perl in that $ and @ are
3499 also handled as literals inside the quotes. In Perl, they will cause variable
3500 interpolation. Note the following examples:
3501
3502 Pattern PCRE matches Perl matches
3503
3504 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3505 \Qabc\$xyz\E abc\$xyz abc\$xyz
3506 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3507
3508 For compatibility with Perl, \Q...\E sequences are recognized inside character
3509 classes as well as outside them.
3510
3511 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3512 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3513 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3514 signed/unsigned warnings.
3515
3516 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3517 option for pcretest, so I've replaced it by a simple function that does just
3518 that job.
3519
3520 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3521 "pcregrep -".
3522
3523 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3524 Java package. This provides some syntactic sugar for simple cases of what my
3525 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3526 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3527 item, you can use this simplified notation. Note that only makes sense with
3528 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3529 greediness, whatever the setting of the PCRE_UNGREEDY option.
3530
3531 13. A change of greediness default within a pattern was not taking effect at
3532 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3533 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3534 was abstracted outside.
3535
3536 14. PCRE now supports the \G assertion. It is true when the current matching
3537 position is at the start point of the match. This differs from \A when the
3538 starting offset is non-zero. Used with the /g option of pcretest (or similar
3539 code), it works in the same way as it does for Perl's /g option. If all
3540 alternatives of a regex begin with \G, the expression is anchored to the start
3541 match position, and the "anchored" flag is set in the compiled expression.
3542
3543 15. Some bugs concerning the handling of certain option changes within patterns
3544 have been fixed. These applied to options other than (?ims). For example,
3545 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3546 been the other way round. Some of this was related to change 7 above.
3547
3548 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3549 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3550 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3551 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3552
3553 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3554 mode. Unlike ".", it always matches newline, whatever the setting of
3555 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3556 assertions. Perl allows it, but it doesn't (in general) work because it can't
3557 calculate the length of the lookbehind. At least, that's the case for Perl
3558 5.8.0 - I've been told they are going to document that it doesn't work in
3559 future.
3560
3561 18. Added an error diagnosis for escapes that PCRE does not support: these are
3562 \L, \l, \N, \P, \p, \U, \u, and \X.
3563
3564 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3565 reading past the end of the pattern in cases such as /[abcd/.
3566
3567 20. PCRE was getting more memory than necessary for patterns with classes that
3568 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3569
3570 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3571 compiling PCRE for use with Virtual Pascal.
3572
3573 22. Small fix to the Makefile to make it work properly if the build is done
3574 outside the source tree.
3575
3576 23. Added a new extension: a condition to go with recursion. If a conditional
3577 subpattern starts with (?(R) the "true" branch is used if recursion has
3578 happened, whereas the "false" branch is used only at the top level.
3579
3580 24. When there was a very long string of literal characters (over 255 bytes
3581 without UTF support, over 250 bytes with UTF support), the computation of how
3582 much memory was required could be incorrect, leading to segfaults or other
3583 strange effects.
3584
3585 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3586 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3587 there was a subsequent back reference to those brackets. This meant that, for
3588 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3589 possible to check for precisely this case. All we can do is abandon the
3590 optimization if .* occurs inside capturing brackets when there are any back
3591 references whatsoever. (See below for a better fix that came later.)
3592
3593 26. The handling of the optimization for finding the first character of a
3594 non-anchored pattern, and for finding a character that is required later in the
3595 match were failing in some cases. This didn't break the matching; it just
3596 failed to optimize when it could. The way this is done has been re-implemented.
3597
3598 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3599
3600 28. Added a new feature that provides some of the functionality that Perl
3601 provides with (?{...}). The facility is termed a "callout". The way it is done
3602 in PCRE is for the caller to provide an optional function, by setting
3603 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3604 global variable. By default it is unset, which disables all calling out. To get
3605 the function called, the regex must include (?C) at appropriate points. This
3606 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3607 This provides a means of identifying different callout points. When PCRE
3608 reaches such a point in the regex, if pcre_callout has been set, the external
3609 function is called. It is provided with data in a structure called
3610 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3611 matching continues; if it returns a non-zero value, the match at the current
3612 point fails. However, backtracking will occur if possible. [This was changed
3613 later and other features added - see item 49 below.]
3614
3615 29. pcretest is upgraded to test the callout functionality. It provides a
3616 callout function that displays information. By default, it shows the start of
3617 the match and the current position in the text. There are some new data escapes
3618 to vary what happens:
3619
3620 \C+ in addition, show current contents of captured substrings
3621 \C- do not supply a callout function
3622 \C!n return 1 when callout number n is reached
3623 \C!n!m return 1 when callout number n is reached for the mth time
3624
3625 30. If pcregrep was called with the -l option and just a single file name, it
3626 output "<stdin>" if a match was found, instead of the file name.
3627
3628 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3629 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3630 pcre_exec(). This saves a malloc/free per call. The default value of
3631 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3632 when configuring.
3633
3634 32. The default maximum size of a compiled pattern is 64K. There have been a
3635 few cases of people hitting this limit. The code now uses macros to handle the
3636 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3637 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3638 configuring. Tests 2 and 5 work only with 2-byte links because they output
3639 debugging information about compiled patterns.
3640
3641 33. Internal code re-arrangements:
3642
3643 (a) Moved the debugging function for printing out a compiled regex into
3644 its own source file (printint.c) and used #include to pull it into
3645 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3646 separate copies.
3647
3648 (b) Defined the list of op-code names for debugging as a macro in
3649 internal.h so that it is next to the definition of the opcodes.
3650
3651 (c) Defined a table of op-code lengths for simpler skipping along compiled
3652 code. This is again a macro in internal.h so that it is next to the
3653 definition of the opcodes.
3654
3655 34. Added support for recursive calls to individual subpatterns, along the
3656 lines of Robin Houston's patch (but implemented somewhat differently).
3657
3658 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3659 allow it to read and process whole directories in Win32. This code was
3660 contributed by Lionel Fourquaux; it has not been tested by me.
3661
3662 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3663 used to name a group. Names consist of alphanumerics and underscores, and must
3664 be unique. Back references use the syntax (?P=name) and recursive calls use
3665 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3666 numbers. The function pcre_fullinfo() can be used after compilation to extract
3667 a name/number map. There are three relevant calls:
3668
3669 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3670 PCRE_INFO_NAMECOUNT yields the number of entries
3671 PCRE_INFO_NAMETABLE yields a pointer to the map.
3672
3673 The map is a vector of fixed-size entries. The size of each entry depends on
3674 the length of the longest name used. The first two bytes of each entry are the
3675 group number, most significant byte first. There follows the corresponding
3676 name, zero terminated. The names are in alphabetical order.
3677
3678 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3679 case instead of 255. Making it the same both with and without UTF-8 support
3680 means that the same test output works with both.
3681
3682 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3683 calling malloc() with a zero argument.
3684
3685 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3686 optimization. I've improved things by keeping a bitmap of backreferences with
3687 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3688 fact referenced, the optimization can be applied. It is unlikely that a
3689 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3690 the match to follow \n) will appear inside brackets with a number greater than
3691 31, but if it does, any back reference > 31 suppresses the optimization.
3692
3693 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3694 of disabling numbered capturing parentheses. Any opening parenthesis that is
3695 not followed by ? behaves as if it were followed by ?: but named parentheses
3696 can still be used for capturing (and they will acquire numbers in the usual
3697 way).
3698
3699 41. Redesigned the return codes from the match() function into yes/no/error so
3700 that errors can be passed back from deep inside the nested calls. A malloc
3701 failure while inside a recursive subpattern call now causes the
3702 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3703
3704 42. It is now possible to set a limit on the number of times the match()
3705 function is called in a call to pcre_exec(). This facility makes it possible to
3706 limit the amount of recursion and backtracking, though not in a directly
3707 obvious way, because the match() function is used in a number of different
3708 circumstances. The count starts from zero for each position in the subject
3709 string (for non-anchored patterns). The default limit is, for compatibility, a
3710 large number, namely 10 000 000. You can change this in two ways:
3711
3712 (a) When configuring PCRE before making, you can use --with-match-limit=n
3713 to set a default value for the compiled library.
3714
3715 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3716 a different value is set. See 45 below.
3717
3718 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3719
3720 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3721 of things that can be changed at compile time. The first argument specifies
3722 what is wanted and the second points to where the information is to be placed.
3723 The current list of available information is:
3724
3725 PCRE_CONFIG_UTF8
3726
3727 The output is an integer that is set to one if UTF-8 support is available;
3728 otherwise it is set to zero.
3729
3730 PCRE_CONFIG_NEWLINE
3731
3732 The output is an integer that it set to the value of the code that is used for
3733 newline. It is either LF (10) or CR (13).
3734
3735 PCRE_CONFIG_LINK_SIZE
3736
3737 The output is an integer that contains the number of bytes used for internal
3738 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3739
3740 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3741
3742 The output is an integer that contains the threshold above which the POSIX
3743 interface uses malloc() for output vectors. See item 31 above.
3744
3745 PCRE_CONFIG_MATCH_LIMIT
3746
3747 The output is an unsigned integer that contains the default limit of the number
3748 of match() calls in a pcre_exec() execution. See 42 above.
3749
3750 44. pcretest has been upgraded by the addition of the -C option. This causes it
3751 to extract all the available output from the new pcre_config() function, and to
3752 output it. The program then exits immediately.
3753
3754 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3755 order to support additional features. One way would have been to define
3756 pcre_exec2() (for example) with extra arguments, but this would not have been
3757 extensible, and would also have required all calls to the original function to
3758 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3759 is used for passing in "extra" data from pcre_study().
3760
3761 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3762 contains the following fields:
3763
3764 flags a bitmap indicating which of the following fields are set
3765 study_data opaque data from pcre_study()
3766 match_limit a way of specifying a limit on match() calls for a specific
3767 call to pcre_exec()
3768 callout_data data for callouts (see 49 below)
3769
3770 The flag bits are also defined in pcre.h, and are
3771
3772 PCRE_EXTRA_STUDY_DATA
3773 PCRE_EXTRA_MATCH_LIMIT
3774 PCRE_EXTRA_CALLOUT_DATA
3775
3776 The pcre_study() function now returns one of these new pcre_extra blocks, with
3777 the actual study data pointed to by the study_data field, and the
3778 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3779 before. That is, this change is entirely upwards-compatible and requires no
3780 change to existing code.
3781
3782 If you want to pass in additional data to pcre_exec(), you can either place it
3783 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3784 block.
3785
3786 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3787 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3788 times with different match limits, until it finds the minimum value needed for
3789 pcre_exec() to complete. The value is then output. This can be instructive; for
3790 most simple matches the number is quite small, but for pathological cases it
3791 gets very large very quickly.
3792
3793 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3794 returns the size of the data block pointed to by the study_data field in a
3795 pcre_extra block, that is, the value that was passed as the argument to
3796 pcre_malloc() when PCRE was getting memory in which to place the information
3797 created by pcre_study(). The fourth argument should point to a size_t variable.
3798 pcretest has been extended so that this information is shown after a successful
3799 pcre_study() call when information about the compiled regex is being displayed.
3800
3801 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3802 because what follows is always an absolute path. (Later: it turns out that this
3803 is more than cosmetic for MinGW, because it doesn't like empty path
3804 components.)
3805
3806 49. Some changes have been made to the callout feature (see 28 above):
3807
3808 (i) A callout function now has three choices for what it returns:
3809
3810 0 => success, carry on matching
3811 > 0 => failure at this point, but backtrack if possible
3812 < 0 => serious error, return this value from pcre_exec()
3813
3814 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3815 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3816 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3817 use by callout functions. It will never be used by PCRE itself.
3818
3819 (ii) The pcre_extra structure (see 45 above) has a void * field called
3820 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3821 pcre_callout_block structure has a field of the same name. The contents of
3822 the field passed in the pcre_extra structure are passed to the callout
3823 function in the corresponding field in the callout block. This makes it
3824 easier to use the same callout-containing regex from multiple threads. For
3825 testing, the pcretest program has a new data escape
3826
3827 \C*n pass the number n (may be negative) as callout_data
3828
3829 If the callout function in pcretest receives a non-zero value as
3830 callout_data, it returns that value.
3831
3832 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3833 there were some redundant $(CFLAGS) in commands that are now specified as
3834 $(LINK), which already includes $(CFLAGS).
3835
3836 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3837 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3838 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3839 one-byte characters throughout. Note that case-insensitive matching applies
3840 only to characters whose values are less than 256. PCRE doesn't support the
3841 notion of cases for higher-valued characters.
3842
3843 (i) A character class whose characters are all within 0-255 is handled as
3844 a bit map, and the map is inverted for negative classes. Previously, a
3845 character > 255 always failed to match such a class; however it should
3846 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3847
3848 (ii) A negated character class with a single character < 255 is coded as
3849 "not this character" (OP_NOT). This wasn't working properly when the test
3850 character was multibyte, either singly or repeated.
3851
3852 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3853 mode, for example: \x{100}{2,3}.
3854
3855 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3856 singly or repeated) now correctly test multibyte characters. However,
3857 PCRE doesn't recognize any characters with values greater than 255 as
3858 digits, spaces, or word characters. Such characters always match \D, \S,
3859 and \W, and never match \d, \s, or \w.
3860
3861 (v) Classes may now contain characters and character ranges with values
3862 greater than 255. For example: [ab\x{100}-\x{400}].
3863
3864 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3865 PCRE in UTF-8 mode.
3866
3867 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3868 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3869 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3870 value.)
3871
3872 53. The single man page has become too large. I have therefore split it up into
3873 a number of separate man pages. These also give rise to individual HTML pages;
3874 these are now put in a separate directory, and there is an index.html page that
3875 lists them all. Some hyperlinking between the pages has been installed.
3876
3877 54. Added convenience functions for handling named capturing parentheses.
3878
3879 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3880 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3881 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3882 are faulted.
3883
3884 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3885 calling configure. These values are used when compiling the dftables.c program
3886 which is run to generate the source of the default character tables. They
3887 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3888 you will need to set these values.
3889
3890 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3891
3892
3893 Version 3.9 02-Jan-02
3894 ---------------------
3895
3896 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3897
3898 2. If --disable-static was given, the building process failed when trying to
3899 build pcretest and pcregrep. (For some reason it was using libtool to compile
3900 them, which is not right, as they aren't part of the library.)
3901
3902
3903 Version 3.8 18-Dec-01
3904 ---------------------
3905
3906 1. The experimental UTF-8 code was completely screwed up. It was packing the
3907 bytes in the wrong order. How dumb can you get?
3908
3909
3910 Version 3.7 29-Oct-01
3911 ---------------------
3912
3913 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3914 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3915 this didn't happen under Solaris 8, where I normally test things.
3916
3917 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3918 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3919 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3920 this fix an hour or so after the initial 3.7 release.)
3921
3922
3923 Version 3.6 23-Oct-01
3924 ---------------------
3925
3926 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3927 offsets passed as NULL with zero offset count.
3928
3929 2. The config.guess and config.sub files had not been updated when I moved to
3930 the latest autoconf.
3931
3932
3933 Version 3.5 15-Aug-01
3934 ---------------------
3935
3936 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3937 had been forgotten.
3938
3939 2. By using declared but undefined structures, we can avoid using "void"
3940 definitions in pcre.h while keeping the internal definitions of the structures
3941 private.
3942
3943 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3944 user point of view, this means that both static and shared libraries are built
3945 by default, but this can be individually controlled. More of the work of
3946 handling this static/shared cases is now inside libtool instead of PCRE's make
3947 file.
3948
3949 4. The pcretest utility is now installed along with pcregrep because it is
3950 useful for users (to test regexs) and by doing this, it automatically gets
3951 relinked by libtool. The documentation has been turned into a man page, so
3952 there are now .1, .txt, and .html versions in /doc.
3953
3954 5. Upgrades to pcregrep:
3955 (i) Added long-form option names like gnu grep.
3956 (ii) Added --help to list all options with an explanatory phrase.
3957 (iii) Added -r, --recursive to recurse into sub-directories.
3958 (iv) Added -f, --file to read patterns from a file.
3959
3960 6. pcre_exec() was referring to its "code" argument before testing that
3961 argument for NULL (and giving an error if it was NULL).
3962
3963 7. Upgraded Makefile.in to allow for compiling in a different directory from
3964 the source directory.
3965
3966 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3967 options bits, the pointer it was passed was to an int instead of to an unsigned
3968 long int. This mattered only on 64-bit systems.
3969
3970 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3971 generated) instead of pcre.in, which it its source. Also made the same change
3972 in several of the .c files.
3973
3974 10. A new release of gcc defines printf() as a macro, which broke pcretest
3975 because it had an ifdef in the middle of a string argument for printf(). Fixed
3976 by using separate calls to printf().
3977
3978 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3979 script, to force use of CR or LF instead of \n in the source. On non-Unix
3980 systems, the value can be set in config.h.
3981
3982 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3983 absolute limit. Changed the text of the error message to make this clear, and
3984 likewise updated the man page.
3985
3986 13. The limit of 99 on the number of capturing subpatterns has been removed.
3987 The new limit is 65535, which I hope will not be a "real" limit.
3988
3989
3990 Version 3.4 22-Aug-00
3991 ---------------------
3992
3993 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3994
3995 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3996
3997
3998 Version 3.3 01-Aug-00
3999 ---------------------
4000
4001 1. If an octal character was given, but the value was greater than \377, it
4002 was not getting masked to the least significant bits, as documented. This could
4003 lead to crashes in some systems.
4004
4005 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4006 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4007
4008 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4009 These just pass their arguments on to (pcre_free)(), but they are provided
4010 because some uses of PCRE bind it to non-C systems that can call its functions,
4011 but cannot call free() or pcre_free() directly.
4012
4013 4. Add "make test" as a synonym for "make check". Corrected some comments in
4014 the Makefile.
4015
4016 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4017 Makefile.
4018
4019 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4020 command called pgrep for grepping around the active processes.
4021
4022 7. Added the beginnings of support for UTF-8 character strings.
4023
4024 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4025 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4026 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4027 out for the ar command.)
4028
4029
4030 Version 3.2 12-May-00
4031 ---------------------
4032
4033 This is purely a bug fixing release.
4034
4035 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4036 of ZA. This was just one example of several cases that could provoke this bug,
4037 which was introduced by change 9 of version 2.00. The code for breaking
4038 infinite loops after an iteration that matches an empty string was't working
4039 correctly.
4040
4041 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4042 when matched against abbab (for example). After matching an empty string, it
4043 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4044 caused it to match further down the string than it should.
4045
4046 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4047 was there because it doesn't seem to be needed, and it causes trouble on some
4048 systems, as it is not a Standard C header. It has been removed.
4049
4050 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4051 were reported on the Macintosh. The changes were from
4052
4053 while ((c = *(++ptr)) != 0 && c != '\n');
4054 to
4055 while ((c = *(++ptr)) != 0 && c != '\n') ;
4056
4057 Totally extraordinary, but if that's what it takes...
4058
4059 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4060 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4061 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4062 assumes the way PCRE uses memmove() (always moving upwards).
4063
4064 6. PCRE is being used in one environment where strchr() is not available. There
4065 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4066 faster code anyway.
4067
4068
4069 Version 3.1 09-Feb-00
4070 ---------------------
4071
4072 The only change in this release is the fixing of some bugs in Makefile.in for
4073 the "install" target:
4074
4075 (1) It was failing to install pcreposix.h.
4076
4077 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4078
4079
4080 Version 3.0 01-Feb-00
4081 ---------------------
4082
4083 1. Add support for the /+ modifier to perltest (to output $` like it does in
4084 pcretest).
4085
4086 2. Add support for the /g modifier to perltest.
4087
4088 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4089 matches null strings.
4090
4091 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4092 pattern. Perl treats empty patterns specially - it reuses the most recent
4093 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4094 effect.
4095
4096 5. The POSIX interface was broken in that it was just handing over the POSIX
4097 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4098 required a bigger vector, with some working space on the end. This means that
4099 the POSIX wrapper now has to get and free some memory, and copy the results.
4100
4101 6. Added some simple autoconf support, placing the test data and the
4102 documentation in separate directories, re-organizing some of the
4103 information files, and making it build pcre-config (a GNU standard). Also added
4104 libtool support for building PCRE as a shared library, which is now the
4105 default.
4106
4107 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4108 09 are not valid octal constants. Single digits will be used for minor values
4109 less than 10.
4110
4111 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4112 existing programs that set these in the POSIX interface can use PCRE without
4113 modification.
4114
4115 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4116 return all that pcre_info() returns, plus additional data. The pcre_info()
4117 function is retained for compatibility, but is considered to be obsolete.
4118
4119 10. Added experimental recursion feature (?R) to handle one common case that
4120 Perl 5.6 will be able to do with (?p{...}).
4121
4122 11. Added support for POSIX character classes like [:alpha:], which Perl is
4123 adopting.
4124
4125
4126 Version 2.08 31-Aug-99
4127 ----------------------
4128
4129 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4130 trying to match at the startoffset position, but instead was moving forward to
4131 the next newline as if a previous match had failed.
4132
4133 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4134 and could get into a loop if a null string was matched other than at the start
4135 of the subject.
4136
4137 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4138 be distinguished at compile time, and for completeness also added PCRE_DATE.
4139
4140 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4141 in GnuWin32 environments.
4142
4143
4144 Version 2.07 29-Jul-99
4145 ----------------------
4146
4147 1. The documentation is now supplied in plain text form and HTML as well as in
4148 the form of man page sources.
4149
4150 2. C++ compilers don't like assigning (void *) values to other pointer types.
4151 In particular this affects malloc(). Although there is no problem in Standard
4152 C, I've put in casts to keep C++ compilers happy.
4153
4154 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4155 should be (const char *).
4156
4157 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4158 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4159 However, I haven't made this a standard facility. The documentation doesn't
4160 mention it, and the Makefile doesn't support it.
4161
4162 5. The Makefile now contains an "install" target, with editable destinations at
4163 the top of the file. The pcretest program is not installed.
4164
4165 6. pgrep -V now gives the PCRE version number and date.
4166
4167 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4168 causing the entire string to be ignored, instead of just the last character.
4169
4170 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4171 non-matching string, it can take a very, very long time, even for strings of
4172 quite modest length, because of the nested recursion. PCRE now does better in
4173 some of these cases. It does this by remembering the last required literal
4174 character in the pattern, and pre-searching the subject to ensure it is present
4175 before running the real match. In other words, it applies a heuristic to detect
4176 some types of certain failure quickly, and in the above example, if presented
4177 with a string that has no trailing " it gives "no match" very quickly.
4178
4179 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4180 other alternatives are tried instead.
4181
4182
4183 Version 2.06 09-Jun-99
4184 ----------------------
4185
4186 1. Change pcretest's output for amount of store used to show just the code
4187 space, because the remainder (the data block) varies in size between 32-bit and
4188 64-bit systems.
4189
4190 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4191 start matching at. This allows lookbehinds to work when searching for multiple
4192 occurrences in a string.
4193
4194 3. Added additional options to pcretest for testing multiple occurrences:
4195
4196 /+ outputs the rest of the string that follows a match
4197 /g loops for multiple occurrences, using the new startoffset argument
4198 /G loops for multiple occurrences by passing an incremented pointer
4199
4200 4. PCRE wasn't doing the "first character" optimization for patterns starting
4201 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4202 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4203 the letter 'x'. On long subject strings, this gives a significant speed-up.
4204
4205
4206 Version 2.05 21-Apr-99
4207 ----------------------
4208
4209 1. Changed the type of magic_number from int to long int so that it works
4210 properly on 16-bit systems.
4211
4212 2. Fixed a bug which caused patterns starting with .* not to work correctly
4213 when the subject string contained newline characters. PCRE was assuming
4214 anchoring for such patterns in all cases, which is not correct because .* will
4215 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4216 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4217 must be retried after every newline in the subject.
4218
4219
4220 Version 2.04 18-Feb-99
4221 ----------------------
4222
4223 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4224 computation of the store needed to hold the pattern was incorrect (too large).
4225 If such patterns were nested a few deep, this could multiply and become a real
4226 problem.
4227
4228 2. Added /M option to pcretest to show the memory requirement of a specific
4229 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4230
4231 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4232 compiled in such a way that the backtracking after subsequent failure was
4233 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4234 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4235
4236
4237 Version 2.03 02-Feb-99
4238 ----------------------
4239
4240 1. Fixed typo and small mistake in man page.
4241
4242 2. Added 4th condition (GPL supersedes if conflict) and created separate
4243 LICENCE file containing the conditions.
4244
4245 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4246 Perl, that is the internal \ allows the delimiter to be included in the
4247 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4248 the final delimiter, add \ to the end of the pattern (to test the error).
4249
4250 4. Added the convenience functions for extracting substrings after a successful
4251 match. Updated pcretest to make it able to test these functions.
4252
4253
4254 Version 2.02 14-Jan-99
4255 ----------------------
4256
4257 1. Initialized the working variables associated with each extraction so that
4258 their saving and restoring doesn't refer to uninitialized store.
4259
4260 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4261 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4262 fix the problem.
4263
4264 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4265 calls, and wasn't printing the correct value for compiling calls. Increased the
4266 default value of LOOPREPEAT, and the number of significant figures in the
4267 times.
4268
4269 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4270
4271 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4272 a building problem on Windows NT with a FAT file system.
4273
4274
4275 Version 2.01 21-Oct-98
4276 ----------------------
4277
4278 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4279 to character tables built by pcre_maketables() in the current locale. If NULL
4280 is passed, the default tables are used.
4281
4282
4283 Version 2.00 24-Sep-98
4284 ----------------------
4285
4286 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4287 it any more.
4288
4289 2. Allow quantification of (?>) groups, and make it work correctly.
4290
4291 3. The first character computation wasn't working for (?>) groups.
4292
4293 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4294 end of the subject) and add 5.005's \z, which really does match only at the
4295 very end of the subject.
4296
4297 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4298
4299 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4300 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4301 localized options. All options to pcre_study() were also removed.
4302
4303 7. Add other new features from 5.005:
4304
4305 $(?<= positive lookbehind
4306 $(?<! negative lookbehind
4307 (?imsx-imsx) added the unsetting capability
4308 such a setting is global if at outer level; local otherwise
4309 (?imsx-imsx:) non-capturing groups with option setting
4310 (?(cond)re|re) conditional pattern matching
4311
4312 A backreference to itself in a repeated group matches the previous
4313 captured string.
4314
4315 8. General tidying up of studying (both automatic and via "study")
4316 consequential on the addition of new assertions.
4317
4318 9. As in 5.005, unlimited repeated groups that could match an empty substring
4319 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4320 runtime if any iteration does actually match an empty substring.
4321
4322 10. Include the RunTest script in the distribution.
4323
4324 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4325 discrepancies, some of which were old and were also with respect to 5.004. They
4326 have now been fixed.
4327
4328
4329 Version 1.09 28-Apr-98
4330 ----------------------
4331
4332 1. A negated single character class followed by a quantifier with a minimum
4333 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4334 program crashes, or just wrong answers. This did not apply to negated classes
4335 containing more than one character, or to minima other than one.
4336
4337
4338 Version 1.08 27-Mar-98
4339 ----------------------
4340
4341 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4342
4343 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4344 latter must appear before anything that relies on it in the pattern.
4345
4346
4347 Version 1.07 16-Feb-98
4348 ----------------------
4349
4350 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4351 repeat of a potentially empty string).
4352
4353
4354 Version 1.06 23-Jan-98
4355 ----------------------
4356
4357 1. Added Markus Oberhumer's little patches for C++.
4358
4359 2. Literal strings longer than 255 characters were broken.
4360
4361
4362 Version 1.05 23-Dec-97
4363 ----------------------
4364
4365 1. Negated character classes containing more than one character were failing if
4366 PCRE_CASELESS was set at run time.
4367
4368
4369 Version 1.04 19-Dec-97
4370 ----------------------
4371
4372 1. Corrected the man page, where some "const" qualifiers had been omitted.
4373
4374 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4375 input syntax.
4376
4377 3. Fixed memory leak which occurred when a regex with back references was
4378 matched with an offsets vector that wasn't big enough. The temporary memory
4379 that is used in this case wasn't being freed if the match failed.
4380
4381 4. Tidied pcretest to ensure it frees memory that it gets.
4382
4383 5. Temporary memory was being obtained in the case where the passed offsets
4384 vector was exactly big enough.
4385
4386 6. Corrected definition of offsetof() from change 5 below.
4387
4388 7. I had screwed up change 6 below and broken the rules for the use of
4389 setjmp(). Now fixed.
4390
4391
4392 Version 1.03 18-Dec-97
4393 ----------------------
4394
4395 1. A erroneous regex with a missing opening parenthesis was correctly
4396 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4397 on some systems.
4398
4399 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4400 it was reported that one broken compiler failed on the former because "code" is
4401 also an independent variable.
4402
4403 3. The erroneous regex a[]b caused an array overrun reference.
4404
4405 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4406 fail on data ending with that character. (It was going on too far, and checking
4407 the next character, typically a binary zero.) This was specific to the
4408 optimized code for single-character negative classes.
4409
4410 5. Added a contributed patch from the TIN world which does the following:
4411
4412 + Add an undef for memmove, in case the the system defines a macro for it.
4413
4414 + Add a definition of offsetof(), in case there isn't one. (I don't know
4415 the reason behind this - offsetof() is part of the ANSI standard - but
4416 it does no harm).
4417
4418 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4419 most of the places where whitespace preceded '#'. I have given up and
4420 allowed the remaining 2 cases to be at the margin.
4421
4422 + Rename some variables in pcre to eliminate shadowing. This seems very
4423 pedantic, but does no harm, of course.
4424
4425 6. Moved the call to setjmp() into its own function, to get rid of warnings
4426 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4427
4428 7. Constructs such as \d{8,} were compiling into the equivalent of
4429 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4430 outcome, but in this particular case used more store than had been allocated,
4431 which caused the bug to be discovered because it threw up an internal error.
4432
4433 8. The debugging code in both pcre and pcretest for outputting the compiled
4434 form of a regex was going wrong in the case of back references followed by
4435 curly-bracketed repeats.
4436
4437
4438 Version 1.02 12-Dec-97
4439 ----------------------
4440
4441 1. Typos in pcre.3 and comments in the source fixed.
4442
4443 2. Applied a contributed patch to get rid of places where it used to remove
4444 'const' from variables, and fixed some signed/unsigned and uninitialized
4445 variable warnings.
4446
4447 3. Added the "runtest" target to Makefile.
4448
4449 4. Set default compiler flag to -O2 rather than just -O.
4450
4451
4452 Version 1.01 19-Nov-97
4453 ----------------------
4454
4455 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4456 like /([ab]*)*/, that is, for classes with more than one character in them.
4457
4458 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4459 as /((?>a*))*/ (a PCRE_EXTRA facility).
4460
4461
4462 Version 1.00 18-Nov-97
4463 ----------------------
4464
4465 1. Added compile-time macros to support systems such as SunOS4 which don't have
4466 memmove() or strerror() but have other things that can be used instead.
4467
4468 2. Arranged that "make clean" removes the executables.
4469
4470
4471 Version 0.99 27-Oct-97
4472 ----------------------
4473
4474 1. Fixed bug in code for optimizing classes with only one character. It was
4475 initializing a 32-byte map regardless, which could cause it to run off the end
4476 of the memory it had got.
4477
4478 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4479
4480
4481 Version 0.98 22-Oct-97
4482 ----------------------
4483
4484 1. Fixed bug in code for handling temporary memory usage when there are more
4485 back references than supplied space in the ovector. This could cause segfaults.
4486
4487
4488 Version 0.97 21-Oct-97
4489 ----------------------
4490
4491 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4492
4493 2. Optimized negated single characters not to use a bit map.
4494
4495 3. Brought error texts together as macro definitions; clarified some of them;
4496 fixed one that was wrong - it said "range out of order" when it meant "invalid
4497 escape sequence".
4498
4499 4. Changed some char * arguments to const char *.
4500
4501 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4502
4503 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4504 pcretest.
4505
4506
4507 Version 0.96 16-Oct-97
4508 ----------------------
4509
4510 1. Added a simple "pgrep" utility to the distribution.
4511
4512 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4513 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4514 where "ddd" means "one or more decimal digits".
4515
4516 3. Fixed serious bug. If a pattern had a back reference, but the call to
4517 pcre_exec() didn't supply a large enough ovector to record the related
4518 identifying subpattern, the match always failed. PCRE now remembers the number
4519 of the largest back reference, and gets some temporary memory in which to save
4520 the offsets during matching if necessary, in order to ensure that
4521 backreferences always work.
4522
4523 4. Increased the compatibility with Perl in a number of ways:
4524
4525 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4526 to request this handling. The option can be set at compile or exec time.
4527
4528 (b) $ matches before a terminating newline by default; an option
4529 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4530 mode). The option can be set at compile or exec time.
4531
4532 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4533 the same as Perl's. If the decimal number it represents is less than 10
4534 or there aren't that many previous left capturing parentheses, an octal
4535 escape is read. Inside a character class, it's always an octal escape,
4536 even if it is a single digit.
4537
4538 (d) An escaped but undefined alphabetic character is taken as a literal,
4539 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4540 escapes.
4541
4542 (e) {0} is now permitted. (The previous item is removed from the compiled
4543 pattern).
4544
4545 5. Changed all the names of code files so that the basic parts are no longer
4546 than 10 characters, and abolished the teeny "globals.c" file.
4547
4548 6. Changed the handling of character classes; they are now done with a 32-byte
4549 bit map always.
4550
4551 7. Added the -d and /D options to pcretest to make it possible to look at the
4552 internals of compilation without having to recompile pcre.
4553
4554
4555 Version 0.95 23-Sep-97
4556 ----------------------
4557
4558 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4559 \x20 at the start of a run of normal characters. These were being treated as
4560 real characters, instead of the source characters being re-checked.
4561
4562
4563 Version 0.94 18-Sep-97
4564 ----------------------
4565
4566 1. The functions are now thread-safe, with the caveat that the global variables
4567 containing pointers to malloc() and free() or alternative functions are the
4568 same for all threads.
4569
4570 2. Get pcre_study() to generate a bitmap of initial characters for non-
4571 anchored patterns when this is possible, and use it if passed to pcre_exec().
4572
4573
4574 Version 0.93 15-Sep-97
4575 ----------------------
4576
4577 1. /(b)|(:+)/ was computing an incorrect first character.
4578
4579 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4580 but not actually doing anything yet.
4581
4582 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4583 as Perl does (e.g. [-az] or [az-]).
4584
4585 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4586 all possible positions.
4587
4588 5. Split up into different modules to avoid including unneeded functions in a
4589 compiled binary. However, compile and exec are still in one module. The "study"
4590 function is split off.
4591
4592 6. The character tables are now in a separate module whose source is generated
4593 by an auxiliary program - but can then be edited by hand if required. There are
4594 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4595 toupper() in the code.
4596
4597 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4598 make them global. Abolish the function for setting them, as the caller can now
4599 set them directly.
4600
4601
4602 Version 0.92 11-Sep-97
4603 ----------------------
4604
4605 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4606 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4607
4608 2. Caseless matching was not working in character classes if the characters in
4609 the pattern were in upper case.
4610
4611 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4612
4613 4. Make PCRE_ANCHORED public and accept as a compile option.
4614
4615 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4616 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4617 pass them.
4618
4619 6. Give an error if bad option bits passed at compile or run time.
4620
4621 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4622 pcretest to cause it to pass that flag.
4623
4624 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4625 options, and the first character, if set.
4626
4627 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4628
4629
4630 Version 0.91 10-Sep-97
4631 ----------------------
4632
4633 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4634 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4635
4636 2. PCRE was looping on encountering an indefinitely repeated back reference to
4637 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4638 Perl does - treats the match as successful.
4639
4640 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12