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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 836 - (show annotations)
Wed Dec 28 17:16:11 2011 UTC (7 years, 9 months ago) by ph10
File size: 224327 byte(s)
Error occurred while calculating annotation data.
Merging all the changes from the pcre16 branch into the trunk.
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 the code for calculating the fixed length of lookbehind
11 assertions.
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
30 Version 8.21 12-Dec-2011
31 ------------------------
32
33 1. Updating the JIT compiler.
34
35 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
36 are added as well.
37
38 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
39 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
40 calling _pcre_jit_exec. Some extra comments are added.
41
42 4. (*MARK) settings inside atomic groups that do not contain any capturing
43 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
44 was introduced by change 18 for 8.20.
45
46 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
47 ECMA-262 standard.
48
49 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
50 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
51 This bug was probably introduced by change 9 of 8.13.
52
53 7. While fixing 6 above, I noticed that a number of other items were being
54 incorrectly rejected as "not fixed length". This arose partly because newer
55 opcodes had not been added to the fixed-length checking code. I have (a)
56 corrected the bug and added tests for these items, and (b) arranged for an
57 error to occur if an unknown opcode is encountered while checking for fixed
58 length instead of just assuming "not fixed length". The items that were
59 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
60 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
61 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
62
63 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
64 being incorrectly compiled and would have given unpredicatble results.
65
66 9. A possessively repeated subpattern with minimum repeat count greater than
67 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
68 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
69 the first (A) could occur when it should not.
70
71 10. Add a cast and remove a redundant test from the code.
72
73 11. JIT should use pcre_malloc/pcre_free for allocation.
74
75 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
76 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
77 is anything other than /usr, -L is still shown).
78
79 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
80
81 14. Perl does not support \N without a following name in a [] class; PCRE now
82 also gives an error.
83
84 15. If a forward reference was repeated with an upper limit of around 2000,
85 it caused the error "internal error: overran compiling workspace". The
86 maximum number of forward references (including repeats) was limited by the
87 internal workspace, and dependent on the LINK_SIZE. The code has been
88 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
89 the default depends on LINK_SIZE. There is a new upper limit (for safety)
90 of around 200,000 forward references. While doing this, I also speeded up
91 the filling in of repeated forward references.
92
93 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
94 incorrectly expecting the subject to contain another "a" after the start.
95
96 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
97 in the match, the SKIP should be ignored. This was not happening; instead
98 the SKIP was being treated as NOMATCH. For patterns such as
99 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
100 tested.
101
102 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
103 now much more compatible with Perl, in particular in cases where the result
104 is a non-match for a non-anchored pattern. For example, if
105 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
106 "m", where previously it did not return a name. A side effect of this
107 change is that for partial matches, the last encountered mark name is
108 returned, as for non matches. A number of tests that were previously not
109 Perl-compatible have been moved into the Perl-compatible test files. The
110 refactoring has had the pleasing side effect of removing one argument from
111 the match() function, thus reducing its stack requirements.
112
113 19. If the /S+ option was used in pcretest to study a pattern using JIT,
114 subsequent uses of /S (without +) incorrectly behaved like /S+.
115
116 21. Retrieve executable code size support for the JIT compiler and fixing
117 some warnings.
118
119 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
120 not work when the shorter character appeared right at the end of the
121 subject string.
122
123 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
124 systems.
125
126 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
127 output it when the /M option is used in pcretest.
128
129 25. The CheckMan script was not being included in the distribution. Also, added
130 an explicit "perl" to run Perl scripts from the PrepareRelease script
131 because this is reportedly needed in Windows.
132
133 26. If study data was being save in a file and studying had not found a set of
134 "starts with" bytes for the pattern, the data written to the file (though
135 never used) was taken from uninitialized memory and so caused valgrind to
136 complain.
137
138 27. Updated RunTest.bat as provided by Sheri Pierce.
139
140 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
141
142 29. Computation of memory usage for the table of capturing group names was
143 giving an unnecessarily large value.
144
145
146 Version 8.20 21-Oct-2011
147 ------------------------
148
149 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
150 a POSIX class. After further experiments with Perl, which convinced me that
151 Perl has bugs and confusions, a closing square bracket is no longer allowed
152 in a POSIX name. This bug also affected patterns with classes that started
153 with full stops.
154
155 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
156 captured substring, but while checking the failing first alternative,
157 substring 1 is temporarily captured. If the output vector supplied to
158 pcre_exec() was not big enough for this capture, the yield of the function
159 was still zero ("insufficient space for captured substrings"). This cannot
160 be totally fixed without adding another stack variable, which seems a lot
161 of expense for a edge case. However, I have improved the situation in cases
162 such as /(a)(b)x|abc/ matched against "abc", where the return code
163 indicates that fewer than the maximum number of slots in the ovector have
164 been set.
165
166 3. Related to (2) above: when there are more back references in a pattern than
167 slots in the output vector, pcre_exec() uses temporary memory during
168 matching, and copies in the captures as far as possible afterwards. It was
169 using the entire output vector, but this conflicts with the specification
170 that only 2/3 is used for passing back captured substrings. Now it uses
171 only the first 2/3, for compatibility. This is, of course, another edge
172 case.
173
174 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
175 main code base, and can be used by building with --enable-jit. When this is
176 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
177 runtime --no-jit option is given.
178
179 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
180 ovector, the return from the function was zero, implying that there were
181 other matches that did not fit. The correct "exactly full" value is now
182 returned.
183
184 6. If a subpattern that was called recursively or as a subroutine contained
185 (*PRUNE) or any other control that caused it to give a non-standard return,
186 invalid errors such as "Error -26 (nested recursion at the same subject
187 position)" or even infinite loops could occur.
188
189 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
190 computing the minimum length on reaching *ACCEPT, and so ended up with the
191 wrong value of 1 rather than 0. Further investigation indicates that
192 computing a minimum subject length in the presence of *ACCEPT is difficult
193 (think back references, subroutine calls), and so I have changed the code
194 so that no minimum is registered for a pattern that contains *ACCEPT.
195
196 8. If (*THEN) was present in the first (true) branch of a conditional group,
197 it was not handled as intended. [But see 16 below.]
198
199 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
200 Sheri Pierce.
201
202 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
203 the first byte in a match must be "a".
204
205 11. Change 17 for 8.13 increased the recursion depth for patterns like
206 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
207 pattern contains any instances of (*THEN). If it does not, the old
208 optimizations are restored. It would be nice to do this on a per-group
209 basis, but at the moment that is not feasible.
210
211 12. In some environments, the output of pcretest -C is CRLF terminated. This
212 broke RunTest's code that checks for the link size. A single white space
213 character after the value is now allowed for.
214
215 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
216 For "fr", it uses the Windows-specific input and output files.
217
218 14. If (*THEN) appeared in a group that was called recursively or as a
219 subroutine, it did not work as intended. [But see next item.]
220
221 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
222 pattern fragments (but not containing any | characters). If A and B are
223 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
224 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
225 D. In other words, Perl considers parentheses that do not contain any |
226 characters to be part of a surrounding alternative, whereas PCRE was
227 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
228 differently. PCRE now behaves in the same way as Perl, except in the case
229 of subroutine/recursion calls such as (?1) which have in any case always
230 been different (but PCRE had them first :-).
231
232 16. Related to 15 above: Perl does not treat the | in a conditional group as
233 creating alternatives. Such a group is treated in the same way as an
234 ordinary group without any | characters when processing (*THEN). PCRE has
235 been changed to match Perl's behaviour.
236
237 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
238 RunGrepTest script failed.
239
240 18. Change 22 for version 13 caused atomic groups to use more stack. This is
241 inevitable for groups that contain captures, but it can lead to a lot of
242 stack use in large patterns. The old behaviour has been restored for atomic
243 groups that do not contain any capturing parentheses.
244
245 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
246 suppress the check for a minimum subject length at run time. (If it was
247 given to pcre_exec() or pcre_dfa_exec() it did work.)
248
249 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
250 fail to work when decoding hex characters in data strings in EBCDIC
251 environments.
252
253 21. It appears that in at least one Mac OS environment, the isxdigit() function
254 is implemented as a macro that evaluates to its argument more than once,
255 contravening the C 90 Standard (I haven't checked a later standard). There
256 was an instance in pcretest which caused it to go wrong when processing
257 \x{...} escapes in subject strings. The has been rewritten to avoid using
258 things like p++ in the argument of isxdigit().
259
260
261 Version 8.13 16-Aug-2011
262 ------------------------
263
264 1. The Unicode data tables have been updated to Unicode 6.0.0.
265
266 2. Two minor typos in pcre_internal.h have been fixed.
267
268 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
269 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
270 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
271
272 4. There were a number of related bugs in the code for matching backrefences
273 caselessly in UTF-8 mode when codes for the characters concerned were
274 different numbers of bytes. For example, U+023A and U+2C65 are an upper
275 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
276 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
277 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
278 2-byte code at the end of the subject (it thought there wasn't enough data
279 left).
280
281 5. Comprehensive information about what went wrong is now returned by
282 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
283 as the output vector has at least 2 elements. The offset of the start of
284 the failing character and a reason code are placed in the vector.
285
286 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
287 now returned is for the first byte of the failing character, instead of the
288 last byte inspected. This is an incompatible change, but I hope it is small
289 enough not to be a problem. It makes the returned offset consistent with
290 pcre_exec() and pcre_dfa_exec().
291
292 7. pcretest now gives a text phrase as well as the error number when
293 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
294 failure, the offset and reason code are output.
295
296 8. When \R was used with a maximizing quantifier it failed to skip backwards
297 over a \r\n pair if the subsequent match failed. Instead, it just skipped
298 back over a single character (\n). This seems wrong (because it treated the
299 two characters as a single entity when going forwards), conflicts with the
300 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
301 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
302 has been changed.
303
304 9. Some internal refactoring has changed the processing so that the handling
305 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
306 time (the PCRE_DOTALL option was changed this way some time ago: version
307 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
308 which was always a bit of a fudge. It also means that there is one less
309 argument for the match() function, which reduces its stack requirements
310 slightly. This change also fixes an incompatibility with Perl: the pattern
311 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
312
313 10. More internal refactoring has drastically reduced the number of recursive
314 calls to match() for possessively repeated groups such as (abc)++ when
315 using pcre_exec().
316
317 11. While implementing 10, a number of bugs in the handling of groups were
318 discovered and fixed:
319
320 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
321 (a|)*(?1) gave a compile-time internal error.
322 ((a|)+)+ did not notice that the outer group could match an empty string.
323 (^a|^)+ was not marked as anchored.
324 (.*a|.*)+ was not marked as matching at start or after a newline.
325
326 12. Yet more internal refactoring has removed another argument from the match()
327 function. Special calls to this function are now indicated by setting a
328 value in a variable in the "match data" data block.
329
330 13. Be more explicit in pcre_study() instead of relying on "default" for
331 opcodes that mean there is no starting character; this means that when new
332 ones are added and accidentally left out of pcre_study(), testing should
333 pick them up.
334
335 14. The -s option of pcretest has been documented for ages as being an old
336 synonym of -m (show memory usage). I have changed it to mean "force study
337 for every regex", that is, assume /S for every regex. This is similar to -i
338 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
339 using it. It makes it easier to run collections of tests with and without
340 study enabled, and thereby test pcre_study() more easily. All the standard
341 tests are now run with and without -s (but some patterns can be marked as
342 "never study" - see 20 below).
343
344 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
345 restoration of the capturing data to the outer values was not happening
346 correctly.
347
348 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
349 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
350 pattern had matched an empty string, and so incorrectly returned a no
351 match.
352
353 17. There was optimizing code for the last branch of non-capturing parentheses,
354 and also for the obeyed branch of a conditional subexpression, which used
355 tail recursion to cut down on stack usage. Unfortunately, now that there is
356 the possibility of (*THEN) occurring in these branches, tail recursion is
357 no longer possible because the return has to be checked for (*THEN). These
358 two optimizations have therefore been removed. [But see 8.20/11 above.]
359
360 18. If a pattern containing \R was studied, it was assumed that \R always
361 matched two bytes, thus causing the minimum subject length to be
362 incorrectly computed because \R can also match just one byte.
363
364 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
365 was incorrectly computed.
366
367 20. If /S is present twice on a test pattern in pcretest input, it now
368 *disables* studying, thereby overriding the use of -s on the command line
369 (see 14 above). This is necessary for one or two tests to keep the output
370 identical in both cases.
371
372 21. When (*ACCEPT) was used in an assertion that matched an empty string and
373 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
374
375 22. When an atomic group that contained a capturing parenthesis was
376 successfully matched, but the branch in which it appeared failed, the
377 capturing was not being forgotten if a higher numbered group was later
378 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
379 group 1 to "a", when in fact it should be unset. This applied to multi-
380 branched capturing and non-capturing groups, repeated or not, and also to
381 positive assertions (capturing in negative assertions does not happen
382 in PCRE) and also to nested atomic groups.
383
384 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
385 subject after a captured substring, to make it easier to tell which of a
386 number of identical substrings has been captured.
387
388 24. The way atomic groups are processed by pcre_exec() has been changed so that
389 if they are repeated, backtracking one repetition now resets captured
390 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
391 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
392 "aaa". Previously, it would have been "a". As part of this code
393 refactoring, the way recursive calls are handled has also been changed.
394
395 25. If an assertion condition captured any substrings, they were not passed
396 back unless some other capturing happened later. For example, if
397 (?(?=(a))a) was matched against "a", no capturing was returned.
398
399 26. When studying a pattern that contained subroutine calls or assertions,
400 the code for finding the minimum length of a possible match was handling
401 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
402 group 1 called group 2 while simultaneously a separate group 2 called group
403 1). A stack overflow occurred in this case. I have fixed this by limiting
404 the recursion depth to 10.
405
406 27. Updated RunTest.bat in the distribution to the version supplied by Tom
407 Fortmann. This supports explicit test numbers on the command line, and has
408 argument validation and error reporting.
409
410 28. An instance of \X with an unlimited repeat could fail if at any point the
411 first character it looked at was a mark character.
412
413 29. Some minor code refactoring concerning Unicode properties and scripts
414 should reduce the stack requirement of match() slightly.
415
416 30. Added the '=' option to pcretest to check the setting of unused capturing
417 slots at the end of the pattern, which are documented as being -1, but are
418 not included in the return count.
419
420 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
421 compiled something random. Now it gives a compile-time error (as does
422 Perl).
423
424 32. A *MARK encountered during the processing of a positive assertion is now
425 recorded and passed back (compatible with Perl).
426
427 33. If --only-matching or --colour was set on a pcregrep call whose pattern
428 had alternative anchored branches, the search for a second match in a line
429 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
430 matched the line "0102" twice. The same bug affected patterns that started
431 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
432 twice.
433
434 34. Previously, PCRE did not allow quantification of assertions. However, Perl
435 does, and because of capturing effects, quantifying parenthesized
436 assertions may at times be useful. Quantifiers are now allowed for
437 parenthesized assertions.
438
439 35. A minor code tidy in pcre_compile() when checking options for \R usage.
440
441 36. \g was being checked for fancy things in a character class, when it should
442 just be a literal "g".
443
444 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
445 appearance of a nested POSIX class supersedes an apparent external class.
446 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
447 unescaped square brackets may also appear as part of class names. For
448 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
449 more like Perl. (But see 8.20/1 above.)
450
451 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
452 was because it thought it was \N{name}, which is not supported).
453
454 39. Add minix to OS list not supporting the -S option in pcretest.
455
456 40. PCRE tries to detect cases of infinite recursion at compile time, but it
457 cannot analyze patterns in sufficient detail to catch mutual recursions
458 such as ((?1))((?2)). There is now a runtime test that gives an error if a
459 subgroup is called recursively as a subpattern for a second time at the
460 same position in the subject string. In previous releases this might have
461 been caught by the recursion limit, or it might have run out of stack.
462
463 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
464 happen only once. PCRE was, however incorrectly giving a compile time error
465 "recursive call could loop indefinitely" because it cannot analyze the
466 pattern in sufficient detail. The compile time test no longer happens when
467 PCRE is compiling a conditional subpattern, but actual runaway loops are
468 now caught at runtime (see 40 above).
469
470 42. It seems that Perl allows any characters other than a closing parenthesis
471 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
472 has been changed to be the same.
473
474 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
475 as not to get warnings when autogen.sh is called. Also changed
476 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
477
478 44. To help people who use pcregrep to scan files containing exceedingly long
479 lines, the following changes have been made:
480
481 (a) The default value of the buffer size parameter has been increased from
482 8K to 20K. (The actual buffer used is three times this size.)
483
484 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
485 PCRE is built.
486
487 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
488 to be set at run time.
489
490 (d) Numerical values in pcregrep options can be followed by K or M, for
491 example --buffer-size=50K.
492
493 (e) If a line being scanned overflows pcregrep's buffer, an error is now
494 given and the return code is set to 2.
495
496 45. Add a pointer to the latest mark to the callout data block.
497
498 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
499 partial match of an empty string instead of no match. This was specific to
500 the use of ".".
501
502 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
503 complete match instead of a partial match. This bug was dependent on both
504 the PCRE_UTF8 and PCRE_DOTALL options being set.
505
506 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
507 starting byte set, because \b was not being ignored.
508
509
510 Version 8.12 15-Jan-2011
511 ------------------------
512
513 1. Fixed some typos in the markup of the man pages, and wrote a script that
514 checks for such things as part of the documentation building process.
515
516 2. On a big-endian 64-bit system, pcregrep did not correctly process the
517 --match-limit and --recursion-limit options (added for 8.11). In
518 particular, this made one of the standard tests fail. (The integer value
519 went into the wrong half of a long int.)
520
521 3. If the --colour option was given to pcregrep with -v (invert match), it
522 did strange things, either producing crazy output, or crashing. It should,
523 of course, ignore a request for colour when reporting lines that do not
524 match.
525
526 4. Another pcregrep bug caused similar problems if --colour was specified with
527 -M (multiline) and the pattern match finished with a line ending.
528
529 5. In pcregrep, when a pattern that ended with a literal newline sequence was
530 matched in multiline mode, the following line was shown as part of the
531 match. This seems wrong, so I have changed it.
532
533 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
534 the check for further matches in the same line (so they could be coloured)
535 to overrun the end of the current line. If another match was found, it was
536 incorrectly shown (and then shown again when found in the next line).
537
538 7. If pcregrep was compiled under Windows, there was a reference to the
539 function pcregrep_exit() before it was defined. I am assuming this was
540 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
541 reported by a user. I've moved the definition above the reference.
542
543
544 Version 8.11 10-Dec-2010
545 ------------------------
546
547 1. (*THEN) was not working properly if there were untried alternatives prior
548 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
549 backtracked to try for "b" instead of moving to the next alternative branch
550 at the same level (in this case, to look for "c"). The Perl documentation
551 is clear that when (*THEN) is backtracked onto, it goes to the "next
552 alternative in the innermost enclosing group".
553
554 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
555 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
556 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
557 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
558 (*THEN).
559
560 3. If \s appeared in a character class, it removed the VT character from
561 the class, even if it had been included by some previous item, for example
562 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
563 of \s, but is part of the POSIX "space" class.)
564
565 4. A partial match never returns an empty string (because you can always
566 match an empty string at the end of the subject); however the checking for
567 an empty string was starting at the "start of match" point. This has been
568 changed to the "earliest inspected character" point, because the returned
569 data for a partial match starts at this character. This means that, for
570 example, /(?<=abc)def/ gives a partial match for the subject "abc"
571 (previously it gave "no match").
572
573 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
574 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
575 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
576 has an implication that the given string is incomplete (because a partial
577 match is preferred over a full match). For this reason, these items now
578 give a partial match in this situation. [Aside: previously, the one case
579 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
580 match rather than a full match, which was wrong by the old rules, but is
581 now correct.]
582
583 6. There was a bug in the handling of #-introduced comments, recognized when
584 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
585 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
586 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
587 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
588 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
589 places in pcre_compile().
590
591 7. Related to (6) above, when pcre_compile() was skipping #-introduced
592 comments when looking ahead for named forward references to subpatterns,
593 the only newline sequence it recognized was NL. It now handles newlines
594 according to the set newline convention.
595
596 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
597 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
598 cater for a lack of strerror(). These oversights have been fixed.
599
600 9. Added --match-limit and --recursion-limit to pcregrep.
601
602 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
603
604 11. When the -o option was used, pcregrep was setting a return code of 1, even
605 when matches were found, and --line-buffered was not being honoured.
606
607 12. Added an optional parentheses number to the -o and --only-matching options
608 of pcregrep.
609
610 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
611 can match an empty string. The code to do it in pcretest and pcredemo
612 needed fixing:
613
614 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
615 only one byte after an empty string match just before CRLF (this case
616 just got forgotten; "any" and "anycrlf" were OK).
617
618 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
619 mode when an empty string match preceded an ASCII character followed by
620 a non-ASCII character. (The code for advancing by one character rather
621 than one byte was nonsense.)
622
623 (c) The pcredemo.c sample program did not have any code at all to handle
624 the cases when CRLF is a valid newline sequence.
625
626 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
627 as a starting offset was within the subject string. There is now a new
628 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
629 negative or greater than the length of the string. In order to test this,
630 pcretest is extended to allow the setting of negative starting offsets.
631
632 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
633 starting offset points to the beginning of a UTF-8 character was
634 unnecessarily clumsy. I tidied it up.
635
636 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
637 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
638
639 17. Nobody had reported that the --include_dir option, which was added in
640 release 7.7 should have been called --include-dir (hyphen, not underscore)
641 for compatibility with GNU grep. I have changed it to --include-dir, but
642 left --include_dir as an undocumented synonym, and the same for
643 --exclude-dir, though that is not available in GNU grep, at least as of
644 release 2.5.4.
645
646 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
647 characters from a string of bytes) have been redefined so as not to use
648 loops, in order to improve performance in some environments. At the same
649 time, I abstracted some of the common code into auxiliary macros to save
650 repetition (this should not affect the compiled code).
651
652 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
653 compile-time error is now given if \c is not followed by an ASCII
654 character, that is, a byte less than 128. (In EBCDIC mode, the code is
655 different, and any byte value is allowed.)
656
657 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
658 START_OPTIMIZE option, which is now allowed at compile time - but just
659 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
660 to pcregrep and other applications that have no direct access to PCRE
661 options. The new /Y option in pcretest sets this option when calling
662 pcre_compile().
663
664 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
665 back references. Groups containing recursive back references were forced to
666 be atomic by that change, but in the case of named groups, the amount of
667 memory required was incorrectly computed, leading to "Failed: internal
668 error: code overflow". This has been fixed.
669
670 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
671 pcretest.c, to avoid build problems in some Borland environments.
672
673
674 Version 8.10 25-Jun-2010
675 ------------------------
676
677 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
678 THEN.
679
680 2. (*ACCEPT) was not working when inside an atomic group.
681
682 3. Inside a character class, \B is treated as a literal by default, but
683 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
684 causes the error). The code is unchanged, but I tidied the documentation.
685
686 4. Inside a character class, PCRE always treated \R and \X as literals,
687 whereas Perl faults them if its -w option is set. I have changed PCRE so
688 that it faults them when PCRE_EXTRA is set.
689
690 5. Added support for \N, which always matches any character other than
691 newline. (It is the same as "." when PCRE_DOTALL is not set.)
692
693 6. When compiling pcregrep with newer versions of gcc which may have
694 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
695 declared with attribute warn_unused_result" were given. Just casting the
696 result to (void) does not stop the warnings; a more elaborate fudge is
697 needed. I've used a macro to implement this.
698
699 7. Minor change to pcretest.c to avoid a compiler warning.
700
701 8. Added four artifical Unicode properties to help with an option to make
702 \s etc use properties (see next item). The new properties are: Xan
703 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
704
705 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
706 use Unicode properties. (*UCP) at the start of a pattern can be used to set
707 this option. Modified pcretest to add /W to test this facility. Added
708 REG_UCP to make it available via the POSIX interface.
709
710 10. Added --line-buffered to pcregrep.
711
712 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
713 studied, and the match started with a letter with a code point greater than
714 127 whose first byte was different to the first byte of the other case of
715 the letter, the other case of this starting letter was not recognized
716 (#976).
717
718 12. If a pattern that was studied started with a repeated Unicode property
719 test, for example, \p{Nd}+, there was the theoretical possibility of
720 setting up an incorrect bitmap of starting bytes, but fortunately it could
721 not have actually happened in practice until change 8 above was made (it
722 added property types that matched character-matching opcodes).
723
724 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
725 possible starting bytes for non-anchored patterns.
726
727 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
728 \R, and also a number of cases that involve Unicode properties, both
729 explicit and implicit when PCRE_UCP is set.
730
731 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
732 input, it could crash or give wrong results if characters with values
733 greater than 0xc0 were present in the subject string. (Detail: it assumed
734 UTF-8 input when processing these items.)
735
736 16. Added a lot of (int) casts to avoid compiler warnings in systems where
737 size_t is 64-bit (#991).
738
739 17. Added a check for running out of memory when PCRE is compiled with
740 --disable-stack-for-recursion (#990).
741
742 18. If the last data line in a file for pcretest does not have a newline on
743 the end, a newline was missing in the output.
744
745 19. The default pcre_chartables.c file recognizes only ASCII characters (values
746 less than 128) in its various bitmaps. However, there is a facility for
747 generating tables according to the current locale when PCRE is compiled. It
748 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
749 characters, are recognized by isspace() and therefore were getting set in
750 these tables, and indeed these tables seem to approximate to ISO 8859. This
751 caused a problem in UTF-8 mode when pcre_study() was used to create a list
752 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
753 which of course cannot start UTF-8 characters. I have changed the code so
754 that only real ASCII characters (less than 128) and the correct starting
755 bytes for UTF-8 encodings are set for characters greater than 127 when in
756 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
757 altogether.)
758
759 20. Added the /T option to pcretest so as to be able to run tests with non-
760 standard character tables, thus making it possible to include the tests
761 used for 19 above in the standard set of tests.
762
763 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
764 reference to a subpattern the other side of a comment that contains an
765 opening parenthesis caused either an internal compiling error, or a
766 reference to the wrong subpattern.
767
768
769 Version 8.02 19-Mar-2010
770 ------------------------
771
772 1. The Unicode data tables have been updated to Unicode 5.2.0.
773
774 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
775 configured.
776
777 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
778 original author of that file, following a query about its status.
779
780 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
781 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
782
783 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
784 quantifier applied to a forward-referencing subroutine call, could compile
785 incorrect code or give the error "internal error: previously-checked
786 referenced subpattern not found".
787
788 6. Both MS Visual Studio and Symbian OS have problems with initializing
789 variables to point to external functions. For these systems, therefore,
790 pcre_malloc etc. are now initialized to local functions that call the
791 relevant global functions.
792
793 7. There were two entries missing in the vectors called coptable and poptable
794 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
795 I've fixed the data, and added a kludgy way of testing at compile time that
796 the lengths are correct (equal to the number of opcodes).
797
798 8. Following on from 7, I added a similar kludge to check the length of the
799 eint vector in pcreposix.c.
800
801 9. Error texts for pcre_compile() are held as one long string to avoid too
802 much relocation at load time. To find a text, the string is searched,
803 counting zeros. There was no check for running off the end of the string,
804 which could happen if a new error number was added without updating the
805 string.
806
807 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
808
809 11. \K was not working if it appeared in an atomic group or in a group that
810 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
811 \K is "not well defined" if used in an assertion. PCRE now accepts it if
812 the assertion is positive, but not if it is negative.
813
814 12. Change 11 fortuitously reduced the size of the stack frame used in the
815 "match()" function of pcre_exec.c by one pointer. Forthcoming
816 implementation of support for (*MARK) will need an extra pointer on the
817 stack; I have reserved it now, so that the stack frame size does not
818 decrease.
819
820 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
821 item in branch that calls a recursion is a subroutine call - as in the
822 second branch in the above example - was incorrectly given the compile-
823 time error "recursive call could loop indefinitely" because pcre_compile()
824 was not correctly checking the subroutine for matching a non-empty string.
825
826 14. The checks for overrunning compiling workspace could trigger after an
827 overrun had occurred. This is a "should never occur" error, but it can be
828 triggered by pathological patterns such as hundreds of nested parentheses.
829 The checks now trigger 100 bytes before the end of the workspace.
830
831 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
832
833
834 Version 8.01 19-Jan-2010
835 ------------------------
836
837 1. If a pattern contained a conditional subpattern with only one branch (in
838 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
839 computed the wrong minimum data length (which is of course zero for such
840 subpatterns). This could cause incorrect "no match" results.
841
842 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
843 the pattern is reset in the first branch, pcre_compile() failed with
844 "internal error: code overflow at offset...". This happened only when
845 the reset was to the original external option setting. (An optimization
846 abstracts leading options settings into an external setting, which was the
847 cause of this.)
848
849 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
850 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
851 assertion pattern did not match (meaning that the assertion was true), it
852 was incorrectly treated as false if the SKIP had been reached during the
853 matching. This also applied to assertions used as conditions.
854
855 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
856 assertion subpattern, including such a pattern used as a condition,
857 unpredictable results occurred, instead of the error return
858 PCRE_ERROR_DFA_UITEM.
859
860 5. The C++ GlobalReplace function was not working like Perl for the special
861 situation when an empty string is matched. It now does the fancy magic
862 stuff that is necessary.
863
864 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
865 removed. (These were left over from very, very early versions of PCRE.)
866
867 7. Some cosmetic changes to the code to make life easier when compiling it
868 as part of something else:
869
870 (a) Change DEBUG to PCRE_DEBUG.
871
872 (b) In pcre_compile(), rename the member of the "branch_chain" structure
873 called "current" as "current_branch", to prevent a collision with the
874 Linux macro when compiled as a kernel module.
875
876 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
877 prevent a collision with the Linux macro when compiled as a kernel
878 module.
879
880 8. In pcre_compile() there are some checks for integer overflows that used to
881 cast potentially large values to (double). This has been changed to that
882 when building, a check for int64_t is made, and if it is found, it is used
883 instead, thus avoiding the use of floating point arithmetic. (There is no
884 other use of FP in PCRE.) If int64_t is not found, the fallback is to
885 double.
886
887 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
888 2005 (difference between two addresses compared to an unsigned value).
889
890 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
891 custom one, because of the following reported problem in Windows:
892
893 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
894 under Win32.
895 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
896 therefore missing the function definition.
897 - The compiler thus generates a "C" signature for the test function.
898 - The linker fails to find the "C" function.
899 - PCRE fails to configure if asked to do so against libbz2.
900
901 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
902 messages were output:
903
904 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
905 rerunning libtoolize, to keep the correct libtool macros in-tree.
906 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
907
908 I have done both of these things.
909
910 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
911 most of the time, it *can* run out if it is given a pattern that contains a
912 runaway infinite recursion. I updated the discussion in the pcrestack man
913 page.
914
915 13. Now that we have gone to the x.xx style of version numbers, the minor
916 version may start with zero. Using 08 or 09 is a bad idea because users
917 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
918 interpreted as invalid octal numbers. I've updated the previous comment in
919 configure.ac, and also added a check that gives an error if 08 or 09 are
920 used.
921
922 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
923 causing partial matching to fail when the end of the subject matched \W
924 in a UTF-8 pattern where \W was quantified with a minimum of 3.
925
926 15. There were some discrepancies between the declarations in pcre_internal.h
927 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
928 their definitions. The declarations used "const uschar *" and the
929 definitions used USPTR. Even though USPTR is normally defined as "const
930 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
931 reported that: "This difference in casting confuses some C++ compilers, for
932 example, SunCC recognizes above declarations as different functions and
933 generates broken code for hbpcre." I have changed the declarations to use
934 USPTR.
935
936 16. GNU libtool is named differently on some systems. The autogen.sh script now
937 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
938 (FreeBSD).
939
940 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
941 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
942 comment: "Figure out how to create a longlong from a string: strtoll and
943 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
944 instance, but it only takes 2 args instead of 3!"
945
946 18. A subtle bug concerned with back references has been fixed by a change of
947 specification, with a corresponding code fix. A pattern such as
948 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
949 refers, was giving matches when it shouldn't. For example, xa=xaaa would
950 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
951 same bug. Such groups have to be quantified to be useful, or contained
952 inside another quantified group. (If there's no repetition, the reference
953 can never match.) The problem arises because, having left the group and
954 moved on to the rest of the pattern, a later failure that backtracks into
955 the group uses the captured value from the final iteration of the group
956 rather than the correct earlier one. I have fixed this in PCRE by forcing
957 any group that contains a reference to itself to be an atomic group; that
958 is, there cannot be any backtracking into it once it has completed. This is
959 similar to recursive and subroutine calls.
960
961
962 Version 8.00 19-Oct-09
963 ----------------------
964
965 1. The table for translating pcre_compile() error codes into POSIX error codes
966 was out-of-date, and there was no check on the pcre_compile() error code
967 being within the table. This could lead to an OK return being given in
968 error.
969
970 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
971 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
972 in a Windows environment.
973
974 3. The pcregrep --count option prints the count for each file even when it is
975 zero, as does GNU grep. However, pcregrep was also printing all files when
976 --files-with-matches was added. Now, when both options are given, it prints
977 counts only for those files that have at least one match. (GNU grep just
978 prints the file name in this circumstance, but including the count seems
979 more useful - otherwise, why use --count?) Also ensured that the
980 combination -clh just lists non-zero counts, with no names.
981
982 4. The long form of the pcregrep -F option was incorrectly implemented as
983 --fixed_strings instead of --fixed-strings. This is an incompatible change,
984 but it seems right to fix it, and I didn't think it was worth preserving
985 the old behaviour.
986
987 5. The command line items --regex=pattern and --regexp=pattern were not
988 recognized by pcregrep, which required --regex pattern or --regexp pattern
989 (with a space rather than an '='). The man page documented the '=' forms,
990 which are compatible with GNU grep; these now work.
991
992 6. No libpcreposix.pc file was created for pkg-config; there was just
993 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
994
995 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
996 when UCP support is not needed, by modifying the Python script that
997 generates it from Unicode data files. This should not matter if the module
998 is correctly used as a library, but I received one complaint about 50K of
999 unwanted data. My guess is that the person linked everything into his
1000 program rather than using a library. Anyway, it does no harm.
1001
1002 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1003 was a minimum greater than 1 for a wide character in a possessive
1004 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1005 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1006 character. Chaos in the form of incorrect output or a compiling loop could
1007 result.
1008
1009 9. The restrictions on what a pattern can contain when partial matching is
1010 requested for pcre_exec() have been removed. All patterns can now be
1011 partially matched by this function. In addition, if there are at least two
1012 slots in the offset vector, the offset of the earliest inspected character
1013 for the match and the offset of the end of the subject are set in them when
1014 PCRE_ERROR_PARTIAL is returned.
1015
1016 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1017 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1018 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1019 and may be more useful for multi-segment matching.
1020
1021 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1022 used to be given if ever the end of the subject was reached; now it is
1023 given only if matching could not proceed because another character was
1024 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1025 string "Z", which now yields "no match" instead of "partial match". In the
1026 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1027 final character ended with (*FAIL).
1028
1029 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1030 if the pattern had a "must contain" character that was already found in the
1031 earlier partial match, unless partial matching was again requested. For
1032 example, with the pattern /dog.(body)?/, the "must contain" character is
1033 "g". If the first part-match was for the string "dog", restarting with
1034 "sbody" failed. This bug has been fixed.
1035
1036 13. The string returned by pcre_dfa_exec() after a partial match has been
1037 changed so that it starts at the first inspected character rather than the
1038 first character of the match. This makes a difference only if the pattern
1039 starts with a lookbehind assertion or \b or \B (\K is not supported by
1040 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1041 matching functions compatible, and I think it's the right thing to do.
1042
1043 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1044 so that the demonstration program is easily available in environments where
1045 PCRE has not been installed from source.
1046
1047 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1048 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1049 library.
1050
1051 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1052 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1053 is not the first non-POSIX option to be added. Clearly some people find
1054 these options useful.
1055
1056 17. If a caller to the POSIX matching function regexec() passes a non-zero
1057 value for nmatch with a NULL value for pmatch, the value of
1058 nmatch is forced to zero.
1059
1060 18. RunGrepTest did not have a test for the availability of the -u option of
1061 the diff command, as RunTest does. It now checks in the same way as
1062 RunTest, and also checks for the -b option.
1063
1064 19. If an odd number of negated classes containing just a single character
1065 interposed, within parentheses, between a forward reference to a named
1066 subpattern and the definition of the subpattern, compilation crashed with
1067 an internal error, complaining that it could not find the referenced
1068 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1069 [The bug was that it was starting one character too far in when skipping
1070 over the character class, thus treating the ] as data rather than
1071 terminating the class. This meant it could skip too much.]
1072
1073 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1074 /g option in pcretest when the pattern contains \K, which makes it possible
1075 to have an empty string match not at the start, even when the pattern is
1076 anchored. Updated pcretest and pcredemo to use this option.
1077
1078 21. If the maximum number of capturing subpatterns in a recursion was greater
1079 than the maximum at the outer level, the higher number was returned, but
1080 with unset values at the outer level. The correct (outer level) value is
1081 now given.
1082
1083 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1084 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1085 make it do so. The string so far is captured, making this feature
1086 compatible with Perl.
1087
1088 23. The tests have been re-organized, adding tests 11 and 12, to make it
1089 possible to check the Perl 5.10 features against Perl 5.10.
1090
1091 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1092 pattern matches a fixed length string. PCRE did not allow this; now it
1093 does. Neither allows recursion.
1094
1095 25. I finally figured out how to implement a request to provide the minimum
1096 length of subject string that was needed in order to match a given pattern.
1097 (It was back references and recursion that I had previously got hung up
1098 on.) This code has now been added to pcre_study(); it finds a lower bound
1099 to the length of subject needed. It is not necessarily the greatest lower
1100 bound, but using it to avoid searching strings that are too short does give
1101 some useful speed-ups. The value is available to calling programs via
1102 pcre_fullinfo().
1103
1104 26. While implementing 25, I discovered to my embarrassment that pcretest had
1105 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1106 study optimizations had never been tested with that matching function.
1107 Oops. What is worse, even when it was passed study data, there was a bug in
1108 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1109 were also very few tests of studied patterns with pcre_dfa_exec().
1110
1111 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1112 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1113 on the other side of the coin, they are no longer allowed to have different
1114 names, because these cannot be distinguished in PCRE, and this has caused
1115 confusion. (This is a difference from Perl.)
1116
1117 28. When duplicate subpattern names are present (necessarily with different
1118 numbers, as required by 27 above), and a test is made by name in a
1119 conditional pattern, either for a subpattern having been matched, or for
1120 recursion in such a pattern, all the associated numbered subpatterns are
1121 tested, and the overall condition is true if the condition is true for any
1122 one of them. This is the way Perl works, and is also more like the way
1123 testing by number works.
1124
1125
1126 Version 7.9 11-Apr-09
1127 ---------------------
1128
1129 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1130 (pcretest), all targets were linked against these libraries. This included
1131 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1132 libraries. This caused unwanted dependencies to be created. This problem
1133 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1134 pcretest is linked with readline.
1135
1136 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1137 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1138 moved outside it again, because FALSE and TRUE are already defined in AIX,
1139 but BOOL is not.
1140
1141 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1142 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1143
1144 4. The pcregrep documentation said spaces were inserted as well as colons (or
1145 hyphens) following file names and line numbers when outputting matching
1146 lines. This is not true; no spaces are inserted. I have also clarified the
1147 wording for the --colour (or --color) option.
1148
1149 5. In pcregrep, when --colour was used with -o, the list of matching strings
1150 was not coloured; this is different to GNU grep, so I have changed it to be
1151 the same.
1152
1153 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1154 each matching line was coloured. Now it goes on to look for further matches
1155 of any of the test patterns, which is the same behaviour as GNU grep.
1156
1157 7. A pattern that could match an empty string could cause pcregrep to loop; it
1158 doesn't make sense to accept an empty string match in pcregrep, so I have
1159 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1160 seems to be how GNU grep behaves.
1161
1162 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1163 start or after a newline", because the conditional assertion was not being
1164 correctly handled. The rule now is that both the assertion and what follows
1165 in the first alternative must satisfy the test.
1166
1167 9. If auto-callout was enabled in a pattern with a conditional group whose
1168 condition was an assertion, PCRE could crash during matching, both with
1169 pcre_exec() and pcre_dfa_exec().
1170
1171 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1172 used for matching.
1173
1174 11. Unicode property support in character classes was not working for
1175 characters (bytes) greater than 127 when not in UTF-8 mode.
1176
1177 12. Added the -M command line option to pcretest.
1178
1179 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1180
1181 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1182
1183 16. Added comments and documentation about mis-use of no_arg in the C++
1184 wrapper.
1185
1186 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1187 from Martin Jerabek that uses macro names for all relevant character and
1188 string constants.
1189
1190 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1191 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1192 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1193 these, but not everybody uses configure.
1194
1195 19. A conditional group that had only one branch was not being correctly
1196 recognized as an item that could match an empty string. This meant that an
1197 enclosing group might also not be so recognized, causing infinite looping
1198 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1199 with the subject "ab", where knowledge that the repeated group can match
1200 nothing is needed in order to break the loop.
1201
1202 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1203 exec(), but without supplying a callout function, matching went wrong.
1204
1205 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1206 leak if the size of the offset vector was greater than 30. When the vector
1207 is smaller, the saved offsets during recursion go onto a local stack
1208 vector, but for larger vectors malloc() is used. It was failing to free
1209 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1210 error, in fact).
1211
1212 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1213 heapframe that is used only when UTF-8 support is enabled. This caused no
1214 problem, but was untidy.
1215
1216 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1217 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1218 included within another project.
1219
1220 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1221 slightly modified by me:
1222
1223 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1224 not building pcregrep.
1225
1226 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1227 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1228
1229 25. Forward references, both numeric and by name, in patterns that made use of
1230 duplicate group numbers, could behave incorrectly or give incorrect errors,
1231 because when scanning forward to find the reference group, PCRE was not
1232 taking into account the duplicate group numbers. A pattern such as
1233 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1234
1235 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1236 the feature of a custom pointer more persuasive (as requested by a user).
1237
1238 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1239 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1240 pre-defined.
1241
1242 28. Added support for (*UTF8) at the start of a pattern.
1243
1244 29. Arrange for flags added by the "release type" setting in CMake to be shown
1245 in the configuration summary.
1246
1247
1248 Version 7.8 05-Sep-08
1249 ---------------------
1250
1251 1. Replaced UCP searching code with optimized version as implemented for Ad
1252 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1253 stage table and inline lookup instead of a function, giving speed ups of 2
1254 to 5 times on some simple patterns that I tested. Permission was given to
1255 distribute the MultiStage2.py script that generates the tables (it's not in
1256 the tarball, but is in the Subversion repository).
1257
1258 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1259 scripts.
1260
1261 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1262 a group with a zero qualifier. The result of the study could be incorrect,
1263 or the function might crash, depending on the pattern.
1264
1265 4. Caseless matching was not working for non-ASCII characters in back
1266 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1267 It now works when Unicode Property Support is available.
1268
1269 5. In pcretest, an escape such as \x{de} in the data was always generating
1270 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1271 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1272 truncation.
1273
1274 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1275
1276 7. Added two (int) casts to pcregrep when printing the difference of two
1277 pointers, in case they are 64-bit values.
1278
1279 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1280 test 2 if it fails.
1281
1282 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1283 and a #define of that name to empty if it is not externally set. This is to
1284 allow users of MSVC to set it if necessary.
1285
1286 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1287 the convenience functions in the pcre_get.c source file.
1288
1289 11. An option change at the start of a pattern that had top-level alternatives
1290 could cause overwriting and/or a crash. This command provoked a crash in
1291 some environments:
1292
1293 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1294
1295 This potential security problem was recorded as CVE-2008-2371.
1296
1297 12. For a pattern where the match had to start at the beginning or immediately
1298 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1299 pcre_dfa_exec() could read past the end of the passed subject if there was
1300 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1301 pcretest so that it places the subject at the end of its malloc-ed buffer.
1302
1303 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1304 exec() might read past the end of the data buffer in UTF-8 mode.
1305
1306 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1307 the data contained the byte 0x85 as part of a UTF-8 character within its
1308 first line. This applied both to normal and DFA matching.
1309
1310 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1311 /^[^d]*?$/8 failed to match "abc".
1312
1313 16. Added a missing copyright notice to pcrecpp_internal.h.
1314
1315 17. Make it more clear in the documentation that values returned from
1316 pcre_exec() in ovector are byte offsets, not character counts.
1317
1318 18. Tidied a few places to stop certain compilers from issuing warnings.
1319
1320 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1321 supplied by Stefan Weber. I made a further small update for 7.8 because
1322 there is a change of source arrangements: the pcre_searchfuncs.c module is
1323 replaced by pcre_ucd.c.
1324
1325
1326 Version 7.7 07-May-08
1327 ---------------------
1328
1329 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1330 a string to a long long, pretend we don't even have a long long." This is
1331 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1332
1333 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1334 pre-7.6 versions, which defined a global no_arg variable instead of putting
1335 it in the RE class. (See also #8 below.)
1336
1337 3. Remove a line of dead code, identified by coverity and reported by Nuno
1338 Lopes.
1339
1340 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1341
1342 (1) The include/exclude patterns were being applied to the whole pathnames
1343 of files, instead of just to the final components.
1344
1345 (2) If there was more than one level of directory, the subdirectories were
1346 skipped unless they satisfied the include/exclude conditions. This is
1347 inconsistent with GNU grep (and could even be seen as contrary to the
1348 pcregrep specification - which I improved to make it absolutely clear).
1349 The action now is always to scan all levels of directory, and just
1350 apply the include/exclude patterns to regular files.
1351
1352 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1353 --exclude_dir in the tests to avoid scanning .svn directories.
1354
1355 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1356 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1357 doesn't support NULs in patterns.
1358
1359 7. Added some missing "const"s to declarations of static tables in
1360 pcre_compile.c and pcre_dfa_exec.c.
1361
1362 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1363 caused by fix #2 above. (Subsequently also a second patch to fix the
1364 first patch. And a third patch - this was a messy problem.)
1365
1366 9. Applied Craig's patch to remove the use of push_back().
1367
1368 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1369 matching function regexec().
1370
1371 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1372 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1373 references. PCRE supports relative numbers with this syntax (I don't think
1374 Oniguruma does).
1375
1376 12. Previously, a group with a zero repeat such as (...){0} was completely
1377 omitted from the compiled regex. However, this means that if the group
1378 was called as a subroutine from elsewhere in the pattern, things went wrong
1379 (an internal error was given). Such groups are now left in the compiled
1380 pattern, with a new opcode that causes them to be skipped at execution
1381 time.
1382
1383 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1384 to the way PCRE behaves:
1385
1386 (a) A lone ] character is dis-allowed (Perl treats it as data).
1387
1388 (b) A back reference to an unmatched subpattern matches an empty string
1389 (Perl fails the current match path).
1390
1391 (c) A data ] in a character class must be notated as \] because if the
1392 first data character in a class is ], it defines an empty class. (In
1393 Perl it is not possible to have an empty class.) The empty class []
1394 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1395 The negative empty class [^] matches any one character, independently
1396 of the DOTALL setting.
1397
1398 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1399 non-existent subpattern following a character class starting with ']' and
1400 containing () gave an internal compiling error instead of "reference to
1401 non-existent subpattern". Fortunately, when the pattern did exist, the
1402 compiled code was correct. (When scanning forwards to check for the
1403 existencd of the subpattern, it was treating the data ']' as terminating
1404 the class, so got the count wrong. When actually compiling, the reference
1405 was subsequently set up correctly.)
1406
1407 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1408 it was being rejected as not supported by pcre_dfa_exec(), even though
1409 other assertions are supported. I have made pcre_dfa_exec() support
1410 (*FAIL).
1411
1412 16. The implementation of 13c above involved the invention of a new opcode,
1413 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1414 cannot be changed at match time, I realized I could make a small
1415 improvement to matching performance by compiling OP_ALLANY instead of
1416 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1417 on the OP_ANY path.
1418
1419 17. Compiling pcretest on Windows with readline support failed without the
1420 following two fixes: (1) Make the unistd.h include conditional on
1421 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1422
1423 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1424 ncurses library to be included for pcretest when ReadLine support is
1425 requested, but also to allow for it to be overridden. This patch came from
1426 Daniel Bergström.
1427
1428 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1429 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1430 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1431 spotting this.
1432
1433
1434 Version 7.6 28-Jan-08
1435 ---------------------
1436
1437 1. A character class containing a very large number of characters with
1438 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1439 overflow.
1440
1441 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1442 HAVE_LONG_LONG is not defined.
1443
1444 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1445 bring them up to date and include new features. This patch includes:
1446
1447 - Fixed PH's badly added libz and libbz2 support.
1448 - Fixed a problem with static linking.
1449 - Added pcredemo. [But later removed - see 7 below.]
1450 - Fixed dftables problem and added an option.
1451 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1452 HAVE_LONG_LONG.
1453 - Added readline support for pcretest.
1454 - Added an listing of the option settings after cmake has run.
1455
1456 4. A user submitted a patch to Makefile that makes it easy to create
1457 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1458 Makefile.am that cause it to include this special target, without
1459 affecting anything else. Note that the same mingw target plus all
1460 the other distribution libraries and programs are now supported
1461 when configuring with CMake (see 6 below) instead of with
1462 Configure/Make.
1463
1464 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1465 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1466 exported in the Windows port". It has not yet been confirmed that the patch
1467 solves the problem, but it does no harm.
1468
1469 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1470 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1471 with CMake, and also correct the comment about stack recursion.
1472
1473 7. Remove the automatic building of pcredemo from the ./configure system and
1474 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1475 of a program that users should build themselves after PCRE is installed, so
1476 building it automatically is not really right. What is more, it gave
1477 trouble in some build environments.
1478
1479 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1480
1481
1482 Version 7.5 10-Jan-08
1483 ---------------------
1484
1485 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1486 values in parens when parsing an RE using the C++ wrapper."
1487
1488 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1489 Characters greater than 255 were excluded from the class instead of being
1490 included.
1491
1492 3. The same bug as (2) above applied to negated POSIX classes such as
1493 [:^space:].
1494
1495 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1496 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1497 I have changed it.
1498
1499 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1500 first named subpattern) and a construct such as (?&a) would reference the
1501 first named subpattern whose name started with "a" (in other words, the
1502 length check was missing). Both these problems are fixed. "Subpattern name
1503 expected" is now given for (?&) (a zero-length name), and this patch also
1504 makes it give the same error for \k'' (previously it complained that that
1505 was a reference to a non-existent subpattern).
1506
1507 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1508 this is right because (?- can be followed by option settings as well as by
1509 digits. I have, however, made the messages clearer.
1510
1511 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1512 than the number used in the conditional) now cause a compile-time error.
1513 This is actually not compatible with Perl, which accepts such patterns, but
1514 treats the conditional as always being FALSE (as PCRE used to), but it
1515 seems to me that giving a diagnostic is better.
1516
1517 8. Change "alphameric" to the more common word "alphanumeric" in comments
1518 and messages.
1519
1520 9. Fix two occurrences of "backslash" in comments that should have been
1521 "backspace".
1522
1523 10. Remove two redundant lines of code that can never be obeyed (their function
1524 was moved elsewhere).
1525
1526 11. The program that makes PCRE's Unicode character property table had a bug
1527 which caused it to generate incorrect table entries for sequences of
1528 characters that have the same character type, but are in different scripts.
1529 It amalgamated them into a single range, with the script of the first of
1530 them. In other words, some characters were in the wrong script. There were
1531 thirteen such cases, affecting characters in the following ranges:
1532
1533 U+002b0 - U+002c1
1534 U+0060c - U+0060d
1535 U+0061e - U+00612
1536 U+0064b - U+0065e
1537 U+0074d - U+0076d
1538 U+01800 - U+01805
1539 U+01d00 - U+01d77
1540 U+01d9b - U+01dbf
1541 U+0200b - U+0200f
1542 U+030fc - U+030fe
1543 U+03260 - U+0327f
1544 U+0fb46 - U+0fbb1
1545 U+10450 - U+1049d
1546
1547 12. The -o option (show only the matching part of a line) for pcregrep was not
1548 compatible with GNU grep in that, if there was more than one match in a
1549 line, it showed only the first of them. It now behaves in the same way as
1550 GNU grep.
1551
1552 13. If the -o and -v options were combined for pcregrep, it printed a blank
1553 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1554 does the same. The return code can be used to tell if there were any
1555 non-matching lines.
1556
1557 14. Added --file-offsets and --line-offsets to pcregrep.
1558
1559 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1560 infinitely looping recursion. The bug was that positive lookaheads were not
1561 being skipped when checking for a possible empty match (negative lookaheads
1562 and both kinds of lookbehind were skipped).
1563
1564 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1565 inclusion of <windows.h> to before rather than after the definition of
1566 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1567
1568 17. Specifying a possessive quantifier with a specific limit for a Unicode
1569 character property caused pcre_compile() to compile bad code, which led at
1570 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1571 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1572 caused the error; without that there was no problem.
1573
1574 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1575
1576 19. Added --enable-pcretest-libreadline.
1577
1578 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1579 RE::GlobalReplace(). As a result, the number of replacements returned was
1580 double what it should be. I removed one of the increments, but Craig sent a
1581 later patch that removed the other one (the right fix) and added unit tests
1582 that check the return values (which was not done before).
1583
1584 21. Several CMake things:
1585
1586 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1587 the names libpcre and libpcreposix, not just pcre and pcreposix.
1588
1589 (2) The above change means that pcretest and pcregrep are now correctly
1590 linked with the newly-built libraries, not previously installed ones.
1591
1592 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1593
1594 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1595 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1596 UTF-8 newline character). The key issue is that the pattern starts .*;
1597 this means that the match must be either at the beginning, or after a
1598 newline. The bug was in the code for advancing after a failed match and
1599 checking that the new position followed a newline. It was not taking
1600 account of UTF-8 characters correctly.
1601
1602 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1603 character classes. PCRE was not treating the sequence [:...:] as a
1604 character class unless the ... were all letters. Perl, however, seems to
1605 allow any characters between [: and :], though of course it rejects as
1606 unknown any "names" that contain non-letters, because all the known class
1607 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1608 for example, whereas PCRE did not - it did not recognize a POSIX character
1609 class. This seemed a bit dangerous, so the code has been changed to be
1610 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1611 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1612 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1613 Perl does, and where it didn't before.
1614
1615 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1616 Windows environments %n is disabled by default.
1617
1618
1619 Version 7.4 21-Sep-07
1620 ---------------------
1621
1622 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1623 means that a class such as [\s] counted as "explicit reference to CR or
1624 LF". That isn't really right - the whole point of the change was to try to
1625 help when there was an actual mention of one of the two characters. So now
1626 the change happens only if \r or \n (or a literal CR or LF) character is
1627 encountered.
1628
1629 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1630 of both had grown to the point where there were only 3 bits left.
1631 Fortunately, there was spare space in the data structure, and so I have
1632 moved the internal flags into a new 16-bit field to free up more option
1633 bits.
1634
1635 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1636 but did not set the internal JCHANGED flag - either of these is enough to
1637 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1638 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1639 start sets both bits.
1640
1641 4. Added options (at build time, compile time, exec time) to change \R from
1642 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1643
1644 5. doc/pcresyntax.html was missing from the distribution.
1645
1646 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1647 compatibility, even though it is no longer used.
1648
1649 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1650 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1651 windows.h file is present (where different names are used). [This was
1652 reversed later after testing - see 16 below.]
1653
1654 8. Changed all #include <config.h> to #include "config.h". There were also
1655 some further <pcre.h> cases that I changed to "pcre.h".
1656
1657 9. When pcregrep was used with the --colour option, it missed the line ending
1658 sequence off the lines that it output.
1659
1660 10. It was pointed out to me that arrays of string pointers cause lots of
1661 relocations when a shared library is dynamically loaded. A technique of
1662 using a single long string with a table of offsets can drastically reduce
1663 these. I have refactored PCRE in four places to do this. The result is
1664 dramatic:
1665
1666 Originally: 290
1667 After changing UCP table: 187
1668 After changing error message table: 43
1669 After changing table of "verbs" 36
1670 After changing table of Posix names 22
1671
1672 Thanks to the folks working on Gregex for glib for this insight.
1673
1674 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1675 unicode-properties was also set.
1676
1677 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1678
1679 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1680 checked only for CRLF.
1681
1682 14. Added casts to pcretest.c to avoid compiler warnings.
1683
1684 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1685
1686 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1687 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1688 entirely. This removes changes made in 7 above.
1689
1690 17. The CMake files have been updated, and there is now more information about
1691 building with CMake in the NON-UNIX-USE document.
1692
1693
1694 Version 7.3 28-Aug-07
1695 ---------------------
1696
1697 1. In the rejigging of the build system that eventually resulted in 7.1, the
1698 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1699 brackets there is not right, since it causes compilers to look for an
1700 installed pcre.h, not the version that is in the source that is being
1701 compiled (which of course may be different). I have changed it back to:
1702
1703 #include "pcre.h"
1704
1705 I have a vague recollection that the change was concerned with compiling in
1706 different directories, but in the new build system, that is taken care of
1707 by the VPATH setting the Makefile.
1708
1709 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1710 when the subject happened to end in the byte 0x85 (e.g. if the last
1711 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1712 characters but of course it shouldn't be taken as a newline when it is part
1713 of another character. The bug was that, for an unlimited repeat of . in
1714 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1715 characters when looking for a newline.
1716
1717 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1718
1719 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1720 in debug output.
1721
1722 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1723 long printing in the pcrecpp unittest when running under MinGW.
1724
1725 6. ESC_K was left out of the EBCDIC table.
1726
1727 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1728 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1729 limit also applies to "virtual nesting" when a pattern is recursive, and in
1730 this case 1000 isn't so big. I have been able to remove this limit at the
1731 expense of backing off one optimization in certain circumstances. Normally,
1732 when pcre_exec() would call its internal match() function recursively and
1733 immediately return the result unconditionally, it uses a "tail recursion"
1734 feature to save stack. However, when a subpattern that can match an empty
1735 string has an unlimited repetition quantifier, it no longer makes this
1736 optimization. That gives it a stack frame in which to save the data for
1737 checking that an empty string has been matched. Previously this was taken
1738 from the 1000-entry workspace that had been reserved. So now there is no
1739 explicit limit, but more stack is used.
1740
1741 8. Applied Daniel's patches to solve problems with the import/export magic
1742 syntax that is required for Windows, and which was going wrong for the
1743 pcreposix and pcrecpp parts of the library. These were overlooked when this
1744 problem was solved for the main library.
1745
1746 9. There were some crude static tests to avoid integer overflow when computing
1747 the size of patterns that contain repeated groups with explicit upper
1748 limits. As the maximum quantifier is 65535, the maximum group length was
1749 set at 30,000 so that the product of these two numbers did not overflow a
1750 32-bit integer. However, it turns out that people want to use groups that
1751 are longer than 30,000 bytes (though not repeat them that many times).
1752 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1753 made it possible to implement the integer overflow checks in a much more
1754 dynamic way, which I have now done. The artificial limitation on group
1755 length has been removed - we now have only the limit on the total length of
1756 the compiled pattern, which depends on the LINK_SIZE setting.
1757
1758 10. Fixed a bug in the documentation for get/copy named substring when
1759 duplicate names are permitted. If none of the named substrings are set, the
1760 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1761 empty string.
1762
1763 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1764 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1765 because the ] is interpreted as the first data character and the
1766 terminating ] is not found. PCRE has been made compatible with Perl in this
1767 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1768 cause memory overwriting.
1769
1770 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1771 string has been matched (to stop an infinite loop). It was not recognizing
1772 a conditional subpattern that could match an empty string if that
1773 subpattern was within another subpattern. For example, it looped when
1774 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1775 condition was not nested. This bug has been fixed.
1776
1777 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1778 past the start of the subject in the presence of bytes with the top bit
1779 set, for example "\x8aBCD".
1780
1781 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1782 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1783
1784 14. Optimized (?!) to (*FAIL).
1785
1786 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1787 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1788 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1789 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1790 does: it's just the validity check that is more restrictive.
1791
1792 16. Inserted checks for integer overflows during escape sequence (backslash)
1793 processing, and also fixed erroneous offset values for syntax errors during
1794 backslash processing.
1795
1796 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1797 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1798
1799 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1800 caused an overrun.
1801
1802 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1803 something other than just ASCII characters) inside a group that had an
1804 unlimited repeat caused a loop at compile time (while checking to see
1805 whether the group could match an empty string).
1806
1807 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1808 [\P{Any}] did so. (Error in the code for printing property names.)
1809
1810 21. An orphan \E inside a character class could cause a crash.
1811
1812 22. A repeated capturing bracket such as (A)? could cause a wild memory
1813 reference during compilation.
1814
1815 23. There are several functions in pcre_compile() that scan along a compiled
1816 expression for various reasons (e.g. to see if it's fixed length for look
1817 behind). There were bugs in these functions when a repeated \p or \P was
1818 present in the pattern. These operators have additional parameters compared
1819 with \d, etc, and these were not being taken into account when moving along
1820 the compiled data. Specifically:
1821
1822 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1823 length.
1824
1825 (b) An item such as \pL+ within a repeated group could cause crashes or
1826 loops.
1827
1828 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1829 "reference to non-existent subpattern" error.
1830
1831 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1832
1833 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1834 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1835
1836 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1837
1838 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1839 character were causing crashes (broken optimization).
1840
1841 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1842 \p or \P) caused a compile-time loop.
1843
1844 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1845 break. For example, the unstudied pattern [\r\n]A does not match the string
1846 "\r\nA" because change 7.0/46 below moves the current point on by two
1847 characters after failing to match at the start. However, the pattern \nA
1848 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1849 the same is true. There doesn't seem any very clean way out of this, but
1850 what I have chosen to do makes the common cases work: PCRE now takes note
1851 of whether there can be an explicit match for \r or \n anywhere in the
1852 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1853 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1854 pattern has explicit CR or LF references.
1855
1856 29. Added (*CR) etc for changing newline setting at start of pattern.
1857
1858
1859 Version 7.2 19-Jun-07
1860 ---------------------
1861
1862 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1863 which is apparently normally available under Windows.
1864
1865 2. Re-jig the pcregrep tests with different newline settings in an attempt
1866 to make them independent of the local environment's newline setting.
1867
1868 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1869
1870 4. Some of the "internals" tests were previously cut out when the link size
1871 was not 2, because the output contained actual offsets. The recent new
1872 "Z" feature of pcretest means that these can be cut out, making the tests
1873 usable with all link sizes.
1874
1875 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1876 stack recursion. This gives a massive performance boost under BSD, but just
1877 a small improvement under Linux. However, it saves one field in the frame
1878 in all cases.
1879
1880 6. Added more features from the forthcoming Perl 5.10:
1881
1882 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1883 recursion call. It refers to the nth most recently opened parentheses.
1884
1885 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1886 to be opened parentheses.
1887
1888 (c) Conditions that refer to capturing parentheses can be specified
1889 relatively, for example, (?(-2)... or (?(+3)...
1890
1891 (d) \K resets the start of the current match so that everything before
1892 is not part of it.
1893
1894 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1895
1896 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1897 reference syntax.
1898
1899 (g) (?| introduces a group in which the numbering of parentheses in each
1900 alternative starts with the same number.
1901
1902 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1903
1904 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1905 PCRE_INFO_JCHANGED.
1906
1907 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1908 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1909 for detecting groups that can match an empty string.
1910
1911 9. A pattern with a very large number of alternatives (more than several
1912 hundred) was running out of internal workspace during the pre-compile
1913 phase, where pcre_compile() figures out how much memory will be needed. A
1914 bit of new cunning has reduced the workspace needed for groups with
1915 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1916 workspace instead of running out of the 4096 that are available.
1917
1918 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1919
1920 11. Applied patch from Google to remove an optimization that didn't quite work.
1921 The report of the bug said:
1922
1923 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1924 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1925 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1926
1927 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1928 it matched the wrong number of bytes.
1929
1930
1931 Version 7.1 24-Apr-07
1932 ---------------------
1933
1934 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1935 that is more "standard", making use of automake and other Autotools. There
1936 is some re-arrangement of the files and adjustment of comments consequent
1937 on this.
1938
1939 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1940 for recursive directory scanning broke on some systems because the files
1941 are not scanned in any specific order and on different systems the order
1942 was different. A call to "sort" has been inserted into RunGrepTest for the
1943 approprate test as a short-term fix. In the longer term there may be an
1944 alternative.
1945
1946 3. I had an email from Eric Raymond about problems translating some of PCRE's
1947 man pages to HTML (despite the fact that I distribute HTML pages, some
1948 people do their own conversions for various reasons). The problems
1949 concerned the use of low-level troff macros .br and .in. I have therefore
1950 removed all such uses from the man pages (some were redundant, some could
1951 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1952 HTML has been updated to handle .nf/.fi and to complain if it encounters
1953 .br or .in.
1954
1955 4. Updated comments in configure.ac that get placed in config.h.in and also
1956 arranged for config.h to be included in the distribution, with the name
1957 config.h.generic, for the benefit of those who have to compile without
1958 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1959
1960 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1961 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1962 makevp.bat for the current PCRE, using the additional files
1963 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1964
1965 6. A Windows user reported a minor discrepancy with test 2, which turned out
1966 to be caused by a trailing space on an input line that had got lost in his
1967 copy. The trailing space was an accident, so I've just removed it.
1968
1969 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1970 that is needed.
1971
1972 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1973 as "const" (a) because they are and (b) because it helps the PHP
1974 maintainers who have recently made a script to detect big data structures
1975 in the php code that should be moved to the .rodata section. I remembered
1976 to update Builducptable as well, so it won't revert if ucptable.h is ever
1977 re-created.
1978
1979 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1980 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1981 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1982 support is not required. This saves 1.5-2K of code, which is important in
1983 some applications.
1984
1985 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1986 so as not to refer to the tables, even though these functions will never be
1987 called when UTF-8 support is disabled. Otherwise there are problems with a
1988 shared library.
1989
1990 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1991
1992 (a) It was defining its arguments as char * instead of void *.
1993
1994 (b) It was assuming that all moves were upwards in memory; this was true
1995 a long time ago when I wrote it, but is no longer the case.
1996
1997 The emulated memove() is provided for those environments that have neither
1998 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1999 is clearly not the case, as these two bugs were recently reported.
2000
2001 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2002 and Detrail to create the HTML documentation, the .txt form of the man
2003 pages, and it removes trailing spaces from listed files. It also creates
2004 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2005 case, it wraps all the #defines with #ifndefs. This script should be run
2006 before "make dist".
2007
2008 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2009 with Unicode property support.
2010
2011 (a) For a maximizing quantifier, if the two different cases of the
2012 character were of different lengths in their UTF-8 codings (there are
2013 some cases like this - I found 11), and the matching function had to
2014 back up over a mixture of the two cases, it incorrectly assumed they
2015 were both the same length.
2016
2017 (b) When PCRE was configured to use the heap rather than the stack for
2018 recursion during matching, it was not correctly preserving the data for
2019 the other case of a UTF-8 character when checking ahead for a match
2020 while processing a minimizing repeat. If the check also involved
2021 matching a wide character, but failed, corruption could cause an
2022 erroneous result when trying to check for a repeat of the original
2023 character.
2024
2025 13. Some tidying changes to the testing mechanism:
2026
2027 (a) The RunTest script now detects the internal link size and whether there
2028 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2029 values substituted by "configure". (The RunGrepTest script already did
2030 this for UTF-8.) The configure.ac script no longer substitutes the
2031 relevant variables.
2032
2033 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2034 with length and offset values. This means that the output is different
2035 for different internal link sizes. Test 2 is skipped for link sizes
2036 other than 2 because of this, bypassing the problem. Unfortunately,
2037 there was also a test in test 3 (the locale tests) that used /B and
2038 failed for link sizes other than 2. Rather than cut the whole test out,
2039 I have added a new /Z option to pcretest that replaces the length and
2040 offset values with spaces. This is now used to make test 3 independent
2041 of link size. (Test 2 will be tidied up later.)
2042
2043 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2044 segmentation fault instead of returning the appropriate error message.
2045
2046 15. In multiline mode when the newline sequence was set to "any", the pattern
2047 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2048 This doesn't seem right; it now treats the CRLF combination as the line
2049 ending, and so does not match in that case. It's only a pattern such as ^$
2050 that would hit this one: something like ^ABC$ would have failed after \r
2051 and then tried again after \r\n.
2052
2053 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2054 in an attempt to make files that differ only in their line terminators
2055 compare equal. This works on Linux.
2056
2057 17. Under certain error circumstances pcregrep might try to free random memory
2058 as it exited. This is now fixed, thanks to valgrind.
2059
2060 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2061 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2062 was because its rules for how to advance for /g after matching an empty
2063 string at the end of a line did not allow for this case. They now check for
2064 it specially.
2065
2066 20. pcretest is supposed to handle patterns and data of any length, by
2067 extending its buffers when necessary. It was getting this wrong when the
2068 buffer for a data line had to be extended.
2069
2070 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2071 CRLF as a newline sequence.
2072
2073 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2074 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2075 I have nevertheless tidied it up.
2076
2077 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2078
2079 24. Added a man page for pcre-config.
2080
2081
2082 Version 7.0 19-Dec-06
2083 ---------------------
2084
2085 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2086 moving to gcc 4.1.1.
2087
2088 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2089 sys/time.h, which is documented as needed for this function. It doesn't
2090 seem to matter on Linux, but it showed up on some releases of OS X.
2091
2092 3. It seems that there are systems where bytes whose values are greater than
2093 127 match isprint() in the "C" locale. The "C" locale should be the
2094 default when a C program starts up. In most systems, only ASCII printing
2095 characters match isprint(). This difference caused the output from pcretest
2096 to vary, making some of the tests fail. I have changed pcretest so that:
2097
2098 (a) When it is outputting text in the compiled version of a pattern, bytes
2099 other than 32-126 are always shown as hex escapes.
2100
2101 (b) When it is outputting text that is a matched part of a subject string,
2102 it does the same, unless a different locale has been set for the match
2103 (using the /L modifier). In this case, it uses isprint() to decide.
2104
2105 4. Fixed a major bug that caused incorrect computation of the amount of memory
2106 required for a compiled pattern when options that changed within the
2107 pattern affected the logic of the preliminary scan that determines the
2108 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2109 that the computed length was too small. The symptoms of this bug were
2110 either the PCRE error "internal error: code overflow" from pcre_compile(),
2111 or a glibc crash with a message such as "pcretest: free(): invalid next
2112 size (fast)". Examples of patterns that provoked this bug (shown in
2113 pcretest format) are:
2114
2115 /(?-x: )/x
2116 /(?x)(?-x: \s*#\s*)/
2117 /((?i)[\x{c0}])/8
2118 /(?i:[\x{c0}])/8
2119
2120 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2121 is now done differently.
2122
2123 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2124 wrapper classes; (b) implement a new function in the C++ scanner that is
2125 more efficient than the old way of doing things because it avoids levels of
2126 recursion in the regex matching; (c) add a paragraph to the documentation
2127 for the FullMatch() function.
2128
2129 6. The escape sequence \n was being treated as whatever was defined as
2130 "newline". Not only was this contrary to the documentation, which states
2131 that \n is character 10 (hex 0A), but it also went horribly wrong when
2132 "newline" was defined as CRLF. This has been fixed.
2133
2134 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2135 was being set to -1 for the "end of line" case (supposedly a value that no
2136 character can have). Though this value is never used (the check for end of
2137 line is "zero bytes in current character"), it caused compiler complaints.
2138 I've changed it to 0xffffffff.
2139
2140 8. In pcre_version.c, the version string was being built by a sequence of
2141 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2142 string (as it is for production releases) called a macro with an empty
2143 argument. The C standard says the result of this is undefined. The gcc
2144 compiler treats it as an empty string (which was what was wanted) but it is
2145 reported that Visual C gives an error. The source has been hacked around to
2146 avoid this problem.
2147
2148 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2149 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2150 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2151 of them did).
2152
2153 10. Originally, pcretest opened its input and output without "b"; then I was
2154 told that "b" was needed in some environments, so it was added for release
2155 5.0 to both the input and output. (It makes no difference on Unix-like
2156 systems.) Later I was told that it is wrong for the input on Windows. I've
2157 now abstracted the modes into two macros, to make it easier to fiddle with
2158 them, and removed "b" from the input mode under Windows.
2159
2160 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2161
2162 12. Added -help and --help to pcretest as an official way of being reminded
2163 of the options.
2164
2165 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2166 and pcrecpp.cc because they annoy compilers at high warning levels.
2167
2168 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2169
2170 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2171 scripts are not C programs :-) and which was not noticed because it works
2172 on Linux.
2173
2174 16. pcretest is supposed to handle any length of pattern and data line (as one
2175 line or as a continued sequence of lines) by extending its input buffer if
2176 necessary. This feature was broken for very long pattern lines, leading to
2177 a string of junk being passed to pcre_compile() if the pattern was longer
2178 than about 50K.
2179
2180 17. I have done a major re-factoring of the way pcre_compile() computes the
2181 amount of memory needed for a compiled pattern. Previously, there was code
2182 that made a preliminary scan of the pattern in order to do this. That was
2183 OK when PCRE was new, but as the facilities have expanded, it has become
2184 harder and harder to keep it in step with the real compile phase, and there
2185 have been a number of bugs (see for example, 4 above). I have now found a
2186 cunning way of running the real compile function in a "fake" mode that
2187 enables it to compute how much memory it would need, while actually only
2188 ever using a few hundred bytes of working memory and without too many
2189 tests of the mode. This should make future maintenance and development
2190 easier. A side effect of this work is that the limit of 200 on the nesting
2191 depth of parentheses has been removed (though this was never a serious
2192 limitation, I suspect). However, there is a downside: pcre_compile() now
2193 runs more slowly than before (30% or more, depending on the pattern). I
2194 hope this isn't a big issue. There is no effect on runtime performance.
2195
2196 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2197 newline (only possible for the last line of a file) and it was a
2198 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2199
2200 19. Added additional timing features to pcretest. (1) The -tm option now times
2201 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2202 separate command line item, by a number that specifies the number of
2203 repeats to use when timing. The default is 50000; this gives better
2204 precision, but takes uncomfortably long for very large patterns.
2205
2206 20. Extended pcre_study() to be more clever in cases where a branch of a
2207 subpattern has no definite first character. For example, (a*|b*)[cd] would
2208 previously give no result from pcre_study(). Now it recognizes that the
2209 first character must be a, b, c, or d.
2210
2211 21. There was an incorrect error "recursive call could loop indefinitely" if
2212 a subpattern (or the entire pattern) that was being tested for matching an
2213 empty string contained only one non-empty item after a nested subpattern.
2214 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2215 incorrectly, because the \d was being skipped in the check.
2216
2217 22. The pcretest program now has a new pattern option /B and a command line
2218 option -b, which is equivalent to adding /B to every pattern. This causes
2219 it to show the compiled bytecode, without the additional information that
2220 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2221 is the same as /B/I).
2222
2223 23. A new optimization is now able automatically to treat some sequences such
2224 as a*b as a*+b. More specifically, if something simple (such as a character
2225 or a simple class like \d) has an unlimited quantifier, and is followed by
2226 something that cannot possibly match the quantified thing, the quantifier
2227 is automatically "possessified".
2228
2229 24. A recursive reference to a subpattern whose number was greater than 39
2230 went wrong under certain circumstances in UTF-8 mode. This bug could also
2231 have affected the operation of pcre_study().
2232
2233 25. Realized that a little bit of performance could be had by replacing
2234 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2235
2236 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2237
2238 27. Possessive quantifiers such as a++ were previously implemented by turning
2239 them into atomic groups such as ($>a+). Now they have their own opcodes,
2240 which improves performance. This includes the automatically created ones
2241 from 23 above.
2242
2243 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2244 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2245 the first matched character to be a colon. This applied both to named and
2246 numbered groups.
2247
2248 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2249
2250 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2251 building PCRE on Windows easier, so I have included it in the distribution.
2252
2253 31. There is now a check in pcretest against a ridiculously large number being
2254 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2255 loop, the loop is abandoned.
2256
2257 32. Forward references to subpatterns in conditions such as (?(2)...) where
2258 subpattern 2 is defined later cause pcre_compile() to search forwards in
2259 the pattern for the relevant set of parentheses. This search went wrong
2260 when there were unescaped parentheses in a character class, parentheses
2261 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2262
2263 33. "Subroutine" calls and backreferences were previously restricted to
2264 referencing subpatterns earlier in the regex. This restriction has now
2265 been removed.
2266
2267 34. Added a number of extra features that are going to be in Perl 5.10. On the
2268 whole, these are just syntactic alternatives for features that PCRE had
2269 previously implemented using the Python syntax or my own invention. The
2270 other formats are all retained for compatibility.
2271
2272 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2273 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2274 also .NET compatible.
2275
2276 (b) A recursion or subroutine call to a named group can now be defined as
2277 (?&name) as well as (?P>name).
2278
2279 (c) A backreference to a named group can now be defined as \k<name> or
2280 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2281 5.10, are also .NET compatible.
2282
2283 (d) A conditional reference to a named group can now use the syntax
2284 (?(<name>) or (?('name') as well as (?(name).
2285
2286 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2287 groups (named and numbered) that are never evaluated inline, but can be
2288 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2289 is always false. There may be only one alternative in such a group.
2290
2291 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2292 as the simple (?(R). The condition is true only if the most recent
2293 recursion is that of the given number or name. It does not search out
2294 through the entire recursion stack.
2295
2296 (g) The escape \gN or \g{N} has been added, where N is a positive or
2297 negative number, specifying an absolute or relative reference.
2298
2299 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2300 some "unreachable code" warnings.
2301
2302 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2303 things, this adds five new scripts.
2304
2305 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2306 There were also incompatibilities regarding the handling of \Q..\E inside
2307 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2308 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2309
2310 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2311 matches an empty string, and forcibly breaks the loop. There were bugs in
2312 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2313 against aaaa the result was just "a" rather than "aaaa", for example. Two
2314 separate and independent bugs (that affected different cases) have been
2315 fixed.
2316
2317 39. Refactored the code to abolish the use of different opcodes for small
2318 capturing bracket numbers. This is a tidy that I avoided doing when I
2319 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2320 The new approach is not only tidier, it makes it possible to reduce the
2321 memory needed to fix the previous bug (38).
2322
2323 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2324 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2325 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2326 mode.
2327
2328 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2329 report.
2330
2331 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2332 copy construction and assignment in the C++ wrapper.
2333
2334 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2335 couple of bugs that could have given wrong results in the "--newline=crlf"
2336 case.
2337
2338 44. Added a number of casts and did some reorganization of signed/unsigned int
2339 variables following suggestions from Dair Grant. Also renamed the variable
2340 "this" as "item" because it is a C++ keyword.
2341
2342 45. Arranged for dftables to add
2343
2344 #include "pcre_internal.h"
2345
2346 to pcre_chartables.c because without it, gcc 4.x may remove the array
2347 definition from the final binary if PCRE is built into a static library and
2348 dead code stripping is activated.
2349
2350 46. For an unanchored pattern, if a match attempt fails at the start of a
2351 newline sequence, and the newline setting is CRLF or ANY, and the next two
2352 characters are CRLF, advance by two characters instead of one.
2353
2354
2355 Version 6.7 04-Jul-06
2356 ---------------------
2357
2358 1. In order to handle tests when input lines are enormously long, pcretest has
2359 been re-factored so that it automatically extends its buffers when
2360 necessary. The code is crude, but this _is_ just a test program. The
2361 default size has been increased from 32K to 50K.
2362
2363 2. The code in pcre_study() was using the value of the re argument before
2364 testing it for NULL. (Of course, in any sensible call of the function, it
2365 won't be NULL.)
2366
2367 3. The memmove() emulation function in pcre_internal.h, which is used on
2368 systems that lack both memmove() and bcopy() - that is, hardly ever -
2369 was missing a "static" storage class specifier.
2370
2371 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2372 containing an extended class (one that cannot be represented by a bitmap
2373 because it contains high-valued characters or Unicode property items, e.g.
2374 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2375 pattern, but PCRE should not loop if you do not (it no longer does).
2376 [Detail: two cases were found: (a) a repeated subpattern containing an
2377 extended class; (b) a recursive reference to a subpattern that followed a
2378 previous extended class. It wasn't skipping over the extended class
2379 correctly when UTF-8 mode was not set.]
2380
2381 5. A negated single-character class was not being recognized as fixed-length
2382 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2383 compile error "lookbehind assertion is not fixed length".
2384
2385 6. The RunPerlTest auxiliary script was showing an unexpected difference
2386 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2387 write a Perl script that can interpret lines of an input file either as
2388 byte characters or as UTF-8, which is what "perltest" was being required to
2389 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2390 can't do is switch easily at run time between having the "use utf8;" pragma
2391 or not. In the end, I fudged it by using the RunPerlTest script to insert
2392 "use utf8;" explicitly for the UTF-8 tests.
2393
2394 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2395 the end of the subject string, contrary to the documentation and to what
2396 Perl does. This was true of both matching functions. Now it matches only at
2397 the start of the subject and immediately after *internal* newlines.
2398
2399 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2400 a pointer to an int instead of a pointer to an unsigned long int. This
2401 caused problems on 64-bit systems.
2402
2403 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2404 instance of the 'standard' template library not being so standard".
2405
2406 10. There was no check on the number of named subpatterns nor the maximum
2407 length of a subpattern name. The product of these values is used to compute
2408 the size of the memory block for a compiled pattern. By supplying a very
2409 long subpattern name and a large number of named subpatterns, the size
2410 computation could be caused to overflow. This is now prevented by limiting
2411 the length of names to 32 characters, and the number of named subpatterns
2412 to 10,000.
2413
2414 11. Subpatterns that are repeated with specific counts have to be replicated in
2415 the compiled pattern. The size of memory for this was computed from the
2416 length of the subpattern and the repeat count. The latter is limited to
2417 65535, but there was no limit on the former, meaning that integer overflow
2418 could in principle occur. The compiled length of a repeated subpattern is
2419 now limited to 30,000 bytes in order to prevent this.
2420
2421 12. Added the optional facility to have named substrings with the same name.
2422
2423 13. Added the ability to use a named substring as a condition, using the
2424 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2425 are numbers (not recommended). Forward references are permitted.
2426
2427 14. Added forward references in named backreferences (if you see what I mean).
2428
2429 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2430 pattern could run off the end of the subject. For example, the pattern
2431 "(?s)(.{1,5})"8 did this with the subject "ab".
2432
2433 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2434 PCRE_CASELESS was set when matching characters that were quantified with ?
2435 or *.
2436
2437 17. A character class other than a single negated character that had a minimum
2438 but no maximum quantifier - for example [ab]{6,} - was not handled
2439 correctly by pce_dfa_exec(). It would match only one character.
2440
2441 18. A valid (though odd) pattern that looked like a POSIX character
2442 class but used an invalid character after [ (for example [[,abc,]]) caused
2443 pcre_compile() to give the error "Failed: internal error: code overflow" or
2444 in some cases to crash with a glibc free() error. This could even happen if
2445 the pattern terminated after [[ but there just happened to be a sequence of
2446 letters, a binary zero, and a closing ] in the memory that followed.
2447
2448 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2449 over the years. Originally (before any Unicode support), just the bottom 8
2450 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2451 output from "man perlunicode" includes this:
2452
2453 The regular expression compiler produces polymorphic opcodes. That
2454 is, the pattern adapts to the data and automatically switches to
2455 the Unicode character scheme when presented with Unicode data--or
2456 instead uses a traditional byte scheme when presented with byte
2457 data.
2458
2459 Sadly, a wide octal escape does not cause a switch, and in a string with
2460 no other multibyte characters, these octal escapes are treated as before.
2461 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2462 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2463 Unicode string.
2464
2465 I have not perpetrated such confusion in PCRE. Up till now, it took just
2466 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2467 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2468 translate to the appropriate multibyte character.
2469
2470 29. Applied some refactoring to reduce the number of warnings from Microsoft
2471 and Borland compilers. This has included removing the fudge introduced
2472 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2473 a warning about an unused variable.
2474
2475 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2476 characters since release 4.0, because Perl (from release 5.004) does not.
2477 [Or at least, is documented not to: some releases seem to be in conflict
2478 with the documentation.] However, when a pattern was studied with
2479 pcre_study() and all its branches started with \s, PCRE still included VT
2480 as a possible starting character. Of course, this did no harm; it just
2481 caused an unnecessary match attempt.
2482
2483 22. Removed a now-redundant internal flag bit that recorded the fact that case
2484 dependency changed within the pattern. This was once needed for "required
2485 byte" processing, but is no longer used. This recovers a now-scarce options
2486 bit. Also moved the least significant internal flag bit to the most-
2487 significant bit of the word, which was not previously used (hangover from
2488 the days when it was an int rather than a uint) to free up another bit for
2489 the future.
2490
2491 23. Added support for CRLF line endings as well as CR and LF. As well as the
2492 default being selectable at build time, it can now be changed at runtime
2493 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2494 specify that it is scanning data with non-default line endings.
2495
2496 24. Changed the definition of CXXLINK to make it agree with the definition of
2497 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2498
2499 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2500 recursions. This makes a big different to stack usage for some patterns.
2501
2502 26. If a subpattern containing a named recursion or subroutine reference such
2503 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2504 the space required for the compiled pattern went wrong and gave too small a
2505 value. Depending on the environment, this could lead to "Failed: internal
2506 error: code overflow at offset 49" or "glibc detected double free or
2507 corruption" errors.
2508
2509 27. Applied patches from Google (a) to support the new newline modes and (b) to
2510 advance over multibyte UTF-8 characters in GlobalReplace.
2511
2512 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2513 difference for some implementation of PCRE in some Windows version.
2514
2515 29. Added some extra testing facilities to pcretest:
2516
2517 \q<number> in a data line sets the "match limit" value
2518 \Q<number> in a data line sets the "match recursion limt" value
2519 -S <number> sets the stack size, where <number> is in megabytes
2520
2521 The -S option isn't available for Windows.
2522
2523
2524 Version 6.6 06-Feb-06
2525 ---------------------
2526
2527 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2528 in pcreposix.h. I have copied the definition from pcre.h.
2529
2530 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2531 because pcre.h is no longer a built file.
2532
2533 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2534 not normally included in the compiled code.
2535
2536
2537 Version 6.5 01-Feb-06
2538 ---------------------
2539
2540 1. When using the partial match feature with pcre_dfa_exec(), it was not
2541 anchoring the second and subsequent partial matches at the new starting
2542 point. This could lead to incorrect results. For example, with the pattern
2543 /1234/, partially matching against "123" and then "a4" gave a match.
2544
2545 2. Changes to pcregrep:
2546
2547 (a) All non-match returns from pcre_exec() were being treated as failures
2548 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2549 error message is output. Some extra information is given for the
2550 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2551 probably the only errors that are likely to be caused by users (by
2552 specifying a regex that has nested indefinite repeats, for instance).
2553 If there are more than 20 of these errors, pcregrep is abandoned.
2554
2555 (b) A binary zero was treated as data while matching, but terminated the
2556 output line if it was written out. This has been fixed: binary zeroes
2557 are now no different to any other data bytes.
2558
2559 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2560 used to set a locale for matching. The --locale=xxxx long option has
2561 been added (no short equivalent) to specify a locale explicitly on the
2562 pcregrep command, overriding the environment variables.
2563
2564 (d) When -B was used with -n, some line numbers in the output were one less
2565 than they should have been.
2566
2567 (e) Added the -o (--only-matching) option.
2568
2569 (f) If -A or -C was used with -c (count only), some lines of context were
2570 accidentally printed for the final match.
2571
2572 (g) Added the -H (--with-filename) option.
2573
2574 (h) The combination of options -rh failed to suppress file names for files
2575 that were found from directory arguments.
2576
2577 (i) Added the -D (--devices) and -d (--directories) options.
2578
2579 (j) Added the -F (--fixed-strings) option.
2580
2581 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2582
2583 (l) Added the --colo(u)r option.
2584
2585 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2586 is not present by default.
2587
2588 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2589 items such as (?R) or (?1), when the recursion could match a number of
2590 alternatives. If it matched one of the alternatives, but subsequently,
2591 outside the recursion, there was a failure, the code tried to back up into
2592 the recursion. However, because of the way PCRE is implemented, this is not
2593 possible, and the result was an incorrect result from the match.
2594
2595 In order to prevent this happening, the specification of recursion has
2596 been changed so that all such subpatterns are automatically treated as
2597 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2598
2599 4. I had overlooked the fact that, in some locales, there are characters for
2600 which isalpha() is true but neither isupper() nor islower() are true. In
2601 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2602 and ordfeminine) are like this. This affected the treatment of \w and \W
2603 when they appeared in character classes, but not when they appeared outside
2604 a character class. The bit map for "word" characters is now created
2605 separately from the results of isalnum() instead of just taking it from the
2606 upper, lower, and digit maps. (Plus the underscore character, of course.)
2607
2608 5. The above bug also affected the handling of POSIX character classes such as
2609 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2610 permanent tables. Instead, the bit maps for such a class were previously
2611 created as the appropriate unions of the upper, lower, and digit bitmaps.
2612 Now they are created by subtraction from the [[:word:]] class, which has
2613 its own bitmap.
2614
2615 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2616 It is created by subtracting the vertical space characters (\x09, \x0a,
2617 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2618 subtraction was done in the overall bitmap for a character class, meaning
2619 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2620 be recognized. This bug has been fixed.
2621
2622 7. Patches from the folks at Google:
2623
2624 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2625 real life, but is still worth protecting against".
2626
2627 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2628 regular expressions".
2629
2630 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2631 have it.
2632
2633 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2634 "configure" and the latter not, in order to fix a problem somebody had
2635 with compiling the Arg class on HP-UX.
2636
2637 (e) Improve the error-handling of the C++ wrapper a little bit.
2638
2639 (f) New tests for checking recursion limiting.
2640
2641 8. The pcre_memmove() function, which is used only if the environment does not
2642 have a standard memmove() function (and is therefore rarely compiled),
2643 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2644 returning a result (though PCRE never actually uses the result).
2645
2646 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2647 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2648 returned instead of calling malloc() with an overflowing number that would
2649 most likely cause subsequent chaos.
2650
2651 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2652
2653 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2654 with this option is matched, the nmatch and pmatch options of regexec() are
2655 ignored.
2656
2657 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2658 provided in case anyone wants to the the POSIX interface with UTF-8
2659 strings.
2660
2661 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2662 C++ linking (needed for some HP-UX environments).
2663
2664 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2665 (unused parameter) and in the pcre_printint() function (omitted "default"
2666 switch label when the default is to do nothing).
2667
2668 15. Added some code to make it possible, when PCRE is compiled as a C++
2669 library, to replace subject pointers for pcre_exec() with a smart pointer
2670 class, thus making it possible to process discontinuous strings.
2671
2672 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2673 much the same function. They were added by different people who were trying
2674 to make PCRE easy to compile on non-Unix systems. It has been suggested
2675 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2676 for compiling on Windows systems. I have therefore replaced it with
2677 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2678 defaults to "extern" for C or "extern C" for C++, which works fine on
2679 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2680 SCOPE with something explicit in config.h. In addition:
2681
2682 (a) pcreposix.h still had just "extern" instead of either of these macros;
2683 I have replaced it with PCRE_DATA_SCOPE.
2684
2685 (b) Functions such as _pcre_xclass(), which are internal to the library,
2686 but external in the C sense, all had PCRE_EXPORT in their definitions.
2687 This is apparently wrong for the Windows case, so I have removed it.
2688 (It makes no difference on Unix-like systems.)
2689
2690 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2691 of recursive calls to match(). This is different to MATCH_LIMIT because
2692 that limits the total number of calls to match(), not all of which increase
2693 the depth of recursion. Limiting the recursion depth limits the amount of
2694 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2695 when PCRE is compiled, and changed at run time. A patch from Google adds
2696 this functionality to the C++ interface.
2697
2698 18. Changes to the handling of Unicode character properties:
2699
2700 (a) Updated the table to Unicode 4.1.0.
2701
2702 (b) Recognize characters that are not in the table as "Cn" (undefined).
2703
2704 (c) I revised the way the table is implemented to a much improved format
2705 which includes recognition of ranges. It now supports the ranges that
2706 are defined in UnicodeData.txt, and it also amalgamates other
2707 characters into ranges. This has reduced the number of entries in the
2708 table from around 16,000 to around 3,000, thus reducing its size
2709 considerably. I realized I did not need to use a tree structure after
2710 all - a binary chop search is just as efficient. Having reduced the
2711 number of entries, I extended their size from 6 bytes to 8 bytes to
2712 allow for more data.
2713
2714 (d) Added support for Unicode script names via properties such as \p{Han}.
2715
2716 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2717 matching that character.
2718
2719 20. When matching a repeated Unicode property with a minimum greater than zero,
2720 (for example \pL{2,}), PCRE could look past the end of the subject if it
2721 reached it while seeking the minimum number of characters. This could
2722 happen only if some of the characters were more than one byte long, because
2723 there is a check for at least the minimum number of bytes.
2724
2725 21. Refactored the implementation of \p and \P so as to be more general, to
2726 allow for more different types of property in future. This has changed the
2727 compiled form incompatibly. Anybody with saved compiled patterns that use
2728 \p or \P will have to recompile them.
2729
2730 22. Added "Any" and "L&" to the supported property types.
2731
2732 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2733 but give a compile time error if the value is greater than 0xff.
2734
2735 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2736 accidentally not being installed or uninstalled.
2737
2738 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2739 made were to insert the current release number. This seemed silly, because
2740 it made things harder for people building PCRE on systems that don't run
2741 "configure". I have turned pcre.h into a distributed file, no longer built
2742 by "configure", with the version identification directly included. There is
2743 no longer a pcre.h.in file.
2744
2745 However, this change necessitated a change to the pcre-config script as
2746 well. It is built from pcre-config.in, and one of the substitutions was the
2747 release number. I have updated configure.ac so that ./configure now finds
2748 the release number by grepping pcre.h.
2749
2750 26. Added the ability to run the tests under valgrind.
2751
2752
2753 Version 6.4 05-Sep-05
2754 ---------------------
2755
2756 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2757 "--" to be printed when multiple files were scanned, even when none of the
2758 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2759 consider it to be a bug, and have restored the previous behaviour.
2760
2761 2. A couple of code tidies to get rid of compiler warnings.
2762
2763 3. The pcretest program used to cheat by referring to symbols in the library
2764 whose names begin with _pcre_. These are internal symbols that are not
2765 really supposed to be visible externally, and in some environments it is
2766 possible to suppress them. The cheating is now confined to including
2767 certain files from the library's source, which is a bit cleaner.
2768
2769 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2770 file's purpose clearer.
2771
2772 5. Reorganized pcre_ucp_findchar().
2773
2774
2775 Version 6.3 15-Aug-05
2776 ---------------------
2777
2778 1. The file libpcre.pc.in did not have general read permission in the tarball.
2779
2780 2. There were some problems when building without C++ support:
2781
2782 (a) If C++ support was not built, "make install" and "make test" still
2783 tried to test it.
2784
2785 (b) There were problems when the value of CXX was explicitly set. Some
2786 changes have been made to try to fix these, and ...
2787
2788 (c) --disable-cpp can now be used to explicitly disable C++ support.
2789
2790 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2791 backslash in a target when C++ was disabled. This confuses some
2792 versions of "make", apparently. Using an intermediate variable solves
2793 this. (Same for CPP_LOBJ.)
2794
2795 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2796 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2797 necessary on certain architectures.
2798
2799 4. Added a setting of -export-symbols-regex to the link command to remove
2800 those symbols that are exported in the C sense, but actually are local
2801 within the library, and not documented. Their names all begin with
2802 "_pcre_". This is not a perfect job, because (a) we have to except some
2803 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2804 available (and never for static libraries). I have made a note to try to
2805 find a way round (a) in the future.
2806
2807
2808 Version 6.2 01-Aug-05
2809 ---------------------
2810
2811 1. There was no test for integer overflow of quantifier values. A construction
2812 such as {1111111111111111} would give undefined results. What is worse, if
2813 a minimum quantifier for a parenthesized subpattern overflowed and became
2814 negative, the calculation of the memory size went wrong. This could have
2815 led to memory overwriting.
2816
2817 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2818
2819 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2820 operating environments where this matters.
2821
2822 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2823 PCRE options from within the C++ wrapper.
2824
2825 5. Named capturing subpatterns were not being correctly counted when a pattern
2826 was compiled. This caused two problems: (a) If there were more than 100
2827 such subpatterns, the calculation of the memory needed for the whole
2828 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2829 back references of the form \12, where the number was greater than 9, were
2830 not recognized as back references, even though there were sufficient
2831 previous subpatterns.
2832
2833 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2834 versions of gcc, e.g. 2.95.4.
2835
2836
2837 Version 6.1 21-Jun-05
2838 ---------------------
2839
2840 1. There was one reference to the variable "posix" in pcretest.c that was not
2841 surrounded by "#if !defined NOPOSIX".
2842
2843 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2844 the cross-check on the old pcre_info() function, for the benefit of the
2845 cut-down version of PCRE that is currently imported into Exim.
2846
2847 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2848 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2849 patterns that start (?i) but I don't think that matters. The use of (?i) is
2850 just an example; this all applies to the other options as well.
2851
2852 4. Since libtool seems to echo the compile commands it is issuing, the output
2853 from "make" can be reduced a bit by putting "@" in front of each libtool
2854 compile command.
2855
2856 5. Patch from the folks at Google for configure.in to be a bit more thorough
2857 in checking for a suitable C++ installation before trying to compile the
2858 C++ stuff. This should fix a reported problem when a compiler was present,
2859 but no suitable headers.
2860
2861 6. The man pages all had just "PCRE" as their title. I have changed them to
2862 be the relevant file name. I have also arranged that these names are
2863 retained in the file doc/pcre.txt, which is a concatenation in text format
2864 of all the man pages except the little individual ones for each function.
2865
2866 7. The NON-UNIX-USE file had not been updated for the different set of source
2867 files that come with release 6. I also added a few comments about the C++
2868 wrapper.
2869
2870
2871 Version 6.0 07-Jun-05
2872 ---------------------
2873
2874 1. Some minor internal re-organization to help with my DFA experiments.
2875
2876 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2877 didn't matter for the library itself when fully configured, but did matter
2878 when compiling without UCP support, or within Exim, where the ucp files are
2879 not imported.
2880
2881 3. Refactoring of the library code to split up the various functions into
2882 different source modules. The addition of the new DFA matching code (see
2883 below) to a single monolithic source would have made it really too
2884 unwieldy, quite apart from causing all the code to be include in a
2885 statically linked application, when only some functions are used. This is
2886 relevant even without the DFA addition now that patterns can be compiled in
2887 one application and matched in another.
2888
2889 The downside of splitting up is that there have to be some external
2890 functions and data tables that are used internally in different modules of
2891 the library but which are not part of the API. These have all had their
2892 names changed to start with "_pcre_" so that they are unlikely to clash
2893 with other external names.
2894
2895 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2896 a different (DFA) algorithm. Although it is slower than the original
2897 function, it does have some advantages for certain types of matching
2898 problem.
2899
2900 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2901 including restarting after a partial match.
2902
2903 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2904 defined when compiling for Windows was sent to me. I have put it into the
2905 code, though I have no means of testing or verifying it.
2906
2907 7. Added the pcre_refcount() auxiliary function.
2908
2909 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2910 match before or at the first newline in the subject string. In pcretest,
2911 the /f option on a pattern can be used to set this.
2912
2913 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2914 would behave wrongly. This has been present in PCRE since release 4.0.
2915
2916 10. A number of changes to the pcregrep command:
2917
2918 (a) Refactored how -x works; insert ^(...)$ instead of setting
2919 PCRE_ANCHORED and checking the length, in preparation for adding
2920 something similar for -w.
2921
2922 (b) Added the -w (match as a word) option.
2923
2924 (c) Refactored the way lines are read and buffered so as to have more
2925 than one at a time available.
2926
2927 (d) Implemented a pcregrep test script.
2928
2929 (e) Added the -M (multiline match) option. This allows patterns to match
2930 over several lines of the subject. The buffering ensures that at least
2931 8K, or the rest of the document (whichever is the shorter) is available
2932 for matching (and similarly the previous 8K for lookbehind assertions).
2933
2934 (f) Changed the --help output so that it now says
2935
2936 -w, --word-regex(p)
2937
2938 instead of two lines, one with "regex" and the other with "regexp"
2939 because that confused at least one person since the short forms are the
2940 same. (This required a bit of code, as the output is generated
2941 automatically from a table. It wasn't just a text change.)
2942
2943 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2944 option but starts with a hyphen. Could be a pattern or a path name
2945 starting with a hyphen, for instance.
2946
2947 (h) "-" can be given as a file name to represent stdin.
2948
2949 (i) When file names are being printed, "(standard input)" is used for
2950 the standard input, for compatibility with GNU grep. Previously
2951 "<stdin>" was used.
2952
2953 (j) The option --label=xxx can be used to supply a name to be used for
2954 stdin when file names are being printed. There is no short form.
2955
2956 (k) Re-factored the options decoding logic because we are going to add
2957 two more options that take data. Such options can now be given in four
2958 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2959
2960 (l) Added the -A, -B, and -C options for requesting that lines of context
2961 around matches be printed.
2962
2963 (m) Added the -L option to print the names of files that do not contain
2964 any matching lines, that is, the complement of -l.
2965
2966 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2967 continue to scan other files.
2968
2969 (o) The -s option was incorrectly implemented. For compatibility with other
2970 greps, it now suppresses the error message for a non-existent or non-
2971 accessible file (but not the return code). There is a new option called
2972 -q that suppresses the output of matching lines, which was what -s was
2973 previously doing.
2974
2975 (p) Added --include and --exclude options to specify files for inclusion
2976 and exclusion when recursing.
2977
2978 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2979 Hopefully, it now does.
2980
2981 12. Missing cast in pcre_study().
2982
2983 13. Added an "uninstall" target to the makefile.
2984
2985 14. Replaced "extern" in the function prototypes in Makefile.in with
2986 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2987 world, but is set differently for Windows.
2988
2989 15. Added a second compiling function called pcre_compile2(). The only
2990 difference is that it has an extra argument, which is a pointer to an
2991 integer error code. When there is a compile-time failure, this is set
2992 non-zero, in addition to the error test pointer being set to point to an
2993 error message. The new argument may be NULL if no error number is required
2994 (but then you may as well call pcre_compile(), which is now just a
2995 wrapper). This facility is provided because some applications need a
2996 numeric error indication, but it has also enabled me to tidy up the way
2997 compile-time errors are handled in the POSIX wrapper.
2998
2999 16. Added VPATH=.libs to the makefile; this should help when building with one
3000 prefix path and installing with another. (Or so I'm told by someone who
3001 knows more about this stuff than I do.)
3002
3003 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3004 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3005 match everything, including newlines. This is not POSIX-compatible, but
3006 somebody wanted the feature. From pcretest it can be activated by using
3007 both the P and the s flags.
3008
3009 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3010
3011 19. libpcre.pc was being incorrectly installed as executable.
3012
3013 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3014 it now also looks for '\r' so that it will work unmodified on Windows.
3015
3016 21. Added Google's contributed C++ wrapper to the distribution.
3017
3018 22. Added some untidy missing memory free() calls in pcretest, to keep
3019 Electric Fence happy when testing.
3020
3021
3022
3023 Version 5.0 13-Sep-04
3024 ---------------------
3025
3026 1. Internal change: literal characters are no longer packed up into items
3027 containing multiple characters in a single byte-string. Each character
3028 is now matched using a separate opcode. However, there may be more than one
3029 byte in the character in UTF-8 mode.
3030
3031 2. The pcre_callout_block structure has two new fields: pattern_position and
3032 next_item_length. These contain the offset in the pattern to the next match
3033 item, and its length, respectively.
3034
3035 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3036 insertion of callouts before each pattern item. Added the /C option to
3037 pcretest to make use of this.
3038
3039 4. On the advice of a Windows user, the lines
3040
3041 #if defined(_WIN32) || defined(WIN32)
3042 _setmode( _fileno( stdout ), 0x8000 );
3043 #endif /* defined(_WIN32) || defined(WIN32) */
3044
3045 have been added to the source of pcretest. This apparently does useful
3046 magic in relation to line terminators.
3047
3048 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3049 for the benefit of those environments where the "b" makes a difference.
3050
3051 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3052 to know about it. I have put a hack into configure.in that adds in code
3053 to set GCC=yes if CC=icc. This seems to end up at a point in the
3054 generated configure script that is early enough to affect the setting of
3055 compiler options, which is what is needed, but I have no means of testing
3056 whether it really works. (The user who reported this had patched the
3057 generated configure script, which of course I cannot do.)
3058
3059 LATER: After change 22 below (new libtool files), the configure script
3060 seems to know about icc (and also ecc). Therefore, I have commented out
3061 this hack in configure.in.
3062
3063 7. Added support for pkg-config (2 patches were sent in).
3064
3065 8. Negated POSIX character classes that used a combination of internal tables
3066 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3067 [[:^ascii]]. Typically, they would match almost any characters. The other
3068 POSIX classes were not broken in this way.
3069
3070 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3071 to find the match, as PCRE was deluded into thinking that the match had to
3072 start at the start point or following a newline. The same bug applied to
3073 patterns with negative forward assertions or any backward assertions
3074 preceding ".*" at the start, unless the pattern required a fixed first
3075 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3076
3077 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3078 starting at the last subject character, bytes beyond the end of the subject
3079 string were read.
3080
3081 11. Renamed the variable "class" as "classbits" to make life easier for C++
3082 users. (Previously there was a macro definition, but it apparently wasn't
3083 enough.)
3084
3085 12. Added the new field "tables" to the extra data so that tables can be passed
3086 in at exec time, or the internal tables can be re-selected. This allows
3087 a compiled regex to be saved and re-used at a later time by a different
3088 program that might have everything at different addresses.
3089
3090 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3091 -R library as well as a -L library.
3092
3093 14. The debugging options of pcretest (-d on the command line or D on a
3094 pattern) showed incorrect output for anything following an extended class
3095 that contained multibyte characters and which was followed by a quantifier.
3096
3097 15. Added optional support for general category Unicode character properties
3098 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3099 support. It adds about 90K to the size of the library. The meanings of the
3100 inbuilt class escapes such as \d and \s have NOT been changed.
3101
3102 16. Updated pcredemo.c to include calls to free() to release the memory for the
3103 compiled pattern.
3104
3105 17. The generated file chartables.c was being created in the source directory
3106 instead of in the building directory. This caused the build to fail if the
3107 source directory was different from the building directory, and was
3108 read-only.
3109
3110 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3111 file. No doubt somebody will tell me if they don't make sense... Also added
3112 Dan Mooney's comments about building on OpenVMS.
3113
3114 19. Added support for partial matching via the PCRE_PARTIAL option for
3115 pcre_exec() and the \P data escape in pcretest.
3116
3117 20. Extended pcretest with 3 new pattern features:
3118
3119 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3120 write the compiled pattern to the file whose name is "rest-of-line".
3121 This is a straight binary dump of the data, with the saved pointer to
3122 the character tables forced to be NULL. The study data, if any, is
3123 written too. After writing, pcretest reads a new pattern.
3124
3125 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3126 compiled pattern from the given file. There must not be any
3127 occurrences of "<" in the file name (pretty unlikely); if there are,
3128 pcretest will instead treat the initial "<" as a pattern delimiter.
3129 After reading in the pattern, pcretest goes on to read data lines as
3130 usual.
3131
3132 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3133 and 16-bit fields in a compiled pattern, to simulate a pattern that
3134 was compiled on a host of opposite endianness.
3135
3136 21. The pcre-exec() function can now cope with patterns that were compiled on
3137 hosts of opposite endianness, with this restriction:
3138
3139 As for any compiled expression that is saved and used later, the tables
3140 pointer field cannot be preserved; the extra_data field in the arguments
3141 to pcre_exec() should be used to pass in a tables address if a value
3142 other than the default internal tables were used at compile time.
3143
3144 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3145 now diagnosed as an error. Previously, most of the time, a negative number
3146 would have been treated as zero, but if in addition "ovector" was passed as
3147 NULL, a crash could occur.
3148
3149 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3150 new versions from the libtool 1.5 distribution (the last one is a copy of
3151 a file called libtool.m4). This seems to have fixed the need to patch
3152 "configure" to support Darwin 1.3 (which I used to do). However, I still
3153 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3154 workstation).
3155
3156 24. Changed the PCRE licence to be the more standard "BSD" licence.
3157
3158
3159 Version 4.5 01-Dec-03
3160 ---------------------
3161
3162 1. There has been some re-arrangement of the code for the match() function so
3163 that it can be compiled in a version that does not call itself recursively.
3164 Instead, it keeps those local variables that need separate instances for
3165 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3166 needs to "recurse". Keeping track of where control must go is done by means
3167 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3168 hide most of the details from the main code, and operates only if
3169 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3170 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3171 operating.
3172
3173 To make it easier for callers to provide specially tailored get/free
3174 functions for this usage, two new functions, pcre_stack_malloc, and
3175 pcre_stack_free, are used. They are always called in strict stacking order,
3176 and the size of block requested is always the same.
3177
3178 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3179 PCRE has been compiled to use the stack or the heap for recursion. The
3180 -C option of pcretest uses this to show which version is compiled.
3181
3182 A new data escape \S, is added to pcretest; it causes the amounts of store
3183 obtained and freed by both kinds of malloc/free at match time to be added
3184 to the output.
3185
3186 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3187 what's available on my current Linux desktop machine.
3188
3189 3. When matching a UTF-8 string, the test for a valid string at the start has
3190 been extended. If start_offset is not zero, PCRE now checks that it points
3191 to a byte that is the start of a UTF-8 character. If not, it returns
3192 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3193 this is necessary because there may be backward assertions in the pattern.
3194 When matching the same subject several times, it may save resources to use
3195 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3196
3197 4. The code for checking the validity of UTF-8 strings has been tightened so
3198 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3199 containing "overlong sequences".
3200
3201 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3202 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3203 should have been "&", but it just so happened that all the cases this let
3204 through by mistake were picked up later in the function.
3205
3206 6. I had used a variable called "isblank" - this is a C99 function, causing
3207 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3208
3209 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3210 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3211 so the version is shown; (c) stop "make test" echoing "./RunTest".
3212
3213 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3214
3215 9. The prototype for memmove() for systems that don't have it was using
3216 size_t, but the inclusion of the header that defines size_t was later. I've
3217 moved the #includes for the C headers earlier to avoid this.
3218
3219 10. Added some adjustments to the code to make it easier to compiler on certain
3220 special systems:
3221
3222 (a) Some "const" qualifiers were missing.
3223 (b) Added the macro EXPORT before all exported functions; by default this
3224 is defined to be empty.
3225 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3226 that it reads its output file name as an argument instead of writing
3227 to the standard output and assuming this can be redirected.
3228
3229 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3230 class containing characters with values greater than 255, PCRE compilation
3231 went into a loop.
3232
3233 12. A recursive reference to a subpattern that was within another subpattern
3234 that had a minimum quantifier of zero caused PCRE to crash. For example,
3235 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3236 recursion. If the recursively-called subpattern itself had a zero repeat,
3237 that was OK.
3238
3239 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3240 buffer into which it was copied (for escape processing) was still set at
3241 1024, so long lines caused crashes.
3242
3243 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3244 "internal error: code overflow...". This applied to any character class
3245 that was followed by a possessive quantifier.
3246
3247 15. Modified the Makefile to add libpcre.la as a prerequisite for
3248 libpcreposix.la because I was told this is needed for a parallel build to
3249 work.
3250
3251 16. If a pattern that contained .* following optional items at the start was
3252 studied, the wrong optimizing data was generated, leading to matching
3253 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3254 matching string must start with a or b or c. The correct conclusion for
3255 this pattern is that a match can start with any character.
3256
3257
3258 Version 4.4 13-Aug-03
3259 ---------------------
3260
3261 1. In UTF-8 mode, a character class containing characters with values between
3262 127 and 255 was not handled correctly if the compiled pattern was studied.
3263 In fixing this, I have also improved the studying algorithm for such
3264 classes (slightly).
3265
3266 2. Three internal functions had redundant arguments passed to them. Removal
3267 might give a very teeny performance improvement.
3268
3269 3. Documentation bug: the value of the capture_top field in a callout is *one
3270 more than* the number of the hightest numbered captured substring.
3271
3272 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3273 in incorrectly linking with a previously installed version. They now link
3274 explicitly with libpcre.la.
3275
3276 5. configure.in no longer needs to recognize Cygwin specially.
3277
3278 6. A problem in pcre.in for Windows platforms is fixed.
3279
3280 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3281 pcretest, it used to include the size of the study block as part of its
3282 output. Unfortunately, the structure contains a field that has a different
3283 size on different hardware architectures. This meant that the tests that
3284 showed this size failed. As the block is currently always of a fixed size,
3285 this information isn't actually particularly useful in pcretest output, so
3286 I have just removed it.
3287
3288 8. Three pre-processor statements accidentally did not start in column 1.
3289 Sadly, there are *still* compilers around that complain, even though
3290 standard C has not required this for well over a decade. Sigh.
3291
3292 9. In pcretest, the code for checking callouts passed small integers in the
3293 callout_data field, which is a void * field. However, some picky compilers
3294 complained about the casts involved for this on 64-bit systems. Now
3295 pcretest passes the address of the small integer instead, which should get
3296 rid of the warnings.
3297
3298 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3299 both compile and run time, and gives an error if an invalid UTF-8 sequence
3300 is found. There is a option for disabling this check in cases where the
3301 string is known to be correct and/or the maximum performance is wanted.
3302
3303 11. In response to a bug report, I changed one line in Makefile.in from
3304
3305 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3306 to
3307 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3308
3309 to look similar to other lines, but I have no way of telling whether this
3310 is the right thing to do, as I do not use Windows. No doubt I'll get told
3311 if it's wrong...
3312
3313
3314 Version 4.3 21-May-03
3315 ---------------------
3316
3317 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3318 Makefile.
3319
3320 2. Some refactoring to improve the quality of the code:
3321
3322 (i) The utf8_table... variables are now declared "const".
3323
3324 (ii) The code for \cx, which used the "case flipping" table to upper case
3325 lower case letters, now just substracts 32. This is ASCII-specific,
3326 but the whole concept of \cx is ASCII-specific, so it seems
3327 reasonable.
3328
3329 (iii) PCRE was using its character types table to recognize decimal and
3330 hexadecimal digits in the pattern. This is silly, because it handles
3331 only 0-9, a-f, and A-F, but the character types table is locale-
3332 specific, which means strange things might happen. A private
3333 table is now used for this - though it costs 256 bytes, a table is
3334 much faster than multiple explicit tests. Of course, the standard
3335 character types table is still used for matching digits in subject
3336 strings against \d.
3337
3338 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3339 ending in _t are). So I've renamed it as ESC_tee.
3340
3341 3. The first argument for regexec() in the POSIX wrapper should have been
3342 defined as "const".
3343
3344 4. Changed pcretest to use malloc() for its buffers so that they can be
3345 Electric Fenced for debugging.
3346
3347 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3348 to read one or more bytes before the start of the subject string. Often this
3349 had no effect on PCRE's behaviour, but in some circumstances it could
3350 provoke a segmentation fault.
3351
3352 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3353 to try to read one or more bytes before the start of the subject string.
3354
3355 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3356 UTF-8 support could misbehave in various ways if the subject string
3357 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3358 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3359 back over UTF-8 characters.)
3360
3361
3362 Version 4.2 14-Apr-03
3363 ---------------------
3364
3365 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3366
3367 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3368 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3369 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3370 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3371 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3372 and BUILD_EXEEXT
3373 Note: automatic setting of the BUILD variables is not yet working
3374 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3375 compile-time but not at link-time
3376 [LINK]: use for linking executables only
3377 make different versions for Windows and non-Windows
3378 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3379 libraries
3380 [LINK_FOR_BUILD]: new variable
3381 [OBJEXT]: use throughout
3382 [EXEEXT]: use throughout
3383 <winshared>: new target
3384 <wininstall>: new target
3385 <dftables.o>: use native compiler
3386 <dftables>: use native linker
3387 <install>: handle Windows platform correctly
3388 <clean>: ditto
3389 <check>: ditto
3390 copy DLL to top builddir before testing
3391
3392 As part of these changes, -no-undefined was removed again. This was reported
3393 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3394 in any case.
3395
3396 3. Some tidies to get rid of compiler warnings:
3397
3398 . In the match_data structure, match_limit was an unsigned long int, whereas
3399 match_call_count was an int. I've made them both unsigned long ints.
3400
3401 . In pcretest the fact that a const uschar * doesn't automatically cast to
3402 a void * provoked a warning.
3403
3404 . Turning on some more compiler warnings threw up some "shadow" variables
3405 and a few more missing casts.
3406
3407 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3408 option, a class that contained a single character with a value between 128
3409 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3410
3411 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3412 option, a class that contained several characters, but with at least one
3413 whose value was between 128 and 255 caused PCRE to crash.
3414
3415
3416 Version 4.1 12-Mar-03
3417 ---------------------
3418
3419 1. Compiling with gcc -pedantic found a couple of places where casts were
3420 needed, and a string in dftables.c that was longer than standard compilers are
3421 required to support.
3422
3423 2. Compiling with Sun's compiler found a few more places where the code could
3424 be tidied up in order to avoid warnings.
3425
3426 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3427 first of these names is deprecated in the latest Autoconf in favour of the name
3428 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3429 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3430 analogy I have changed it to CFLAGS_FOR_BUILD.
3431
3432 4. Added -no-undefined to the linking command in the Makefile, because this is
3433 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3434 linking step for the pcreposix library.
3435
3436 5. PCRE was failing to diagnose the case of two named groups with the same
3437 name.
3438
3439 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3440 literal character that is needed in the subject for a match, and scans along to
3441 ensure that it is present before embarking on the full matching process. This
3442 saves time in cases of nested unlimited repeats that are never going to match.
3443 Problem: the scan can take a lot of time if the subject is very long (e.g.
3444 megabytes), thus penalizing straightforward matches. It is now done only if the
3445 amount of subject to be scanned is less than 1000 bytes.
3446
3447 7. A lesser problem with the same optimization is that it was recording the
3448 first character of an anchored pattern as "needed", thus provoking a search
3449 right along the subject, even when the first match of the pattern was going to
3450 fail. The "needed" character is now not set for anchored patterns, unless it
3451 follows something in the pattern that is of non-fixed length. Thus, it still
3452 fulfils its original purpose of finding quick non-matches in cases of nested
3453 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3454
3455
3456 Version 4.0 17-Feb-03
3457 ---------------------
3458
3459 1. If a comment in an extended regex that started immediately after a meta-item
3460 extended to the end of string, PCRE compiled incorrect data. This could lead to
3461 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3462
3463 2. Moved to autoconf 2.53 and libtool 1.4.2.
3464
3465 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3466 the special perltest8 script is no longer needed - all the tests can be run
3467 from a single perltest script.
3468
3469 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3470 by \s. It has now been removed in PCRE. This means it isn't recognized as
3471 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3472 class [:space:] *does* include VT, thereby creating a mess.
3473
3474 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3475 space and tab.
3476
3477 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3478 its new features into the main test script, reducing the number of scripts.
3479
3480 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3481 were backward compatible, and made the (?i) apply to the whole pattern, as if
3482 /i were given. Now it behaves more logically, and applies the option setting
3483 only to what follows. PCRE has been changed to follow suit. However, if it
3484 finds options settings right at the start of the pattern, it extracts them into
3485 the global options, as before. Thus, they show up in the info data.
3486
3487 8. Added support for the \Q...\E escape sequence. Characters in between are
3488 treated as literals. This is slightly different from Perl in that $ and @ are
3489 also handled as literals inside the quotes. In Perl, they will cause variable
3490 interpolation. Note the following examples:
3491
3492 Pattern PCRE matches Perl matches
3493
3494 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3495 \Qabc\$xyz\E abc\$xyz abc\$xyz
3496 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3497
3498 For compatibility with Perl, \Q...\E sequences are recognized inside character
3499 classes as well as outside them.
3500
3501 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3502 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3503 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3504 signed/unsigned warnings.
3505
3506 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3507 option for pcretest, so I've replaced it by a simple function that does just
3508 that job.
3509
3510 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3511 "pcregrep -".
3512
3513 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3514 Java package. This provides some syntactic sugar for simple cases of what my
3515 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3516 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3517 item, you can use this simplified notation. Note that only makes sense with
3518 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3519 greediness, whatever the setting of the PCRE_UNGREEDY option.
3520
3521 13. A change of greediness default within a pattern was not taking effect at
3522 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3523 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3524 was abstracted outside.
3525
3526 14. PCRE now supports the \G assertion. It is true when the current matching
3527 position is at the start point of the match. This differs from \A when the
3528 starting offset is non-zero. Used with the /g option of pcretest (or similar
3529 code), it works in the same way as it does for Perl's /g option. If all
3530 alternatives of a regex begin with \G, the expression is anchored to the start
3531 match position, and the "anchored" flag is set in the compiled expression.
3532
3533 15. Some bugs concerning the handling of certain option changes within patterns
3534 have been fixed. These applied to options other than (?ims). For example,
3535 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3536 been the other way round. Some of this was related to change 7 above.
3537
3538 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3539 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3540 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3541 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3542
3543 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3544 mode. Unlike ".", it always matches newline, whatever the setting of
3545 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3546 assertions. Perl allows it, but it doesn't (in general) work because it can't
3547 calculate the length of the lookbehind. At least, that's the case for Perl
3548 5.8.0 - I've been told they are going to document that it doesn't work in
3549 future.
3550
3551 18. Added an error diagnosis for escapes that PCRE does not support: these are
3552 \L, \l, \N, \P, \p, \U, \u, and \X.
3553
3554 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3555 reading past the end of the pattern in cases such as /[abcd/.
3556
3557 20. PCRE was getting more memory than necessary for patterns with classes that
3558 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3559
3560 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3561 compiling PCRE for use with Virtual Pascal.
3562
3563 22. Small fix to the Makefile to make it work properly if the build is done
3564 outside the source tree.
3565
3566 23. Added a new extension: a condition to go with recursion. If a conditional
3567 subpattern starts with (?(R) the "true" branch is used if recursion has
3568 happened, whereas the "false" branch is used only at the top level.
3569
3570 24. When there was a very long string of literal characters (over 255 bytes
3571 without UTF support, over 250 bytes with UTF support), the computation of how
3572 much memory was required could be incorrect, leading to segfaults or other
3573 strange effects.
3574
3575 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3576 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3577 there was a subsequent back reference to those brackets. This meant that, for
3578 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3579 possible to check for precisely this case. All we can do is abandon the
3580 optimization if .* occurs inside capturing brackets when there are any back
3581 references whatsoever. (See below for a better fix that came later.)
3582
3583 26. The handling of the optimization for finding the first character of a
3584 non-anchored pattern, and for finding a character that is required later in the
3585 match were failing in some cases. This didn't break the matching; it just
3586 failed to optimize when it could. The way this is done has been re-implemented.
3587
3588 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3589
3590 28. Added a new feature that provides some of the functionality that Perl
3591 provides with (?{...}). The facility is termed a "callout". The way it is done
3592 in PCRE is for the caller to provide an optional function, by setting
3593 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3594 global variable. By default it is unset, which disables all calling out. To get
3595 the function called, the regex must include (?C) at appropriate points. This
3596 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3597 This provides a means of identifying different callout points. When PCRE
3598 reaches such a point in the regex, if pcre_callout has been set, the external
3599 function is called. It is provided with data in a structure called
3600 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3601 matching continues; if it returns a non-zero value, the match at the current
3602 point fails. However, backtracking will occur if possible. [This was changed
3603 later and other features added - see item 49 below.]
3604
3605 29. pcretest is upgraded to test the callout functionality. It provides a
3606 callout function that displays information. By default, it shows the start of
3607 the match and the current position in the text. There are some new data escapes
3608 to vary what happens:
3609
3610 \C+ in addition, show current contents of captured substrings
3611 \C- do not supply a callout function
3612 \C!n return 1 when callout number n is reached
3613 \C!n!m return 1 when callout number n is reached for the mth time
3614
3615 30. If pcregrep was called with the -l option and just a single file name, it
3616 output "<stdin>" if a match was found, instead of the file name.
3617
3618 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3619 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3620 pcre_exec(). This saves a malloc/free per call. The default value of
3621 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3622 when configuring.
3623
3624 32. The default maximum size of a compiled pattern is 64K. There have been a
3625 few cases of people hitting this limit. The code now uses macros to handle the
3626 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3627 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3628 configuring. Tests 2 and 5 work only with 2-byte links because they output
3629 debugging information about compiled patterns.
3630
3631 33. Internal code re-arrangements:
3632
3633 (a) Moved the debugging function for printing out a compiled regex into
3634 its own source file (printint.c) and used #include to pull it into
3635 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3636 separate copies.
3637
3638 (b) Defined the list of op-code names for debugging as a macro in
3639 internal.h so that it is next to the definition of the opcodes.
3640
3641 (c) Defined a table of op-code lengths for simpler skipping along compiled
3642 code. This is again a macro in internal.h so that it is next to the
3643 definition of the opcodes.
3644
3645 34. Added support for recursive calls to individual subpatterns, along the
3646 lines of Robin Houston's patch (but implemented somewhat differently).
3647
3648 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3649 allow it to read and process whole directories in Win32. This code was
3650 contributed by Lionel Fourquaux; it has not been tested by me.
3651
3652 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3653 used to name a group. Names consist of alphanumerics and underscores, and must
3654 be unique. Back references use the syntax (?P=name) and recursive calls use
3655 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3656 numbers. The function pcre_fullinfo() can be used after compilation to extract
3657 a name/number map. There are three relevant calls:
3658
3659 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3660 PCRE_INFO_NAMECOUNT yields the number of entries
3661 PCRE_INFO_NAMETABLE yields a pointer to the map.
3662
3663 The map is a vector of fixed-size entries. The size of each entry depends on
3664 the length of the longest name used. The first two bytes of each entry are the
3665 group number, most significant byte first. There follows the corresponding
3666 name, zero terminated. The names are in alphabetical order.
3667
3668 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3669 case instead of 255. Making it the same both with and without UTF-8 support
3670 means that the same test output works with both.
3671
3672 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3673 calling malloc() with a zero argument.
3674
3675 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3676 optimization. I've improved things by keeping a bitmap of backreferences with
3677 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3678 fact referenced, the optimization can be applied. It is unlikely that a
3679 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3680 the match to follow \n) will appear inside brackets with a number greater than
3681 31, but if it does, any back reference > 31 suppresses the optimization.
3682
3683 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3684 of disabling numbered capturing parentheses. Any opening parenthesis that is
3685 not followed by ? behaves as if it were followed by ?: but named parentheses
3686 can still be used for capturing (and they will acquire numbers in the usual
3687 way).
3688
3689 41. Redesigned the return codes from the match() function into yes/no/error so
3690 that errors can be passed back from deep inside the nested calls. A malloc
3691 failure while inside a recursive subpattern call now causes the
3692 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3693
3694 42. It is now possible to set a limit on the number of times the match()
3695 function is called in a call to pcre_exec(). This facility makes it possible to
3696 limit the amount of recursion and backtracking, though not in a directly
3697 obvious way, because the match() function is used in a number of different
3698 circumstances. The count starts from zero for each position in the subject
3699 string (for non-anchored patterns). The default limit is, for compatibility, a
3700 large number, namely 10 000 000. You can change this in two ways:
3701
3702 (a) When configuring PCRE before making, you can use --with-match-limit=n
3703 to set a default value for the compiled library.
3704
3705 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3706 a different value is set. See 45 below.
3707
3708 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3709
3710 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3711 of things that can be changed at compile time. The first argument specifies
3712 what is wanted and the second points to where the information is to be placed.
3713 The current list of available information is:
3714
3715 PCRE_CONFIG_UTF8
3716
3717 The output is an integer that is set to one if UTF-8 support is available;
3718 otherwise it is set to zero.
3719
3720 PCRE_CONFIG_NEWLINE
3721
3722 The output is an integer that it set to the value of the code that is used for
3723 newline. It is either LF (10) or CR (13).
3724
3725 PCRE_CONFIG_LINK_SIZE
3726
3727 The output is an integer that contains the number of bytes used for internal
3728 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3729
3730 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3731
3732 The output is an integer that contains the threshold above which the POSIX
3733 interface uses malloc() for output vectors. See item 31 above.
3734
3735 PCRE_CONFIG_MATCH_LIMIT
3736
3737 The output is an unsigned integer that contains the default limit of the number
3738 of match() calls in a pcre_exec() execution. See 42 above.
3739
3740 44. pcretest has been upgraded by the addition of the -C option. This causes it
3741 to extract all the available output from the new pcre_config() function, and to
3742 output it. The program then exits immediately.
3743
3744 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3745 order to support additional features. One way would have been to define
3746 pcre_exec2() (for example) with extra arguments, but this would not have been
3747 extensible, and would also have required all calls to the original function to
3748 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3749 is used for passing in "extra" data from pcre_study().
3750
3751 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3752 contains the following fields:
3753
3754 flags a bitmap indicating which of the following fields are set
3755 study_data opaque data from pcre_study()
3756 match_limit a way of specifying a limit on match() calls for a specific
3757 call to pcre_exec()
3758 callout_data data for callouts (see 49 below)
3759
3760 The flag bits are also defined in pcre.h, and are
3761
3762 PCRE_EXTRA_STUDY_DATA
3763 PCRE_EXTRA_MATCH_LIMIT
3764 PCRE_EXTRA_CALLOUT_DATA
3765
3766 The pcre_study() function now returns one of these new pcre_extra blocks, with
3767 the actual study data pointed to by the study_data field, and the
3768 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3769 before. That is, this change is entirely upwards-compatible and requires no
3770 change to existing code.
3771
3772 If you want to pass in additional data to pcre_exec(), you can either place it
3773 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3774 block.
3775
3776 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3777 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3778 times with different match limits, until it finds the minimum value needed for
3779 pcre_exec() to complete. The value is then output. This can be instructive; for
3780 most simple matches the number is quite small, but for pathological cases it
3781 gets very large very quickly.
3782
3783 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3784 returns the size of the data block pointed to by the study_data field in a
3785 pcre_extra block, that is, the value that was passed as the argument to
3786 pcre_malloc() when PCRE was getting memory in which to place the information
3787 created by pcre_study(). The fourth argument should point to a size_t variable.
3788 pcretest has been extended so that this information is shown after a successful
3789 pcre_study() call when information about the compiled regex is being displayed.
3790
3791 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3792 because what follows is always an absolute path. (Later: it turns out that this
3793 is more than cosmetic for MinGW, because it doesn't like empty path
3794 components.)
3795
3796 49. Some changes have been made to the callout feature (see 28 above):
3797
3798 (i) A callout function now has three choices for what it returns:
3799
3800 0 => success, carry on matching
3801 > 0 => failure at this point, but backtrack if possible
3802 < 0 => serious error, return this value from pcre_exec()
3803
3804 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3805 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3806 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3807 use by callout functions. It will never be used by PCRE itself.
3808
3809 (ii) The pcre_extra structure (see 45 above) has a void * field called
3810 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3811 pcre_callout_block structure has a field of the same name. The contents of
3812 the field passed in the pcre_extra structure are passed to the callout
3813 function in the corresponding field in the callout block. This makes it
3814 easier to use the same callout-containing regex from multiple threads. For
3815 testing, the pcretest program has a new data escape
3816
3817 \C*n pass the number n (may be negative) as callout_data
3818
3819 If the callout function in pcretest receives a non-zero value as
3820 callout_data, it returns that value.
3821
3822 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3823 there were some redundant $(CFLAGS) in commands that are now specified as
3824 $(LINK), which already includes $(CFLAGS).
3825
3826 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3827 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3828 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3829 one-byte characters throughout. Note that case-insensitive matching applies
3830 only to characters whose values are less than 256. PCRE doesn't support the
3831 notion of cases for higher-valued characters.
3832
3833 (i) A character class whose characters are all within 0-255 is handled as
3834 a bit map, and the map is inverted for negative classes. Previously, a
3835 character > 255 always failed to match such a class; however it should
3836 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3837
3838 (ii) A negated character class with a single character < 255 is coded as
3839 "not this character" (OP_NOT). This wasn't working properly when the test
3840 character was multibyte, either singly or repeated.
3841
3842 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3843 mode, for example: \x{100}{2,3}.
3844
3845 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3846 singly or repeated) now correctly test multibyte characters. However,
3847 PCRE doesn't recognize any characters with values greater than 255 as
3848 digits, spaces, or word characters. Such characters always match \D, \S,
3849 and \W, and never match \d, \s, or \w.
3850
3851 (v) Classes may now contain characters and character ranges with values
3852 greater than 255. For example: [ab\x{100}-\x{400}].
3853
3854 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3855 PCRE in UTF-8 mode.
3856
3857 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3858 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3859 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3860 value.)
3861
3862 53. The single man page has become too large. I have therefore split it up into
3863 a number of separate man pages. These also give rise to individual HTML pages;
3864 these are now put in a separate directory, and there is an index.html page that
3865 lists them all. Some hyperlinking between the pages has been installed.
3866
3867 54. Added convenience functions for handling named capturing parentheses.
3868
3869 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3870 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3871 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3872 are faulted.
3873
3874 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3875 calling configure. These values are used when compiling the dftables.c program
3876 which is run to generate the source of the default character tables. They
3877 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3878 you will need to set these values.
3879
3880 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3881
3882
3883 Version 3.9 02-Jan-02
3884 ---------------------
3885
3886 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3887
3888 2. If --disable-static was given, the building process failed when trying to
3889 build pcretest and pcregrep. (For some reason it was using libtool to compile
3890 them, which is not right, as they aren't part of the library.)
3891
3892
3893 Version 3.8 18-Dec-01
3894 ---------------------
3895
3896 1. The experimental UTF-8 code was completely screwed up. It was packing the
3897 bytes in the wrong order. How dumb can you get?
3898
3899
3900 Version 3.7 29-Oct-01
3901 ---------------------
3902
3903 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3904 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3905 this didn't happen under Solaris 8, where I normally test things.
3906
3907 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3908 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3909 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3910 this fix an hour or so after the initial 3.7 release.)
3911
3912
3913 Version 3.6 23-Oct-01
3914 ---------------------
3915
3916 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3917 offsets passed as NULL with zero offset count.
3918
3919 2. The config.guess and config.sub files had not been updated when I moved to
3920 the latest autoconf.
3921
3922
3923 Version 3.5 15-Aug-01
3924 ---------------------
3925
3926 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3927 had been forgotten.
3928
3929 2. By using declared but undefined structures, we can avoid using "void"
3930 definitions in pcre.h while keeping the internal definitions of the structures
3931 private.
3932
3933 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3934 user point of view, this means that both static and shared libraries are built
3935 by default, but this can be individually controlled. More of the work of
3936 handling this static/shared cases is now inside libtool instead of PCRE's make
3937 file.
3938
3939 4. The pcretest utility is now installed along with pcregrep because it is
3940 useful for users (to test regexs) and by doing this, it automatically gets
3941 relinked by libtool. The documentation has been turned into a man page, so
3942 there are now .1, .txt, and .html versions in /doc.
3943
3944 5. Upgrades to pcregrep:
3945 (i) Added long-form option names like gnu grep.
3946 (ii) Added --help to list all options with an explanatory phrase.
3947 (iii) Added -r, --recursive to recurse into sub-directories.
3948 (iv) Added -f, --file to read patterns from a file.
3949
3950 6. pcre_exec() was referring to its "code" argument before testing that
3951 argument for NULL (and giving an error if it was NULL).
3952
3953 7. Upgraded Makefile.in to allow for compiling in a different directory from
3954 the source directory.
3955
3956 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3957 options bits, the pointer it was passed was to an int instead of to an unsigned
3958 long int. This mattered only on 64-bit systems.
3959
3960 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3961 generated) instead of pcre.in, which it its source. Also made the same change
3962 in several of the .c files.
3963
3964 10. A new release of gcc defines printf() as a macro, which broke pcretest
3965 because it had an ifdef in the middle of a string argument for printf(). Fixed
3966 by using separate calls to printf().
3967
3968 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3969 script, to force use of CR or LF instead of \n in the source. On non-Unix
3970 systems, the value can be set in config.h.
3971
3972 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3973 absolute limit. Changed the text of the error message to make this clear, and
3974 likewise updated the man page.
3975
3976 13. The limit of 99 on the number of capturing subpatterns has been removed.
3977 The new limit is 65535, which I hope will not be a "real" limit.
3978
3979
3980 Version 3.4 22-Aug-00
3981 ---------------------
3982
3983 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3984
3985 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3986
3987
3988 Version 3.3 01-Aug-00
3989 ---------------------
3990
3991 1. If an octal character was given, but the value was greater than \377, it
3992 was not getting masked to the least significant bits, as documented. This could
3993 lead to crashes in some systems.
3994
3995 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3996 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3997
3998 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3999 These just pass their arguments on to (pcre_free)(), but they are provided
4000 because some uses of PCRE bind it to non-C systems that can call its functions,
4001 but cannot call free() or pcre_free() directly.
4002
4003 4. Add "make test" as a synonym for "make check". Corrected some comments in
4004 the Makefile.
4005
4006 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4007 Makefile.
4008
4009 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4010 command called pgrep for grepping around the active processes.
4011
4012 7. Added the beginnings of support for UTF-8 character strings.
4013
4014 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4015 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4016 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4017 out for the ar command.)
4018
4019
4020 Version 3.2 12-May-00
4021 ---------------------
4022
4023 This is purely a bug fixing release.
4024
4025 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4026 of ZA. This was just one example of several cases that could provoke this bug,
4027 which was introduced by change 9 of version 2.00. The code for breaking
4028 infinite loops after an iteration that matches an empty string was't working
4029 correctly.
4030
4031 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4032 when matched against abbab (for example). After matching an empty string, it
4033 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4034 caused it to match further down the string than it should.
4035
4036 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4037 was there because it doesn't seem to be needed, and it causes trouble on some
4038 systems, as it is not a Standard C header. It has been removed.
4039
4040 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4041 were reported on the Macintosh. The changes were from
4042
4043 while ((c = *(++ptr)) != 0 && c != '\n');
4044 to
4045 while ((c = *(++ptr)) != 0 && c != '\n') ;
4046
4047 Totally extraordinary, but if that's what it takes...
4048
4049 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4050 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4051 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4052 assumes the way PCRE uses memmove() (always moving upwards).
4053
4054 6. PCRE is being used in one environment where strchr() is not available. There
4055 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4056 faster code anyway.
4057
4058
4059 Version 3.1 09-Feb-00
4060 ---------------------
4061
4062 The only change in this release is the fixing of some bugs in Makefile.in for
4063 the "install" target:
4064
4065 (1) It was failing to install pcreposix.h.
4066
4067 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4068
4069
4070 Version 3.0 01-Feb-00
4071 ---------------------
4072
4073 1. Add support for the /+ modifier to perltest (to output $` like it does in
4074 pcretest).
4075
4076 2. Add support for the /g modifier to perltest.
4077
4078 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4079 matches null strings.
4080
4081 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4082 pattern. Perl treats empty patterns specially - it reuses the most recent
4083 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4084 effect.
4085
4086 5. The POSIX interface was broken in that it was just handing over the POSIX
4087 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4088 required a bigger vector, with some working space on the end. This means that
4089 the POSIX wrapper now has to get and free some memory, and copy the results.
4090
4091 6. Added some simple autoconf support, placing the test data and the
4092 documentation in separate directories, re-organizing some of the
4093 information files, and making it build pcre-config (a GNU standard). Also added
4094 libtool support for building PCRE as a shared library, which is now the
4095 default.
4096
4097 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4098 09 are not valid octal constants. Single digits will be used for minor values
4099 less than 10.
4100
4101 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4102 existing programs that set these in the POSIX interface can use PCRE without
4103 modification.
4104
4105 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4106 return all that pcre_info() returns, plus additional data. The pcre_info()
4107 function is retained for compatibility, but is considered to be obsolete.
4108
4109 10. Added experimental recursion feature (?R) to handle one common case that
4110 Perl 5.6 will be able to do with (?p{...}).
4111
4112 11. Added support for POSIX character classes like [:alpha:], which Perl is
4113 adopting.
4114
4115
4116 Version 2.08 31-Aug-99
4117 ----------------------
4118
4119 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4120 trying to match at the startoffset position, but instead was moving forward to
4121 the next newline as if a previous match had failed.
4122
4123 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4124 and could get into a loop if a null string was matched other than at the start
4125 of the subject.
4126
4127 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4128 be distinguished at compile time, and for completeness also added PCRE_DATE.
4129
4130 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4131 in GnuWin32 environments.
4132
4133
4134 Version 2.07 29-Jul-99
4135 ----------------------
4136
4137 1. The documentation is now supplied in plain text form and HTML as well as in
4138 the form of man page sources.
4139
4140 2. C++ compilers don't like assigning (void *) values to other pointer types.
4141 In particular this affects malloc(). Although there is no problem in Standard
4142 C, I've put in casts to keep C++ compilers happy.
4143
4144 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4145 should be (const char *).
4146
4147 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4148 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4149 However, I haven't made this a standard facility. The documentation doesn't
4150 mention it, and the Makefile doesn't support it.
4151
4152 5. The Makefile now contains an "install" target, with editable destinations at
4153 the top of the file. The pcretest program is not installed.
4154
4155 6. pgrep -V now gives the PCRE version number and date.
4156
4157 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4158 causing the entire string to be ignored, instead of just the last character.
4159
4160 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4161 non-matching string, it can take a very, very long time, even for strings of
4162 quite modest length, because of the nested recursion. PCRE now does better in
4163 some of these cases. It does this by remembering the last required literal
4164 character in the pattern, and pre-searching the subject to ensure it is present
4165 before running the real match. In other words, it applies a heuristic to detect
4166 some types of certain failure quickly, and in the above example, if presented
4167 with a string that has no trailing " it gives "no match" very quickly.
4168
4169 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4170 other alternatives are tried instead.
4171
4172
4173 Version 2.06 09-Jun-99
4174 ----------------------
4175
4176 1. Change pcretest's output for amount of store used to show just the code
4177 space, because the remainder (the data block) varies in size between 32-bit and
4178 64-bit systems.
4179
4180 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4181 start matching at. This allows lookbehinds to work when searching for multiple
4182 occurrences in a string.
4183
4184 3. Added additional options to pcretest for testing multiple occurrences:
4185
4186 /+ outputs the rest of the string that follows a match
4187 /g loops for multiple occurrences, using the new startoffset argument
4188 /G loops for multiple occurrences by passing an incremented pointer
4189
4190 4. PCRE wasn't doing the "first character" optimization for patterns starting
4191 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4192 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4193 the letter 'x'. On long subject strings, this gives a significant speed-up.
4194
4195
4196 Version 2.05 21-Apr-99
4197 ----------------------
4198
4199 1. Changed the type of magic_number from int to long int so that it works
4200 properly on 16-bit systems.
4201
4202 2. Fixed a bug which caused patterns starting with .* not to work correctly
4203 when the subject string contained newline characters. PCRE was assuming
4204 anchoring for such patterns in all cases, which is not correct because .* will
4205 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4206 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4207 must be retried after every newline in the subject.
4208
4209
4210 Version 2.04 18-Feb-99
4211 ----------------------
4212
4213 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4214 computation of the store needed to hold the pattern was incorrect (too large).
4215 If such patterns were nested a few deep, this could multiply and become a real
4216 problem.
4217
4218 2. Added /M option to pcretest to show the memory requirement of a specific
4219 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4220
4221 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4222 compiled in such a way that the backtracking after subsequent failure was
4223 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4224 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4225
4226
4227 Version 2.03 02-Feb-99
4228 ----------------------
4229
4230 1. Fixed typo and small mistake in man page.
4231
4232 2. Added 4th condition (GPL supersedes if conflict) and created separate
4233 LICENCE file containing the conditions.
4234
4235 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4236 Perl, that is the internal \ allows the delimiter to be included in the
4237 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4238 the final delimiter, add \ to the end of the pattern (to test the error).
4239
4240 4. Added the convenience functions for extracting substrings after a successful
4241 match. Updated pcretest to make it able to test these functions.
4242
4243
4244 Version 2.02 14-Jan-99
4245 ----------------------
4246
4247 1. Initialized the working variables associated with each extraction so that
4248 their saving and restoring doesn't refer to uninitialized store.
4249
4250 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4251 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4252 fix the problem.
4253
4254 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4255 calls, and wasn't printing the correct value for compiling calls. Increased the
4256 default value of LOOPREPEAT, and the number of significant figures in the
4257 times.
4258
4259 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4260
4261 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4262 a building problem on Windows NT with a FAT file system.
4263
4264
4265 Version 2.01 21-Oct-98
4266 ----------------------
4267
4268 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4269 to character tables built by pcre_maketables() in the current locale. If NULL
4270 is passed, the default tables are used.
4271
4272
4273 Version 2.00 24-Sep-98
4274 ----------------------
4275
4276 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4277 it any more.
4278
4279 2. Allow quantification of (?>) groups, and make it work correctly.
4280
4281 3. The first character computation wasn't working for (?>) groups.
4282
4283 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4284 end of the subject) and add 5.005's \z, which really does match only at the
4285 very end of the subject.
4286
4287 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4288
4289 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4290 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4291 localized options. All options to pcre_study() were also removed.
4292
4293 7. Add other new features from 5.005:
4294
4295 $(?<= positive lookbehind
4296 $(?<! negative lookbehind
4297 (?imsx-imsx) added the unsetting capability
4298 such a setting is global if at outer level; local otherwise
4299 (?imsx-imsx:) non-capturing groups with option setting
4300 (?(cond)re|re) conditional pattern matching
4301
4302 A backreference to itself in a repeated group matches the previous
4303 captured string.
4304
4305 8. General tidying up of studying (both automatic and via "study")
4306 consequential on the addition of new assertions.
4307
4308 9. As in 5.005, unlimited repeated groups that could match an empty substring
4309 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4310 runtime if any iteration does actually match an empty substring.
4311
4312 10. Include the RunTest script in the distribution.
4313
4314 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4315 discrepancies, some of which were old and were also with respect to 5.004. They
4316 have now been fixed.
4317
4318
4319 Version 1.09 28-Apr-98
4320 ----------------------
4321
4322 1. A negated single character class followed by a quantifier with a minimum
4323 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4324 program crashes, or just wrong answers. This did not apply to negated classes
4325 containing more than one character, or to minima other than one.
4326
4327
4328 Version 1.08 27-Mar-98
4329 ----------------------
4330
4331 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4332
4333 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4334 latter must appear before anything that relies on it in the pattern.
4335
4336
4337 Version 1.07 16-Feb-98
4338 ----------------------
4339
4340 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4341 repeat of a potentially empty string).
4342
4343
4344 Version 1.06 23-Jan-98
4345 ----------------------
4346
4347 1. Added Markus Oberhumer's little patches for C++.
4348
4349 2. Literal strings longer than 255 characters were broken.
4350
4351
4352 Version 1.05 23-Dec-97
4353 ----------------------
4354
4355 1. Negated character classes containing more than one character were failing if
4356 PCRE_CASELESS was set at run time.
4357
4358
4359 Version 1.04 19-Dec-97
4360 ----------------------
4361
4362 1. Corrected the man page, where some "const" qualifiers had been omitted.
4363
4364 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4365 input syntax.
4366
4367 3. Fixed memory leak which occurred when a regex with back references was
4368 matched with an offsets vector that wasn't big enough. The temporary memory
4369 that is used in this case wasn't being freed if the match failed.
4370
4371 4. Tidied pcretest to ensure it frees memory that it gets.
4372
4373 5. Temporary memory was being obtained in the case where the passed offsets
4374 vector was exactly big enough.
4375
4376 6. Corrected definition of offsetof() from change 5 below.
4377
4378 7. I had screwed up change 6 below and broken the rules for the use of
4379 setjmp(). Now fixed.
4380
4381
4382 Version 1.03 18-Dec-97
4383 ----------------------
4384
4385 1. A erroneous regex with a missing opening parenthesis was correctly
4386 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4387 on some systems.
4388
4389 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4390 it was reported that one broken compiler failed on the former because "code" is
4391 also an independent variable.
4392
4393 3. The erroneous regex a[]b caused an array overrun reference.
4394
4395 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4396 fail on data ending with that character. (It was going on too far, and checking
4397 the next character, typically a binary zero.) This was specific to the
4398 optimized code for single-character negative classes.
4399
4400 5. Added a contributed patch from the TIN world which does the following:
4401
4402 + Add an undef for memmove, in case the the system defines a macro for it.
4403
4404 + Add a definition of offsetof(), in case there isn't one. (I don't know
4405 the reason behind this - offsetof() is part of the ANSI standard - but
4406 it does no harm).
4407
4408 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4409 most of the places where whitespace preceded '#'. I have given up and
4410 allowed the remaining 2 cases to be at the margin.
4411
4412 + Rename some variables in pcre to eliminate shadowing. This seems very
4413 pedantic, but does no harm, of course.
4414
4415 6. Moved the call to setjmp() into its own function, to get rid of warnings
4416 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4417
4418 7. Constructs such as \d{8,} were compiling into the equivalent of
4419 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4420 outcome, but in this particular case used more store than had been allocated,
4421 which caused the bug to be discovered because it threw up an internal error.
4422
4423 8. The debugging code in both pcre and pcretest for outputting the compiled
4424 form of a regex was going wrong in the case of back references followed by
4425 curly-bracketed repeats.
4426
4427
4428 Version 1.02 12-Dec-97
4429 ----------------------
4430
4431 1. Typos in pcre.3 and comments in the source fixed.
4432
4433 2. Applied a contributed patch to get rid of places where it used to remove
4434 'const' from variables, and fixed some signed/unsigned and uninitialized
4435 variable warnings.
4436
4437 3. Added the "runtest" target to Makefile.
4438
4439 4. Set default compiler flag to -O2 rather than just -O.
4440
4441
4442 Version 1.01 19-Nov-97
4443 ----------------------
4444
4445 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4446 like /([ab]*)*/, that is, for classes with more than one character in them.
4447
4448 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4449 as /((?>a*))*/ (a PCRE_EXTRA facility).
4450
4451
4452 Version 1.00 18-Nov-97
4453 ----------------------
4454
4455 1. Added compile-time macros to support systems such as SunOS4 which don't have
4456 memmove() or strerror() but have other things that can be used instead.
4457
4458 2. Arranged that "make clean" removes the executables.
4459
4460
4461 Version 0.99 27-Oct-97
4462 ----------------------
4463
4464 1. Fixed bug in code for optimizing classes with only one character. It was
4465 initializing a 32-byte map regardless, which could cause it to run off the end
4466 of the memory it had got.
4467
4468 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4469
4470
4471 Version 0.98 22-Oct-97
4472 ----------------------
4473
4474 1. Fixed bug in code for handling temporary memory usage when there are more
4475 back references than supplied space in the ovector. This could cause segfaults.
4476
4477
4478 Version 0.97 21-Oct-97
4479 ----------------------
4480
4481 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4482
4483 2. Optimized negated single characters not to use a bit map.
4484
4485 3. Brought error texts together as macro definitions; clarified some of them;
4486 fixed one that was wrong - it said "range out of order" when it meant "invalid
4487 escape sequence".
4488
4489 4. Changed some char * arguments to const char *.
4490
4491 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4492
4493 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4494 pcretest.
4495
4496
4497 Version 0.96 16-Oct-97
4498 ----------------------
4499
4500 1. Added a simple "pgrep" utility to the distribution.
4501
4502 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4503 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4504 where "ddd" means "one or more decimal digits".
4505
4506 3. Fixed serious bug. If a pattern had a back reference, but the call to
4507 pcre_exec() didn't supply a large enough ovector to record the related
4508 identifying subpattern, the match always failed. PCRE now remembers the number
4509 of the largest back reference, and gets some temporary memory in which to save
4510 the offsets during matching if necessary, in order to ensure that
4511 backreferences always work.
4512
4513 4. Increased the compatibility with Perl in a number of ways:
4514
4515 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4516 to request this handling. The option can be set at compile or exec time.
4517
4518 (b) $ matches before a terminating newline by default; an option
4519 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4520 mode). The option can be set at compile or exec time.
4521
4522 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4523 the same as Perl's. If the decimal number it represents is less than 10
4524 or there aren't that many previous left capturing parentheses, an octal
4525 escape is read. Inside a character class, it's always an octal escape,
4526 even if it is a single digit.
4527
4528 (d) An escaped but undefined alphabetic character is taken as a literal,
4529 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4530 escapes.
4531
4532 (e) {0} is now permitted. (The previous item is removed from the compiled
4533 pattern).
4534
4535 5. Changed all the names of code files so that the basic parts are no longer
4536 than 10 characters, and abolished the teeny "globals.c" file.
4537
4538 6. Changed the handling of character classes; they are now done with a 32-byte
4539 bit map always.
4540
4541 7. Added the -d and /D options to pcretest to make it possible to look at the
4542 internals of compilation without having to recompile pcre.
4543
4544
4545 Version 0.95 23-Sep-97
4546 ----------------------
4547
4548 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4549 \x20 at the start of a run of normal characters. These were being treated as
4550 real characters, instead of the source characters being re-checked.
4551
4552
4553 Version 0.94 18-Sep-97
4554 ----------------------
4555
4556 1. The functions are now thread-safe, with the caveat that the global variables
4557 containing pointers to malloc() and free() or alternative functions are the
4558 same for all threads.
4559
4560 2. Get pcre_study() to generate a bitmap of initial characters for non-
4561 anchored patterns when this is possible, and use it if passed to pcre_exec().
4562
4563
4564 Version 0.93 15-Sep-97
4565 ----------------------
4566
4567 1. /(b)|(:+)/ was computing an incorrect first character.
4568
4569 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4570 but not actually doing anything yet.
4571
4572 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4573 as Perl does (e.g. [-az] or [az-]).
4574
4575 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4576 all possible positions.
4577
4578 5. Split up into different modules to avoid including unneeded functions in a
4579 compiled binary. However, compile and exec are still in one module. The "study"
4580 function is split off.
4581
4582 6. The character tables are now in a separate module whose source is generated
4583 by an auxiliary program - but can then be edited by hand if required. There are
4584 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4585 toupper() in the code.
4586
4587 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4588 make them global. Abolish the function for setting them, as the caller can now
4589 set them directly.
4590
4591
4592 Version 0.92 11-Sep-97
4593 ----------------------
4594
4595 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4596 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4597
4598 2. Caseless matching was not working in character classes if the characters in
4599 the pattern were in upper case.
4600
4601 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4602
4603 4. Make PCRE_ANCHORED public and accept as a compile option.
4604
4605 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4606 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4607 pass them.
4608
4609 6. Give an error if bad option bits passed at compile or run time.
4610
4611 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4612 pcretest to cause it to pass that flag.
4613
4614 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4615 options, and the first character, if set.
4616
4617 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4618
4619
4620 Version 0.91 10-Sep-97
4621 ----------------------
4622
4623 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4624 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4625
4626 2. PCRE was looping on encountering an indefinitely repeated back reference to
4627 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4628 Perl does - treats the match as successful.
4629
4630 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5