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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 860 - (show annotations)
Mon Jan 9 20:12:58 2012 UTC (7 years, 10 months ago) by zherczeg
File size: 224633 byte(s)
rename PCRE_SCHAR16 to PCRE_UCHAR16 and JIT compiler update
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.30
5 ------------
6
7 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
8 name is defined in ctype.h.
9
10 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
11 only in quite long subpatterns.
12
13 3. Removed the function pcre_info(), which has been obsolete and deprecated
14 since it was replaced by pcre_fullinfo() in February 2000.
15
16 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
17 match a (*MARK), and the match failed at the start of the subject, a
18 reference to memory before the start of the subject could occur. This bug
19 was introduced by fix 17 of release 8.21.
20
21 5. A reference to an unset group with zero minimum repetition was giving
22 totally wrong answers (in non-JavaScript-compatibility mode). For example,
23 /(another)?(\1?)test/ matched against "hello world test". This bug was
24 introduced in release 8.13.
25
26 6. Add support for 16-bit character strings (a large amount of work involving
27 many changes and refactorings).
28
29 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31 from a file.
32
33 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34 rounding is not applied in this particular case).
35
36
37 Version 8.21 12-Dec-2011
38 ------------------------
39
40 1. Updating the JIT compiler.
41
42 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
43 are added as well.
44
45 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
46 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
47 calling _pcre_jit_exec. Some extra comments are added.
48
49 4. (*MARK) settings inside atomic groups that do not contain any capturing
50 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
51 was introduced by change 18 for 8.20.
52
53 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
54 ECMA-262 standard.
55
56 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
57 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
58 This bug was probably introduced by change 9 of 8.13.
59
60 7. While fixing 6 above, I noticed that a number of other items were being
61 incorrectly rejected as "not fixed length". This arose partly because newer
62 opcodes had not been added to the fixed-length checking code. I have (a)
63 corrected the bug and added tests for these items, and (b) arranged for an
64 error to occur if an unknown opcode is encountered while checking for fixed
65 length instead of just assuming "not fixed length". The items that were
66 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
67 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
68 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
69
70 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
71 being incorrectly compiled and would have given unpredicatble results.
72
73 9. A possessively repeated subpattern with minimum repeat count greater than
74 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
75 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
76 the first (A) could occur when it should not.
77
78 10. Add a cast and remove a redundant test from the code.
79
80 11. JIT should use pcre_malloc/pcre_free for allocation.
81
82 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
83 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
84 is anything other than /usr, -L is still shown).
85
86 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
87
88 14. Perl does not support \N without a following name in a [] class; PCRE now
89 also gives an error.
90
91 15. If a forward reference was repeated with an upper limit of around 2000,
92 it caused the error "internal error: overran compiling workspace". The
93 maximum number of forward references (including repeats) was limited by the
94 internal workspace, and dependent on the LINK_SIZE. The code has been
95 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
96 the default depends on LINK_SIZE. There is a new upper limit (for safety)
97 of around 200,000 forward references. While doing this, I also speeded up
98 the filling in of repeated forward references.
99
100 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
101 incorrectly expecting the subject to contain another "a" after the start.
102
103 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
104 in the match, the SKIP should be ignored. This was not happening; instead
105 the SKIP was being treated as NOMATCH. For patterns such as
106 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
107 tested.
108
109 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
110 now much more compatible with Perl, in particular in cases where the result
111 is a non-match for a non-anchored pattern. For example, if
112 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
113 "m", where previously it did not return a name. A side effect of this
114 change is that for partial matches, the last encountered mark name is
115 returned, as for non matches. A number of tests that were previously not
116 Perl-compatible have been moved into the Perl-compatible test files. The
117 refactoring has had the pleasing side effect of removing one argument from
118 the match() function, thus reducing its stack requirements.
119
120 19. If the /S+ option was used in pcretest to study a pattern using JIT,
121 subsequent uses of /S (without +) incorrectly behaved like /S+.
122
123 21. Retrieve executable code size support for the JIT compiler and fixing
124 some warnings.
125
126 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
127 not work when the shorter character appeared right at the end of the
128 subject string.
129
130 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
131 systems.
132
133 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
134 output it when the /M option is used in pcretest.
135
136 25. The CheckMan script was not being included in the distribution. Also, added
137 an explicit "perl" to run Perl scripts from the PrepareRelease script
138 because this is reportedly needed in Windows.
139
140 26. If study data was being save in a file and studying had not found a set of
141 "starts with" bytes for the pattern, the data written to the file (though
142 never used) was taken from uninitialized memory and so caused valgrind to
143 complain.
144
145 27. Updated RunTest.bat as provided by Sheri Pierce.
146
147 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
148
149 29. Computation of memory usage for the table of capturing group names was
150 giving an unnecessarily large value.
151
152
153 Version 8.20 21-Oct-2011
154 ------------------------
155
156 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
157 a POSIX class. After further experiments with Perl, which convinced me that
158 Perl has bugs and confusions, a closing square bracket is no longer allowed
159 in a POSIX name. This bug also affected patterns with classes that started
160 with full stops.
161
162 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
163 captured substring, but while checking the failing first alternative,
164 substring 1 is temporarily captured. If the output vector supplied to
165 pcre_exec() was not big enough for this capture, the yield of the function
166 was still zero ("insufficient space for captured substrings"). This cannot
167 be totally fixed without adding another stack variable, which seems a lot
168 of expense for a edge case. However, I have improved the situation in cases
169 such as /(a)(b)x|abc/ matched against "abc", where the return code
170 indicates that fewer than the maximum number of slots in the ovector have
171 been set.
172
173 3. Related to (2) above: when there are more back references in a pattern than
174 slots in the output vector, pcre_exec() uses temporary memory during
175 matching, and copies in the captures as far as possible afterwards. It was
176 using the entire output vector, but this conflicts with the specification
177 that only 2/3 is used for passing back captured substrings. Now it uses
178 only the first 2/3, for compatibility. This is, of course, another edge
179 case.
180
181 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
182 main code base, and can be used by building with --enable-jit. When this is
183 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
184 runtime --no-jit option is given.
185
186 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
187 ovector, the return from the function was zero, implying that there were
188 other matches that did not fit. The correct "exactly full" value is now
189 returned.
190
191 6. If a subpattern that was called recursively or as a subroutine contained
192 (*PRUNE) or any other control that caused it to give a non-standard return,
193 invalid errors such as "Error -26 (nested recursion at the same subject
194 position)" or even infinite loops could occur.
195
196 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
197 computing the minimum length on reaching *ACCEPT, and so ended up with the
198 wrong value of 1 rather than 0. Further investigation indicates that
199 computing a minimum subject length in the presence of *ACCEPT is difficult
200 (think back references, subroutine calls), and so I have changed the code
201 so that no minimum is registered for a pattern that contains *ACCEPT.
202
203 8. If (*THEN) was present in the first (true) branch of a conditional group,
204 it was not handled as intended. [But see 16 below.]
205
206 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
207 Sheri Pierce.
208
209 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
210 the first byte in a match must be "a".
211
212 11. Change 17 for 8.13 increased the recursion depth for patterns like
213 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
214 pattern contains any instances of (*THEN). If it does not, the old
215 optimizations are restored. It would be nice to do this on a per-group
216 basis, but at the moment that is not feasible.
217
218 12. In some environments, the output of pcretest -C is CRLF terminated. This
219 broke RunTest's code that checks for the link size. A single white space
220 character after the value is now allowed for.
221
222 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
223 For "fr", it uses the Windows-specific input and output files.
224
225 14. If (*THEN) appeared in a group that was called recursively or as a
226 subroutine, it did not work as intended. [But see next item.]
227
228 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
229 pattern fragments (but not containing any | characters). If A and B are
230 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
231 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
232 D. In other words, Perl considers parentheses that do not contain any |
233 characters to be part of a surrounding alternative, whereas PCRE was
234 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
235 differently. PCRE now behaves in the same way as Perl, except in the case
236 of subroutine/recursion calls such as (?1) which have in any case always
237 been different (but PCRE had them first :-).
238
239 16. Related to 15 above: Perl does not treat the | in a conditional group as
240 creating alternatives. Such a group is treated in the same way as an
241 ordinary group without any | characters when processing (*THEN). PCRE has
242 been changed to match Perl's behaviour.
243
244 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
245 RunGrepTest script failed.
246
247 18. Change 22 for version 13 caused atomic groups to use more stack. This is
248 inevitable for groups that contain captures, but it can lead to a lot of
249 stack use in large patterns. The old behaviour has been restored for atomic
250 groups that do not contain any capturing parentheses.
251
252 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
253 suppress the check for a minimum subject length at run time. (If it was
254 given to pcre_exec() or pcre_dfa_exec() it did work.)
255
256 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
257 fail to work when decoding hex characters in data strings in EBCDIC
258 environments.
259
260 21. It appears that in at least one Mac OS environment, the isxdigit() function
261 is implemented as a macro that evaluates to its argument more than once,
262 contravening the C 90 Standard (I haven't checked a later standard). There
263 was an instance in pcretest which caused it to go wrong when processing
264 \x{...} escapes in subject strings. The has been rewritten to avoid using
265 things like p++ in the argument of isxdigit().
266
267
268 Version 8.13 16-Aug-2011
269 ------------------------
270
271 1. The Unicode data tables have been updated to Unicode 6.0.0.
272
273 2. Two minor typos in pcre_internal.h have been fixed.
274
275 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
276 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
277 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
278
279 4. There were a number of related bugs in the code for matching backrefences
280 caselessly in UTF-8 mode when codes for the characters concerned were
281 different numbers of bytes. For example, U+023A and U+2C65 are an upper
282 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
283 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
284 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
285 2-byte code at the end of the subject (it thought there wasn't enough data
286 left).
287
288 5. Comprehensive information about what went wrong is now returned by
289 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
290 as the output vector has at least 2 elements. The offset of the start of
291 the failing character and a reason code are placed in the vector.
292
293 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
294 now returned is for the first byte of the failing character, instead of the
295 last byte inspected. This is an incompatible change, but I hope it is small
296 enough not to be a problem. It makes the returned offset consistent with
297 pcre_exec() and pcre_dfa_exec().
298
299 7. pcretest now gives a text phrase as well as the error number when
300 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
301 failure, the offset and reason code are output.
302
303 8. When \R was used with a maximizing quantifier it failed to skip backwards
304 over a \r\n pair if the subsequent match failed. Instead, it just skipped
305 back over a single character (\n). This seems wrong (because it treated the
306 two characters as a single entity when going forwards), conflicts with the
307 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
308 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
309 has been changed.
310
311 9. Some internal refactoring has changed the processing so that the handling
312 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
313 time (the PCRE_DOTALL option was changed this way some time ago: version
314 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
315 which was always a bit of a fudge. It also means that there is one less
316 argument for the match() function, which reduces its stack requirements
317 slightly. This change also fixes an incompatibility with Perl: the pattern
318 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
319
320 10. More internal refactoring has drastically reduced the number of recursive
321 calls to match() for possessively repeated groups such as (abc)++ when
322 using pcre_exec().
323
324 11. While implementing 10, a number of bugs in the handling of groups were
325 discovered and fixed:
326
327 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
328 (a|)*(?1) gave a compile-time internal error.
329 ((a|)+)+ did not notice that the outer group could match an empty string.
330 (^a|^)+ was not marked as anchored.
331 (.*a|.*)+ was not marked as matching at start or after a newline.
332
333 12. Yet more internal refactoring has removed another argument from the match()
334 function. Special calls to this function are now indicated by setting a
335 value in a variable in the "match data" data block.
336
337 13. Be more explicit in pcre_study() instead of relying on "default" for
338 opcodes that mean there is no starting character; this means that when new
339 ones are added and accidentally left out of pcre_study(), testing should
340 pick them up.
341
342 14. The -s option of pcretest has been documented for ages as being an old
343 synonym of -m (show memory usage). I have changed it to mean "force study
344 for every regex", that is, assume /S for every regex. This is similar to -i
345 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
346 using it. It makes it easier to run collections of tests with and without
347 study enabled, and thereby test pcre_study() more easily. All the standard
348 tests are now run with and without -s (but some patterns can be marked as
349 "never study" - see 20 below).
350
351 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
352 restoration of the capturing data to the outer values was not happening
353 correctly.
354
355 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
356 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
357 pattern had matched an empty string, and so incorrectly returned a no
358 match.
359
360 17. There was optimizing code for the last branch of non-capturing parentheses,
361 and also for the obeyed branch of a conditional subexpression, which used
362 tail recursion to cut down on stack usage. Unfortunately, now that there is
363 the possibility of (*THEN) occurring in these branches, tail recursion is
364 no longer possible because the return has to be checked for (*THEN). These
365 two optimizations have therefore been removed. [But see 8.20/11 above.]
366
367 18. If a pattern containing \R was studied, it was assumed that \R always
368 matched two bytes, thus causing the minimum subject length to be
369 incorrectly computed because \R can also match just one byte.
370
371 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
372 was incorrectly computed.
373
374 20. If /S is present twice on a test pattern in pcretest input, it now
375 *disables* studying, thereby overriding the use of -s on the command line
376 (see 14 above). This is necessary for one or two tests to keep the output
377 identical in both cases.
378
379 21. When (*ACCEPT) was used in an assertion that matched an empty string and
380 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
381
382 22. When an atomic group that contained a capturing parenthesis was
383 successfully matched, but the branch in which it appeared failed, the
384 capturing was not being forgotten if a higher numbered group was later
385 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
386 group 1 to "a", when in fact it should be unset. This applied to multi-
387 branched capturing and non-capturing groups, repeated or not, and also to
388 positive assertions (capturing in negative assertions does not happen
389 in PCRE) and also to nested atomic groups.
390
391 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
392 subject after a captured substring, to make it easier to tell which of a
393 number of identical substrings has been captured.
394
395 24. The way atomic groups are processed by pcre_exec() has been changed so that
396 if they are repeated, backtracking one repetition now resets captured
397 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
398 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
399 "aaa". Previously, it would have been "a". As part of this code
400 refactoring, the way recursive calls are handled has also been changed.
401
402 25. If an assertion condition captured any substrings, they were not passed
403 back unless some other capturing happened later. For example, if
404 (?(?=(a))a) was matched against "a", no capturing was returned.
405
406 26. When studying a pattern that contained subroutine calls or assertions,
407 the code for finding the minimum length of a possible match was handling
408 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
409 group 1 called group 2 while simultaneously a separate group 2 called group
410 1). A stack overflow occurred in this case. I have fixed this by limiting
411 the recursion depth to 10.
412
413 27. Updated RunTest.bat in the distribution to the version supplied by Tom
414 Fortmann. This supports explicit test numbers on the command line, and has
415 argument validation and error reporting.
416
417 28. An instance of \X with an unlimited repeat could fail if at any point the
418 first character it looked at was a mark character.
419
420 29. Some minor code refactoring concerning Unicode properties and scripts
421 should reduce the stack requirement of match() slightly.
422
423 30. Added the '=' option to pcretest to check the setting of unused capturing
424 slots at the end of the pattern, which are documented as being -1, but are
425 not included in the return count.
426
427 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
428 compiled something random. Now it gives a compile-time error (as does
429 Perl).
430
431 32. A *MARK encountered during the processing of a positive assertion is now
432 recorded and passed back (compatible with Perl).
433
434 33. If --only-matching or --colour was set on a pcregrep call whose pattern
435 had alternative anchored branches, the search for a second match in a line
436 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
437 matched the line "0102" twice. The same bug affected patterns that started
438 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
439 twice.
440
441 34. Previously, PCRE did not allow quantification of assertions. However, Perl
442 does, and because of capturing effects, quantifying parenthesized
443 assertions may at times be useful. Quantifiers are now allowed for
444 parenthesized assertions.
445
446 35. A minor code tidy in pcre_compile() when checking options for \R usage.
447
448 36. \g was being checked for fancy things in a character class, when it should
449 just be a literal "g".
450
451 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
452 appearance of a nested POSIX class supersedes an apparent external class.
453 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
454 unescaped square brackets may also appear as part of class names. For
455 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
456 more like Perl. (But see 8.20/1 above.)
457
458 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
459 was because it thought it was \N{name}, which is not supported).
460
461 39. Add minix to OS list not supporting the -S option in pcretest.
462
463 40. PCRE tries to detect cases of infinite recursion at compile time, but it
464 cannot analyze patterns in sufficient detail to catch mutual recursions
465 such as ((?1))((?2)). There is now a runtime test that gives an error if a
466 subgroup is called recursively as a subpattern for a second time at the
467 same position in the subject string. In previous releases this might have
468 been caught by the recursion limit, or it might have run out of stack.
469
470 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
471 happen only once. PCRE was, however incorrectly giving a compile time error
472 "recursive call could loop indefinitely" because it cannot analyze the
473 pattern in sufficient detail. The compile time test no longer happens when
474 PCRE is compiling a conditional subpattern, but actual runaway loops are
475 now caught at runtime (see 40 above).
476
477 42. It seems that Perl allows any characters other than a closing parenthesis
478 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
479 has been changed to be the same.
480
481 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
482 as not to get warnings when autogen.sh is called. Also changed
483 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
484
485 44. To help people who use pcregrep to scan files containing exceedingly long
486 lines, the following changes have been made:
487
488 (a) The default value of the buffer size parameter has been increased from
489 8K to 20K. (The actual buffer used is three times this size.)
490
491 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
492 PCRE is built.
493
494 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
495 to be set at run time.
496
497 (d) Numerical values in pcregrep options can be followed by K or M, for
498 example --buffer-size=50K.
499
500 (e) If a line being scanned overflows pcregrep's buffer, an error is now
501 given and the return code is set to 2.
502
503 45. Add a pointer to the latest mark to the callout data block.
504
505 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
506 partial match of an empty string instead of no match. This was specific to
507 the use of ".".
508
509 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
510 complete match instead of a partial match. This bug was dependent on both
511 the PCRE_UTF8 and PCRE_DOTALL options being set.
512
513 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
514 starting byte set, because \b was not being ignored.
515
516
517 Version 8.12 15-Jan-2011
518 ------------------------
519
520 1. Fixed some typos in the markup of the man pages, and wrote a script that
521 checks for such things as part of the documentation building process.
522
523 2. On a big-endian 64-bit system, pcregrep did not correctly process the
524 --match-limit and --recursion-limit options (added for 8.11). In
525 particular, this made one of the standard tests fail. (The integer value
526 went into the wrong half of a long int.)
527
528 3. If the --colour option was given to pcregrep with -v (invert match), it
529 did strange things, either producing crazy output, or crashing. It should,
530 of course, ignore a request for colour when reporting lines that do not
531 match.
532
533 4. Another pcregrep bug caused similar problems if --colour was specified with
534 -M (multiline) and the pattern match finished with a line ending.
535
536 5. In pcregrep, when a pattern that ended with a literal newline sequence was
537 matched in multiline mode, the following line was shown as part of the
538 match. This seems wrong, so I have changed it.
539
540 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
541 the check for further matches in the same line (so they could be coloured)
542 to overrun the end of the current line. If another match was found, it was
543 incorrectly shown (and then shown again when found in the next line).
544
545 7. If pcregrep was compiled under Windows, there was a reference to the
546 function pcregrep_exit() before it was defined. I am assuming this was
547 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
548 reported by a user. I've moved the definition above the reference.
549
550
551 Version 8.11 10-Dec-2010
552 ------------------------
553
554 1. (*THEN) was not working properly if there were untried alternatives prior
555 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
556 backtracked to try for "b" instead of moving to the next alternative branch
557 at the same level (in this case, to look for "c"). The Perl documentation
558 is clear that when (*THEN) is backtracked onto, it goes to the "next
559 alternative in the innermost enclosing group".
560
561 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
562 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
563 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
564 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
565 (*THEN).
566
567 3. If \s appeared in a character class, it removed the VT character from
568 the class, even if it had been included by some previous item, for example
569 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
570 of \s, but is part of the POSIX "space" class.)
571
572 4. A partial match never returns an empty string (because you can always
573 match an empty string at the end of the subject); however the checking for
574 an empty string was starting at the "start of match" point. This has been
575 changed to the "earliest inspected character" point, because the returned
576 data for a partial match starts at this character. This means that, for
577 example, /(?<=abc)def/ gives a partial match for the subject "abc"
578 (previously it gave "no match").
579
580 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
581 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
582 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
583 has an implication that the given string is incomplete (because a partial
584 match is preferred over a full match). For this reason, these items now
585 give a partial match in this situation. [Aside: previously, the one case
586 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
587 match rather than a full match, which was wrong by the old rules, but is
588 now correct.]
589
590 6. There was a bug in the handling of #-introduced comments, recognized when
591 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
592 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
593 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
594 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
595 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
596 places in pcre_compile().
597
598 7. Related to (6) above, when pcre_compile() was skipping #-introduced
599 comments when looking ahead for named forward references to subpatterns,
600 the only newline sequence it recognized was NL. It now handles newlines
601 according to the set newline convention.
602
603 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
604 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
605 cater for a lack of strerror(). These oversights have been fixed.
606
607 9. Added --match-limit and --recursion-limit to pcregrep.
608
609 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
610
611 11. When the -o option was used, pcregrep was setting a return code of 1, even
612 when matches were found, and --line-buffered was not being honoured.
613
614 12. Added an optional parentheses number to the -o and --only-matching options
615 of pcregrep.
616
617 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
618 can match an empty string. The code to do it in pcretest and pcredemo
619 needed fixing:
620
621 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
622 only one byte after an empty string match just before CRLF (this case
623 just got forgotten; "any" and "anycrlf" were OK).
624
625 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
626 mode when an empty string match preceded an ASCII character followed by
627 a non-ASCII character. (The code for advancing by one character rather
628 than one byte was nonsense.)
629
630 (c) The pcredemo.c sample program did not have any code at all to handle
631 the cases when CRLF is a valid newline sequence.
632
633 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
634 as a starting offset was within the subject string. There is now a new
635 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
636 negative or greater than the length of the string. In order to test this,
637 pcretest is extended to allow the setting of negative starting offsets.
638
639 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
640 starting offset points to the beginning of a UTF-8 character was
641 unnecessarily clumsy. I tidied it up.
642
643 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
644 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
645
646 17. Nobody had reported that the --include_dir option, which was added in
647 release 7.7 should have been called --include-dir (hyphen, not underscore)
648 for compatibility with GNU grep. I have changed it to --include-dir, but
649 left --include_dir as an undocumented synonym, and the same for
650 --exclude-dir, though that is not available in GNU grep, at least as of
651 release 2.5.4.
652
653 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
654 characters from a string of bytes) have been redefined so as not to use
655 loops, in order to improve performance in some environments. At the same
656 time, I abstracted some of the common code into auxiliary macros to save
657 repetition (this should not affect the compiled code).
658
659 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
660 compile-time error is now given if \c is not followed by an ASCII
661 character, that is, a byte less than 128. (In EBCDIC mode, the code is
662 different, and any byte value is allowed.)
663
664 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
665 START_OPTIMIZE option, which is now allowed at compile time - but just
666 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
667 to pcregrep and other applications that have no direct access to PCRE
668 options. The new /Y option in pcretest sets this option when calling
669 pcre_compile().
670
671 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
672 back references. Groups containing recursive back references were forced to
673 be atomic by that change, but in the case of named groups, the amount of
674 memory required was incorrectly computed, leading to "Failed: internal
675 error: code overflow". This has been fixed.
676
677 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
678 pcretest.c, to avoid build problems in some Borland environments.
679
680
681 Version 8.10 25-Jun-2010
682 ------------------------
683
684 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
685 THEN.
686
687 2. (*ACCEPT) was not working when inside an atomic group.
688
689 3. Inside a character class, \B is treated as a literal by default, but
690 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
691 causes the error). The code is unchanged, but I tidied the documentation.
692
693 4. Inside a character class, PCRE always treated \R and \X as literals,
694 whereas Perl faults them if its -w option is set. I have changed PCRE so
695 that it faults them when PCRE_EXTRA is set.
696
697 5. Added support for \N, which always matches any character other than
698 newline. (It is the same as "." when PCRE_DOTALL is not set.)
699
700 6. When compiling pcregrep with newer versions of gcc which may have
701 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
702 declared with attribute warn_unused_result" were given. Just casting the
703 result to (void) does not stop the warnings; a more elaborate fudge is
704 needed. I've used a macro to implement this.
705
706 7. Minor change to pcretest.c to avoid a compiler warning.
707
708 8. Added four artifical Unicode properties to help with an option to make
709 \s etc use properties (see next item). The new properties are: Xan
710 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
711
712 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
713 use Unicode properties. (*UCP) at the start of a pattern can be used to set
714 this option. Modified pcretest to add /W to test this facility. Added
715 REG_UCP to make it available via the POSIX interface.
716
717 10. Added --line-buffered to pcregrep.
718
719 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
720 studied, and the match started with a letter with a code point greater than
721 127 whose first byte was different to the first byte of the other case of
722 the letter, the other case of this starting letter was not recognized
723 (#976).
724
725 12. If a pattern that was studied started with a repeated Unicode property
726 test, for example, \p{Nd}+, there was the theoretical possibility of
727 setting up an incorrect bitmap of starting bytes, but fortunately it could
728 not have actually happened in practice until change 8 above was made (it
729 added property types that matched character-matching opcodes).
730
731 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
732 possible starting bytes for non-anchored patterns.
733
734 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
735 \R, and also a number of cases that involve Unicode properties, both
736 explicit and implicit when PCRE_UCP is set.
737
738 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
739 input, it could crash or give wrong results if characters with values
740 greater than 0xc0 were present in the subject string. (Detail: it assumed
741 UTF-8 input when processing these items.)
742
743 16. Added a lot of (int) casts to avoid compiler warnings in systems where
744 size_t is 64-bit (#991).
745
746 17. Added a check for running out of memory when PCRE is compiled with
747 --disable-stack-for-recursion (#990).
748
749 18. If the last data line in a file for pcretest does not have a newline on
750 the end, a newline was missing in the output.
751
752 19. The default pcre_chartables.c file recognizes only ASCII characters (values
753 less than 128) in its various bitmaps. However, there is a facility for
754 generating tables according to the current locale when PCRE is compiled. It
755 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
756 characters, are recognized by isspace() and therefore were getting set in
757 these tables, and indeed these tables seem to approximate to ISO 8859. This
758 caused a problem in UTF-8 mode when pcre_study() was used to create a list
759 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
760 which of course cannot start UTF-8 characters. I have changed the code so
761 that only real ASCII characters (less than 128) and the correct starting
762 bytes for UTF-8 encodings are set for characters greater than 127 when in
763 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
764 altogether.)
765
766 20. Added the /T option to pcretest so as to be able to run tests with non-
767 standard character tables, thus making it possible to include the tests
768 used for 19 above in the standard set of tests.
769
770 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
771 reference to a subpattern the other side of a comment that contains an
772 opening parenthesis caused either an internal compiling error, or a
773 reference to the wrong subpattern.
774
775
776 Version 8.02 19-Mar-2010
777 ------------------------
778
779 1. The Unicode data tables have been updated to Unicode 5.2.0.
780
781 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
782 configured.
783
784 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
785 original author of that file, following a query about its status.
786
787 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
788 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
789
790 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
791 quantifier applied to a forward-referencing subroutine call, could compile
792 incorrect code or give the error "internal error: previously-checked
793 referenced subpattern not found".
794
795 6. Both MS Visual Studio and Symbian OS have problems with initializing
796 variables to point to external functions. For these systems, therefore,
797 pcre_malloc etc. are now initialized to local functions that call the
798 relevant global functions.
799
800 7. There were two entries missing in the vectors called coptable and poptable
801 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
802 I've fixed the data, and added a kludgy way of testing at compile time that
803 the lengths are correct (equal to the number of opcodes).
804
805 8. Following on from 7, I added a similar kludge to check the length of the
806 eint vector in pcreposix.c.
807
808 9. Error texts for pcre_compile() are held as one long string to avoid too
809 much relocation at load time. To find a text, the string is searched,
810 counting zeros. There was no check for running off the end of the string,
811 which could happen if a new error number was added without updating the
812 string.
813
814 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
815
816 11. \K was not working if it appeared in an atomic group or in a group that
817 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
818 \K is "not well defined" if used in an assertion. PCRE now accepts it if
819 the assertion is positive, but not if it is negative.
820
821 12. Change 11 fortuitously reduced the size of the stack frame used in the
822 "match()" function of pcre_exec.c by one pointer. Forthcoming
823 implementation of support for (*MARK) will need an extra pointer on the
824 stack; I have reserved it now, so that the stack frame size does not
825 decrease.
826
827 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
828 item in branch that calls a recursion is a subroutine call - as in the
829 second branch in the above example - was incorrectly given the compile-
830 time error "recursive call could loop indefinitely" because pcre_compile()
831 was not correctly checking the subroutine for matching a non-empty string.
832
833 14. The checks for overrunning compiling workspace could trigger after an
834 overrun had occurred. This is a "should never occur" error, but it can be
835 triggered by pathological patterns such as hundreds of nested parentheses.
836 The checks now trigger 100 bytes before the end of the workspace.
837
838 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
839
840
841 Version 8.01 19-Jan-2010
842 ------------------------
843
844 1. If a pattern contained a conditional subpattern with only one branch (in
845 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
846 computed the wrong minimum data length (which is of course zero for such
847 subpatterns). This could cause incorrect "no match" results.
848
849 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
850 the pattern is reset in the first branch, pcre_compile() failed with
851 "internal error: code overflow at offset...". This happened only when
852 the reset was to the original external option setting. (An optimization
853 abstracts leading options settings into an external setting, which was the
854 cause of this.)
855
856 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
857 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
858 assertion pattern did not match (meaning that the assertion was true), it
859 was incorrectly treated as false if the SKIP had been reached during the
860 matching. This also applied to assertions used as conditions.
861
862 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
863 assertion subpattern, including such a pattern used as a condition,
864 unpredictable results occurred, instead of the error return
865 PCRE_ERROR_DFA_UITEM.
866
867 5. The C++ GlobalReplace function was not working like Perl for the special
868 situation when an empty string is matched. It now does the fancy magic
869 stuff that is necessary.
870
871 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
872 removed. (These were left over from very, very early versions of PCRE.)
873
874 7. Some cosmetic changes to the code to make life easier when compiling it
875 as part of something else:
876
877 (a) Change DEBUG to PCRE_DEBUG.
878
879 (b) In pcre_compile(), rename the member of the "branch_chain" structure
880 called "current" as "current_branch", to prevent a collision with the
881 Linux macro when compiled as a kernel module.
882
883 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
884 prevent a collision with the Linux macro when compiled as a kernel
885 module.
886
887 8. In pcre_compile() there are some checks for integer overflows that used to
888 cast potentially large values to (double). This has been changed to that
889 when building, a check for int64_t is made, and if it is found, it is used
890 instead, thus avoiding the use of floating point arithmetic. (There is no
891 other use of FP in PCRE.) If int64_t is not found, the fallback is to
892 double.
893
894 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
895 2005 (difference between two addresses compared to an unsigned value).
896
897 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
898 custom one, because of the following reported problem in Windows:
899
900 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
901 under Win32.
902 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
903 therefore missing the function definition.
904 - The compiler thus generates a "C" signature for the test function.
905 - The linker fails to find the "C" function.
906 - PCRE fails to configure if asked to do so against libbz2.
907
908 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
909 messages were output:
910
911 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
912 rerunning libtoolize, to keep the correct libtool macros in-tree.
913 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
914
915 I have done both of these things.
916
917 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
918 most of the time, it *can* run out if it is given a pattern that contains a
919 runaway infinite recursion. I updated the discussion in the pcrestack man
920 page.
921
922 13. Now that we have gone to the x.xx style of version numbers, the minor
923 version may start with zero. Using 08 or 09 is a bad idea because users
924 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
925 interpreted as invalid octal numbers. I've updated the previous comment in
926 configure.ac, and also added a check that gives an error if 08 or 09 are
927 used.
928
929 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
930 causing partial matching to fail when the end of the subject matched \W
931 in a UTF-8 pattern where \W was quantified with a minimum of 3.
932
933 15. There were some discrepancies between the declarations in pcre_internal.h
934 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
935 their definitions. The declarations used "const uschar *" and the
936 definitions used USPTR. Even though USPTR is normally defined as "const
937 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
938 reported that: "This difference in casting confuses some C++ compilers, for
939 example, SunCC recognizes above declarations as different functions and
940 generates broken code for hbpcre." I have changed the declarations to use
941 USPTR.
942
943 16. GNU libtool is named differently on some systems. The autogen.sh script now
944 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
945 (FreeBSD).
946
947 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
948 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
949 comment: "Figure out how to create a longlong from a string: strtoll and
950 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
951 instance, but it only takes 2 args instead of 3!"
952
953 18. A subtle bug concerned with back references has been fixed by a change of
954 specification, with a corresponding code fix. A pattern such as
955 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
956 refers, was giving matches when it shouldn't. For example, xa=xaaa would
957 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
958 same bug. Such groups have to be quantified to be useful, or contained
959 inside another quantified group. (If there's no repetition, the reference
960 can never match.) The problem arises because, having left the group and
961 moved on to the rest of the pattern, a later failure that backtracks into
962 the group uses the captured value from the final iteration of the group
963 rather than the correct earlier one. I have fixed this in PCRE by forcing
964 any group that contains a reference to itself to be an atomic group; that
965 is, there cannot be any backtracking into it once it has completed. This is
966 similar to recursive and subroutine calls.
967
968
969 Version 8.00 19-Oct-09
970 ----------------------
971
972 1. The table for translating pcre_compile() error codes into POSIX error codes
973 was out-of-date, and there was no check on the pcre_compile() error code
974 being within the table. This could lead to an OK return being given in
975 error.
976
977 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
978 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
979 in a Windows environment.
980
981 3. The pcregrep --count option prints the count for each file even when it is
982 zero, as does GNU grep. However, pcregrep was also printing all files when
983 --files-with-matches was added. Now, when both options are given, it prints
984 counts only for those files that have at least one match. (GNU grep just
985 prints the file name in this circumstance, but including the count seems
986 more useful - otherwise, why use --count?) Also ensured that the
987 combination -clh just lists non-zero counts, with no names.
988
989 4. The long form of the pcregrep -F option was incorrectly implemented as
990 --fixed_strings instead of --fixed-strings. This is an incompatible change,
991 but it seems right to fix it, and I didn't think it was worth preserving
992 the old behaviour.
993
994 5. The command line items --regex=pattern and --regexp=pattern were not
995 recognized by pcregrep, which required --regex pattern or --regexp pattern
996 (with a space rather than an '='). The man page documented the '=' forms,
997 which are compatible with GNU grep; these now work.
998
999 6. No libpcreposix.pc file was created for pkg-config; there was just
1000 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1001
1002 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1003 when UCP support is not needed, by modifying the Python script that
1004 generates it from Unicode data files. This should not matter if the module
1005 is correctly used as a library, but I received one complaint about 50K of
1006 unwanted data. My guess is that the person linked everything into his
1007 program rather than using a library. Anyway, it does no harm.
1008
1009 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1010 was a minimum greater than 1 for a wide character in a possessive
1011 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1012 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1013 character. Chaos in the form of incorrect output or a compiling loop could
1014 result.
1015
1016 9. The restrictions on what a pattern can contain when partial matching is
1017 requested for pcre_exec() have been removed. All patterns can now be
1018 partially matched by this function. In addition, if there are at least two
1019 slots in the offset vector, the offset of the earliest inspected character
1020 for the match and the offset of the end of the subject are set in them when
1021 PCRE_ERROR_PARTIAL is returned.
1022
1023 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1024 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1025 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1026 and may be more useful for multi-segment matching.
1027
1028 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1029 used to be given if ever the end of the subject was reached; now it is
1030 given only if matching could not proceed because another character was
1031 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1032 string "Z", which now yields "no match" instead of "partial match". In the
1033 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1034 final character ended with (*FAIL).
1035
1036 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1037 if the pattern had a "must contain" character that was already found in the
1038 earlier partial match, unless partial matching was again requested. For
1039 example, with the pattern /dog.(body)?/, the "must contain" character is
1040 "g". If the first part-match was for the string "dog", restarting with
1041 "sbody" failed. This bug has been fixed.
1042
1043 13. The string returned by pcre_dfa_exec() after a partial match has been
1044 changed so that it starts at the first inspected character rather than the
1045 first character of the match. This makes a difference only if the pattern
1046 starts with a lookbehind assertion or \b or \B (\K is not supported by
1047 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1048 matching functions compatible, and I think it's the right thing to do.
1049
1050 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1051 so that the demonstration program is easily available in environments where
1052 PCRE has not been installed from source.
1053
1054 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1055 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1056 library.
1057
1058 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1059 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1060 is not the first non-POSIX option to be added. Clearly some people find
1061 these options useful.
1062
1063 17. If a caller to the POSIX matching function regexec() passes a non-zero
1064 value for nmatch with a NULL value for pmatch, the value of
1065 nmatch is forced to zero.
1066
1067 18. RunGrepTest did not have a test for the availability of the -u option of
1068 the diff command, as RunTest does. It now checks in the same way as
1069 RunTest, and also checks for the -b option.
1070
1071 19. If an odd number of negated classes containing just a single character
1072 interposed, within parentheses, between a forward reference to a named
1073 subpattern and the definition of the subpattern, compilation crashed with
1074 an internal error, complaining that it could not find the referenced
1075 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1076 [The bug was that it was starting one character too far in when skipping
1077 over the character class, thus treating the ] as data rather than
1078 terminating the class. This meant it could skip too much.]
1079
1080 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1081 /g option in pcretest when the pattern contains \K, which makes it possible
1082 to have an empty string match not at the start, even when the pattern is
1083 anchored. Updated pcretest and pcredemo to use this option.
1084
1085 21. If the maximum number of capturing subpatterns in a recursion was greater
1086 than the maximum at the outer level, the higher number was returned, but
1087 with unset values at the outer level. The correct (outer level) value is
1088 now given.
1089
1090 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1091 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1092 make it do so. The string so far is captured, making this feature
1093 compatible with Perl.
1094
1095 23. The tests have been re-organized, adding tests 11 and 12, to make it
1096 possible to check the Perl 5.10 features against Perl 5.10.
1097
1098 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1099 pattern matches a fixed length string. PCRE did not allow this; now it
1100 does. Neither allows recursion.
1101
1102 25. I finally figured out how to implement a request to provide the minimum
1103 length of subject string that was needed in order to match a given pattern.
1104 (It was back references and recursion that I had previously got hung up
1105 on.) This code has now been added to pcre_study(); it finds a lower bound
1106 to the length of subject needed. It is not necessarily the greatest lower
1107 bound, but using it to avoid searching strings that are too short does give
1108 some useful speed-ups. The value is available to calling programs via
1109 pcre_fullinfo().
1110
1111 26. While implementing 25, I discovered to my embarrassment that pcretest had
1112 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1113 study optimizations had never been tested with that matching function.
1114 Oops. What is worse, even when it was passed study data, there was a bug in
1115 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1116 were also very few tests of studied patterns with pcre_dfa_exec().
1117
1118 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1119 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1120 on the other side of the coin, they are no longer allowed to have different
1121 names, because these cannot be distinguished in PCRE, and this has caused
1122 confusion. (This is a difference from Perl.)
1123
1124 28. When duplicate subpattern names are present (necessarily with different
1125 numbers, as required by 27 above), and a test is made by name in a
1126 conditional pattern, either for a subpattern having been matched, or for
1127 recursion in such a pattern, all the associated numbered subpatterns are
1128 tested, and the overall condition is true if the condition is true for any
1129 one of them. This is the way Perl works, and is also more like the way
1130 testing by number works.
1131
1132
1133 Version 7.9 11-Apr-09
1134 ---------------------
1135
1136 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1137 (pcretest), all targets were linked against these libraries. This included
1138 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1139 libraries. This caused unwanted dependencies to be created. This problem
1140 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1141 pcretest is linked with readline.
1142
1143 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1144 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1145 moved outside it again, because FALSE and TRUE are already defined in AIX,
1146 but BOOL is not.
1147
1148 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1149 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1150
1151 4. The pcregrep documentation said spaces were inserted as well as colons (or
1152 hyphens) following file names and line numbers when outputting matching
1153 lines. This is not true; no spaces are inserted. I have also clarified the
1154 wording for the --colour (or --color) option.
1155
1156 5. In pcregrep, when --colour was used with -o, the list of matching strings
1157 was not coloured; this is different to GNU grep, so I have changed it to be
1158 the same.
1159
1160 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1161 each matching line was coloured. Now it goes on to look for further matches
1162 of any of the test patterns, which is the same behaviour as GNU grep.
1163
1164 7. A pattern that could match an empty string could cause pcregrep to loop; it
1165 doesn't make sense to accept an empty string match in pcregrep, so I have
1166 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1167 seems to be how GNU grep behaves.
1168
1169 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1170 start or after a newline", because the conditional assertion was not being
1171 correctly handled. The rule now is that both the assertion and what follows
1172 in the first alternative must satisfy the test.
1173
1174 9. If auto-callout was enabled in a pattern with a conditional group whose
1175 condition was an assertion, PCRE could crash during matching, both with
1176 pcre_exec() and pcre_dfa_exec().
1177
1178 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1179 used for matching.
1180
1181 11. Unicode property support in character classes was not working for
1182 characters (bytes) greater than 127 when not in UTF-8 mode.
1183
1184 12. Added the -M command line option to pcretest.
1185
1186 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1187
1188 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1189
1190 16. Added comments and documentation about mis-use of no_arg in the C++
1191 wrapper.
1192
1193 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1194 from Martin Jerabek that uses macro names for all relevant character and
1195 string constants.
1196
1197 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1198 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1199 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1200 these, but not everybody uses configure.
1201
1202 19. A conditional group that had only one branch was not being correctly
1203 recognized as an item that could match an empty string. This meant that an
1204 enclosing group might also not be so recognized, causing infinite looping
1205 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1206 with the subject "ab", where knowledge that the repeated group can match
1207 nothing is needed in order to break the loop.
1208
1209 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1210 exec(), but without supplying a callout function, matching went wrong.
1211
1212 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1213 leak if the size of the offset vector was greater than 30. When the vector
1214 is smaller, the saved offsets during recursion go onto a local stack
1215 vector, but for larger vectors malloc() is used. It was failing to free
1216 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1217 error, in fact).
1218
1219 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1220 heapframe that is used only when UTF-8 support is enabled. This caused no
1221 problem, but was untidy.
1222
1223 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1224 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1225 included within another project.
1226
1227 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1228 slightly modified by me:
1229
1230 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1231 not building pcregrep.
1232
1233 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1234 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1235
1236 25. Forward references, both numeric and by name, in patterns that made use of
1237 duplicate group numbers, could behave incorrectly or give incorrect errors,
1238 because when scanning forward to find the reference group, PCRE was not
1239 taking into account the duplicate group numbers. A pattern such as
1240 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1241
1242 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1243 the feature of a custom pointer more persuasive (as requested by a user).
1244
1245 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1246 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1247 pre-defined.
1248
1249 28. Added support for (*UTF8) at the start of a pattern.
1250
1251 29. Arrange for flags added by the "release type" setting in CMake to be shown
1252 in the configuration summary.
1253
1254
1255 Version 7.8 05-Sep-08
1256 ---------------------
1257
1258 1. Replaced UCP searching code with optimized version as implemented for Ad
1259 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1260 stage table and inline lookup instead of a function, giving speed ups of 2
1261 to 5 times on some simple patterns that I tested. Permission was given to
1262 distribute the MultiStage2.py script that generates the tables (it's not in
1263 the tarball, but is in the Subversion repository).
1264
1265 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1266 scripts.
1267
1268 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1269 a group with a zero qualifier. The result of the study could be incorrect,
1270 or the function might crash, depending on the pattern.
1271
1272 4. Caseless matching was not working for non-ASCII characters in back
1273 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1274 It now works when Unicode Property Support is available.
1275
1276 5. In pcretest, an escape such as \x{de} in the data was always generating
1277 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1278 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1279 truncation.
1280
1281 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1282
1283 7. Added two (int) casts to pcregrep when printing the difference of two
1284 pointers, in case they are 64-bit values.
1285
1286 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1287 test 2 if it fails.
1288
1289 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1290 and a #define of that name to empty if it is not externally set. This is to
1291 allow users of MSVC to set it if necessary.
1292
1293 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1294 the convenience functions in the pcre_get.c source file.
1295
1296 11. An option change at the start of a pattern that had top-level alternatives
1297 could cause overwriting and/or a crash. This command provoked a crash in
1298 some environments:
1299
1300 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1301
1302 This potential security problem was recorded as CVE-2008-2371.
1303
1304 12. For a pattern where the match had to start at the beginning or immediately
1305 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1306 pcre_dfa_exec() could read past the end of the passed subject if there was
1307 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1308 pcretest so that it places the subject at the end of its malloc-ed buffer.
1309
1310 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1311 exec() might read past the end of the data buffer in UTF-8 mode.
1312
1313 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1314 the data contained the byte 0x85 as part of a UTF-8 character within its
1315 first line. This applied both to normal and DFA matching.
1316
1317 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1318 /^[^d]*?$/8 failed to match "abc".
1319
1320 16. Added a missing copyright notice to pcrecpp_internal.h.
1321
1322 17. Make it more clear in the documentation that values returned from
1323 pcre_exec() in ovector are byte offsets, not character counts.
1324
1325 18. Tidied a few places to stop certain compilers from issuing warnings.
1326
1327 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1328 supplied by Stefan Weber. I made a further small update for 7.8 because
1329 there is a change of source arrangements: the pcre_searchfuncs.c module is
1330 replaced by pcre_ucd.c.
1331
1332
1333 Version 7.7 07-May-08
1334 ---------------------
1335
1336 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1337 a string to a long long, pretend we don't even have a long long." This is
1338 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1339
1340 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1341 pre-7.6 versions, which defined a global no_arg variable instead of putting
1342 it in the RE class. (See also #8 below.)
1343
1344 3. Remove a line of dead code, identified by coverity and reported by Nuno
1345 Lopes.
1346
1347 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1348
1349 (1) The include/exclude patterns were being applied to the whole pathnames
1350 of files, instead of just to the final components.
1351
1352 (2) If there was more than one level of directory, the subdirectories were
1353 skipped unless they satisfied the include/exclude conditions. This is
1354 inconsistent with GNU grep (and could even be seen as contrary to the
1355 pcregrep specification - which I improved to make it absolutely clear).
1356 The action now is always to scan all levels of directory, and just
1357 apply the include/exclude patterns to regular files.
1358
1359 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1360 --exclude_dir in the tests to avoid scanning .svn directories.
1361
1362 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1363 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1364 doesn't support NULs in patterns.
1365
1366 7. Added some missing "const"s to declarations of static tables in
1367 pcre_compile.c and pcre_dfa_exec.c.
1368
1369 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1370 caused by fix #2 above. (Subsequently also a second patch to fix the
1371 first patch. And a third patch - this was a messy problem.)
1372
1373 9. Applied Craig's patch to remove the use of push_back().
1374
1375 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1376 matching function regexec().
1377
1378 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1379 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1380 references. PCRE supports relative numbers with this syntax (I don't think
1381 Oniguruma does).
1382
1383 12. Previously, a group with a zero repeat such as (...){0} was completely
1384 omitted from the compiled regex. However, this means that if the group
1385 was called as a subroutine from elsewhere in the pattern, things went wrong
1386 (an internal error was given). Such groups are now left in the compiled
1387 pattern, with a new opcode that causes them to be skipped at execution
1388 time.
1389
1390 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1391 to the way PCRE behaves:
1392
1393 (a) A lone ] character is dis-allowed (Perl treats it as data).
1394
1395 (b) A back reference to an unmatched subpattern matches an empty string
1396 (Perl fails the current match path).
1397
1398 (c) A data ] in a character class must be notated as \] because if the
1399 first data character in a class is ], it defines an empty class. (In
1400 Perl it is not possible to have an empty class.) The empty class []
1401 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1402 The negative empty class [^] matches any one character, independently
1403 of the DOTALL setting.
1404
1405 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1406 non-existent subpattern following a character class starting with ']' and
1407 containing () gave an internal compiling error instead of "reference to
1408 non-existent subpattern". Fortunately, when the pattern did exist, the
1409 compiled code was correct. (When scanning forwards to check for the
1410 existencd of the subpattern, it was treating the data ']' as terminating
1411 the class, so got the count wrong. When actually compiling, the reference
1412 was subsequently set up correctly.)
1413
1414 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1415 it was being rejected as not supported by pcre_dfa_exec(), even though
1416 other assertions are supported. I have made pcre_dfa_exec() support
1417 (*FAIL).
1418
1419 16. The implementation of 13c above involved the invention of a new opcode,
1420 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1421 cannot be changed at match time, I realized I could make a small
1422 improvement to matching performance by compiling OP_ALLANY instead of
1423 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1424 on the OP_ANY path.
1425
1426 17. Compiling pcretest on Windows with readline support failed without the
1427 following two fixes: (1) Make the unistd.h include conditional on
1428 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1429
1430 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1431 ncurses library to be included for pcretest when ReadLine support is
1432 requested, but also to allow for it to be overridden. This patch came from
1433 Daniel Bergström.
1434
1435 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1436 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1437 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1438 spotting this.
1439
1440
1441 Version 7.6 28-Jan-08
1442 ---------------------
1443
1444 1. A character class containing a very large number of characters with
1445 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1446 overflow.
1447
1448 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1449 HAVE_LONG_LONG is not defined.
1450
1451 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1452 bring them up to date and include new features. This patch includes:
1453
1454 - Fixed PH's badly added libz and libbz2 support.
1455 - Fixed a problem with static linking.
1456 - Added pcredemo. [But later removed - see 7 below.]
1457 - Fixed dftables problem and added an option.
1458 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1459 HAVE_LONG_LONG.
1460 - Added readline support for pcretest.
1461 - Added an listing of the option settings after cmake has run.
1462
1463 4. A user submitted a patch to Makefile that makes it easy to create
1464 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1465 Makefile.am that cause it to include this special target, without
1466 affecting anything else. Note that the same mingw target plus all
1467 the other distribution libraries and programs are now supported
1468 when configuring with CMake (see 6 below) instead of with
1469 Configure/Make.
1470
1471 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1472 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1473 exported in the Windows port". It has not yet been confirmed that the patch
1474 solves the problem, but it does no harm.
1475
1476 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1477 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1478 with CMake, and also correct the comment about stack recursion.
1479
1480 7. Remove the automatic building of pcredemo from the ./configure system and
1481 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1482 of a program that users should build themselves after PCRE is installed, so
1483 building it automatically is not really right. What is more, it gave
1484 trouble in some build environments.
1485
1486 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1487
1488
1489 Version 7.5 10-Jan-08
1490 ---------------------
1491
1492 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1493 values in parens when parsing an RE using the C++ wrapper."
1494
1495 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1496 Characters greater than 255 were excluded from the class instead of being
1497 included.
1498
1499 3. The same bug as (2) above applied to negated POSIX classes such as
1500 [:^space:].
1501
1502 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1503 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1504 I have changed it.
1505
1506 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1507 first named subpattern) and a construct such as (?&a) would reference the
1508 first named subpattern whose name started with "a" (in other words, the
1509 length check was missing). Both these problems are fixed. "Subpattern name
1510 expected" is now given for (?&) (a zero-length name), and this patch also
1511 makes it give the same error for \k'' (previously it complained that that
1512 was a reference to a non-existent subpattern).
1513
1514 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1515 this is right because (?- can be followed by option settings as well as by
1516 digits. I have, however, made the messages clearer.
1517
1518 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1519 than the number used in the conditional) now cause a compile-time error.
1520 This is actually not compatible with Perl, which accepts such patterns, but
1521 treats the conditional as always being FALSE (as PCRE used to), but it
1522 seems to me that giving a diagnostic is better.
1523
1524 8. Change "alphameric" to the more common word "alphanumeric" in comments
1525 and messages.
1526
1527 9. Fix two occurrences of "backslash" in comments that should have been
1528 "backspace".
1529
1530 10. Remove two redundant lines of code that can never be obeyed (their function
1531 was moved elsewhere).
1532
1533 11. The program that makes PCRE's Unicode character property table had a bug
1534 which caused it to generate incorrect table entries for sequences of
1535 characters that have the same character type, but are in different scripts.
1536 It amalgamated them into a single range, with the script of the first of
1537 them. In other words, some characters were in the wrong script. There were
1538 thirteen such cases, affecting characters in the following ranges:
1539
1540 U+002b0 - U+002c1
1541 U+0060c - U+0060d
1542 U+0061e - U+00612
1543 U+0064b - U+0065e
1544 U+0074d - U+0076d
1545 U+01800 - U+01805
1546 U+01d00 - U+01d77
1547 U+01d9b - U+01dbf
1548 U+0200b - U+0200f
1549 U+030fc - U+030fe
1550 U+03260 - U+0327f
1551 U+0fb46 - U+0fbb1
1552 U+10450 - U+1049d
1553
1554 12. The -o option (show only the matching part of a line) for pcregrep was not
1555 compatible with GNU grep in that, if there was more than one match in a
1556 line, it showed only the first of them. It now behaves in the same way as
1557 GNU grep.
1558
1559 13. If the -o and -v options were combined for pcregrep, it printed a blank
1560 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1561 does the same. The return code can be used to tell if there were any
1562 non-matching lines.
1563
1564 14. Added --file-offsets and --line-offsets to pcregrep.
1565
1566 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1567 infinitely looping recursion. The bug was that positive lookaheads were not
1568 being skipped when checking for a possible empty match (negative lookaheads
1569 and both kinds of lookbehind were skipped).
1570
1571 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1572 inclusion of <windows.h> to before rather than after the definition of
1573 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1574
1575 17. Specifying a possessive quantifier with a specific limit for a Unicode
1576 character property caused pcre_compile() to compile bad code, which led at
1577 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1578 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1579 caused the error; without that there was no problem.
1580
1581 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1582
1583 19. Added --enable-pcretest-libreadline.
1584
1585 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1586 RE::GlobalReplace(). As a result, the number of replacements returned was
1587 double what it should be. I removed one of the increments, but Craig sent a
1588 later patch that removed the other one (the right fix) and added unit tests
1589 that check the return values (which was not done before).
1590
1591 21. Several CMake things:
1592
1593 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1594 the names libpcre and libpcreposix, not just pcre and pcreposix.
1595
1596 (2) The above change means that pcretest and pcregrep are now correctly
1597 linked with the newly-built libraries, not previously installed ones.
1598
1599 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1600
1601 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1602 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1603 UTF-8 newline character). The key issue is that the pattern starts .*;
1604 this means that the match must be either at the beginning, or after a
1605 newline. The bug was in the code for advancing after a failed match and
1606 checking that the new position followed a newline. It was not taking
1607 account of UTF-8 characters correctly.
1608
1609 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1610 character classes. PCRE was not treating the sequence [:...:] as a
1611 character class unless the ... were all letters. Perl, however, seems to
1612 allow any characters between [: and :], though of course it rejects as
1613 unknown any "names" that contain non-letters, because all the known class
1614 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1615 for example, whereas PCRE did not - it did not recognize a POSIX character
1616 class. This seemed a bit dangerous, so the code has been changed to be
1617 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1618 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1619 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1620 Perl does, and where it didn't before.
1621
1622 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1623 Windows environments %n is disabled by default.
1624
1625
1626 Version 7.4 21-Sep-07
1627 ---------------------
1628
1629 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1630 means that a class such as [\s] counted as "explicit reference to CR or
1631 LF". That isn't really right - the whole point of the change was to try to
1632 help when there was an actual mention of one of the two characters. So now
1633 the change happens only if \r or \n (or a literal CR or LF) character is
1634 encountered.
1635
1636 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1637 of both had grown to the point where there were only 3 bits left.
1638 Fortunately, there was spare space in the data structure, and so I have
1639 moved the internal flags into a new 16-bit field to free up more option
1640 bits.
1641
1642 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1643 but did not set the internal JCHANGED flag - either of these is enough to
1644 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1645 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1646 start sets both bits.
1647
1648 4. Added options (at build time, compile time, exec time) to change \R from
1649 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1650
1651 5. doc/pcresyntax.html was missing from the distribution.
1652
1653 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1654 compatibility, even though it is no longer used.
1655
1656 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1657 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1658 windows.h file is present (where different names are used). [This was
1659 reversed later after testing - see 16 below.]
1660
1661 8. Changed all #include <config.h> to #include "config.h". There were also
1662 some further <pcre.h> cases that I changed to "pcre.h".
1663
1664 9. When pcregrep was used with the --colour option, it missed the line ending
1665 sequence off the lines that it output.
1666
1667 10. It was pointed out to me that arrays of string pointers cause lots of
1668 relocations when a shared library is dynamically loaded. A technique of
1669 using a single long string with a table of offsets can drastically reduce
1670 these. I have refactored PCRE in four places to do this. The result is
1671 dramatic:
1672
1673 Originally: 290
1674 After changing UCP table: 187
1675 After changing error message table: 43
1676 After changing table of "verbs" 36
1677 After changing table of Posix names 22
1678
1679 Thanks to the folks working on Gregex for glib for this insight.
1680
1681 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1682 unicode-properties was also set.
1683
1684 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1685
1686 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1687 checked only for CRLF.
1688
1689 14. Added casts to pcretest.c to avoid compiler warnings.
1690
1691 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1692
1693 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1694 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1695 entirely. This removes changes made in 7 above.
1696
1697 17. The CMake files have been updated, and there is now more information about
1698 building with CMake in the NON-UNIX-USE document.
1699
1700
1701 Version 7.3 28-Aug-07
1702 ---------------------
1703
1704 1. In the rejigging of the build system that eventually resulted in 7.1, the
1705 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1706 brackets there is not right, since it causes compilers to look for an
1707 installed pcre.h, not the version that is in the source that is being
1708 compiled (which of course may be different). I have changed it back to:
1709
1710 #include "pcre.h"
1711
1712 I have a vague recollection that the change was concerned with compiling in
1713 different directories, but in the new build system, that is taken care of
1714 by the VPATH setting the Makefile.
1715
1716 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1717 when the subject happened to end in the byte 0x85 (e.g. if the last
1718 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1719 characters but of course it shouldn't be taken as a newline when it is part
1720 of another character. The bug was that, for an unlimited repeat of . in
1721 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1722 characters when looking for a newline.
1723
1724 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1725
1726 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1727 in debug output.
1728
1729 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1730 long printing in the pcrecpp unittest when running under MinGW.
1731
1732 6. ESC_K was left out of the EBCDIC table.
1733
1734 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1735 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1736 limit also applies to "virtual nesting" when a pattern is recursive, and in
1737 this case 1000 isn't so big. I have been able to remove this limit at the
1738 expense of backing off one optimization in certain circumstances. Normally,
1739 when pcre_exec() would call its internal match() function recursively and
1740 immediately return the result unconditionally, it uses a "tail recursion"
1741 feature to save stack. However, when a subpattern that can match an empty
1742 string has an unlimited repetition quantifier, it no longer makes this
1743 optimization. That gives it a stack frame in which to save the data for
1744 checking that an empty string has been matched. Previously this was taken
1745 from the 1000-entry workspace that had been reserved. So now there is no
1746 explicit limit, but more stack is used.
1747
1748 8. Applied Daniel's patches to solve problems with the import/export magic
1749 syntax that is required for Windows, and which was going wrong for the
1750 pcreposix and pcrecpp parts of the library. These were overlooked when this
1751 problem was solved for the main library.
1752
1753 9. There were some crude static tests to avoid integer overflow when computing
1754 the size of patterns that contain repeated groups with explicit upper
1755 limits. As the maximum quantifier is 65535, the maximum group length was
1756 set at 30,000 so that the product of these two numbers did not overflow a
1757 32-bit integer. However, it turns out that people want to use groups that
1758 are longer than 30,000 bytes (though not repeat them that many times).
1759 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1760 made it possible to implement the integer overflow checks in a much more
1761 dynamic way, which I have now done. The artificial limitation on group
1762 length has been removed - we now have only the limit on the total length of
1763 the compiled pattern, which depends on the LINK_SIZE setting.
1764
1765 10. Fixed a bug in the documentation for get/copy named substring when
1766 duplicate names are permitted. If none of the named substrings are set, the
1767 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1768 empty string.
1769
1770 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1771 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1772 because the ] is interpreted as the first data character and the
1773 terminating ] is not found. PCRE has been made compatible with Perl in this
1774 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1775 cause memory overwriting.
1776
1777 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1778 string has been matched (to stop an infinite loop). It was not recognizing
1779 a conditional subpattern that could match an empty string if that
1780 subpattern was within another subpattern. For example, it looped when
1781 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1782 condition was not nested. This bug has been fixed.
1783
1784 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1785 past the start of the subject in the presence of bytes with the top bit
1786 set, for example "\x8aBCD".
1787
1788 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1789 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1790
1791 14. Optimized (?!) to (*FAIL).
1792
1793 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1794 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1795 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1796 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1797 does: it's just the validity check that is more restrictive.
1798
1799 16. Inserted checks for integer overflows during escape sequence (backslash)
1800 processing, and also fixed erroneous offset values for syntax errors during
1801 backslash processing.
1802
1803 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1804 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1805
1806 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1807 caused an overrun.
1808
1809 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1810 something other than just ASCII characters) inside a group that had an
1811 unlimited repeat caused a loop at compile time (while checking to see
1812 whether the group could match an empty string).
1813
1814 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1815 [\P{Any}] did so. (Error in the code for printing property names.)
1816
1817 21. An orphan \E inside a character class could cause a crash.
1818
1819 22. A repeated capturing bracket such as (A)? could cause a wild memory
1820 reference during compilation.
1821
1822 23. There are several functions in pcre_compile() that scan along a compiled
1823 expression for various reasons (e.g. to see if it's fixed length for look
1824 behind). There were bugs in these functions when a repeated \p or \P was
1825 present in the pattern. These operators have additional parameters compared
1826 with \d, etc, and these were not being taken into account when moving along
1827 the compiled data. Specifically:
1828
1829 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1830 length.
1831
1832 (b) An item such as \pL+ within a repeated group could cause crashes or
1833 loops.
1834
1835 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1836 "reference to non-existent subpattern" error.
1837
1838 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1839
1840 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1841 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1842
1843 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1844
1845 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1846 character were causing crashes (broken optimization).
1847
1848 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1849 \p or \P) caused a compile-time loop.
1850
1851 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1852 break. For example, the unstudied pattern [\r\n]A does not match the string
1853 "\r\nA" because change 7.0/46 below moves the current point on by two
1854 characters after failing to match at the start. However, the pattern \nA
1855 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1856 the same is true. There doesn't seem any very clean way out of this, but
1857 what I have chosen to do makes the common cases work: PCRE now takes note
1858 of whether there can be an explicit match for \r or \n anywhere in the
1859 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1860 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1861 pattern has explicit CR or LF references.
1862
1863 29. Added (*CR) etc for changing newline setting at start of pattern.
1864
1865
1866 Version 7.2 19-Jun-07
1867 ---------------------
1868
1869 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1870 which is apparently normally available under Windows.
1871
1872 2. Re-jig the pcregrep tests with different newline settings in an attempt
1873 to make them independent of the local environment's newline setting.
1874
1875 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1876
1877 4. Some of the "internals" tests were previously cut out when the link size
1878 was not 2, because the output contained actual offsets. The recent new
1879 "Z" feature of pcretest means that these can be cut out, making the tests
1880 usable with all link sizes.
1881
1882 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1883 stack recursion. This gives a massive performance boost under BSD, but just
1884 a small improvement under Linux. However, it saves one field in the frame
1885 in all cases.
1886
1887 6. Added more features from the forthcoming Perl 5.10:
1888
1889 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1890 recursion call. It refers to the nth most recently opened parentheses.
1891
1892 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1893 to be opened parentheses.
1894
1895 (c) Conditions that refer to capturing parentheses can be specified
1896 relatively, for example, (?(-2)... or (?(+3)...
1897
1898 (d) \K resets the start of the current match so that everything before
1899 is not part of it.
1900
1901 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1902
1903 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1904 reference syntax.
1905
1906 (g) (?| introduces a group in which the numbering of parentheses in each
1907 alternative starts with the same number.
1908
1909 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1910
1911 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1912 PCRE_INFO_JCHANGED.
1913
1914 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1915 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1916 for detecting groups that can match an empty string.
1917
1918 9. A pattern with a very large number of alternatives (more than several
1919 hundred) was running out of internal workspace during the pre-compile
1920 phase, where pcre_compile() figures out how much memory will be needed. A
1921 bit of new cunning has reduced the workspace needed for groups with
1922 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1923 workspace instead of running out of the 4096 that are available.
1924
1925 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1926
1927 11. Applied patch from Google to remove an optimization that didn't quite work.
1928 The report of the bug said:
1929
1930 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1931 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1932 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1933
1934 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1935 it matched the wrong number of bytes.
1936
1937
1938 Version 7.1 24-Apr-07
1939 ---------------------
1940
1941 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1942 that is more "standard", making use of automake and other Autotools. There
1943 is some re-arrangement of the files and adjustment of comments consequent
1944 on this.
1945
1946 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1947 for recursive directory scanning broke on some systems because the files
1948 are not scanned in any specific order and on different systems the order
1949 was different. A call to "sort" has been inserted into RunGrepTest for the
1950 approprate test as a short-term fix. In the longer term there may be an
1951 alternative.
1952
1953 3. I had an email from Eric Raymond about problems translating some of PCRE's
1954 man pages to HTML (despite the fact that I distribute HTML pages, some
1955 people do their own conversions for various reasons). The problems
1956 concerned the use of low-level troff macros .br and .in. I have therefore
1957 removed all such uses from the man pages (some were redundant, some could
1958 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1959 HTML has been updated to handle .nf/.fi and to complain if it encounters
1960 .br or .in.
1961
1962 4. Updated comments in configure.ac that get placed in config.h.in and also
1963 arranged for config.h to be included in the distribution, with the name
1964 config.h.generic, for the benefit of those who have to compile without
1965 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1966
1967 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1968 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1969 makevp.bat for the current PCRE, using the additional files
1970 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1971
1972 6. A Windows user reported a minor discrepancy with test 2, which turned out
1973 to be caused by a trailing space on an input line that had got lost in his
1974 copy. The trailing space was an accident, so I've just removed it.
1975
1976 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1977 that is needed.
1978
1979 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1980 as "const" (a) because they are and (b) because it helps the PHP
1981 maintainers who have recently made a script to detect big data structures
1982 in the php code that should be moved to the .rodata section. I remembered
1983 to update Builducptable as well, so it won't revert if ucptable.h is ever
1984 re-created.
1985
1986 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1987 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1988 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1989 support is not required. This saves 1.5-2K of code, which is important in
1990 some applications.
1991
1992 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1993 so as not to refer to the tables, even though these functions will never be
1994 called when UTF-8 support is disabled. Otherwise there are problems with a
1995 shared library.
1996
1997 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1998
1999 (a) It was defining its arguments as char * instead of void *.
2000
2001 (b) It was assuming that all moves were upwards in memory; this was true
2002 a long time ago when I wrote it, but is no longer the case.
2003
2004 The emulated memove() is provided for those environments that have neither
2005 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2006 is clearly not the case, as these two bugs were recently reported.
2007
2008 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2009 and Detrail to create the HTML documentation, the .txt form of the man
2010 pages, and it removes trailing spaces from listed files. It also creates
2011 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2012 case, it wraps all the #defines with #ifndefs. This script should be run
2013 before "make dist".
2014
2015 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2016 with Unicode property support.
2017
2018 (a) For a maximizing quantifier, if the two different cases of the
2019 character were of different lengths in their UTF-8 codings (there are
2020 some cases like this - I found 11), and the matching function had to
2021 back up over a mixture of the two cases, it incorrectly assumed they
2022 were both the same length.
2023
2024 (b) When PCRE was configured to use the heap rather than the stack for
2025 recursion during matching, it was not correctly preserving the data for
2026 the other case of a UTF-8 character when checking ahead for a match
2027 while processing a minimizing repeat. If the check also involved
2028 matching a wide character, but failed, corruption could cause an
2029 erroneous result when trying to check for a repeat of the original
2030 character.
2031
2032 13. Some tidying changes to the testing mechanism:
2033
2034 (a) The RunTest script now detects the internal link size and whether there
2035 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2036 values substituted by "configure". (The RunGrepTest script already did
2037 this for UTF-8.) The configure.ac script no longer substitutes the
2038 relevant variables.
2039
2040 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2041 with length and offset values. This means that the output is different
2042 for different internal link sizes. Test 2 is skipped for link sizes
2043 other than 2 because of this, bypassing the problem. Unfortunately,
2044 there was also a test in test 3 (the locale tests) that used /B and
2045 failed for link sizes other than 2. Rather than cut the whole test out,
2046 I have added a new /Z option to pcretest that replaces the length and
2047 offset values with spaces. This is now used to make test 3 independent
2048 of link size. (Test 2 will be tidied up later.)
2049
2050 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2051 segmentation fault instead of returning the appropriate error message.
2052
2053 15. In multiline mode when the newline sequence was set to "any", the pattern
2054 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2055 This doesn't seem right; it now treats the CRLF combination as the line
2056 ending, and so does not match in that case. It's only a pattern such as ^$
2057 that would hit this one: something like ^ABC$ would have failed after \r
2058 and then tried again after \r\n.
2059
2060 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2061 in an attempt to make files that differ only in their line terminators
2062 compare equal. This works on Linux.
2063
2064 17. Under certain error circumstances pcregrep might try to free random memory
2065 as it exited. This is now fixed, thanks to valgrind.
2066
2067 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2068 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2069 was because its rules for how to advance for /g after matching an empty
2070 string at the end of a line did not allow for this case. They now check for
2071 it specially.
2072
2073 20. pcretest is supposed to handle patterns and data of any length, by
2074 extending its buffers when necessary. It was getting this wrong when the
2075 buffer for a data line had to be extended.
2076
2077 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2078 CRLF as a newline sequence.
2079
2080 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2081 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2082 I have nevertheless tidied it up.
2083
2084 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2085
2086 24. Added a man page for pcre-config.
2087
2088
2089 Version 7.0 19-Dec-06
2090 ---------------------
2091
2092 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2093 moving to gcc 4.1.1.
2094
2095 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2096 sys/time.h, which is documented as needed for this function. It doesn't
2097 seem to matter on Linux, but it showed up on some releases of OS X.
2098
2099 3. It seems that there are systems where bytes whose values are greater than
2100 127 match isprint() in the "C" locale. The "C" locale should be the
2101 default when a C program starts up. In most systems, only ASCII printing
2102 characters match isprint(). This difference caused the output from pcretest
2103 to vary, making some of the tests fail. I have changed pcretest so that:
2104
2105 (a) When it is outputting text in the compiled version of a pattern, bytes
2106 other than 32-126 are always shown as hex escapes.
2107
2108 (b) When it is outputting text that is a matched part of a subject string,
2109 it does the same, unless a different locale has been set for the match
2110 (using the /L modifier). In this case, it uses isprint() to decide.
2111
2112 4. Fixed a major bug that caused incorrect computation of the amount of memory
2113 required for a compiled pattern when options that changed within the
2114 pattern affected the logic of the preliminary scan that determines the
2115 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2116 that the computed length was too small. The symptoms of this bug were
2117 either the PCRE error "internal error: code overflow" from pcre_compile(),
2118 or a glibc crash with a message such as "pcretest: free(): invalid next
2119 size (fast)". Examples of patterns that provoked this bug (shown in
2120 pcretest format) are:
2121
2122 /(?-x: )/x
2123 /(?x)(?-x: \s*#\s*)/
2124 /((?i)[\x{c0}])/8
2125 /(?i:[\x{c0}])/8
2126
2127 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2128 is now done differently.
2129
2130 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2131 wrapper classes; (b) implement a new function in the C++ scanner that is
2132 more efficient than the old way of doing things because it avoids levels of
2133 recursion in the regex matching; (c) add a paragraph to the documentation
2134 for the FullMatch() function.
2135
2136 6. The escape sequence \n was being treated as whatever was defined as
2137 "newline". Not only was this contrary to the documentation, which states
2138 that \n is character 10 (hex 0A), but it also went horribly wrong when
2139 "newline" was defined as CRLF. This has been fixed.
2140
2141 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2142 was being set to -1 for the "end of line" case (supposedly a value that no
2143 character can have). Though this value is never used (the check for end of
2144 line is "zero bytes in current character"), it caused compiler complaints.
2145 I've changed it to 0xffffffff.
2146
2147 8. In pcre_version.c, the version string was being built by a sequence of
2148 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2149 string (as it is for production releases) called a macro with an empty
2150 argument. The C standard says the result of this is undefined. The gcc
2151 compiler treats it as an empty string (which was what was wanted) but it is
2152 reported that Visual C gives an error. The source has been hacked around to
2153 avoid this problem.
2154
2155 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2156 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2157 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2158 of them did).
2159
2160 10. Originally, pcretest opened its input and output without "b"; then I was
2161 told that "b" was needed in some environments, so it was added for release
2162 5.0 to both the input and output. (It makes no difference on Unix-like
2163 systems.) Later I was told that it is wrong for the input on Windows. I've
2164 now abstracted the modes into two macros, to make it easier to fiddle with
2165 them, and removed "b" from the input mode under Windows.
2166
2167 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2168
2169 12. Added -help and --help to pcretest as an official way of being reminded
2170 of the options.
2171
2172 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2173 and pcrecpp.cc because they annoy compilers at high warning levels.
2174
2175 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2176
2177 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2178 scripts are not C programs :-) and which was not noticed because it works
2179 on Linux.
2180
2181 16. pcretest is supposed to handle any length of pattern and data line (as one
2182 line or as a continued sequence of lines) by extending its input buffer if
2183 necessary. This feature was broken for very long pattern lines, leading to
2184 a string of junk being passed to pcre_compile() if the pattern was longer
2185 than about 50K.
2186
2187 17. I have done a major re-factoring of the way pcre_compile() computes the
2188 amount of memory needed for a compiled pattern. Previously, there was code
2189 that made a preliminary scan of the pattern in order to do this. That was
2190 OK when PCRE was new, but as the facilities have expanded, it has become
2191 harder and harder to keep it in step with the real compile phase, and there
2192 have been a number of bugs (see for example, 4 above). I have now found a
2193 cunning way of running the real compile function in a "fake" mode that
2194 enables it to compute how much memory it would need, while actually only
2195 ever using a few hundred bytes of working memory and without too many
2196 tests of the mode. This should make future maintenance and development
2197 easier. A side effect of this work is that the limit of 200 on the nesting
2198 depth of parentheses has been removed (though this was never a serious
2199 limitation, I suspect). However, there is a downside: pcre_compile() now
2200 runs more slowly than before (30% or more, depending on the pattern). I
2201 hope this isn't a big issue. There is no effect on runtime performance.
2202
2203 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2204 newline (only possible for the last line of a file) and it was a
2205 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2206
2207 19. Added additional timing features to pcretest. (1) The -tm option now times
2208 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2209 separate command line item, by a number that specifies the number of
2210 repeats to use when timing. The default is 50000; this gives better
2211 precision, but takes uncomfortably long for very large patterns.
2212
2213 20. Extended pcre_study() to be more clever in cases where a branch of a
2214 subpattern has no definite first character. For example, (a*|b*)[cd] would
2215 previously give no result from pcre_study(). Now it recognizes that the
2216 first character must be a, b, c, or d.
2217
2218 21. There was an incorrect error "recursive call could loop indefinitely" if
2219 a subpattern (or the entire pattern) that was being tested for matching an
2220 empty string contained only one non-empty item after a nested subpattern.
2221 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2222 incorrectly, because the \d was being skipped in the check.
2223
2224 22. The pcretest program now has a new pattern option /B and a command line
2225 option -b, which is equivalent to adding /B to every pattern. This causes
2226 it to show the compiled bytecode, without the additional information that
2227 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2228 is the same as /B/I).
2229
2230 23. A new optimization is now able automatically to treat some sequences such
2231 as a*b as a*+b. More specifically, if something simple (such as a character
2232 or a simple class like \d) has an unlimited quantifier, and is followed by
2233 something that cannot possibly match the quantified thing, the quantifier
2234 is automatically "possessified".
2235
2236 24. A recursive reference to a subpattern whose number was greater than 39
2237 went wrong under certain circumstances in UTF-8 mode. This bug could also
2238 have affected the operation of pcre_study().
2239
2240 25. Realized that a little bit of performance could be had by replacing
2241 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2242
2243 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2244
2245 27. Possessive quantifiers such as a++ were previously implemented by turning
2246 them into atomic groups such as ($>a+). Now they have their own opcodes,
2247 which improves performance. This includes the automatically created ones
2248 from 23 above.
2249
2250 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2251 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2252 the first matched character to be a colon. This applied both to named and
2253 numbered groups.
2254
2255 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2256
2257 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2258 building PCRE on Windows easier, so I have included it in the distribution.
2259
2260 31. There is now a check in pcretest against a ridiculously large number being
2261 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2262 loop, the loop is abandoned.
2263
2264 32. Forward references to subpatterns in conditions such as (?(2)...) where
2265 subpattern 2 is defined later cause pcre_compile() to search forwards in
2266 the pattern for the relevant set of parentheses. This search went wrong
2267 when there were unescaped parentheses in a character class, parentheses
2268 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2269
2270 33. "Subroutine" calls and backreferences were previously restricted to
2271 referencing subpatterns earlier in the regex. This restriction has now
2272 been removed.
2273
2274 34. Added a number of extra features that are going to be in Perl 5.10. On the
2275 whole, these are just syntactic alternatives for features that PCRE had
2276 previously implemented using the Python syntax or my own invention. The
2277 other formats are all retained for compatibility.
2278
2279 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2280 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2281 also .NET compatible.
2282
2283 (b) A recursion or subroutine call to a named group can now be defined as
2284 (?&name) as well as (?P>name).
2285
2286 (c) A backreference to a named group can now be defined as \k<name> or
2287 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2288 5.10, are also .NET compatible.
2289
2290 (d) A conditional reference to a named group can now use the syntax
2291 (?(<name>) or (?('name') as well as (?(name).
2292
2293 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2294 groups (named and numbered) that are never evaluated inline, but can be
2295 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2296 is always false. There may be only one alternative in such a group.
2297
2298 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2299 as the simple (?(R). The condition is true only if the most recent
2300 recursion is that of the given number or name. It does not search out
2301 through the entire recursion stack.
2302
2303 (g) The escape \gN or \g{N} has been added, where N is a positive or
2304 negative number, specifying an absolute or relative reference.
2305
2306 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2307 some "unreachable code" warnings.
2308
2309 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2310 things, this adds five new scripts.
2311
2312 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2313 There were also incompatibilities regarding the handling of \Q..\E inside
2314 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2315 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2316
2317 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2318 matches an empty string, and forcibly breaks the loop. There were bugs in
2319 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2320 against aaaa the result was just "a" rather than "aaaa", for example. Two
2321 separate and independent bugs (that affected different cases) have been
2322 fixed.
2323
2324 39. Refactored the code to abolish the use of different opcodes for small
2325 capturing bracket numbers. This is a tidy that I avoided doing when I
2326 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2327 The new approach is not only tidier, it makes it possible to reduce the
2328 memory needed to fix the previous bug (38).
2329
2330 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2331 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2332 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2333 mode.
2334
2335 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2336 report.
2337
2338 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2339 copy construction and assignment in the C++ wrapper.
2340
2341 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2342 couple of bugs that could have given wrong results in the "--newline=crlf"
2343 case.
2344
2345 44. Added a number of casts and did some reorganization of signed/unsigned int
2346 variables following suggestions from Dair Grant. Also renamed the variable
2347 "this" as "item" because it is a C++ keyword.
2348
2349 45. Arranged for dftables to add
2350
2351 #include "pcre_internal.h"
2352
2353 to pcre_chartables.c because without it, gcc 4.x may remove the array
2354 definition from the final binary if PCRE is built into a static library and
2355 dead code stripping is activated.
2356
2357 46. For an unanchored pattern, if a match attempt fails at the start of a
2358 newline sequence, and the newline setting is CRLF or ANY, and the next two
2359 characters are CRLF, advance by two characters instead of one.
2360
2361
2362 Version 6.7 04-Jul-06
2363 ---------------------
2364
2365 1. In order to handle tests when input lines are enormously long, pcretest has
2366 been re-factored so that it automatically extends its buffers when
2367 necessary. The code is crude, but this _is_ just a test program. The
2368 default size has been increased from 32K to 50K.
2369
2370 2. The code in pcre_study() was using the value of the re argument before
2371 testing it for NULL. (Of course, in any sensible call of the function, it
2372 won't be NULL.)
2373
2374 3. The memmove() emulation function in pcre_internal.h, which is used on
2375 systems that lack both memmove() and bcopy() - that is, hardly ever -
2376 was missing a "static" storage class specifier.
2377
2378 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2379 containing an extended class (one that cannot be represented by a bitmap
2380 because it contains high-valued characters or Unicode property items, e.g.
2381 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2382 pattern, but PCRE should not loop if you do not (it no longer does).
2383 [Detail: two cases were found: (a) a repeated subpattern containing an
2384 extended class; (b) a recursive reference to a subpattern that followed a
2385 previous extended class. It wasn't skipping over the extended class
2386 correctly when UTF-8 mode was not set.]
2387
2388 5. A negated single-character class was not being recognized as fixed-length
2389 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2390 compile error "lookbehind assertion is not fixed length".
2391
2392 6. The RunPerlTest auxiliary script was showing an unexpected difference
2393 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2394 write a Perl script that can interpret lines of an input file either as
2395 byte characters or as UTF-8, which is what "perltest" was being required to
2396 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2397 can't do is switch easily at run time between having the "use utf8;" pragma
2398 or not. In the end, I fudged it by using the RunPerlTest script to insert
2399 "use utf8;" explicitly for the UTF-8 tests.
2400
2401 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2402 the end of the subject string, contrary to the documentation and to what
2403 Perl does. This was true of both matching functions. Now it matches only at
2404 the start of the subject and immediately after *internal* newlines.
2405
2406 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2407 a pointer to an int instead of a pointer to an unsigned long int. This
2408 caused problems on 64-bit systems.
2409
2410 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2411 instance of the 'standard' template library not being so standard".
2412
2413 10. There was no check on the number of named subpatterns nor the maximum
2414 length of a subpattern name. The product of these values is used to compute
2415 the size of the memory block for a compiled pattern. By supplying a very
2416 long subpattern name and a large number of named subpatterns, the size
2417 computation could be caused to overflow. This is now prevented by limiting
2418 the length of names to 32 characters, and the number of named subpatterns
2419 to 10,000.
2420
2421 11. Subpatterns that are repeated with specific counts have to be replicated in
2422 the compiled pattern. The size of memory for this was computed from the
2423 length of the subpattern and the repeat count. The latter is limited to
2424 65535, but there was no limit on the former, meaning that integer overflow
2425 could in principle occur. The compiled length of a repeated subpattern is
2426 now limited to 30,000 bytes in order to prevent this.
2427
2428 12. Added the optional facility to have named substrings with the same name.
2429
2430 13. Added the ability to use a named substring as a condition, using the
2431 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2432 are numbers (not recommended). Forward references are permitted.
2433
2434 14. Added forward references in named backreferences (if you see what I mean).
2435
2436 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2437 pattern could run off the end of the subject. For example, the pattern
2438 "(?s)(.{1,5})"8 did this with the subject "ab".
2439
2440 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2441 PCRE_CASELESS was set when matching characters that were quantified with ?
2442 or *.
2443
2444 17. A character class other than a single negated character that had a minimum
2445 but no maximum quantifier - for example [ab]{6,} - was not handled
2446 correctly by pce_dfa_exec(). It would match only one character.
2447
2448 18. A valid (though odd) pattern that looked like a POSIX character
2449 class but used an invalid character after [ (for example [[,abc,]]) caused
2450 pcre_compile() to give the error "Failed: internal error: code overflow" or
2451 in some cases to crash with a glibc free() error. This could even happen if
2452 the pattern terminated after [[ but there just happened to be a sequence of
2453 letters, a binary zero, and a closing ] in the memory that followed.
2454
2455 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2456 over the years. Originally (before any Unicode support), just the bottom 8
2457 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2458 output from "man perlunicode" includes this:
2459
2460 The regular expression compiler produces polymorphic opcodes. That
2461 is, the pattern adapts to the data and automatically switches to
2462 the Unicode character scheme when presented with Unicode data--or
2463 instead uses a traditional byte scheme when presented with byte
2464 data.
2465
2466 Sadly, a wide octal escape does not cause a switch, and in a string with
2467 no other multibyte characters, these octal escapes are treated as before.
2468 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2469 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2470 Unicode string.
2471
2472 I have not perpetrated such confusion in PCRE. Up till now, it took just
2473 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2474 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2475 translate to the appropriate multibyte character.
2476
2477 29. Applied some refactoring to reduce the number of warnings from Microsoft
2478 and Borland compilers. This has included removing the fudge introduced
2479 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2480 a warning about an unused variable.
2481
2482 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2483 characters since release 4.0, because Perl (from release 5.004) does not.
2484 [Or at least, is documented not to: some releases seem to be in conflict
2485 with the documentation.] However, when a pattern was studied with
2486 pcre_study() and all its branches started with \s, PCRE still included VT
2487 as a possible starting character. Of course, this did no harm; it just
2488 caused an unnecessary match attempt.
2489
2490 22. Removed a now-redundant internal flag bit that recorded the fact that case
2491 dependency changed within the pattern. This was once needed for "required
2492 byte" processing, but is no longer used. This recovers a now-scarce options
2493 bit. Also moved the least significant internal flag bit to the most-
2494 significant bit of the word, which was not previously used (hangover from
2495 the days when it was an int rather than a uint) to free up another bit for
2496 the future.
2497
2498 23. Added support for CRLF line endings as well as CR and LF. As well as the
2499 default being selectable at build time, it can now be changed at runtime
2500 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2501 specify that it is scanning data with non-default line endings.
2502
2503 24. Changed the definition of CXXLINK to make it agree with the definition of
2504 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2505
2506 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2507 recursions. This makes a big different to stack usage for some patterns.
2508
2509 26. If a subpattern containing a named recursion or subroutine reference such
2510 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2511 the space required for the compiled pattern went wrong and gave too small a
2512 value. Depending on the environment, this could lead to "Failed: internal
2513 error: code overflow at offset 49" or "glibc detected double free or
2514 corruption" errors.
2515
2516 27. Applied patches from Google (a) to support the new newline modes and (b) to
2517 advance over multibyte UTF-8 characters in GlobalReplace.
2518
2519 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2520 difference for some implementation of PCRE in some Windows version.
2521
2522 29. Added some extra testing facilities to pcretest:
2523
2524 \q<number> in a data line sets the "match limit" value
2525 \Q<number> in a data line sets the "match recursion limt" value
2526 -S <number> sets the stack size, where <number> is in megabytes
2527
2528 The -S option isn't available for Windows.
2529
2530
2531 Version 6.6 06-Feb-06
2532 ---------------------
2533
2534 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2535 in pcreposix.h. I have copied the definition from pcre.h.
2536
2537 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2538 because pcre.h is no longer a built file.
2539
2540 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2541 not normally included in the compiled code.
2542
2543
2544 Version 6.5 01-Feb-06
2545 ---------------------
2546
2547 1. When using the partial match feature with pcre_dfa_exec(), it was not
2548 anchoring the second and subsequent partial matches at the new starting
2549 point. This could lead to incorrect results. For example, with the pattern
2550 /1234/, partially matching against "123" and then "a4" gave a match.
2551
2552 2. Changes to pcregrep:
2553
2554 (a) All non-match returns from pcre_exec() were being treated as failures
2555 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2556 error message is output. Some extra information is given for the
2557 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2558 probably the only errors that are likely to be caused by users (by
2559 specifying a regex that has nested indefinite repeats, for instance).
2560 If there are more than 20 of these errors, pcregrep is abandoned.
2561
2562 (b) A binary zero was treated as data while matching, but terminated the
2563 output line if it was written out. This has been fixed: binary zeroes
2564 are now no different to any other data bytes.
2565
2566 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2567 used to set a locale for matching. The --locale=xxxx long option has
2568 been added (no short equivalent) to specify a locale explicitly on the
2569 pcregrep command, overriding the environment variables.
2570
2571 (d) When -B was used with -n, some line numbers in the output were one less
2572 than they should have been.
2573
2574 (e) Added the -o (--only-matching) option.
2575
2576 (f) If -A or -C was used with -c (count only), some lines of context were
2577 accidentally printed for the final match.
2578
2579 (g) Added the -H (--with-filename) option.
2580
2581 (h) The combination of options -rh failed to suppress file names for files
2582 that were found from directory arguments.
2583
2584 (i) Added the -D (--devices) and -d (--directories) options.
2585
2586 (j) Added the -F (--fixed-strings) option.
2587
2588 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2589
2590 (l) Added the --colo(u)r option.
2591
2592 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2593 is not present by default.
2594
2595 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2596 items such as (?R) or (?1), when the recursion could match a number of
2597 alternatives. If it matched one of the alternatives, but subsequently,
2598 outside the recursion, there was a failure, the code tried to back up into
2599 the recursion. However, because of the way PCRE is implemented, this is not
2600 possible, and the result was an incorrect result from the match.
2601
2602 In order to prevent this happening, the specification of recursion has
2603 been changed so that all such subpatterns are automatically treated as
2604 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2605
2606 4. I had overlooked the fact that, in some locales, there are characters for
2607 which isalpha() is true but neither isupper() nor islower() are true. In
2608 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2609 and ordfeminine) are like this. This affected the treatment of \w and \W
2610 when they appeared in character classes, but not when they appeared outside
2611 a character class. The bit map for "word" characters is now created
2612 separately from the results of isalnum() instead of just taking it from the
2613 upper, lower, and digit maps. (Plus the underscore character, of course.)
2614
2615 5. The above bug also affected the handling of POSIX character classes such as
2616 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2617 permanent tables. Instead, the bit maps for such a class were previously
2618 created as the appropriate unions of the upper, lower, and digit bitmaps.
2619 Now they are created by subtraction from the [[:word:]] class, which has
2620 its own bitmap.
2621
2622 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2623 It is created by subtracting the vertical space characters (\x09, \x0a,
2624 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2625 subtraction was done in the overall bitmap for a character class, meaning
2626 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2627 be recognized. This bug has been fixed.
2628
2629 7. Patches from the folks at Google:
2630
2631 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2632 real life, but is still worth protecting against".
2633
2634 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2635 regular expressions".
2636
2637 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2638 have it.
2639
2640 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2641 "configure" and the latter not, in order to fix a problem somebody had
2642 with compiling the Arg class on HP-UX.
2643
2644 (e) Improve the error-handling of the C++ wrapper a little bit.
2645
2646 (f) New tests for checking recursion limiting.
2647
2648 8. The pcre_memmove() function, which is used only if the environment does not
2649 have a standard memmove() function (and is therefore rarely compiled),
2650 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2651 returning a result (though PCRE never actually uses the result).
2652
2653 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2654 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2655 returned instead of calling malloc() with an overflowing number that would
2656 most likely cause subsequent chaos.
2657
2658 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2659
2660 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2661 with this option is matched, the nmatch and pmatch options of regexec() are
2662 ignored.
2663
2664 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2665 provided in case anyone wants to the the POSIX interface with UTF-8
2666 strings.
2667
2668 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2669 C++ linking (needed for some HP-UX environments).
2670
2671 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2672 (unused parameter) and in the pcre_printint() function (omitted "default"
2673 switch label when the default is to do nothing).
2674
2675 15. Added some code to make it possible, when PCRE is compiled as a C++
2676 library, to replace subject pointers for pcre_exec() with a smart pointer
2677 class, thus making it possible to process discontinuous strings.
2678
2679 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2680 much the same function. They were added by different people who were trying
2681 to make PCRE easy to compile on non-Unix systems. It has been suggested
2682 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2683 for compiling on Windows systems. I have therefore replaced it with
2684 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2685 defaults to "extern" for C or "extern C" for C++, which works fine on
2686 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2687 SCOPE with something explicit in config.h. In addition:
2688
2689 (a) pcreposix.h still had just "extern" instead of either of these macros;
2690 I have replaced it with PCRE_DATA_SCOPE.
2691
2692 (b) Functions such as _pcre_xclass(), which are internal to the library,
2693 but external in the C sense, all had PCRE_EXPORT in their definitions.
2694 This is apparently wrong for the Windows case, so I have removed it.
2695 (It makes no difference on Unix-like systems.)
2696
2697 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2698 of recursive calls to match(). This is different to MATCH_LIMIT because
2699 that limits the total number of calls to match(), not all of which increase
2700 the depth of recursion. Limiting the recursion depth limits the amount of
2701 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2702 when PCRE is compiled, and changed at run time. A patch from Google adds
2703 this functionality to the C++ interface.
2704
2705 18. Changes to the handling of Unicode character properties:
2706
2707 (a) Updated the table to Unicode 4.1.0.
2708
2709 (b) Recognize characters that are not in the table as "Cn" (undefined).
2710
2711 (c) I revised the way the table is implemented to a much improved format
2712 which includes recognition of ranges. It now supports the ranges that
2713 are defined in UnicodeData.txt, and it also amalgamates other
2714 characters into ranges. This has reduced the number of entries in the
2715 table from around 16,000 to around 3,000, thus reducing its size
2716 considerably. I realized I did not need to use a tree structure after
2717 all - a binary chop search is just as efficient. Having reduced the
2718 number of entries, I extended their size from 6 bytes to 8 bytes to
2719 allow for more data.
2720
2721 (d) Added support for Unicode script names via properties such as \p{Han}.
2722
2723 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2724 matching that character.
2725
2726 20. When matching a repeated Unicode property with a minimum greater than zero,
2727 (for example \pL{2,}), PCRE could look past the end of the subject if it
2728 reached it while seeking the minimum number of characters. This could
2729 happen only if some of the characters were more than one byte long, because
2730 there is a check for at least the minimum number of bytes.
2731
2732 21. Refactored the implementation of \p and \P so as to be more general, to
2733 allow for more different types of property in future. This has changed the
2734 compiled form incompatibly. Anybody with saved compiled patterns that use
2735 \p or \P will have to recompile them.
2736
2737 22. Added "Any" and "L&" to the supported property types.
2738
2739 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2740 but give a compile time error if the value is greater than 0xff.
2741
2742 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2743 accidentally not being installed or uninstalled.
2744
2745 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2746 made were to insert the current release number. This seemed silly, because
2747 it made things harder for people building PCRE on systems that don't run
2748 "configure". I have turned pcre.h into a distributed file, no longer built
2749 by "configure", with the version identification directly included. There is
2750 no longer a pcre.h.in file.
2751
2752 However, this change necessitated a change to the pcre-config script as
2753 well. It is built from pcre-config.in, and one of the substitutions was the
2754 release number. I have updated configure.ac so that ./configure now finds
2755 the release number by grepping pcre.h.
2756
2757 26. Added the ability to run the tests under valgrind.
2758
2759
2760 Version 6.4 05-Sep-05
2761 ---------------------
2762
2763 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2764 "--" to be printed when multiple files were scanned, even when none of the
2765 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2766 consider it to be a bug, and have restored the previous behaviour.
2767
2768 2. A couple of code tidies to get rid of compiler warnings.
2769
2770 3. The pcretest program used to cheat by referring to symbols in the library
2771 whose names begin with _pcre_. These are internal symbols that are not
2772 really supposed to be visible externally, and in some environments it is
2773 possible to suppress them. The cheating is now confined to including
2774 certain files from the library's source, which is a bit cleaner.
2775
2776 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2777 file's purpose clearer.
2778
2779 5. Reorganized pcre_ucp_findchar().
2780
2781
2782 Version 6.3 15-Aug-05
2783 ---------------------
2784
2785 1. The file libpcre.pc.in did not have general read permission in the tarball.
2786
2787 2. There were some problems when building without C++ support:
2788
2789 (a) If C++ support was not built, "make install" and "make test" still
2790 tried to test it.
2791
2792 (b) There were problems when the value of CXX was explicitly set. Some
2793 changes have been made to try to fix these, and ...
2794
2795 (c) --disable-cpp can now be used to explicitly disable C++ support.
2796
2797 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2798 backslash in a target when C++ was disabled. This confuses some
2799 versions of "make", apparently. Using an intermediate variable solves
2800 this. (Same for CPP_LOBJ.)
2801
2802 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2803 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2804 necessary on certain architectures.
2805
2806 4. Added a setting of -export-symbols-regex to the link command to remove
2807 those symbols that are exported in the C sense, but actually are local
2808 within the library, and not documented. Their names all begin with
2809 "_pcre_". This is not a perfect job, because (a) we have to except some
2810 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2811 available (and never for static libraries). I have made a note to try to
2812 find a way round (a) in the future.
2813
2814
2815 Version 6.2 01-Aug-05
2816 ---------------------
2817
2818 1. There was no test for integer overflow of quantifier values. A construction
2819 such as {1111111111111111} would give undefined results. What is worse, if
2820 a minimum quantifier for a parenthesized subpattern overflowed and became
2821 negative, the calculation of the memory size went wrong. This could have
2822 led to memory overwriting.
2823
2824 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2825
2826 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2827 operating environments where this matters.
2828
2829 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2830 PCRE options from within the C++ wrapper.
2831
2832 5. Named capturing subpatterns were not being correctly counted when a pattern
2833 was compiled. This caused two problems: (a) If there were more than 100
2834 such subpatterns, the calculation of the memory needed for the whole
2835 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2836 back references of the form \12, where the number was greater than 9, were
2837 not recognized as back references, even though there were sufficient
2838 previous subpatterns.
2839
2840 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2841 versions of gcc, e.g. 2.95.4.
2842
2843
2844 Version 6.1 21-Jun-05
2845 ---------------------
2846
2847 1. There was one reference to the variable "posix" in pcretest.c that was not
2848 surrounded by "#if !defined NOPOSIX".
2849
2850 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2851 the cross-check on the old pcre_info() function, for the benefit of the
2852 cut-down version of PCRE that is currently imported into Exim.
2853
2854 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2855 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2856 patterns that start (?i) but I don't think that matters. The use of (?i) is
2857 just an example; this all applies to the other options as well.
2858
2859 4. Since libtool seems to echo the compile commands it is issuing, the output
2860 from "make" can be reduced a bit by putting "@" in front of each libtool
2861 compile command.
2862
2863 5. Patch from the folks at Google for configure.in to be a bit more thorough
2864 in checking for a suitable C++ installation before trying to compile the
2865 C++ stuff. This should fix a reported problem when a compiler was present,
2866 but no suitable headers.
2867
2868 6. The man pages all had just "PCRE" as their title. I have changed them to
2869 be the relevant file name. I have also arranged that these names are
2870 retained in the file doc/pcre.txt, which is a concatenation in text format
2871 of all the man pages except the little individual ones for each function.
2872
2873 7. The NON-UNIX-USE file had not been updated for the different set of source
2874 files that come with release 6. I also added a few comments about the C++
2875 wrapper.
2876
2877
2878 Version 6.0 07-Jun-05
2879 ---------------------
2880
2881 1. Some minor internal re-organization to help with my DFA experiments.
2882
2883 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2884 didn't matter for the library itself when fully configured, but did matter
2885 when compiling without UCP support, or within Exim, where the ucp files are
2886 not imported.
2887
2888 3. Refactoring of the library code to split up the various functions into
2889 different source modules. The addition of the new DFA matching code (see
2890 below) to a single monolithic source would have made it really too
2891 unwieldy, quite apart from causing all the code to be include in a
2892 statically linked application, when only some functions are used. This is
2893 relevant even without the DFA addition now that patterns can be compiled in
2894 one application and matched in another.
2895
2896 The downside of splitting up is that there have to be some external
2897 functions and data tables that are used internally in different modules of
2898 the library but which are not part of the API. These have all had their
2899 names changed to start with "_pcre_" so that they are unlikely to clash
2900 with other external names.
2901
2902 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2903 a different (DFA) algorithm. Although it is slower than the original
2904 function, it does have some advantages for certain types of matching
2905 problem.
2906
2907 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2908 including restarting after a partial match.
2909
2910 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2911 defined when compiling for Windows was sent to me. I have put it into the
2912 code, though I have no means of testing or verifying it.
2913
2914 7. Added the pcre_refcount() auxiliary function.
2915
2916 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2917 match before or at the first newline in the subject string. In pcretest,
2918 the /f option on a pattern can be used to set this.
2919
2920 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2921 would behave wrongly. This has been present in PCRE since release 4.0.
2922
2923 10. A number of changes to the pcregrep command:
2924
2925 (a) Refactored how -x works; insert ^(...)$ instead of setting
2926 PCRE_ANCHORED and checking the length, in preparation for adding
2927 something similar for -w.
2928
2929 (b) Added the -w (match as a word) option.
2930
2931 (c) Refactored the way lines are read and buffered so as to have more
2932 than one at a time available.
2933
2934 (d) Implemented a pcregrep test script.
2935
2936 (e) Added the -M (multiline match) option. This allows patterns to match
2937 over several lines of the subject. The buffering ensures that at least
2938 8K, or the rest of the document (whichever is the shorter) is available
2939 for matching (and similarly the previous 8K for lookbehind assertions).
2940
2941 (f) Changed the --help output so that it now says
2942
2943 -w, --word-regex(p)
2944
2945 instead of two lines, one with "regex" and the other with "regexp"
2946 because that confused at least one person since the short forms are the
2947 same. (This required a bit of code, as the output is generated
2948 automatically from a table. It wasn't just a text change.)
2949
2950 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2951 option but starts with a hyphen. Could be a pattern or a path name
2952 starting with a hyphen, for instance.
2953
2954 (h) "-" can be given as a file name to represent stdin.
2955
2956 (i) When file names are being printed, "(standard input)" is used for
2957 the standard input, for compatibility with GNU grep. Previously
2958 "<stdin>" was used.
2959
2960 (j) The option --label=xxx can be used to supply a name to be used for
2961 stdin when file names are being printed. There is no short form.
2962
2963 (k) Re-factored the options decoding logic because we are going to add
2964 two more options that take data. Such options can now be given in four
2965 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2966
2967 (l) Added the -A, -B, and -C options for requesting that lines of context
2968 around matches be printed.
2969
2970 (m) Added the -L option to print the names of files that do not contain
2971 any matching lines, that is, the complement of -l.
2972
2973 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2974 continue to scan other files.
2975
2976 (o) The -s option was incorrectly implemented. For compatibility with other
2977 greps, it now suppresses the error message for a non-existent or non-
2978 accessible file (but not the return code). There is a new option called
2979 -q that suppresses the output of matching lines, which was what -s was
2980 previously doing.
2981
2982 (p) Added --include and --exclude options to specify files for inclusion
2983 and exclusion when recursing.
2984
2985 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2986 Hopefully, it now does.
2987
2988 12. Missing cast in pcre_study().
2989
2990 13. Added an "uninstall" target to the makefile.
2991
2992 14. Replaced "extern" in the function prototypes in Makefile.in with
2993 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2994 world, but is set differently for Windows.
2995
2996 15. Added a second compiling function called pcre_compile2(). The only
2997 difference is that it has an extra argument, which is a pointer to an
2998 integer error code. When there is a compile-time failure, this is set
2999 non-zero, in addition to the error test pointer being set to point to an
3000 error message. The new argument may be NULL if no error number is required
3001 (but then you may as well call pcre_compile(), which is now just a
3002 wrapper). This facility is provided because some applications need a
3003 numeric error indication, but it has also enabled me to tidy up the way
3004 compile-time errors are handled in the POSIX wrapper.
3005
3006 16. Added VPATH=.libs to the makefile; this should help when building with one
3007 prefix path and installing with another. (Or so I'm told by someone who
3008 knows more about this stuff than I do.)
3009
3010 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3011 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3012 match everything, including newlines. This is not POSIX-compatible, but
3013 somebody wanted the feature. From pcretest it can be activated by using
3014 both the P and the s flags.
3015
3016 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3017
3018 19. libpcre.pc was being incorrectly installed as executable.
3019
3020 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3021 it now also looks for '\r' so that it will work unmodified on Windows.
3022
3023 21. Added Google's contributed C++ wrapper to the distribution.
3024
3025 22. Added some untidy missing memory free() calls in pcretest, to keep
3026 Electric Fence happy when testing.
3027
3028
3029
3030 Version 5.0 13-Sep-04
3031 ---------------------
3032
3033 1. Internal change: literal characters are no longer packed up into items
3034 containing multiple characters in a single byte-string. Each character
3035 is now matched using a separate opcode. However, there may be more than one
3036 byte in the character in UTF-8 mode.
3037
3038 2. The pcre_callout_block structure has two new fields: pattern_position and
3039 next_item_length. These contain the offset in the pattern to the next match
3040 item, and its length, respectively.
3041
3042 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3043 insertion of callouts before each pattern item. Added the /C option to
3044 pcretest to make use of this.
3045
3046 4. On the advice of a Windows user, the lines
3047
3048 #if defined(_WIN32) || defined(WIN32)
3049 _setmode( _fileno( stdout ), 0x8000 );
3050 #endif /* defined(_WIN32) || defined(WIN32) */
3051
3052 have been added to the source of pcretest. This apparently does useful
3053 magic in relation to line terminators.
3054
3055 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3056 for the benefit of those environments where the "b" makes a difference.
3057
3058 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3059 to know about it. I have put a hack into configure.in that adds in code
3060 to set GCC=yes if CC=icc. This seems to end up at a point in the
3061 generated configure script that is early enough to affect the setting of
3062 compiler options, which is what is needed, but I have no means of testing
3063 whether it really works. (The user who reported this had patched the
3064 generated configure script, which of course I cannot do.)
3065
3066 LATER: After change 22 below (new libtool files), the configure script
3067 seems to know about icc (and also ecc). Therefore, I have commented out
3068 this hack in configure.in.
3069
3070 7. Added support for pkg-config (2 patches were sent in).
3071
3072 8. Negated POSIX character classes that used a combination of internal tables
3073 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3074 [[:^ascii]]. Typically, they would match almost any characters. The other
3075 POSIX classes were not broken in this way.
3076
3077 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3078 to find the match, as PCRE was deluded into thinking that the match had to
3079 start at the start point or following a newline. The same bug applied to
3080 patterns with negative forward assertions or any backward assertions
3081 preceding ".*" at the start, unless the pattern required a fixed first
3082 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3083
3084 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3085 starting at the last subject character, bytes beyond the end of the subject
3086 string were read.
3087
3088 11. Renamed the variable "class" as "classbits" to make life easier for C++
3089 users. (Previously there was a macro definition, but it apparently wasn't
3090 enough.)
3091
3092 12. Added the new field "tables" to the extra data so that tables can be passed
3093 in at exec time, or the internal tables can be re-selected. This allows
3094 a compiled regex to be saved and re-used at a later time by a different
3095 program that might have everything at different addresses.
3096
3097 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3098 -R library as well as a -L library.
3099
3100 14. The debugging options of pcretest (-d on the command line or D on a
3101 pattern) showed incorrect output for anything following an extended class
3102 that contained multibyte characters and which was followed by a quantifier.
3103
3104 15. Added optional support for general category Unicode character properties
3105 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3106 support. It adds about 90K to the size of the library. The meanings of the
3107 inbuilt class escapes such as \d and \s have NOT been changed.
3108
3109 16. Updated pcredemo.c to include calls to free() to release the memory for the
3110 compiled pattern.
3111
3112 17. The generated file chartables.c was being created in the source directory
3113 instead of in the building directory. This caused the build to fail if the
3114 source directory was different from the building directory, and was
3115 read-only.
3116
3117 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3118 file. No doubt somebody will tell me if they don't make sense... Also added
3119 Dan Mooney's comments about building on OpenVMS.
3120
3121 19. Added support for partial matching via the PCRE_PARTIAL option for
3122 pcre_exec() and the \P data escape in pcretest.
3123
3124 20. Extended pcretest with 3 new pattern features:
3125
3126 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3127 write the compiled pattern to the file whose name is "rest-of-line".
3128 This is a straight binary dump of the data, with the saved pointer to
3129 the character tables forced to be NULL. The study data, if any, is
3130 written too. After writing, pcretest reads a new pattern.
3131
3132 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3133 compiled pattern from the given file. There must not be any
3134 occurrences of "<" in the file name (pretty unlikely); if there are,
3135 pcretest will instead treat the initial "<" as a pattern delimiter.
3136 After reading in the pattern, pcretest goes on to read data lines as
3137 usual.
3138
3139 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3140 and 16-bit fields in a compiled pattern, to simulate a pattern that
3141 was compiled on a host of opposite endianness.
3142
3143 21. The pcre-exec() function can now cope with patterns that were compiled on
3144 hosts of opposite endianness, with this restriction:
3145
3146 As for any compiled expression that is saved and used later, the tables
3147 pointer field cannot be preserved; the extra_data field in the arguments
3148 to pcre_exec() should be used to pass in a tables address if a value
3149 other than the default internal tables were used at compile time.
3150
3151 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3152 now diagnosed as an error. Previously, most of the time, a negative number
3153 would have been treated as zero, but if in addition "ovector" was passed as
3154 NULL, a crash could occur.
3155
3156 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3157 new versions from the libtool 1.5 distribution (the last one is a copy of
3158 a file called libtool.m4). This seems to have fixed the need to patch
3159 "configure" to support Darwin 1.3 (which I used to do). However, I still
3160 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3161 workstation).
3162
3163 24. Changed the PCRE licence to be the more standard "BSD" licence.
3164
3165
3166 Version 4.5 01-Dec-03
3167 ---------------------
3168
3169 1. There has been some re-arrangement of the code for the match() function so
3170 that it can be compiled in a version that does not call itself recursively.
3171 Instead, it keeps those local variables that need separate instances for
3172 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3173 needs to "recurse". Keeping track of where control must go is done by means
3174 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3175 hide most of the details from the main code, and operates only if
3176 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3177 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3178 operating.
3179
3180 To make it easier for callers to provide specially tailored get/free
3181 functions for this usage, two new functions, pcre_stack_malloc, and
3182 pcre_stack_free, are used. They are always called in strict stacking order,
3183 and the size of block requested is always the same.
3184
3185 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3186 PCRE has been compiled to use the stack or the heap for recursion. The
3187 -C option of pcretest uses this to show which version is compiled.
3188
3189 A new data escape \S, is added to pcretest; it causes the amounts of store
3190 obtained and freed by both kinds of malloc/free at match time to be added
3191 to the output.
3192
3193 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3194 what's available on my current Linux desktop machine.
3195
3196 3. When matching a UTF-8 string, the test for a valid string at the start has
3197 been extended. If start_offset is not zero, PCRE now checks that it points
3198 to a byte that is the start of a UTF-8 character. If not, it returns
3199 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3200 this is necessary because there may be backward assertions in the pattern.
3201 When matching the same subject several times, it may save resources to use
3202 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3203
3204 4. The code for checking the validity of UTF-8 strings has been tightened so
3205 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3206 containing "overlong sequences".
3207
3208 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3209 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3210 should have been "&", but it just so happened that all the cases this let
3211 through by mistake were picked up later in the function.
3212
3213 6. I had used a variable called "isblank" - this is a C99 function, causing
3214 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3215
3216 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3217 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3218 so the version is shown; (c) stop "make test" echoing "./RunTest".
3219
3220 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3221
3222 9. The prototype for memmove() for systems that don't have it was using
3223 size_t, but the inclusion of the header that defines size_t was later. I've
3224 moved the #includes for the C headers earlier to avoid this.
3225
3226 10. Added some adjustments to the code to make it easier to compiler on certain
3227 special systems:
3228
3229 (a) Some "const" qualifiers were missing.
3230 (b) Added the macro EXPORT before all exported functions; by default this
3231 is defined to be empty.
3232 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3233 that it reads its output file name as an argument instead of writing
3234 to the standard output and assuming this can be redirected.
3235
3236 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3237 class containing characters with values greater than 255, PCRE compilation
3238 went into a loop.
3239
3240 12. A recursive reference to a subpattern that was within another subpattern
3241 that had a minimum quantifier of zero caused PCRE to crash. For example,
3242 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3243 recursion. If the recursively-called subpattern itself had a zero repeat,
3244 that was OK.
3245
3246 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3247 buffer into which it was copied (for escape processing) was still set at
3248 1024, so long lines caused crashes.
3249
3250 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3251 "internal error: code overflow...". This applied to any character class
3252 that was followed by a possessive quantifier.
3253
3254 15. Modified the Makefile to add libpcre.la as a prerequisite for
3255 libpcreposix.la because I was told this is needed for a parallel build to
3256 work.
3257
3258 16. If a pattern that contained .* following optional items at the start was
3259 studied, the wrong optimizing data was generated, leading to matching
3260 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3261 matching string must start with a or b or c. The correct conclusion for
3262 this pattern is that a match can start with any character.
3263
3264
3265 Version 4.4 13-Aug-03
3266 ---------------------
3267
3268 1. In UTF-8 mode, a character class containing characters with values between
3269 127 and 255 was not handled correctly if the compiled pattern was studied.
3270 In fixing this, I have also improved the studying algorithm for such
3271 classes (slightly).
3272
3273 2. Three internal functions had redundant arguments passed to them. Removal
3274 might give a very teeny performance improvement.
3275
3276 3. Documentation bug: the value of the capture_top field in a callout is *one
3277 more than* the number of the hightest numbered captured substring.
3278
3279 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3280 in incorrectly linking with a previously installed version. They now link
3281 explicitly with libpcre.la.
3282
3283 5. configure.in no longer needs to recognize Cygwin specially.
3284
3285 6. A problem in pcre.in for Windows platforms is fixed.
3286
3287 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3288 pcretest, it used to include the size of the study block as part of its
3289 output. Unfortunately, the structure contains a field that has a different
3290 size on different hardware architectures. This meant that the tests that
3291 showed this size failed. As the block is currently always of a fixed size,
3292 this information isn't actually particularly useful in pcretest output, so
3293 I have just removed it.
3294
3295 8. Three pre-processor statements accidentally did not start in column 1.
3296 Sadly, there are *still* compilers around that complain, even though
3297 standard C has not required this for well over a decade. Sigh.
3298
3299 9. In pcretest, the code for checking callouts passed small integers in the
3300 callout_data field, which is a void * field. However, some picky compilers
3301 complained about the casts involved for this on 64-bit systems. Now
3302 pcretest passes the address of the small integer instead, which should get
3303 rid of the warnings.
3304
3305 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3306 both compile and run time, and gives an error if an invalid UTF-8 sequence
3307 is found. There is a option for disabling this check in cases where the
3308 string is known to be correct and/or the maximum performance is wanted.
3309
3310 11. In response to a bug report, I changed one line in Makefile.in from
3311
3312 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3313 to
3314 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3315
3316 to look similar to other lines, but I have no way of telling whether this
3317 is the right thing to do, as I do not use Windows. No doubt I'll get told
3318 if it's wrong...
3319
3320
3321 Version 4.3 21-May-03
3322 ---------------------
3323
3324 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3325 Makefile.
3326
3327 2. Some refactoring to improve the quality of the code:
3328
3329 (i) The utf8_table... variables are now declared "const".
3330
3331 (ii) The code for \cx, which used the "case flipping" table to upper case
3332 lower case letters, now just substracts 32. This is ASCII-specific,
3333 but the whole concept of \cx is ASCII-specific, so it seems
3334 reasonable.
3335
3336 (iii) PCRE was using its character types table to recognize decimal and
3337 hexadecimal digits in the pattern. This is silly, because it handles
3338 only 0-9, a-f, and A-F, but the character types table is locale-
3339 specific, which means strange things might happen. A private
3340 table is now used for this - though it costs 256 bytes, a table is
3341 much faster than multiple explicit tests. Of course, the standard
3342 character types table is still used for matching digits in subject
3343 strings against \d.
3344
3345 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3346 ending in _t are). So I've renamed it as ESC_tee.
3347
3348 3. The first argument for regexec() in the POSIX wrapper should have been
3349 defined as "const".
3350
3351 4. Changed pcretest to use malloc() for its buffers so that they can be
3352 Electric Fenced for debugging.
3353
3354 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3355 to read one or more bytes before the start of the subject string. Often this
3356 had no effect on PCRE's behaviour, but in some circumstances it could
3357 provoke a segmentation fault.
3358
3359 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3360 to try to read one or more bytes before the start of the subject string.
3361
3362 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3363 UTF-8 support could misbehave in various ways if the subject string
3364 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3365 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3366 back over UTF-8 characters.)
3367
3368
3369 Version 4.2 14-Apr-03
3370 ---------------------
3371
3372 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3373
3374 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3375 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3376 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3377 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3378 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3379 and BUILD_EXEEXT
3380 Note: automatic setting of the BUILD variables is not yet working
3381 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3382 compile-time but not at link-time
3383 [LINK]: use for linking executables only
3384 make different versions for Windows and non-Windows
3385 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3386 libraries
3387 [LINK_FOR_BUILD]: new variable
3388 [OBJEXT]: use throughout
3389 [EXEEXT]: use throughout
3390 <winshared>: new target
3391 <wininstall>: new target
3392 <dftables.o>: use native compiler
3393 <dftables>: use native linker
3394 <install>: handle Windows platform correctly
3395 <clean>: ditto
3396 <check>: ditto
3397 copy DLL to top builddir before testing
3398
3399 As part of these changes, -no-undefined was removed again. This was reported
3400 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3401 in any case.
3402
3403 3. Some tidies to get rid of compiler warnings:
3404
3405 . In the match_data structure, match_limit was an unsigned long int, whereas
3406 match_call_count was an int. I've made them both unsigned long ints.
3407
3408 . In pcretest the fact that a const uschar * doesn't automatically cast to
3409 a void * provoked a warning.
3410
3411 . Turning on some more compiler warnings threw up some "shadow" variables
3412 and a few more missing casts.
3413
3414 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3415 option, a class that contained a single character with a value between 128
3416 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3417
3418 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3419 option, a class that contained several characters, but with at least one
3420 whose value was between 128 and 255 caused PCRE to crash.
3421
3422
3423 Version 4.1 12-Mar-03
3424 ---------------------
3425
3426 1. Compiling with gcc -pedantic found a couple of places where casts were
3427 needed, and a string in dftables.c that was longer than standard compilers are
3428 required to support.
3429
3430 2. Compiling with Sun's compiler found a few more places where the code could
3431 be tidied up in order to avoid warnings.
3432
3433 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3434 first of these names is deprecated in the latest Autoconf in favour of the name
3435 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3436 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3437 analogy I have changed it to CFLAGS_FOR_BUILD.
3438
3439 4. Added -no-undefined to the linking command in the Makefile, because this is
3440 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3441 linking step for the pcreposix library.
3442
3443 5. PCRE was failing to diagnose the case of two named groups with the same
3444 name.
3445
3446 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3447 literal character that is needed in the subject for a match, and scans along to
3448 ensure that it is present before embarking on the full matching process. This
3449 saves time in cases of nested unlimited repeats that are never going to match.
3450 Problem: the scan can take a lot of time if the subject is very long (e.g.
3451 megabytes), thus penalizing straightforward matches. It is now done only if the
3452 amount of subject to be scanned is less than 1000 bytes.
3453
3454 7. A lesser problem with the same optimization is that it was recording the
3455 first character of an anchored pattern as "needed", thus provoking a search
3456 right along the subject, even when the first match of the pattern was going to
3457 fail. The "needed" character is now not set for anchored patterns, unless it
3458 follows something in the pattern that is of non-fixed length. Thus, it still
3459 fulfils its original purpose of finding quick non-matches in cases of nested
3460 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3461
3462
3463 Version 4.0 17-Feb-03
3464 ---------------------
3465
3466 1. If a comment in an extended regex that started immediately after a meta-item
3467 extended to the end of string, PCRE compiled incorrect data. This could lead to
3468 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3469
3470 2. Moved to autoconf 2.53 and libtool 1.4.2.
3471
3472 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3473 the special perltest8 script is no longer needed - all the tests can be run
3474 from a single perltest script.
3475
3476 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3477 by \s. It has now been removed in PCRE. This means it isn't recognized as
3478 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3479 class [:space:] *does* include VT, thereby creating a mess.
3480
3481 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3482 space and tab.
3483
3484 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3485 its new features into the main test script, reducing the number of scripts.
3486
3487 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3488 were backward compatible, and made the (?i) apply to the whole pattern, as if
3489 /i were given. Now it behaves more logically, and applies the option setting
3490 only to what follows. PCRE has been changed to follow suit. However, if it
3491 finds options settings right at the start of the pattern, it extracts them into
3492 the global options, as before. Thus, they show up in the info data.
3493
3494 8. Added support for the \Q...\E escape sequence. Characters in between are
3495 treated as literals. This is slightly different from Perl in that $ and @ are
3496 also handled as literals inside the quotes. In Perl, they will cause variable
3497 interpolation. Note the following examples:
3498
3499 Pattern PCRE matches Perl matches
3500
3501 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3502 \Qabc\$xyz\E abc\$xyz abc\$xyz
3503 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3504
3505 For compatibility with Perl, \Q...\E sequences are recognized inside character
3506 classes as well as outside them.
3507
3508 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3509 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3510 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3511 signed/unsigned warnings.
3512
3513 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3514 option for pcretest, so I've replaced it by a simple function that does just
3515 that job.
3516
3517 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3518 "pcregrep -".
3519
3520 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3521 Java package. This provides some syntactic sugar for simple cases of what my
3522 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3523 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3524 item, you can use this simplified notation. Note that only makes sense with
3525 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3526 greediness, whatever the setting of the PCRE_UNGREEDY option.
3527
3528 13. A change of greediness default within a pattern was not taking effect at
3529 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3530 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3531 was abstracted outside.
3532
3533 14. PCRE now supports the \G assertion. It is true when the current matching
3534 position is at the start point of the match. This differs from \A when the
3535 starting offset is non-zero. Used with the /g option of pcretest (or similar
3536 code), it works in the same way as it does for Perl's /g option. If all
3537 alternatives of a regex begin with \G, the expression is anchored to the start
3538 match position, and the "anchored" flag is set in the compiled expression.
3539
3540 15. Some bugs concerning the handling of certain option changes within patterns
3541 have been fixed. These applied to options other than (?ims). For example,
3542 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3543 been the other way round. Some of this was related to change 7 above.
3544
3545 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3546 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3547 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3548 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3549
3550 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3551 mode. Unlike ".", it always matches newline, whatever the setting of
3552 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3553 assertions. Perl allows it, but it doesn't (in general) work because it can't
3554 calculate the length of the lookbehind. At least, that's the case for Perl
3555 5.8.0 - I've been told they are going to document that it doesn't work in
3556 future.
3557
3558 18. Added an error diagnosis for escapes that PCRE does not support: these are
3559 \L, \l, \N, \P, \p, \U, \u, and \X.
3560
3561 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3562 reading past the end of the pattern in cases such as /[abcd/.
3563
3564 20. PCRE was getting more memory than necessary for patterns with classes that
3565 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3566
3567 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3568 compiling PCRE for use with Virtual Pascal.
3569
3570 22. Small fix to the Makefile to make it work properly if the build is done
3571 outside the source tree.
3572
3573 23. Added a new extension: a condition to go with recursion. If a conditional
3574 subpattern starts with (?(R) the "true" branch is used if recursion has
3575 happened, whereas the "false" branch is used only at the top level.
3576
3577 24. When there was a very long string of literal characters (over 255 bytes
3578 without UTF support, over 250 bytes with UTF support), the computation of how
3579 much memory was required could be incorrect, leading to segfaults or other
3580 strange effects.
3581
3582 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3583 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3584 there was a subsequent back reference to those brackets. This meant that, for
3585 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3586 possible to check for precisely this case. All we can do is abandon the
3587 optimization if .* occurs inside capturing brackets when there are any back
3588 references whatsoever. (See below for a better fix that came later.)
3589
3590 26. The handling of the optimization for finding the first character of a
3591 non-anchored pattern, and for finding a character that is required later in the
3592 match were failing in some cases. This didn't break the matching; it just
3593 failed to optimize when it could. The way this is done has been re-implemented.
3594
3595 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3596
3597 28. Added a new feature that provides some of the functionality that Perl
3598 provides with (?{...}). The facility is termed a "callout". The way it is done
3599 in PCRE is for the caller to provide an optional function, by setting
3600 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3601 global variable. By default it is unset, which disables all calling out. To get
3602 the function called, the regex must include (?C) at appropriate points. This
3603 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3604 This provides a means of identifying different callout points. When PCRE
3605 reaches such a point in the regex, if pcre_callout has been set, the external
3606 function is called. It is provided with data in a structure called
3607 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3608 matching continues; if it returns a non-zero value, the match at the current
3609 point fails. However, backtracking will occur if possible. [This was changed
3610 later and other features added - see item 49 below.]
3611
3612 29. pcretest is upgraded to test the callout functionality. It provides a
3613 callout function that displays information. By default, it shows the start of
3614 the match and the current position in the text. There are some new data escapes
3615 to vary what happens:
3616
3617 \C+ in addition, show current contents of captured substrings
3618 \C- do not supply a callout function
3619 \C!n return 1 when callout number n is reached
3620 \C!n!m return 1 when callout number n is reached for the mth time
3621
3622 30. If pcregrep was called with the -l option and just a single file name, it
3623 output "<stdin>" if a match was found, instead of the file name.
3624
3625 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3626 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3627 pcre_exec(). This saves a malloc/free per call. The default value of
3628 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3629 when configuring.
3630
3631 32. The default maximum size of a compiled pattern is 64K. There have been a
3632 few cases of people hitting this limit. The code now uses macros to handle the
3633 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3634 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3635 configuring. Tests 2 and 5 work only with 2-byte links because they output
3636 debugging information about compiled patterns.
3637
3638 33. Internal code re-arrangements:
3639
3640 (a) Moved the debugging function for printing out a compiled regex into
3641 its own source file (printint.c) and used #include to pull it into
3642 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3643 separate copies.
3644
3645 (b) Defined the list of op-code names for debugging as a macro in
3646 internal.h so that it is next to the definition of the opcodes.
3647
3648 (c) Defined a table of op-code lengths for simpler skipping along compiled
3649 code. This is again a macro in internal.h so that it is next to the
3650 definition of the opcodes.
3651
3652 34. Added support for recursive calls to individual subpatterns, along the
3653 lines of Robin Houston's patch (but implemented somewhat differently).
3654
3655 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3656 allow it to read and process whole directories in Win32. This code was
3657 contributed by Lionel Fourquaux; it has not been tested by me.
3658
3659 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3660 used to name a group. Names consist of alphanumerics and underscores, and must
3661 be unique. Back references use the syntax (?P=name) and recursive calls use
3662 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3663 numbers. The function pcre_fullinfo() can be used after compilation to extract
3664 a name/number map. There are three relevant calls:
3665
3666 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3667 PCRE_INFO_NAMECOUNT yields the number of entries
3668 PCRE_INFO_NAMETABLE yields a pointer to the map.
3669
3670 The map is a vector of fixed-size entries. The size of each entry depends on
3671 the length of the longest name used. The first two bytes of each entry are the
3672 group number, most significant byte first. There follows the corresponding
3673 name, zero terminated. The names are in alphabetical order.
3674
3675 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3676 case instead of 255. Making it the same both with and without UTF-8 support
3677 means that the same test output works with both.
3678
3679 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3680 calling malloc() with a zero argument.
3681
3682 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3683 optimization. I've improved things by keeping a bitmap of backreferences with
3684 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3685 fact referenced, the optimization can be applied. It is unlikely that a
3686 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3687 the match to follow \n) will appear inside brackets with a number greater than
3688 31, but if it does, any back reference > 31 suppresses the optimization.
3689
3690 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3691 of disabling numbered capturing parentheses. Any opening parenthesis that is
3692 not followed by ? behaves as if it were followed by ?: but named parentheses
3693 can still be used for capturing (and they will acquire numbers in the usual
3694 way).
3695
3696 41. Redesigned the return codes from the match() function into yes/no/error so
3697 that errors can be passed back from deep inside the nested calls. A malloc
3698 failure while inside a recursive subpattern call now causes the
3699 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3700
3701 42. It is now possible to set a limit on the number of times the match()
3702 function is called in a call to pcre_exec(). This facility makes it possible to
3703 limit the amount of recursion and backtracking, though not in a directly
3704 obvious way, because the match() function is used in a number of different
3705 circumstances. The count starts from zero for each position in the subject
3706 string (for non-anchored patterns). The default limit is, for compatibility, a
3707 large number, namely 10 000 000. You can change this in two ways:
3708
3709 (a) When configuring PCRE before making, you can use --with-match-limit=n
3710 to set a default value for the compiled library.
3711
3712 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3713 a different value is set. See 45 below.
3714
3715 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3716
3717 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3718 of things that can be changed at compile time. The first argument specifies
3719 what is wanted and the second points to where the information is to be placed.
3720 The current list of available information is:
3721
3722 PCRE_CONFIG_UTF8
3723
3724 The output is an integer that is set to one if UTF-8 support is available;
3725 otherwise it is set to zero.
3726
3727 PCRE_CONFIG_NEWLINE
3728
3729 The output is an integer that it set to the value of the code that is used for
3730 newline. It is either LF (10) or CR (13).
3731
3732 PCRE_CONFIG_LINK_SIZE
3733
3734 The output is an integer that contains the number of bytes used for internal
3735 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3736
3737 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3738
3739 The output is an integer that contains the threshold above which the POSIX
3740 interface uses malloc() for output vectors. See item 31 above.
3741
3742 PCRE_CONFIG_MATCH_LIMIT
3743
3744 The output is an unsigned integer that contains the default limit of the number
3745 of match() calls in a pcre_exec() execution. See 42 above.
3746
3747 44. pcretest has been upgraded by the addition of the -C option. This causes it
3748 to extract all the available output from the new pcre_config() function, and to
3749 output it. The program then exits immediately.
3750
3751 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3752 order to support additional features. One way would have been to define
3753 pcre_exec2() (for example) with extra arguments, but this would not have been
3754 extensible, and would also have required all calls to the original function to
3755 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3756 is used for passing in "extra" data from pcre_study().
3757
3758 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3759 contains the following fields:
3760
3761 flags a bitmap indicating which of the following fields are set
3762 study_data opaque data from pcre_study()
3763 match_limit a way of specifying a limit on match() calls for a specific
3764 call to pcre_exec()
3765 callout_data data for callouts (see 49 below)
3766
3767 The flag bits are also defined in pcre.h, and are
3768
3769 PCRE_EXTRA_STUDY_DATA
3770 PCRE_EXTRA_MATCH_LIMIT
3771 PCRE_EXTRA_CALLOUT_DATA
3772
3773 The pcre_study() function now returns one of these new pcre_extra blocks, with
3774 the actual study data pointed to by the study_data field, and the
3775 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3776 before. That is, this change is entirely upwards-compatible and requires no
3777 change to existing code.
3778
3779 If you want to pass in additional data to pcre_exec(), you can either place it
3780 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3781 block.
3782
3783 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3784 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3785 times with different match limits, until it finds the minimum value needed for
3786 pcre_exec() to complete. The value is then output. This can be instructive; for
3787 most simple matches the number is quite small, but for pathological cases it
3788 gets very large very quickly.
3789
3790 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3791 returns the size of the data block pointed to by the study_data field in a
3792 pcre_extra block, that is, the value that was passed as the argument to
3793 pcre_malloc() when PCRE was getting memory in which to place the information
3794 created by pcre_study(). The fourth argument should point to a size_t variable.
3795 pcretest has been extended so that this information is shown after a successful
3796 pcre_study() call when information about the compiled regex is being displayed.
3797
3798 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3799 because what follows is always an absolute path. (Later: it turns out that this
3800 is more than cosmetic for MinGW, because it doesn't like empty path
3801 components.)
3802
3803 49. Some changes have been made to the callout feature (see 28 above):
3804
3805 (i) A callout function now has three choices for what it returns:
3806
3807 0 => success, carry on matching
3808 > 0 => failure at this point, but backtrack if possible
3809 < 0 => serious error, return this value from pcre_exec()
3810
3811 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3812 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3813 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3814 use by callout functions. It will never be used by PCRE itself.
3815
3816 (ii) The pcre_extra structure (see 45 above) has a void * field called
3817 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3818 pcre_callout_block structure has a field of the same name. The contents of
3819 the field passed in the pcre_extra structure are passed to the callout
3820 function in the corresponding field in the callout block. This makes it
3821 easier to use the same callout-containing regex from multiple threads. For
3822 testing, the pcretest program has a new data escape
3823
3824 \C*n pass the number n (may be negative) as callout_data
3825
3826 If the callout function in pcretest receives a non-zero value as
3827 callout_data, it returns that value.
3828
3829 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3830 there were some redundant $(CFLAGS) in commands that are now specified as
3831 $(LINK), which already includes $(CFLAGS).
3832
3833 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3834 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3835 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3836 one-byte characters throughout. Note that case-insensitive matching applies
3837 only to characters whose values are less than 256. PCRE doesn't support the
3838 notion of cases for higher-valued characters.
3839
3840 (i) A character class whose characters are all within 0-255 is handled as
3841 a bit map, and the map is inverted for negative classes. Previously, a
3842 character > 255 always failed to match such a class; however it should
3843 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3844
3845 (ii) A negated character class with a single character < 255 is coded as
3846 "not this character" (OP_NOT). This wasn't working properly when the test
3847 character was multibyte, either singly or repeated.
3848
3849 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3850 mode, for example: \x{100}{2,3}.
3851
3852 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3853 singly or repeated) now correctly test multibyte characters. However,
3854 PCRE doesn't recognize any characters with values greater than 255 as
3855 digits, spaces, or word characters. Such characters always match \D, \S,
3856 and \W, and never match \d, \s, or \w.
3857
3858 (v) Classes may now contain characters and character ranges with values
3859 greater than 255. For example: [ab\x{100}-\x{400}].
3860
3861 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3862 PCRE in UTF-8 mode.
3863
3864 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3865 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3866 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3867 value.)
3868
3869 53. The single man page has become too large. I have therefore split it up into
3870 a number of separate man pages. These also give rise to individual HTML pages;
3871 these are now put in a separate directory, and there is an index.html page that
3872 lists them all. Some hyperlinking between the pages has been installed.
3873
3874 54. Added convenience functions for handling named capturing parentheses.
3875
3876 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3877 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3878 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3879 are faulted.
3880
3881 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3882 calling configure. These values are used when compiling the dftables.c program
3883 which is run to generate the source of the default character tables. They
3884 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3885 you will need to set these values.
3886
3887 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3888
3889
3890 Version 3.9 02-Jan-02
3891 ---------------------
3892
3893 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3894
3895 2. If --disable-static was given, the building process failed when trying to
3896 build pcretest and pcregrep. (For some reason it was using libtool to compile
3897 them, which is not right, as they aren't part of the library.)
3898
3899
3900 Version 3.8 18-Dec-01
3901 ---------------------
3902
3903 1. The experimental UTF-8 code was completely screwed up. It was packing the
3904 bytes in the wrong order. How dumb can you get?
3905
3906
3907 Version 3.7 29-Oct-01
3908 ---------------------
3909
3910 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3911 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3912 this didn't happen under Solaris 8, where I normally test things.
3913
3914 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3915 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3916 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3917 this fix an hour or so after the initial 3.7 release.)
3918
3919
3920 Version 3.6 23-Oct-01
3921 ---------------------
3922
3923 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3924 offsets passed as NULL with zero offset count.
3925
3926 2. The config.guess and config.sub files had not been updated when I moved to
3927 the latest autoconf.
3928
3929
3930 Version 3.5 15-Aug-01
3931 ---------------------
3932
3933 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3934 had been forgotten.
3935
3936 2. By using declared but undefined structures, we can avoid using "void"
3937 definitions in pcre.h while keeping the internal definitions of the structures
3938 private.
3939
3940 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3941 user point of view, this means that both static and shared libraries are built
3942 by default, but this can be individually controlled. More of the work of
3943 handling this static/shared cases is now inside libtool instead of PCRE's make
3944 file.
3945
3946 4. The pcretest utility is now installed along with pcregrep because it is
3947 useful for users (to test regexs) and by doing this, it automatically gets
3948 relinked by libtool. The documentation has been turned into a man page, so
3949 there are now .1, .txt, and .html versions in /doc.
3950
3951 5. Upgrades to pcregrep:
3952 (i) Added long-form option names like gnu grep.
3953 (ii) Added --help to list all options with an explanatory phrase.
3954 (iii) Added -r, --recursive to recurse into sub-directories.
3955 (iv) Added -f, --file to read patterns from a file.
3956
3957 6. pcre_exec() was referring to its "code" argument before testing that
3958 argument for NULL (and giving an error if it was NULL).
3959
3960 7. Upgraded Makefile.in to allow for compiling in a different directory from
3961 the source directory.
3962
3963 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3964 options bits, the pointer it was passed was to an int instead of to an unsigned
3965 long int. This mattered only on 64-bit systems.
3966
3967 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3968 generated) instead of pcre.in, which it its source. Also made the same change
3969 in several of the .c files.
3970
3971 10. A new release of gcc defines printf() as a macro, which broke pcretest
3972 because it had an ifdef in the middle of a string argument for printf(). Fixed
3973 by using separate calls to printf().
3974
3975 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3976 script, to force use of CR or LF instead of \n in the source. On non-Unix
3977 systems, the value can be set in config.h.
3978
3979 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3980 absolute limit. Changed the text of the error message to make this clear, and
3981 likewise updated the man page.
3982
3983 13. The limit of 99 on the number of capturing subpatterns has been removed.
3984 The new limit is 65535, which I hope will not be a "real" limit.
3985
3986
3987 Version 3.4 22-Aug-00
3988 ---------------------
3989
3990 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3991
3992 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3993
3994
3995 Version 3.3 01-Aug-00
3996 ---------------------
3997
3998 1. If an octal character was given, but the value was greater than \377, it
3999 was not getting masked to the least significant bits, as documented. This could
4000 lead to crashes in some systems.
4001
4002 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4003 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4004
4005 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4006 These just pass their arguments on to (pcre_free)(), but they are provided
4007 because some uses of PCRE bind it to non-C systems that can call its functions,
4008 but cannot call free() or pcre_free() directly.
4009
4010 4. Add "make test" as a synonym for "make check". Corrected some comments in
4011 the Makefile.
4012
4013 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4014 Makefile.
4015
4016 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4017 command called pgrep for grepping around the active processes.
4018
4019 7. Added the beginnings of support for UTF-8 character strings.
4020
4021 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4022 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4023 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4024 out for the ar command.)
4025
4026
4027 Version 3.2 12-May-00
4028 ---------------------
4029
4030 This is purely a bug fixing release.
4031
4032 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4033 of ZA. This was just one example of several cases that could provoke this bug,
4034 which was introduced by change 9 of version 2.00. The code for breaking
4035 infinite loops after an iteration that matches an empty string was't working
4036 correctly.
4037
4038 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4039 when matched against abbab (for example). After matching an empty string, it
4040 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4041 caused it to match further down the string than it should.
4042
4043 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4044 was there because it doesn't seem to be needed, and it causes trouble on some
4045 systems, as it is not a Standard C header. It has been removed.
4046
4047 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4048 were reported on the Macintosh. The changes were from
4049
4050 while ((c = *(++ptr)) != 0 && c != '\n');
4051 to
4052 while ((c = *(++ptr)) != 0 && c != '\n') ;
4053
4054 Totally extraordinary, but if that's what it takes...
4055
4056 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4057 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4058 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4059 assumes the way PCRE uses memmove() (always moving upwards).
4060
4061 6. PCRE is being used in one environment where strchr() is not available. There
4062 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4063 faster code anyway.
4064
4065
4066 Version 3.1 09-Feb-00
4067 ---------------------
4068
4069 The only change in this release is the fixing of some bugs in Makefile.in for
4070 the "install" target:
4071
4072 (1) It was failing to install pcreposix.h.
4073
4074 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4075
4076
4077 Version 3.0 01-Feb-00
4078 ---------------------
4079
4080 1. Add support for the /+ modifier to perltest (to output $` like it does in
4081 pcretest).
4082
4083 2. Add support for the /g modifier to perltest.
4084
4085 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4086 matches null strings.
4087
4088 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4089 pattern. Perl treats empty patterns specially - it reuses the most recent
4090 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4091 effect.
4092
4093 5. The POSIX interface was broken in that it was just handing over the POSIX
4094 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4095 required a bigger vector, with some working space on the end. This means that
4096 the POSIX wrapper now has to get and free some memory, and copy the results.
4097
4098 6. Added some simple autoconf support, placing the test data and the
4099 documentation in separate directories, re-organizing some of the
4100 information files, and making it build pcre-config (a GNU standard). Also added
4101 libtool support for building PCRE as a shared library, which is now the
4102 default.
4103
4104 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4105 09 are not valid octal constants. Single digits will be used for minor values
4106 less than 10.
4107
4108 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4109 existing programs that set these in the POSIX interface can use PCRE without
4110 modification.
4111
4112 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4113 return all that pcre_info() returns, plus additional data. The pcre_info()
4114 function is retained for compatibility, but is considered to be obsolete.
4115
4116 10. Added experimental recursion feature (?R) to handle one common case that
4117 Perl 5.6 will be able to do with (?p{...}).
4118
4119 11. Added support for POSIX character classes like [:alpha:], which Perl is
4120 adopting.
4121
4122
4123 Version 2.08 31-Aug-99
4124 ----------------------
4125
4126 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4127 trying to match at the startoffset position, but instead was moving forward to
4128 the next newline as if a previous match had failed.
4129
4130 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4131 and could get into a loop if a null string was matched other than at the start
4132 of the subject.
4133
4134 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4135 be distinguished at compile time, and for completeness also added PCRE_DATE.
4136
4137 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4138 in GnuWin32 environments.
4139
4140
4141 Version 2.07 29-Jul-99
4142 ----------------------
4143
4144 1. The documentation is now supplied in plain text form and HTML as well as in
4145 the form of man page sources.
4146
4147 2. C++ compilers don't like assigning (void *) values to other pointer types.
4148 In particular this affects malloc(). Although there is no problem in Standard
4149 C, I've put in casts to keep C++ compilers happy.
4150
4151 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4152 should be (const char *).
4153
4154 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4155 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4156 However, I haven't made this a standard facility. The documentation doesn't
4157 mention it, and the Makefile doesn't support it.
4158
4159 5. The Makefile now contains an "install" target, with editable destinations at
4160 the top of the file. The pcretest program is not installed.
4161
4162 6. pgrep -V now gives the PCRE version number and date.
4163
4164 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4165 causing the entire string to be ignored, instead of just the last character.
4166
4167 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4168 non-matching string, it can take a very, very long time, even for strings of
4169 quite modest length, because of the nested recursion. PCRE now does better in
4170 some of these cases. It does this by remembering the last required literal
4171 character in the pattern, and pre-searching the subject to ensure it is present
4172 before running the real match. In other words, it applies a heuristic to detect
4173 some types of certain failure quickly, and in the above example, if presented
4174 with a string that has no trailing " it gives "no match" very quickly.
4175
4176 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4177 other alternatives are tried instead.
4178
4179
4180 Version 2.06 09-Jun-99
4181 ----------------------
4182
4183 1. Change pcretest's output for amount of store used to show just the code
4184 space, because the remainder (the data block) varies in size between 32-bit and
4185 64-bit systems.
4186
4187 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4188 start matching at. This allows lookbehinds to work when searching for multiple
4189 occurrences in a string.
4190
4191 3. Added additional options to pcretest for testing multiple occurrences:
4192
4193 /+ outputs the rest of the string that follows a match
4194 /g loops for multiple occurrences, using the new startoffset argument
4195 /G loops for multiple occurrences by passing an incremented pointer
4196
4197 4. PCRE wasn't doing the "first character" optimization for patterns starting
4198 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4199 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4200 the letter 'x'. On long subject strings, this gives a significant speed-up.
4201
4202
4203 Version 2.05 21-Apr-99
4204 ----------------------
4205
4206 1. Changed the type of magic_number from int to long int so that it works
4207 properly on 16-bit systems.
4208
4209 2. Fixed a bug which caused patterns starting with .* not to work correctly
4210 when the subject string contained newline characters. PCRE was assuming
4211 anchoring for such patterns in all cases, which is not correct because .* will
4212 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4213 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4214 must be retried after every newline in the subject.
4215
4216
4217 Version 2.04 18-Feb-99
4218 ----------------------
4219
4220 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4221 computation of the store needed to hold the pattern was incorrect (too large).
4222 If such patterns were nested a few deep, this could multiply and become a real
4223 problem.
4224
4225 2. Added /M option to pcretest to show the memory requirement of a specific
4226 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4227
4228 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4229 compiled in such a way that the backtracking after subsequent failure was
4230 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4231 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4232
4233
4234 Version 2.03 02-Feb-99
4235 ----------------------
4236
4237 1. Fixed typo and small mistake in man page.
4238
4239 2. Added 4th condition (GPL supersedes if conflict) and created separate
4240 LICENCE file containing the conditions.
4241
4242 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4243 Perl, that is the internal \ allows the delimiter to be included in the
4244 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4245 the final delimiter, add \ to the end of the pattern (to test the error).
4246
4247 4. Added the convenience functions for extracting substrings after a successful
4248 match. Updated pcretest to make it able to test these functions.
4249
4250
4251 Version 2.02 14-Jan-99
4252 ----------------------
4253
4254 1. Initialized the working variables associated with each extraction so that
4255 their saving and restoring doesn't refer to uninitialized store.
4256
4257 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4258 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4259 fix the problem.
4260
4261 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4262 calls, and wasn't printing the correct value for compiling calls. Increased the
4263 default value of LOOPREPEAT, and the number of significant figures in the
4264 times.
4265
4266 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4267
4268 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4269 a building problem on Windows NT with a FAT file system.
4270
4271
4272 Version 2.01 21-Oct-98
4273 ----------------------
4274
4275 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4276 to character tables built by pcre_maketables() in the current locale. If NULL
4277 is passed, the default tables are used.
4278
4279
4280 Version 2.00 24-Sep-98
4281 ----------------------
4282
4283 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4284 it any more.
4285
4286 2. Allow quantification of (?>) groups, and make it work correctly.
4287
4288 3. The first character computation wasn't working for (?>) groups.
4289
4290 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4291 end of the subject) and add 5.005's \z, which really does match only at the
4292 very end of the subject.
4293
4294 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4295
4296 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4297 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4298 localized options. All options to pcre_study() were also removed.
4299
4300 7. Add other new features from 5.005:
4301
4302 $(?<= positive lookbehind
4303 $(?<! negative lookbehind
4304 (?imsx-imsx) added the unsetting capability
4305 such a setting is global if at outer level; local otherwise
4306 (?imsx-imsx:) non-capturing groups with option setting
4307 (?(cond)re|re) conditional pattern matching
4308
4309 A backreference to itself in a repeated group matches the previous
4310 captured string.
4311
4312 8. General tidying up of studying (both automatic and via "study")
4313 consequential on the addition of new assertions.
4314
4315 9. As in 5.005, unlimited repeated groups that could match an empty substring
4316 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4317 runtime if any iteration does actually match an empty substring.
4318
4319 10. Include the RunTest script in the distribution.
4320
4321 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4322 discrepancies, some of which were old and were also with respect to 5.004. They
4323 have now been fixed.
4324
4325
4326 Version 1.09 28-Apr-98
4327 ----------------------
4328
4329 1. A negated single character class followed by a quantifier with a minimum
4330 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4331 program crashes, or just wrong answers. This did not apply to negated classes
4332 containing more than one character, or to minima other than one.
4333
4334
4335 Version 1.08 27-Mar-98
4336 ----------------------
4337
4338 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4339
4340 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4341 latter must appear before anything that relies on it in the pattern.
4342
4343
4344 Version 1.07 16-Feb-98
4345 ----------------------
4346
4347 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4348 repeat of a potentially empty string).
4349
4350
4351 Version 1.06 23-Jan-98
4352 ----------------------
4353
4354 1. Added Markus Oberhumer's little patches for C++.
4355
4356 2. Literal strings longer than 255 characters were broken.
4357
4358
4359 Version 1.05 23-Dec-97
4360 ----------------------
4361
4362 1. Negated character classes containing more than one character were failing if
4363 PCRE_CASELESS was set at run time.
4364
4365
4366 Version 1.04 19-Dec-97
4367 ----------------------
4368
4369 1. Corrected the man page, where some "const" qualifiers had been omitted.
4370
4371 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4372 input syntax.
4373
4374 3. Fixed memory leak which occurred when a regex with back references was
4375 matched with an offsets vector that wasn't big enough. The temporary memory
4376 that is used in this case wasn't being freed if the match failed.
4377
4378 4. Tidied pcretest to ensure it frees memory that it gets.
4379
4380 5. Temporary memory was being obtained in the case where the passed offsets
4381 vector was exactly big enough.
4382
4383 6. Corrected definition of offsetof() from change 5 below.
4384
4385 7. I had screwed up change 6 below and broken the rules for the use of
4386 setjmp(). Now fixed.
4387
4388
4389 Version 1.03 18-Dec-97
4390 ----------------------
4391
4392 1. A erroneous regex with a missing opening parenthesis was correctly
4393 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4394 on some systems.
4395
4396 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4397 it was reported that one broken compiler failed on the former because "code" is
4398 also an independent variable.
4399
4400 3. The erroneous regex a[]b caused an array overrun reference.
4401
4402 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4403 fail on data ending with that character. (It was going on too far, and checking
4404 the next character, typically a binary zero.) This was specific to the
4405 optimized code for single-character negative classes.
4406
4407 5. Added a contributed patch from the TIN world which does the following:
4408
4409 + Add an undef for memmove, in case the the system defines a macro for it.
4410
4411 + Add a definition of offsetof(), in case there isn't one. (I don't know
4412 the reason behind this - offsetof() is part of the ANSI standard - but
4413 it does no harm).
4414
4415 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4416 most of the places where whitespace preceded '#'. I have given up and
4417 allowed the remaining 2 cases to be at the margin.
4418
4419 + Rename some variables in pcre to eliminate shadowing. This seems very
4420 pedantic, but does no harm, of course.
4421
4422 6. Moved the call to setjmp() into its own function, to get rid of warnings
4423 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4424
4425 7. Constructs such as \d{8,} were compiling into the equivalent of
4426 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4427 outcome, but in this particular case used more store than had been allocated,
4428 which caused the bug to be discovered because it threw up an internal error.
4429
4430 8. The debugging code in both pcre and pcretest for outputting the compiled
4431 form of a regex was going wrong in the case of back references followed by
4432 curly-bracketed repeats.
4433
4434
4435 Version 1.02 12-Dec-97
4436 ----------------------
4437
4438 1. Typos in pcre.3 and comments in the source fixed.
4439
4440 2. Applied a contributed patch to get rid of places where it used to remove
4441 'const' from variables, and fixed some signed/unsigned and uninitialized
4442 variable warnings.
4443
4444 3. Added the "runtest" target to Makefile.
4445
4446 4. Set default compiler flag to -O2 rather than just -O.
4447
4448
4449 Version 1.01 19-Nov-97
4450 ----------------------
4451
4452 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4453 like /([ab]*)*/, that is, for classes with more than one character in them.
4454
4455 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4456 as /((?>a*))*/ (a PCRE_EXTRA facility).
4457
4458
4459 Version 1.00 18-Nov-97
4460 ----------------------
4461
4462 1. Added compile-time macros to support systems such as SunOS4 which don't have
4463 memmove() or strerror() but have other things that can be used instead.
4464
4465 2. Arranged that "make clean" removes the executables.
4466
4467
4468 Version 0.99 27-Oct-97
4469 ----------------------
4470
4471 1. Fixed bug in code for optimizing classes with only one character. It was
4472 initializing a 32-byte map regardless, which could cause it to run off the end
4473 of the memory it had got.
4474
4475 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4476
4477
4478 Version 0.98 22-Oct-97
4479 ----------------------
4480
4481 1. Fixed bug in code for handling temporary memory usage when there are more
4482 back references than supplied space in the ovector. This could cause segfaults.
4483
4484
4485 Version 0.97 21-Oct-97
4486 ----------------------
4487
4488 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4489
4490 2. Optimized negated single characters not to use a bit map.
4491
4492 3. Brought error texts together as macro definitions; clarified some of them;
4493 fixed one that was wrong - it said "range out of order" when it meant "invalid
4494 escape sequence".
4495
4496 4. Changed some char * arguments to const char *.
4497
4498 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4499
4500 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4501 pcretest.
4502
4503
4504 Version 0.96 16-Oct-97
4505 ----------------------
4506
4507 1. Added a simple "pgrep" utility to the distribution.
4508
4509 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4510 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4511 where "ddd" means "one or more decimal digits".
4512
4513 3. Fixed serious bug. If a pattern had a back reference, but the call to
4514 pcre_exec() didn't supply a large enough ovector to record the related
4515 identifying subpattern, the match always failed. PCRE now remembers the number
4516 of the largest back reference, and gets some temporary memory in which to save
4517 the offsets during matching if necessary, in order to ensure that
4518 backreferences always work.
4519
4520 4. Increased the compatibility with Perl in a number of ways:
4521
4522 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4523 to request this handling. The option can be set at compile or exec time.
4524
4525 (b) $ matches before a terminating newline by default; an option
4526 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4527 mode). The option can be set at compile or exec time.
4528
4529 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4530 the same as Perl's. If the decimal number it represents is less than 10
4531 or there aren't that many previous left capturing parentheses, an octal
4532 escape is read. Inside a character class, it's always an octal escape,
4533 even if it is a single digit.
4534
4535 (d) An escaped but undefined alphabetic character is taken as a literal,
4536 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4537 escapes.
4538
4539 (e) {0} is now permitted. (The previous item is removed from the compiled
4540 pattern).
4541
4542 5. Changed all the names of code files so that the basic parts are no longer
4543 than 10 characters, and abolished the teeny "globals.c" file.
4544
4545 6. Changed the handling of character classes; they are now done with a 32-byte
4546 bit map always.
4547
4548 7. Added the -d and /D options to pcretest to make it possible to look at the
4549 internals of compilation without having to recompile pcre.
4550
4551
4552 Version 0.95 23-Sep-97
4553 ----------------------
4554
4555 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4556 \x20 at the start of a run of normal characters. These were being treated as
4557 real characters, instead of the source characters being re-checked.
4558
4559
4560 Version 0.94 18-Sep-97
4561 ----------------------
4562
4563 1. The functions are now thread-safe, with the caveat that the global variables
4564 containing pointers to malloc() and free() or alternative functions are the
4565 same for all threads.
4566
4567 2. Get pcre_study() to generate a bitmap of initial characters for non-
4568 anchored patterns when this is possible, and use it if passed to pcre_exec().
4569
4570
4571 Version 0.93 15-Sep-97
4572 ----------------------
4573
4574 1. /(b)|(:+)/ was computing an incorrect first character.
4575
4576 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4577 but not actually doing anything yet.
4578
4579 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4580 as Perl does (e.g. [-az] or [az-]).
4581
4582 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4583 all possible positions.
4584
4585 5. Split up into different modules to avoid including unneeded functions in a
4586 compiled binary. However, compile and exec are still in one module. The "study"
4587 function is split off.
4588
4589 6. The character tables are now in a separate module whose source is generated
4590 by an auxiliary program - but can then be edited by hand if required. There are
4591 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4592 toupper() in the code.
4593
4594 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4595 make them global. Abolish the function for setting them, as the caller can now
4596 set them directly.
4597
4598
4599 Version 0.92 11-Sep-97
4600 ----------------------
4601
4602 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4603 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4604
4605 2. Caseless matching was not working in character classes if the characters in
4606 the pattern were in upper case.
4607
4608 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4609
4610 4. Make PCRE_ANCHORED public and accept as a compile option.
4611
4612 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4613 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4614 pass them.
4615
4616 6. Give an error if bad option bits passed at compile or run time.
4617
4618 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4619 pcretest to cause it to pass that flag.
4620
4621 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4622 options, and the first character, if set.
4623
4624 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4625
4626
4627 Version 0.91 10-Sep-97
4628 ----------------------
4629
4630 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4631 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4632
4633 2. PCRE was looping on encountering an indefinitely repeated back reference to
4634 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4635 Perl does - treats the match as successful.
4636
4637 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5