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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 888 - (show annotations) (download)
Tue Jan 17 14:43:23 2012 UTC (2 years, 6 months ago) by ph10
File size: 225446 byte(s)
Fix MARK bug for assertions.

1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.30
5 ------------
6
7 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
8 name is defined in ctype.h.
9
10 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
11 only in quite long subpatterns.
12
13 3. Removed the function pcre_info(), which has been obsolete and deprecated
14 since it was replaced by pcre_fullinfo() in February 2000.
15
16 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
17 match a (*MARK), and the match failed at the start of the subject, a
18 reference to memory before the start of the subject could occur. This bug
19 was introduced by fix 17 of release 8.21.
20
21 5. A reference to an unset group with zero minimum repetition was giving
22 totally wrong answers (in non-JavaScript-compatibility mode). For example,
23 /(another)?(\1?)test/ matched against "hello world test". This bug was
24 introduced in release 8.13.
25
26 6. Add support for 16-bit character strings (a large amount of work involving
27 many changes and refactorings).
28
29 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31 from a file.
32
33 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34 rounding is not applied in this particular case).
35
36 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
37 if they appear, or are escaped, in patterns.
38
39 10. Get rid of a number of -Wunused-but-set-variable warnings.
40
41 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
42 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
43 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
44 also returns the mark "x". This bug applied to capturing parentheses,
45 non-capturing parentheses, and atomic parentheses. It also applied to some
46 assertions.
47
48 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
49 information out of configure.ac instead of relying on pcre.h.generic, which
50 is not stored in the repository.
51
52
53 Version 8.21 12-Dec-2011
54 ------------------------
55
56 1. Updating the JIT compiler.
57
58 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
59 are added as well.
60
61 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
62 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
63 calling _pcre_jit_exec. Some extra comments are added.
64
65 4. (*MARK) settings inside atomic groups that do not contain any capturing
66 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
67 was introduced by change 18 for 8.20.
68
69 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
70 ECMA-262 standard.
71
72 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
73 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
74 This bug was probably introduced by change 9 of 8.13.
75
76 7. While fixing 6 above, I noticed that a number of other items were being
77 incorrectly rejected as "not fixed length". This arose partly because newer
78 opcodes had not been added to the fixed-length checking code. I have (a)
79 corrected the bug and added tests for these items, and (b) arranged for an
80 error to occur if an unknown opcode is encountered while checking for fixed
81 length instead of just assuming "not fixed length". The items that were
82 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
83 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
84 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
85
86 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
87 being incorrectly compiled and would have given unpredicatble results.
88
89 9. A possessively repeated subpattern with minimum repeat count greater than
90 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
91 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
92 the first (A) could occur when it should not.
93
94 10. Add a cast and remove a redundant test from the code.
95
96 11. JIT should use pcre_malloc/pcre_free for allocation.
97
98 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
99 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
100 is anything other than /usr, -L is still shown).
101
102 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
103
104 14. Perl does not support \N without a following name in a [] class; PCRE now
105 also gives an error.
106
107 15. If a forward reference was repeated with an upper limit of around 2000,
108 it caused the error "internal error: overran compiling workspace". The
109 maximum number of forward references (including repeats) was limited by the
110 internal workspace, and dependent on the LINK_SIZE. The code has been
111 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
112 the default depends on LINK_SIZE. There is a new upper limit (for safety)
113 of around 200,000 forward references. While doing this, I also speeded up
114 the filling in of repeated forward references.
115
116 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
117 incorrectly expecting the subject to contain another "a" after the start.
118
119 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
120 in the match, the SKIP should be ignored. This was not happening; instead
121 the SKIP was being treated as NOMATCH. For patterns such as
122 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
123 tested.
124
125 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
126 now much more compatible with Perl, in particular in cases where the result
127 is a non-match for a non-anchored pattern. For example, if
128 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
129 "m", where previously it did not return a name. A side effect of this
130 change is that for partial matches, the last encountered mark name is
131 returned, as for non matches. A number of tests that were previously not
132 Perl-compatible have been moved into the Perl-compatible test files. The
133 refactoring has had the pleasing side effect of removing one argument from
134 the match() function, thus reducing its stack requirements.
135
136 19. If the /S+ option was used in pcretest to study a pattern using JIT,
137 subsequent uses of /S (without +) incorrectly behaved like /S+.
138
139 21. Retrieve executable code size support for the JIT compiler and fixing
140 some warnings.
141
142 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
143 not work when the shorter character appeared right at the end of the
144 subject string.
145
146 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
147 systems.
148
149 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
150 output it when the /M option is used in pcretest.
151
152 25. The CheckMan script was not being included in the distribution. Also, added
153 an explicit "perl" to run Perl scripts from the PrepareRelease script
154 because this is reportedly needed in Windows.
155
156 26. If study data was being save in a file and studying had not found a set of
157 "starts with" bytes for the pattern, the data written to the file (though
158 never used) was taken from uninitialized memory and so caused valgrind to
159 complain.
160
161 27. Updated RunTest.bat as provided by Sheri Pierce.
162
163 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
164
165 29. Computation of memory usage for the table of capturing group names was
166 giving an unnecessarily large value.
167
168
169 Version 8.20 21-Oct-2011
170 ------------------------
171
172 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
173 a POSIX class. After further experiments with Perl, which convinced me that
174 Perl has bugs and confusions, a closing square bracket is no longer allowed
175 in a POSIX name. This bug also affected patterns with classes that started
176 with full stops.
177
178 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
179 captured substring, but while checking the failing first alternative,
180 substring 1 is temporarily captured. If the output vector supplied to
181 pcre_exec() was not big enough for this capture, the yield of the function
182 was still zero ("insufficient space for captured substrings"). This cannot
183 be totally fixed without adding another stack variable, which seems a lot
184 of expense for a edge case. However, I have improved the situation in cases
185 such as /(a)(b)x|abc/ matched against "abc", where the return code
186 indicates that fewer than the maximum number of slots in the ovector have
187 been set.
188
189 3. Related to (2) above: when there are more back references in a pattern than
190 slots in the output vector, pcre_exec() uses temporary memory during
191 matching, and copies in the captures as far as possible afterwards. It was
192 using the entire output vector, but this conflicts with the specification
193 that only 2/3 is used for passing back captured substrings. Now it uses
194 only the first 2/3, for compatibility. This is, of course, another edge
195 case.
196
197 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
198 main code base, and can be used by building with --enable-jit. When this is
199 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
200 runtime --no-jit option is given.
201
202 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
203 ovector, the return from the function was zero, implying that there were
204 other matches that did not fit. The correct "exactly full" value is now
205 returned.
206
207 6. If a subpattern that was called recursively or as a subroutine contained
208 (*PRUNE) or any other control that caused it to give a non-standard return,
209 invalid errors such as "Error -26 (nested recursion at the same subject
210 position)" or even infinite loops could occur.
211
212 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
213 computing the minimum length on reaching *ACCEPT, and so ended up with the
214 wrong value of 1 rather than 0. Further investigation indicates that
215 computing a minimum subject length in the presence of *ACCEPT is difficult
216 (think back references, subroutine calls), and so I have changed the code
217 so that no minimum is registered for a pattern that contains *ACCEPT.
218
219 8. If (*THEN) was present in the first (true) branch of a conditional group,
220 it was not handled as intended. [But see 16 below.]
221
222 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
223 Sheri Pierce.
224
225 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
226 the first byte in a match must be "a".
227
228 11. Change 17 for 8.13 increased the recursion depth for patterns like
229 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
230 pattern contains any instances of (*THEN). If it does not, the old
231 optimizations are restored. It would be nice to do this on a per-group
232 basis, but at the moment that is not feasible.
233
234 12. In some environments, the output of pcretest -C is CRLF terminated. This
235 broke RunTest's code that checks for the link size. A single white space
236 character after the value is now allowed for.
237
238 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
239 For "fr", it uses the Windows-specific input and output files.
240
241 14. If (*THEN) appeared in a group that was called recursively or as a
242 subroutine, it did not work as intended. [But see next item.]
243
244 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
245 pattern fragments (but not containing any | characters). If A and B are
246 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
247 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
248 D. In other words, Perl considers parentheses that do not contain any |
249 characters to be part of a surrounding alternative, whereas PCRE was
250 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
251 differently. PCRE now behaves in the same way as Perl, except in the case
252 of subroutine/recursion calls such as (?1) which have in any case always
253 been different (but PCRE had them first :-).
254
255 16. Related to 15 above: Perl does not treat the | in a conditional group as
256 creating alternatives. Such a group is treated in the same way as an
257 ordinary group without any | characters when processing (*THEN). PCRE has
258 been changed to match Perl's behaviour.
259
260 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
261 RunGrepTest script failed.
262
263 18. Change 22 for version 13 caused atomic groups to use more stack. This is
264 inevitable for groups that contain captures, but it can lead to a lot of
265 stack use in large patterns. The old behaviour has been restored for atomic
266 groups that do not contain any capturing parentheses.
267
268 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
269 suppress the check for a minimum subject length at run time. (If it was
270 given to pcre_exec() or pcre_dfa_exec() it did work.)
271
272 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
273 fail to work when decoding hex characters in data strings in EBCDIC
274 environments.
275
276 21. It appears that in at least one Mac OS environment, the isxdigit() function
277 is implemented as a macro that evaluates to its argument more than once,
278 contravening the C 90 Standard (I haven't checked a later standard). There
279 was an instance in pcretest which caused it to go wrong when processing
280 \x{...} escapes in subject strings. The has been rewritten to avoid using
281 things like p++ in the argument of isxdigit().
282
283
284 Version 8.13 16-Aug-2011
285 ------------------------
286
287 1. The Unicode data tables have been updated to Unicode 6.0.0.
288
289 2. Two minor typos in pcre_internal.h have been fixed.
290
291 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
292 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
293 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
294
295 4. There were a number of related bugs in the code for matching backrefences
296 caselessly in UTF-8 mode when codes for the characters concerned were
297 different numbers of bytes. For example, U+023A and U+2C65 are an upper
298 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
299 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
300 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
301 2-byte code at the end of the subject (it thought there wasn't enough data
302 left).
303
304 5. Comprehensive information about what went wrong is now returned by
305 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
306 as the output vector has at least 2 elements. The offset of the start of
307 the failing character and a reason code are placed in the vector.
308
309 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
310 now returned is for the first byte of the failing character, instead of the
311 last byte inspected. This is an incompatible change, but I hope it is small
312 enough not to be a problem. It makes the returned offset consistent with
313 pcre_exec() and pcre_dfa_exec().
314
315 7. pcretest now gives a text phrase as well as the error number when
316 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
317 failure, the offset and reason code are output.
318
319 8. When \R was used with a maximizing quantifier it failed to skip backwards
320 over a \r\n pair if the subsequent match failed. Instead, it just skipped
321 back over a single character (\n). This seems wrong (because it treated the
322 two characters as a single entity when going forwards), conflicts with the
323 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
324 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
325 has been changed.
326
327 9. Some internal refactoring has changed the processing so that the handling
328 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
329 time (the PCRE_DOTALL option was changed this way some time ago: version
330 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
331 which was always a bit of a fudge. It also means that there is one less
332 argument for the match() function, which reduces its stack requirements
333 slightly. This change also fixes an incompatibility with Perl: the pattern
334 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
335
336 10. More internal refactoring has drastically reduced the number of recursive
337 calls to match() for possessively repeated groups such as (abc)++ when
338 using pcre_exec().
339
340 11. While implementing 10, a number of bugs in the handling of groups were
341 discovered and fixed:
342
343 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
344 (a|)*(?1) gave a compile-time internal error.
345 ((a|)+)+ did not notice that the outer group could match an empty string.
346 (^a|^)+ was not marked as anchored.
347 (.*a|.*)+ was not marked as matching at start or after a newline.
348
349 12. Yet more internal refactoring has removed another argument from the match()
350 function. Special calls to this function are now indicated by setting a
351 value in a variable in the "match data" data block.
352
353 13. Be more explicit in pcre_study() instead of relying on "default" for
354 opcodes that mean there is no starting character; this means that when new
355 ones are added and accidentally left out of pcre_study(), testing should
356 pick them up.
357
358 14. The -s option of pcretest has been documented for ages as being an old
359 synonym of -m (show memory usage). I have changed it to mean "force study
360 for every regex", that is, assume /S for every regex. This is similar to -i
361 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
362 using it. It makes it easier to run collections of tests with and without
363 study enabled, and thereby test pcre_study() more easily. All the standard
364 tests are now run with and without -s (but some patterns can be marked as
365 "never study" - see 20 below).
366
367 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
368 restoration of the capturing data to the outer values was not happening
369 correctly.
370
371 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
372 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
373 pattern had matched an empty string, and so incorrectly returned a no
374 match.
375
376 17. There was optimizing code for the last branch of non-capturing parentheses,
377 and also for the obeyed branch of a conditional subexpression, which used
378 tail recursion to cut down on stack usage. Unfortunately, now that there is
379 the possibility of (*THEN) occurring in these branches, tail recursion is
380 no longer possible because the return has to be checked for (*THEN). These
381 two optimizations have therefore been removed. [But see 8.20/11 above.]
382
383 18. If a pattern containing \R was studied, it was assumed that \R always
384 matched two bytes, thus causing the minimum subject length to be
385 incorrectly computed because \R can also match just one byte.
386
387 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
388 was incorrectly computed.
389
390 20. If /S is present twice on a test pattern in pcretest input, it now
391 *disables* studying, thereby overriding the use of -s on the command line
392 (see 14 above). This is necessary for one or two tests to keep the output
393 identical in both cases.
394
395 21. When (*ACCEPT) was used in an assertion that matched an empty string and
396 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
397
398 22. When an atomic group that contained a capturing parenthesis was
399 successfully matched, but the branch in which it appeared failed, the
400 capturing was not being forgotten if a higher numbered group was later
401 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
402 group 1 to "a", when in fact it should be unset. This applied to multi-
403 branched capturing and non-capturing groups, repeated or not, and also to
404 positive assertions (capturing in negative assertions does not happen
405 in PCRE) and also to nested atomic groups.
406
407 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
408 subject after a captured substring, to make it easier to tell which of a
409 number of identical substrings has been captured.
410
411 24. The way atomic groups are processed by pcre_exec() has been changed so that
412 if they are repeated, backtracking one repetition now resets captured
413 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
414 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
415 "aaa". Previously, it would have been "a". As part of this code
416 refactoring, the way recursive calls are handled has also been changed.
417
418 25. If an assertion condition captured any substrings, they were not passed
419 back unless some other capturing happened later. For example, if
420 (?(?=(a))a) was matched against "a", no capturing was returned.
421
422 26. When studying a pattern that contained subroutine calls or assertions,
423 the code for finding the minimum length of a possible match was handling
424 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
425 group 1 called group 2 while simultaneously a separate group 2 called group
426 1). A stack overflow occurred in this case. I have fixed this by limiting
427 the recursion depth to 10.
428
429 27. Updated RunTest.bat in the distribution to the version supplied by Tom
430 Fortmann. This supports explicit test numbers on the command line, and has
431 argument validation and error reporting.
432
433 28. An instance of \X with an unlimited repeat could fail if at any point the
434 first character it looked at was a mark character.
435
436 29. Some minor code refactoring concerning Unicode properties and scripts
437 should reduce the stack requirement of match() slightly.
438
439 30. Added the '=' option to pcretest to check the setting of unused capturing
440 slots at the end of the pattern, which are documented as being -1, but are
441 not included in the return count.
442
443 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
444 compiled something random. Now it gives a compile-time error (as does
445 Perl).
446
447 32. A *MARK encountered during the processing of a positive assertion is now
448 recorded and passed back (compatible with Perl).
449
450 33. If --only-matching or --colour was set on a pcregrep call whose pattern
451 had alternative anchored branches, the search for a second match in a line
452 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
453 matched the line "0102" twice. The same bug affected patterns that started
454 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
455 twice.
456
457 34. Previously, PCRE did not allow quantification of assertions. However, Perl
458 does, and because of capturing effects, quantifying parenthesized
459 assertions may at times be useful. Quantifiers are now allowed for
460 parenthesized assertions.
461
462 35. A minor code tidy in pcre_compile() when checking options for \R usage.
463
464 36. \g was being checked for fancy things in a character class, when it should
465 just be a literal "g".
466
467 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
468 appearance of a nested POSIX class supersedes an apparent external class.
469 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
470 unescaped square brackets may also appear as part of class names. For
471 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
472 more like Perl. (But see 8.20/1 above.)
473
474 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
475 was because it thought it was \N{name}, which is not supported).
476
477 39. Add minix to OS list not supporting the -S option in pcretest.
478
479 40. PCRE tries to detect cases of infinite recursion at compile time, but it
480 cannot analyze patterns in sufficient detail to catch mutual recursions
481 such as ((?1))((?2)). There is now a runtime test that gives an error if a
482 subgroup is called recursively as a subpattern for a second time at the
483 same position in the subject string. In previous releases this might have
484 been caught by the recursion limit, or it might have run out of stack.
485
486 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
487 happen only once. PCRE was, however incorrectly giving a compile time error
488 "recursive call could loop indefinitely" because it cannot analyze the
489 pattern in sufficient detail. The compile time test no longer happens when
490 PCRE is compiling a conditional subpattern, but actual runaway loops are
491 now caught at runtime (see 40 above).
492
493 42. It seems that Perl allows any characters other than a closing parenthesis
494 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
495 has been changed to be the same.
496
497 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
498 as not to get warnings when autogen.sh is called. Also changed
499 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
500
501 44. To help people who use pcregrep to scan files containing exceedingly long
502 lines, the following changes have been made:
503
504 (a) The default value of the buffer size parameter has been increased from
505 8K to 20K. (The actual buffer used is three times this size.)
506
507 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
508 PCRE is built.
509
510 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
511 to be set at run time.
512
513 (d) Numerical values in pcregrep options can be followed by K or M, for
514 example --buffer-size=50K.
515
516 (e) If a line being scanned overflows pcregrep's buffer, an error is now
517 given and the return code is set to 2.
518
519 45. Add a pointer to the latest mark to the callout data block.
520
521 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
522 partial match of an empty string instead of no match. This was specific to
523 the use of ".".
524
525 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
526 complete match instead of a partial match. This bug was dependent on both
527 the PCRE_UTF8 and PCRE_DOTALL options being set.
528
529 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
530 starting byte set, because \b was not being ignored.
531
532
533 Version 8.12 15-Jan-2011
534 ------------------------
535
536 1. Fixed some typos in the markup of the man pages, and wrote a script that
537 checks for such things as part of the documentation building process.
538
539 2. On a big-endian 64-bit system, pcregrep did not correctly process the
540 --match-limit and --recursion-limit options (added for 8.11). In
541 particular, this made one of the standard tests fail. (The integer value
542 went into the wrong half of a long int.)
543
544 3. If the --colour option was given to pcregrep with -v (invert match), it
545 did strange things, either producing crazy output, or crashing. It should,
546 of course, ignore a request for colour when reporting lines that do not
547 match.
548
549 4. Another pcregrep bug caused similar problems if --colour was specified with
550 -M (multiline) and the pattern match finished with a line ending.
551
552 5. In pcregrep, when a pattern that ended with a literal newline sequence was
553 matched in multiline mode, the following line was shown as part of the
554 match. This seems wrong, so I have changed it.
555
556 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
557 the check for further matches in the same line (so they could be coloured)
558 to overrun the end of the current line. If another match was found, it was
559 incorrectly shown (and then shown again when found in the next line).
560
561 7. If pcregrep was compiled under Windows, there was a reference to the
562 function pcregrep_exit() before it was defined. I am assuming this was
563 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
564 reported by a user. I've moved the definition above the reference.
565
566
567 Version 8.11 10-Dec-2010
568 ------------------------
569
570 1. (*THEN) was not working properly if there were untried alternatives prior
571 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
572 backtracked to try for "b" instead of moving to the next alternative branch
573 at the same level (in this case, to look for "c"). The Perl documentation
574 is clear that when (*THEN) is backtracked onto, it goes to the "next
575 alternative in the innermost enclosing group".
576
577 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
578 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
579 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
580 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
581 (*THEN).
582
583 3. If \s appeared in a character class, it removed the VT character from
584 the class, even if it had been included by some previous item, for example
585 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
586 of \s, but is part of the POSIX "space" class.)
587
588 4. A partial match never returns an empty string (because you can always
589 match an empty string at the end of the subject); however the checking for
590 an empty string was starting at the "start of match" point. This has been
591 changed to the "earliest inspected character" point, because the returned
592 data for a partial match starts at this character. This means that, for
593 example, /(?<=abc)def/ gives a partial match for the subject "abc"
594 (previously it gave "no match").
595
596 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
597 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
598 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
599 has an implication that the given string is incomplete (because a partial
600 match is preferred over a full match). For this reason, these items now
601 give a partial match in this situation. [Aside: previously, the one case
602 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
603 match rather than a full match, which was wrong by the old rules, but is
604 now correct.]
605
606 6. There was a bug in the handling of #-introduced comments, recognized when
607 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
608 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
609 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
610 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
611 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
612 places in pcre_compile().
613
614 7. Related to (6) above, when pcre_compile() was skipping #-introduced
615 comments when looking ahead for named forward references to subpatterns,
616 the only newline sequence it recognized was NL. It now handles newlines
617 according to the set newline convention.
618
619 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
620 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
621 cater for a lack of strerror(). These oversights have been fixed.
622
623 9. Added --match-limit and --recursion-limit to pcregrep.
624
625 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
626
627 11. When the -o option was used, pcregrep was setting a return code of 1, even
628 when matches were found, and --line-buffered was not being honoured.
629
630 12. Added an optional parentheses number to the -o and --only-matching options
631 of pcregrep.
632
633 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
634 can match an empty string. The code to do it in pcretest and pcredemo
635 needed fixing:
636
637 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
638 only one byte after an empty string match just before CRLF (this case
639 just got forgotten; "any" and "anycrlf" were OK).
640
641 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
642 mode when an empty string match preceded an ASCII character followed by
643 a non-ASCII character. (The code for advancing by one character rather
644 than one byte was nonsense.)
645
646 (c) The pcredemo.c sample program did not have any code at all to handle
647 the cases when CRLF is a valid newline sequence.
648
649 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
650 as a starting offset was within the subject string. There is now a new
651 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
652 negative or greater than the length of the string. In order to test this,
653 pcretest is extended to allow the setting of negative starting offsets.
654
655 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
656 starting offset points to the beginning of a UTF-8 character was
657 unnecessarily clumsy. I tidied it up.
658
659 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
660 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
661
662 17. Nobody had reported that the --include_dir option, which was added in
663 release 7.7 should have been called --include-dir (hyphen, not underscore)
664 for compatibility with GNU grep. I have changed it to --include-dir, but
665 left --include_dir as an undocumented synonym, and the same for
666 --exclude-dir, though that is not available in GNU grep, at least as of
667 release 2.5.4.
668
669 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
670 characters from a string of bytes) have been redefined so as not to use
671 loops, in order to improve performance in some environments. At the same
672 time, I abstracted some of the common code into auxiliary macros to save
673 repetition (this should not affect the compiled code).
674
675 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
676 compile-time error is now given if \c is not followed by an ASCII
677 character, that is, a byte less than 128. (In EBCDIC mode, the code is
678 different, and any byte value is allowed.)
679
680 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
681 START_OPTIMIZE option, which is now allowed at compile time - but just
682 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
683 to pcregrep and other applications that have no direct access to PCRE
684 options. The new /Y option in pcretest sets this option when calling
685 pcre_compile().
686
687 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
688 back references. Groups containing recursive back references were forced to
689 be atomic by that change, but in the case of named groups, the amount of
690 memory required was incorrectly computed, leading to "Failed: internal
691 error: code overflow". This has been fixed.
692
693 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
694 pcretest.c, to avoid build problems in some Borland environments.
695
696
697 Version 8.10 25-Jun-2010
698 ------------------------
699
700 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
701 THEN.
702
703 2. (*ACCEPT) was not working when inside an atomic group.
704
705 3. Inside a character class, \B is treated as a literal by default, but
706 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
707 causes the error). The code is unchanged, but I tidied the documentation.
708
709 4. Inside a character class, PCRE always treated \R and \X as literals,
710 whereas Perl faults them if its -w option is set. I have changed PCRE so
711 that it faults them when PCRE_EXTRA is set.
712
713 5. Added support for \N, which always matches any character other than
714 newline. (It is the same as "." when PCRE_DOTALL is not set.)
715
716 6. When compiling pcregrep with newer versions of gcc which may have
717 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
718 declared with attribute warn_unused_result" were given. Just casting the
719 result to (void) does not stop the warnings; a more elaborate fudge is
720 needed. I've used a macro to implement this.
721
722 7. Minor change to pcretest.c to avoid a compiler warning.
723
724 8. Added four artifical Unicode properties to help with an option to make
725 \s etc use properties (see next item). The new properties are: Xan
726 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
727
728 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
729 use Unicode properties. (*UCP) at the start of a pattern can be used to set
730 this option. Modified pcretest to add /W to test this facility. Added
731 REG_UCP to make it available via the POSIX interface.
732
733 10. Added --line-buffered to pcregrep.
734
735 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
736 studied, and the match started with a letter with a code point greater than
737 127 whose first byte was different to the first byte of the other case of
738 the letter, the other case of this starting letter was not recognized
739 (#976).
740
741 12. If a pattern that was studied started with a repeated Unicode property
742 test, for example, \p{Nd}+, there was the theoretical possibility of
743 setting up an incorrect bitmap of starting bytes, but fortunately it could
744 not have actually happened in practice until change 8 above was made (it
745 added property types that matched character-matching opcodes).
746
747 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
748 possible starting bytes for non-anchored patterns.
749
750 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
751 \R, and also a number of cases that involve Unicode properties, both
752 explicit and implicit when PCRE_UCP is set.
753
754 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
755 input, it could crash or give wrong results if characters with values
756 greater than 0xc0 were present in the subject string. (Detail: it assumed
757 UTF-8 input when processing these items.)
758
759 16. Added a lot of (int) casts to avoid compiler warnings in systems where
760 size_t is 64-bit (#991).
761
762 17. Added a check for running out of memory when PCRE is compiled with
763 --disable-stack-for-recursion (#990).
764
765 18. If the last data line in a file for pcretest does not have a newline on
766 the end, a newline was missing in the output.
767
768 19. The default pcre_chartables.c file recognizes only ASCII characters (values
769 less than 128) in its various bitmaps. However, there is a facility for
770 generating tables according to the current locale when PCRE is compiled. It
771 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
772 characters, are recognized by isspace() and therefore were getting set in
773 these tables, and indeed these tables seem to approximate to ISO 8859. This
774 caused a problem in UTF-8 mode when pcre_study() was used to create a list
775 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
776 which of course cannot start UTF-8 characters. I have changed the code so
777 that only real ASCII characters (less than 128) and the correct starting
778 bytes for UTF-8 encodings are set for characters greater than 127 when in
779 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
780 altogether.)
781
782 20. Added the /T option to pcretest so as to be able to run tests with non-
783 standard character tables, thus making it possible to include the tests
784 used for 19 above in the standard set of tests.
785
786 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
787 reference to a subpattern the other side of a comment that contains an
788 opening parenthesis caused either an internal compiling error, or a
789 reference to the wrong subpattern.
790
791
792 Version 8.02 19-Mar-2010
793 ------------------------
794
795 1. The Unicode data tables have been updated to Unicode 5.2.0.
796
797 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
798 configured.
799
800 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
801 original author of that file, following a query about its status.
802
803 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
804 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
805
806 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
807 quantifier applied to a forward-referencing subroutine call, could compile
808 incorrect code or give the error "internal error: previously-checked
809 referenced subpattern not found".
810
811 6. Both MS Visual Studio and Symbian OS have problems with initializing
812 variables to point to external functions. For these systems, therefore,
813 pcre_malloc etc. are now initialized to local functions that call the
814 relevant global functions.
815
816 7. There were two entries missing in the vectors called coptable and poptable
817 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
818 I've fixed the data, and added a kludgy way of testing at compile time that
819 the lengths are correct (equal to the number of opcodes).
820
821 8. Following on from 7, I added a similar kludge to check the length of the
822 eint vector in pcreposix.c.
823
824 9. Error texts for pcre_compile() are held as one long string to avoid too
825 much relocation at load time. To find a text, the string is searched,
826 counting zeros. There was no check for running off the end of the string,
827 which could happen if a new error number was added without updating the
828 string.
829
830 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
831
832 11. \K was not working if it appeared in an atomic group or in a group that
833 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
834 \K is "not well defined" if used in an assertion. PCRE now accepts it if
835 the assertion is positive, but not if it is negative.
836
837 12. Change 11 fortuitously reduced the size of the stack frame used in the
838 "match()" function of pcre_exec.c by one pointer. Forthcoming
839 implementation of support for (*MARK) will need an extra pointer on the
840 stack; I have reserved it now, so that the stack frame size does not
841 decrease.
842
843 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
844 item in branch that calls a recursion is a subroutine call - as in the
845 second branch in the above example - was incorrectly given the compile-
846 time error "recursive call could loop indefinitely" because pcre_compile()
847 was not correctly checking the subroutine for matching a non-empty string.
848
849 14. The checks for overrunning compiling workspace could trigger after an
850 overrun had occurred. This is a "should never occur" error, but it can be
851 triggered by pathological patterns such as hundreds of nested parentheses.
852 The checks now trigger 100 bytes before the end of the workspace.
853
854 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
855
856
857 Version 8.01 19-Jan-2010
858 ------------------------
859
860 1. If a pattern contained a conditional subpattern with only one branch (in
861 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
862 computed the wrong minimum data length (which is of course zero for such
863 subpatterns). This could cause incorrect "no match" results.
864
865 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
866 the pattern is reset in the first branch, pcre_compile() failed with
867 "internal error: code overflow at offset...". This happened only when
868 the reset was to the original external option setting. (An optimization
869 abstracts leading options settings into an external setting, which was the
870 cause of this.)
871
872 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
873 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
874 assertion pattern did not match (meaning that the assertion was true), it
875 was incorrectly treated as false if the SKIP had been reached during the
876 matching. This also applied to assertions used as conditions.
877
878 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
879 assertion subpattern, including such a pattern used as a condition,
880 unpredictable results occurred, instead of the error return
881 PCRE_ERROR_DFA_UITEM.
882
883 5. The C++ GlobalReplace function was not working like Perl for the special
884 situation when an empty string is matched. It now does the fancy magic
885 stuff that is necessary.
886
887 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
888 removed. (These were left over from very, very early versions of PCRE.)
889
890 7. Some cosmetic changes to the code to make life easier when compiling it
891 as part of something else:
892
893 (a) Change DEBUG to PCRE_DEBUG.
894
895 (b) In pcre_compile(), rename the member of the "branch_chain" structure
896 called "current" as "current_branch", to prevent a collision with the
897 Linux macro when compiled as a kernel module.
898
899 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
900 prevent a collision with the Linux macro when compiled as a kernel
901 module.
902
903 8. In pcre_compile() there are some checks for integer overflows that used to
904 cast potentially large values to (double). This has been changed to that
905 when building, a check for int64_t is made, and if it is found, it is used
906 instead, thus avoiding the use of floating point arithmetic. (There is no
907 other use of FP in PCRE.) If int64_t is not found, the fallback is to
908 double.
909
910 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
911 2005 (difference between two addresses compared to an unsigned value).
912
913 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
914 custom one, because of the following reported problem in Windows:
915
916 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
917 under Win32.
918 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
919 therefore missing the function definition.
920 - The compiler thus generates a "C" signature for the test function.
921 - The linker fails to find the "C" function.
922 - PCRE fails to configure if asked to do so against libbz2.
923
924 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
925 messages were output:
926
927 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
928 rerunning libtoolize, to keep the correct libtool macros in-tree.
929 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
930
931 I have done both of these things.
932
933 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
934 most of the time, it *can* run out if it is given a pattern that contains a
935 runaway infinite recursion. I updated the discussion in the pcrestack man
936 page.
937
938 13. Now that we have gone to the x.xx style of version numbers, the minor
939 version may start with zero. Using 08 or 09 is a bad idea because users
940 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
941 interpreted as invalid octal numbers. I've updated the previous comment in
942 configure.ac, and also added a check that gives an error if 08 or 09 are
943 used.
944
945 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
946 causing partial matching to fail when the end of the subject matched \W
947 in a UTF-8 pattern where \W was quantified with a minimum of 3.
948
949 15. There were some discrepancies between the declarations in pcre_internal.h
950 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
951 their definitions. The declarations used "const uschar *" and the
952 definitions used USPTR. Even though USPTR is normally defined as "const
953 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
954 reported that: "This difference in casting confuses some C++ compilers, for
955 example, SunCC recognizes above declarations as different functions and
956 generates broken code for hbpcre." I have changed the declarations to use
957 USPTR.
958
959 16. GNU libtool is named differently on some systems. The autogen.sh script now
960 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
961 (FreeBSD).
962
963 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
964 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
965 comment: "Figure out how to create a longlong from a string: strtoll and
966 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
967 instance, but it only takes 2 args instead of 3!"
968
969 18. A subtle bug concerned with back references has been fixed by a change of
970 specification, with a corresponding code fix. A pattern such as
971 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
972 refers, was giving matches when it shouldn't. For example, xa=xaaa would
973 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
974 same bug. Such groups have to be quantified to be useful, or contained
975 inside another quantified group. (If there's no repetition, the reference
976 can never match.) The problem arises because, having left the group and
977 moved on to the rest of the pattern, a later failure that backtracks into
978 the group uses the captured value from the final iteration of the group
979 rather than the correct earlier one. I have fixed this in PCRE by forcing
980 any group that contains a reference to itself to be an atomic group; that
981 is, there cannot be any backtracking into it once it has completed. This is
982 similar to recursive and subroutine calls.
983
984
985 Version 8.00 19-Oct-09
986 ----------------------
987
988 1. The table for translating pcre_compile() error codes into POSIX error codes
989 was out-of-date, and there was no check on the pcre_compile() error code
990 being within the table. This could lead to an OK return being given in
991 error.
992
993 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
994 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
995 in a Windows environment.
996
997 3. The pcregrep --count option prints the count for each file even when it is
998 zero, as does GNU grep. However, pcregrep was also printing all files when
999 --files-with-matches was added. Now, when both options are given, it prints
1000 counts only for those files that have at least one match. (GNU grep just
1001 prints the file name in this circumstance, but including the count seems
1002 more useful - otherwise, why use --count?) Also ensured that the
1003 combination -clh just lists non-zero counts, with no names.
1004
1005 4. The long form of the pcregrep -F option was incorrectly implemented as
1006 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1007 but it seems right to fix it, and I didn't think it was worth preserving
1008 the old behaviour.
1009
1010 5. The command line items --regex=pattern and --regexp=pattern were not
1011 recognized by pcregrep, which required --regex pattern or --regexp pattern
1012 (with a space rather than an '='). The man page documented the '=' forms,
1013 which are compatible with GNU grep; these now work.
1014
1015 6. No libpcreposix.pc file was created for pkg-config; there was just
1016 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1017
1018 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1019 when UCP support is not needed, by modifying the Python script that
1020 generates it from Unicode data files. This should not matter if the module
1021 is correctly used as a library, but I received one complaint about 50K of
1022 unwanted data. My guess is that the person linked everything into his
1023 program rather than using a library. Anyway, it does no harm.
1024
1025 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1026 was a minimum greater than 1 for a wide character in a possessive
1027 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1028 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1029 character. Chaos in the form of incorrect output or a compiling loop could
1030 result.
1031
1032 9. The restrictions on what a pattern can contain when partial matching is
1033 requested for pcre_exec() have been removed. All patterns can now be
1034 partially matched by this function. In addition, if there are at least two
1035 slots in the offset vector, the offset of the earliest inspected character
1036 for the match and the offset of the end of the subject are set in them when
1037 PCRE_ERROR_PARTIAL is returned.
1038
1039 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1040 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1041 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1042 and may be more useful for multi-segment matching.
1043
1044 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1045 used to be given if ever the end of the subject was reached; now it is
1046 given only if matching could not proceed because another character was
1047 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1048 string "Z", which now yields "no match" instead of "partial match". In the
1049 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1050 final character ended with (*FAIL).
1051
1052 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1053 if the pattern had a "must contain" character that was already found in the
1054 earlier partial match, unless partial matching was again requested. For
1055 example, with the pattern /dog.(body)?/, the "must contain" character is
1056 "g". If the first part-match was for the string "dog", restarting with
1057 "sbody" failed. This bug has been fixed.
1058
1059 13. The string returned by pcre_dfa_exec() after a partial match has been
1060 changed so that it starts at the first inspected character rather than the
1061 first character of the match. This makes a difference only if the pattern
1062 starts with a lookbehind assertion or \b or \B (\K is not supported by
1063 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1064 matching functions compatible, and I think it's the right thing to do.
1065
1066 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1067 so that the demonstration program is easily available in environments where
1068 PCRE has not been installed from source.
1069
1070 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1071 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1072 library.
1073
1074 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1075 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1076 is not the first non-POSIX option to be added. Clearly some people find
1077 these options useful.
1078
1079 17. If a caller to the POSIX matching function regexec() passes a non-zero
1080 value for nmatch with a NULL value for pmatch, the value of
1081 nmatch is forced to zero.
1082
1083 18. RunGrepTest did not have a test for the availability of the -u option of
1084 the diff command, as RunTest does. It now checks in the same way as
1085 RunTest, and also checks for the -b option.
1086
1087 19. If an odd number of negated classes containing just a single character
1088 interposed, within parentheses, between a forward reference to a named
1089 subpattern and the definition of the subpattern, compilation crashed with
1090 an internal error, complaining that it could not find the referenced
1091 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1092 [The bug was that it was starting one character too far in when skipping
1093 over the character class, thus treating the ] as data rather than
1094 terminating the class. This meant it could skip too much.]
1095
1096 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1097 /g option in pcretest when the pattern contains \K, which makes it possible
1098 to have an empty string match not at the start, even when the pattern is
1099 anchored. Updated pcretest and pcredemo to use this option.
1100
1101 21. If the maximum number of capturing subpatterns in a recursion was greater
1102 than the maximum at the outer level, the higher number was returned, but
1103 with unset values at the outer level. The correct (outer level) value is
1104 now given.
1105
1106 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1107 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1108 make it do so. The string so far is captured, making this feature
1109 compatible with Perl.
1110
1111 23. The tests have been re-organized, adding tests 11 and 12, to make it
1112 possible to check the Perl 5.10 features against Perl 5.10.
1113
1114 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1115 pattern matches a fixed length string. PCRE did not allow this; now it
1116 does. Neither allows recursion.
1117
1118 25. I finally figured out how to implement a request to provide the minimum
1119 length of subject string that was needed in order to match a given pattern.
1120 (It was back references and recursion that I had previously got hung up
1121 on.) This code has now been added to pcre_study(); it finds a lower bound
1122 to the length of subject needed. It is not necessarily the greatest lower
1123 bound, but using it to avoid searching strings that are too short does give
1124 some useful speed-ups. The value is available to calling programs via
1125 pcre_fullinfo().
1126
1127 26. While implementing 25, I discovered to my embarrassment that pcretest had
1128 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1129 study optimizations had never been tested with that matching function.
1130 Oops. What is worse, even when it was passed study data, there was a bug in
1131 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1132 were also very few tests of studied patterns with pcre_dfa_exec().
1133
1134 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1135 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1136 on the other side of the coin, they are no longer allowed to have different
1137 names, because these cannot be distinguished in PCRE, and this has caused
1138 confusion. (This is a difference from Perl.)
1139
1140 28. When duplicate subpattern names are present (necessarily with different
1141 numbers, as required by 27 above), and a test is made by name in a
1142 conditional pattern, either for a subpattern having been matched, or for
1143 recursion in such a pattern, all the associated numbered subpatterns are
1144 tested, and the overall condition is true if the condition is true for any
1145 one of them. This is the way Perl works, and is also more like the way
1146 testing by number works.
1147
1148
1149 Version 7.9 11-Apr-09
1150 ---------------------
1151
1152 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1153 (pcretest), all targets were linked against these libraries. This included
1154 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1155 libraries. This caused unwanted dependencies to be created. This problem
1156 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1157 pcretest is linked with readline.
1158
1159 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1160 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1161 moved outside it again, because FALSE and TRUE are already defined in AIX,
1162 but BOOL is not.
1163
1164 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1165 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1166
1167 4. The pcregrep documentation said spaces were inserted as well as colons (or
1168 hyphens) following file names and line numbers when outputting matching
1169 lines. This is not true; no spaces are inserted. I have also clarified the
1170 wording for the --colour (or --color) option.
1171
1172 5. In pcregrep, when --colour was used with -o, the list of matching strings
1173 was not coloured; this is different to GNU grep, so I have changed it to be
1174 the same.
1175
1176 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1177 each matching line was coloured. Now it goes on to look for further matches
1178 of any of the test patterns, which is the same behaviour as GNU grep.
1179
1180 7. A pattern that could match an empty string could cause pcregrep to loop; it
1181 doesn't make sense to accept an empty string match in pcregrep, so I have
1182 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1183 seems to be how GNU grep behaves.
1184
1185 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1186 start or after a newline", because the conditional assertion was not being
1187 correctly handled. The rule now is that both the assertion and what follows
1188 in the first alternative must satisfy the test.
1189
1190 9. If auto-callout was enabled in a pattern with a conditional group whose
1191 condition was an assertion, PCRE could crash during matching, both with
1192 pcre_exec() and pcre_dfa_exec().
1193
1194 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1195 used for matching.
1196
1197 11. Unicode property support in character classes was not working for
1198 characters (bytes) greater than 127 when not in UTF-8 mode.
1199
1200 12. Added the -M command line option to pcretest.
1201
1202 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1203
1204 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1205
1206 16. Added comments and documentation about mis-use of no_arg in the C++
1207 wrapper.
1208
1209 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1210 from Martin Jerabek that uses macro names for all relevant character and
1211 string constants.
1212
1213 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1214 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1215 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1216 these, but not everybody uses configure.
1217
1218 19. A conditional group that had only one branch was not being correctly
1219 recognized as an item that could match an empty string. This meant that an
1220 enclosing group might also not be so recognized, causing infinite looping
1221 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1222 with the subject "ab", where knowledge that the repeated group can match
1223 nothing is needed in order to break the loop.
1224
1225 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1226 exec(), but without supplying a callout function, matching went wrong.
1227
1228 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1229 leak if the size of the offset vector was greater than 30. When the vector
1230 is smaller, the saved offsets during recursion go onto a local stack
1231 vector, but for larger vectors malloc() is used. It was failing to free
1232 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1233 error, in fact).
1234
1235 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1236 heapframe that is used only when UTF-8 support is enabled. This caused no
1237 problem, but was untidy.
1238
1239 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1240 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1241 included within another project.
1242
1243 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1244 slightly modified by me:
1245
1246 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1247 not building pcregrep.
1248
1249 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1250 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1251
1252 25. Forward references, both numeric and by name, in patterns that made use of
1253 duplicate group numbers, could behave incorrectly or give incorrect errors,
1254 because when scanning forward to find the reference group, PCRE was not
1255 taking into account the duplicate group numbers. A pattern such as
1256 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1257
1258 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1259 the feature of a custom pointer more persuasive (as requested by a user).
1260
1261 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1262 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1263 pre-defined.
1264
1265 28. Added support for (*UTF8) at the start of a pattern.
1266
1267 29. Arrange for flags added by the "release type" setting in CMake to be shown
1268 in the configuration summary.
1269
1270
1271 Version 7.8 05-Sep-08
1272 ---------------------
1273
1274 1. Replaced UCP searching code with optimized version as implemented for Ad
1275 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1276 stage table and inline lookup instead of a function, giving speed ups of 2
1277 to 5 times on some simple patterns that I tested. Permission was given to
1278 distribute the MultiStage2.py script that generates the tables (it's not in
1279 the tarball, but is in the Subversion repository).
1280
1281 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1282 scripts.
1283
1284 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1285 a group with a zero qualifier. The result of the study could be incorrect,
1286 or the function might crash, depending on the pattern.
1287
1288 4. Caseless matching was not working for non-ASCII characters in back
1289 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1290 It now works when Unicode Property Support is available.
1291
1292 5. In pcretest, an escape such as \x{de} in the data was always generating
1293 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1294 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1295 truncation.
1296
1297 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1298
1299 7. Added two (int) casts to pcregrep when printing the difference of two
1300 pointers, in case they are 64-bit values.
1301
1302 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1303 test 2 if it fails.
1304
1305 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1306 and a #define of that name to empty if it is not externally set. This is to
1307 allow users of MSVC to set it if necessary.
1308
1309 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1310 the convenience functions in the pcre_get.c source file.
1311
1312 11. An option change at the start of a pattern that had top-level alternatives
1313 could cause overwriting and/or a crash. This command provoked a crash in
1314 some environments:
1315
1316 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1317
1318 This potential security problem was recorded as CVE-2008-2371.
1319
1320 12. For a pattern where the match had to start at the beginning or immediately
1321 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1322 pcre_dfa_exec() could read past the end of the passed subject if there was
1323 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1324 pcretest so that it places the subject at the end of its malloc-ed buffer.
1325
1326 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1327 exec() might read past the end of the data buffer in UTF-8 mode.
1328
1329 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1330 the data contained the byte 0x85 as part of a UTF-8 character within its
1331 first line. This applied both to normal and DFA matching.
1332
1333 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1334 /^[^d]*?$/8 failed to match "abc".
1335
1336 16. Added a missing copyright notice to pcrecpp_internal.h.
1337
1338 17. Make it more clear in the documentation that values returned from
1339 pcre_exec() in ovector are byte offsets, not character counts.
1340
1341 18. Tidied a few places to stop certain compilers from issuing warnings.
1342
1343 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1344 supplied by Stefan Weber. I made a further small update for 7.8 because
1345 there is a change of source arrangements: the pcre_searchfuncs.c module is
1346 replaced by pcre_ucd.c.
1347
1348
1349 Version 7.7 07-May-08
1350 ---------------------
1351
1352 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1353 a string to a long long, pretend we don't even have a long long." This is
1354 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1355
1356 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1357 pre-7.6 versions, which defined a global no_arg variable instead of putting
1358 it in the RE class. (See also #8 below.)
1359
1360 3. Remove a line of dead code, identified by coverity and reported by Nuno
1361 Lopes.
1362
1363 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1364
1365 (1) The include/exclude patterns were being applied to the whole pathnames
1366 of files, instead of just to the final components.
1367
1368 (2) If there was more than one level of directory, the subdirectories were
1369 skipped unless they satisfied the include/exclude conditions. This is
1370 inconsistent with GNU grep (and could even be seen as contrary to the
1371 pcregrep specification - which I improved to make it absolutely clear).
1372 The action now is always to scan all levels of directory, and just
1373 apply the include/exclude patterns to regular files.
1374
1375 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1376 --exclude_dir in the tests to avoid scanning .svn directories.
1377
1378 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1379 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1380 doesn't support NULs in patterns.
1381
1382 7. Added some missing "const"s to declarations of static tables in
1383 pcre_compile.c and pcre_dfa_exec.c.
1384
1385 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1386 caused by fix #2 above. (Subsequently also a second patch to fix the
1387 first patch. And a third patch - this was a messy problem.)
1388
1389 9. Applied Craig's patch to remove the use of push_back().
1390
1391 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1392 matching function regexec().
1393
1394 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1395 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1396 references. PCRE supports relative numbers with this syntax (I don't think
1397 Oniguruma does).
1398
1399 12. Previously, a group with a zero repeat such as (...){0} was completely
1400 omitted from the compiled regex. However, this means that if the group
1401 was called as a subroutine from elsewhere in the pattern, things went wrong
1402 (an internal error was given). Such groups are now left in the compiled
1403 pattern, with a new opcode that causes them to be skipped at execution
1404 time.
1405
1406 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1407 to the way PCRE behaves:
1408
1409 (a) A lone ] character is dis-allowed (Perl treats it as data).
1410
1411 (b) A back reference to an unmatched subpattern matches an empty string
1412 (Perl fails the current match path).
1413
1414 (c) A data ] in a character class must be notated as \] because if the
1415 first data character in a class is ], it defines an empty class. (In
1416 Perl it is not possible to have an empty class.) The empty class []
1417 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1418 The negative empty class [^] matches any one character, independently
1419 of the DOTALL setting.
1420
1421 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1422 non-existent subpattern following a character class starting with ']' and
1423 containing () gave an internal compiling error instead of "reference to
1424 non-existent subpattern". Fortunately, when the pattern did exist, the
1425 compiled code was correct. (When scanning forwards to check for the
1426 existencd of the subpattern, it was treating the data ']' as terminating
1427 the class, so got the count wrong. When actually compiling, the reference
1428 was subsequently set up correctly.)
1429
1430 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1431 it was being rejected as not supported by pcre_dfa_exec(), even though
1432 other assertions are supported. I have made pcre_dfa_exec() support
1433 (*FAIL).
1434
1435 16. The implementation of 13c above involved the invention of a new opcode,
1436 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1437 cannot be changed at match time, I realized I could make a small
1438 improvement to matching performance by compiling OP_ALLANY instead of
1439 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1440 on the OP_ANY path.
1441
1442 17. Compiling pcretest on Windows with readline support failed without the
1443 following two fixes: (1) Make the unistd.h include conditional on
1444 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1445
1446 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1447 ncurses library to be included for pcretest when ReadLine support is
1448 requested, but also to allow for it to be overridden. This patch came from
1449 Daniel Bergström.
1450
1451 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1452 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1453 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1454 spotting this.
1455
1456
1457 Version 7.6 28-Jan-08
1458 ---------------------
1459
1460 1. A character class containing a very large number of characters with
1461 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1462 overflow.
1463
1464 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1465 HAVE_LONG_LONG is not defined.
1466
1467 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1468 bring them up to date and include new features. This patch includes:
1469
1470 - Fixed PH's badly added libz and libbz2 support.
1471 - Fixed a problem with static linking.
1472 - Added pcredemo. [But later removed - see 7 below.]
1473 - Fixed dftables problem and added an option.
1474 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1475 HAVE_LONG_LONG.
1476 - Added readline support for pcretest.
1477 - Added an listing of the option settings after cmake has run.
1478
1479 4. A user submitted a patch to Makefile that makes it easy to create
1480 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1481 Makefile.am that cause it to include this special target, without
1482 affecting anything else. Note that the same mingw target plus all
1483 the other distribution libraries and programs are now supported
1484 when configuring with CMake (see 6 below) instead of with
1485 Configure/Make.
1486
1487 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1488 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1489 exported in the Windows port". It has not yet been confirmed that the patch
1490 solves the problem, but it does no harm.
1491
1492 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1493 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1494 with CMake, and also correct the comment about stack recursion.
1495
1496 7. Remove the automatic building of pcredemo from the ./configure system and
1497 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1498 of a program that users should build themselves after PCRE is installed, so
1499 building it automatically is not really right. What is more, it gave
1500 trouble in some build environments.
1501
1502 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1503
1504
1505 Version 7.5 10-Jan-08
1506 ---------------------
1507
1508 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1509 values in parens when parsing an RE using the C++ wrapper."
1510
1511 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1512 Characters greater than 255 were excluded from the class instead of being
1513 included.
1514
1515 3. The same bug as (2) above applied to negated POSIX classes such as
1516 [:^space:].
1517
1518 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1519 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1520 I have changed it.
1521
1522 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1523 first named subpattern) and a construct such as (?&a) would reference the
1524 first named subpattern whose name started with "a" (in other words, the
1525 length check was missing). Both these problems are fixed. "Subpattern name
1526 expected" is now given for (?&) (a zero-length name), and this patch also
1527 makes it give the same error for \k'' (previously it complained that that
1528 was a reference to a non-existent subpattern).
1529
1530 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1531 this is right because (?- can be followed by option settings as well as by
1532 digits. I have, however, made the messages clearer.
1533
1534 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1535 than the number used in the conditional) now cause a compile-time error.
1536 This is actually not compatible with Perl, which accepts such patterns, but
1537 treats the conditional as always being FALSE (as PCRE used to), but it
1538 seems to me that giving a diagnostic is better.
1539
1540 8. Change "alphameric" to the more common word "alphanumeric" in comments
1541 and messages.
1542
1543 9. Fix two occurrences of "backslash" in comments that should have been
1544 "backspace".
1545
1546 10. Remove two redundant lines of code that can never be obeyed (their function
1547 was moved elsewhere).
1548
1549 11. The program that makes PCRE's Unicode character property table had a bug
1550 which caused it to generate incorrect table entries for sequences of
1551 characters that have the same character type, but are in different scripts.
1552 It amalgamated them into a single range, with the script of the first of
1553 them. In other words, some characters were in the wrong script. There were
1554 thirteen such cases, affecting characters in the following ranges:
1555
1556 U+002b0 - U+002c1
1557 U+0060c - U+0060d
1558 U+0061e - U+00612
1559 U+0064b - U+0065e
1560 U+0074d - U+0076d
1561 U+01800 - U+01805
1562 U+01d00 - U+01d77
1563 U+01d9b - U+01dbf
1564 U+0200b - U+0200f
1565 U+030fc - U+030fe
1566 U+03260 - U+0327f
1567 U+0fb46 - U+0fbb1
1568 U+10450 - U+1049d
1569
1570 12. The -o option (show only the matching part of a line) for pcregrep was not
1571 compatible with GNU grep in that, if there was more than one match in a
1572 line, it showed only the first of them. It now behaves in the same way as
1573 GNU grep.
1574
1575 13. If the -o and -v options were combined for pcregrep, it printed a blank
1576 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1577 does the same. The return code can be used to tell if there were any
1578 non-matching lines.
1579
1580 14. Added --file-offsets and --line-offsets to pcregrep.
1581
1582 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1583 infinitely looping recursion. The bug was that positive lookaheads were not
1584 being skipped when checking for a possible empty match (negative lookaheads
1585 and both kinds of lookbehind were skipped).
1586
1587 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1588 inclusion of <windows.h> to before rather than after the definition of
1589 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1590
1591 17. Specifying a possessive quantifier with a specific limit for a Unicode
1592 character property caused pcre_compile() to compile bad code, which led at
1593 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1594 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1595 caused the error; without that there was no problem.
1596
1597 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1598
1599 19. Added --enable-pcretest-libreadline.
1600
1601 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1602 RE::GlobalReplace(). As a result, the number of replacements returned was
1603 double what it should be. I removed one of the increments, but Craig sent a
1604 later patch that removed the other one (the right fix) and added unit tests
1605 that check the return values (which was not done before).
1606
1607 21. Several CMake things:
1608
1609 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1610 the names libpcre and libpcreposix, not just pcre and pcreposix.
1611
1612 (2) The above change means that pcretest and pcregrep are now correctly
1613 linked with the newly-built libraries, not previously installed ones.
1614
1615 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1616
1617 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1618 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1619 UTF-8 newline character). The key issue is that the pattern starts .*;
1620 this means that the match must be either at the beginning, or after a
1621 newline. The bug was in the code for advancing after a failed match and
1622 checking that the new position followed a newline. It was not taking
1623 account of UTF-8 characters correctly.
1624
1625 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1626 character classes. PCRE was not treating the sequence [:...:] as a
1627 character class unless the ... were all letters. Perl, however, seems to
1628 allow any characters between [: and :], though of course it rejects as
1629 unknown any "names" that contain non-letters, because all the known class
1630 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1631 for example, whereas PCRE did not - it did not recognize a POSIX character
1632 class. This seemed a bit dangerous, so the code has been changed to be
1633 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1634 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1635 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1636 Perl does, and where it didn't before.
1637
1638 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1639 Windows environments %n is disabled by default.
1640
1641
1642 Version 7.4 21-Sep-07
1643 ---------------------
1644
1645 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1646 means that a class such as [\s] counted as "explicit reference to CR or
1647 LF". That isn't really right - the whole point of the change was to try to
1648 help when there was an actual mention of one of the two characters. So now
1649 the change happens only if \r or \n (or a literal CR or LF) character is
1650 encountered.
1651
1652 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1653 of both had grown to the point where there were only 3 bits left.
1654 Fortunately, there was spare space in the data structure, and so I have
1655 moved the internal flags into a new 16-bit field to free up more option
1656 bits.
1657
1658 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1659 but did not set the internal JCHANGED flag - either of these is enough to
1660 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1661 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1662 start sets both bits.
1663
1664 4. Added options (at build time, compile time, exec time) to change \R from
1665 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1666
1667 5. doc/pcresyntax.html was missing from the distribution.
1668
1669 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1670 compatibility, even though it is no longer used.
1671
1672 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1673 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1674 windows.h file is present (where different names are used). [This was
1675 reversed later after testing - see 16 below.]
1676
1677 8. Changed all #include <config.h> to #include "config.h". There were also
1678 some further <pcre.h> cases that I changed to "pcre.h".
1679
1680 9. When pcregrep was used with the --colour option, it missed the line ending
1681 sequence off the lines that it output.
1682
1683 10. It was pointed out to me that arrays of string pointers cause lots of
1684 relocations when a shared library is dynamically loaded. A technique of
1685 using a single long string with a table of offsets can drastically reduce
1686 these. I have refactored PCRE in four places to do this. The result is
1687 dramatic:
1688
1689 Originally: 290
1690 After changing UCP table: 187
1691 After changing error message table: 43
1692 After changing table of "verbs" 36
1693 After changing table of Posix names 22
1694
1695 Thanks to the folks working on Gregex for glib for this insight.
1696
1697 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1698 unicode-properties was also set.
1699
1700 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1701
1702 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1703 checked only for CRLF.
1704
1705 14. Added casts to pcretest.c to avoid compiler warnings.
1706
1707 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1708
1709 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1710 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1711 entirely. This removes changes made in 7 above.
1712
1713 17. The CMake files have been updated, and there is now more information about
1714 building with CMake in the NON-UNIX-USE document.
1715
1716
1717 Version 7.3 28-Aug-07
1718 ---------------------
1719
1720 1. In the rejigging of the build system that eventually resulted in 7.1, the
1721 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1722 brackets there is not right, since it causes compilers to look for an
1723 installed pcre.h, not the version that is in the source that is being
1724 compiled (which of course may be different). I have changed it back to:
1725
1726 #include "pcre.h"
1727
1728 I have a vague recollection that the change was concerned with compiling in
1729 different directories, but in the new build system, that is taken care of
1730 by the VPATH setting the Makefile.
1731
1732 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1733 when the subject happened to end in the byte 0x85 (e.g. if the last
1734 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1735 characters but of course it shouldn't be taken as a newline when it is part
1736 of another character. The bug was that, for an unlimited repeat of . in
1737 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1738 characters when looking for a newline.
1739
1740 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1741
1742 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1743 in debug output.
1744
1745 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1746 long printing in the pcrecpp unittest when running under MinGW.
1747
1748 6. ESC_K was left out of the EBCDIC table.
1749
1750 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1751 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1752 limit also applies to "virtual nesting" when a pattern is recursive, and in
1753 this case 1000 isn't so big. I have been able to remove this limit at the
1754 expense of backing off one optimization in certain circumstances. Normally,
1755 when pcre_exec() would call its internal match() function recursively and
1756 immediately return the result unconditionally, it uses a "tail recursion"
1757 feature to save stack. However, when a subpattern that can match an empty
1758 string has an unlimited repetition quantifier, it no longer makes this
1759 optimization. That gives it a stack frame in which to save the data for
1760 checking that an empty string has been matched. Previously this was taken
1761 from the 1000-entry workspace that had been reserved. So now there is no
1762 explicit limit, but more stack is used.
1763
1764 8. Applied Daniel's patches to solve problems with the import/export magic
1765 syntax that is required for Windows, and which was going wrong for the
1766 pcreposix and pcrecpp parts of the library. These were overlooked when this
1767 problem was solved for the main library.
1768
1769 9. There were some crude static tests to avoid integer overflow when computing
1770 the size of patterns that contain repeated groups with explicit upper
1771 limits. As the maximum quantifier is 65535, the maximum group length was
1772 set at 30,000 so that the product of these two numbers did not overflow a
1773 32-bit integer. However, it turns out that people want to use groups that
1774 are longer than 30,000 bytes (though not repeat them that many times).
1775 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1776 made it possible to implement the integer overflow checks in a much more
1777 dynamic way, which I have now done. The artificial limitation on group
1778 length has been removed - we now have only the limit on the total length of
1779 the compiled pattern, which depends on the LINK_SIZE setting.
1780
1781 10. Fixed a bug in the documentation for get/copy named substring when
1782 duplicate names are permitted. If none of the named substrings are set, the
1783 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1784 empty string.
1785
1786 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1787 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1788 because the ] is interpreted as the first data character and the
1789 terminating ] is not found. PCRE has been made compatible with Perl in this
1790 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1791 cause memory overwriting.
1792
1793 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1794 string has been matched (to stop an infinite loop). It was not recognizing
1795 a conditional subpattern that could match an empty string if that
1796 subpattern was within another subpattern. For example, it looped when
1797 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1798 condition was not nested. This bug has been fixed.
1799
1800 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1801 past the start of the subject in the presence of bytes with the top bit
1802 set, for example "\x8aBCD".
1803
1804 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1805 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1806
1807 14. Optimized (?!) to (*FAIL).
1808
1809 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1810 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1811 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1812 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1813 does: it's just the validity check that is more restrictive.
1814
1815 16. Inserted checks for integer overflows during escape sequence (backslash)
1816 processing, and also fixed erroneous offset values for syntax errors during
1817 backslash processing.
1818
1819 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1820 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1821
1822 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1823 caused an overrun.
1824
1825 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1826 something other than just ASCII characters) inside a group that had an
1827 unlimited repeat caused a loop at compile time (while checking to see
1828 whether the group could match an empty string).
1829
1830 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1831 [\P{Any}] did so. (Error in the code for printing property names.)
1832
1833 21. An orphan \E inside a character class could cause a crash.
1834
1835 22. A repeated capturing bracket such as (A)? could cause a wild memory
1836 reference during compilation.
1837
1838 23. There are several functions in pcre_compile() that scan along a compiled
1839 expression for various reasons (e.g. to see if it's fixed length for look
1840 behind). There were bugs in these functions when a repeated \p or \P was
1841 present in the pattern. These operators have additional parameters compared
1842 with \d, etc, and these were not being taken into account when moving along
1843 the compiled data. Specifically:
1844
1845 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1846 length.
1847
1848 (b) An item such as \pL+ within a repeated group could cause crashes or
1849 loops.
1850
1851 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1852 "reference to non-existent subpattern" error.
1853
1854 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1855
1856 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1857 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1858
1859 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1860
1861 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1862 character were causing crashes (broken optimization).
1863
1864 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1865 \p or \P) caused a compile-time loop.
1866
1867 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1868 break. For example, the unstudied pattern [\r\n]A does not match the string
1869 "\r\nA" because change 7.0/46 below moves the current point on by two
1870 characters after failing to match at the start. However, the pattern \nA
1871 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1872 the same is true. There doesn't seem any very clean way out of this, but
1873 what I have chosen to do makes the common cases work: PCRE now takes note
1874 of whether there can be an explicit match for \r or \n anywhere in the
1875 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1876 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1877 pattern has explicit CR or LF references.
1878
1879 29. Added (*CR) etc for changing newline setting at start of pattern.
1880
1881
1882 Version 7.2 19-Jun-07
1883 ---------------------
1884
1885 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1886 which is apparently normally available under Windows.
1887
1888 2. Re-jig the pcregrep tests with different newline settings in an attempt
1889 to make them independent of the local environment's newline setting.
1890
1891 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1892
1893 4. Some of the "internals" tests were previously cut out when the link size
1894 was not 2, because the output contained actual offsets. The recent new
1895 "Z" feature of pcretest means that these can be cut out, making the tests
1896 usable with all link sizes.
1897
1898 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1899 stack recursion. This gives a massive performance boost under BSD, but just
1900 a small improvement under Linux. However, it saves one field in the frame
1901 in all cases.
1902
1903 6. Added more features from the forthcoming Perl 5.10:
1904
1905 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1906 recursion call. It refers to the nth most recently opened parentheses.
1907
1908 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1909 to be opened parentheses.
1910
1911 (c) Conditions that refer to capturing parentheses can be specified
1912 relatively, for example, (?(-2)... or (?(+3)...
1913
1914 (d) \K resets the start of the current match so that everything before
1915 is not part of it.
1916
1917 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1918
1919 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1920 reference syntax.
1921
1922 (g) (?| introduces a group in which the numbering of parentheses in each
1923 alternative starts with the same number.
1924
1925 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1926
1927 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1928 PCRE_INFO_JCHANGED.
1929
1930 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1931 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1932 for detecting groups that can match an empty string.
1933
1934 9. A pattern with a very large number of alternatives (more than several
1935 hundred) was running out of internal workspace during the pre-compile
1936 phase, where pcre_compile() figures out how much memory will be needed. A
1937 bit of new cunning has reduced the workspace needed for groups with
1938 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1939 workspace instead of running out of the 4096 that are available.
1940
1941 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1942
1943 11. Applied patch from Google to remove an optimization that didn't quite work.
1944 The report of the bug said:
1945
1946 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1947 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1948 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1949
1950 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1951 it matched the wrong number of bytes.
1952
1953
1954 Version 7.1 24-Apr-07
1955 ---------------------
1956
1957 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1958 that is more "standard", making use of automake and other Autotools. There
1959 is some re-arrangement of the files and adjustment of comments consequent
1960 on this.
1961
1962 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1963 for recursive directory scanning broke on some systems because the files
1964 are not scanned in any specific order and on different systems the order
1965 was different. A call to "sort" has been inserted into RunGrepTest for the
1966 approprate test as a short-term fix. In the longer term there may be an
1967 alternative.
1968
1969 3. I had an email from Eric Raymond about problems translating some of PCRE's
1970 man pages to HTML (despite the fact that I distribute HTML pages, some
1971 people do their own conversions for various reasons). The problems
1972 concerned the use of low-level troff macros .br and .in. I have therefore
1973 removed all such uses from the man pages (some were redundant, some could
1974 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1975 HTML has been updated to handle .nf/.fi and to complain if it encounters
1976 .br or .in.
1977
1978 4. Updated comments in configure.ac that get placed in config.h.in and also
1979 arranged for config.h to be included in the distribution, with the name
1980 config.h.generic, for the benefit of those who have to compile without
1981 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1982
1983 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1984 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1985 makevp.bat for the current PCRE, using the additional files
1986 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1987
1988 6. A Windows user reported a minor discrepancy with test 2, which turned out
1989 to be caused by a trailing space on an input line that had got lost in his
1990 copy. The trailing space was an accident, so I've just removed it.
1991
1992 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1993 that is needed.
1994
1995 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1996 as "const" (a) because they are and (b) because it helps the PHP
1997 maintainers who have recently made a script to detect big data structures
1998 in the php code that should be moved to the .rodata section. I remembered
1999 to update Builducptable as well, so it won't revert if ucptable.h is ever
2000 re-created.
2001
2002 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2003 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2004 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2005 support is not required. This saves 1.5-2K of code, which is important in
2006 some applications.
2007
2008 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2009 so as not to refer to the tables, even though these functions will never be
2010 called when UTF-8 support is disabled. Otherwise there are problems with a
2011 shared library.
2012
2013 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2014
2015 (a) It was defining its arguments as char * instead of void *.
2016
2017 (b) It was assuming that all moves were upwards in memory; this was true
2018 a long time ago when I wrote it, but is no longer the case.
2019
2020 The emulated memove() is provided for those environments that have neither
2021 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2022 is clearly not the case, as these two bugs were recently reported.
2023
2024 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2025 and Detrail to create the HTML documentation, the .txt form of the man
2026 pages, and it removes trailing spaces from listed files. It also creates
2027 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2028 case, it wraps all the #defines with #ifndefs. This script should be run
2029 before "make dist".
2030
2031 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2032 with Unicode property support.
2033
2034 (a) For a maximizing quantifier, if the two different cases of the
2035 character were of different lengths in their UTF-8 codings (there are
2036 some cases like this - I found 11), and the matching function had to
2037 back up over a mixture of the two cases, it incorrectly assumed they
2038 were both the same length.
2039
2040 (b) When PCRE was configured to use the heap rather than the stack for
2041 recursion during matching, it was not correctly preserving the data for
2042 the other case of a UTF-8 character when checking ahead for a match
2043 while processing a minimizing repeat. If the check also involved
2044 matching a wide character, but failed, corruption could cause an
2045 erroneous result when trying to check for a repeat of the original
2046 character.
2047
2048 13. Some tidying changes to the testing mechanism:
2049
2050 (a) The RunTest script now detects the internal link size and whether there
2051 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2052 values substituted by "configure". (The RunGrepTest script already did
2053 this for UTF-8.) The configure.ac script no longer substitutes the
2054 relevant variables.
2055
2056 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2057 with length and offset values. This means that the output is different
2058 for different internal link sizes. Test 2 is skipped for link sizes
2059 other than 2 because of this, bypassing the problem. Unfortunately,
2060 there was also a test in test 3 (the locale tests) that used /B and
2061 failed for link sizes other than 2. Rather than cut the whole test out,
2062 I have added a new /Z option to pcretest that replaces the length and
2063 offset values with spaces. This is now used to make test 3 independent
2064 of link size. (Test 2 will be tidied up later.)
2065
2066 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2067 segmentation fault instead of returning the appropriate error message.
2068
2069 15. In multiline mode when the newline sequence was set to "any", the pattern
2070 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2071 This doesn't seem right; it now treats the CRLF combination as the line
2072 ending, and so does not match in that case. It's only a pattern such as ^$
2073 that would hit this one: something like ^ABC$ would have failed after \r
2074 and then tried again after \r\n.
2075
2076 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2077 in an attempt to make files that differ only in their line terminators
2078 compare equal. This works on Linux.
2079
2080 17. Under certain error circumstances pcregrep might try to free random memory
2081 as it exited. This is now fixed, thanks to valgrind.
2082
2083 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2084 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2085 was because its rules for how to advance for /g after matching an empty
2086 string at the end of a line did not allow for this case. They now check for
2087 it specially.
2088
2089 20. pcretest is supposed to handle patterns and data of any length, by
2090 extending its buffers when necessary. It was getting this wrong when the
2091 buffer for a data line had to be extended.
2092
2093 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2094 CRLF as a newline sequence.
2095
2096 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2097 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2098 I have nevertheless tidied it up.
2099
2100 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2101
2102 24. Added a man page for pcre-config.
2103
2104
2105 Version 7.0 19-Dec-06
2106 ---------------------
2107
2108 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2109 moving to gcc 4.1.1.
2110
2111 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2112 sys/time.h, which is documented as needed for this function. It doesn't
2113 seem to matter on Linux, but it showed up on some releases of OS X.
2114
2115 3. It seems that there are systems where bytes whose values are greater than
2116 127 match isprint() in the "C" locale. The "C" locale should be the
2117 default when a C program starts up. In most systems, only ASCII printing
2118 characters match isprint(). This difference caused the output from pcretest
2119 to vary, making some of the tests fail. I have changed pcretest so that:
2120
2121 (a) When it is outputting text in the compiled version of a pattern, bytes
2122 other than 32-126 are always shown as hex escapes.
2123
2124 (b) When it is outputting text that is a matched part of a subject string,
2125 it does the same, unless a different locale has been set for the match
2126 (using the /L modifier). In this case, it uses isprint() to decide.
2127
2128 4. Fixed a major bug that caused incorrect computation of the amount of memory
2129 required for a compiled pattern when options that changed within the
2130 pattern affected the logic of the preliminary scan that determines the
2131 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2132 that the computed length was too small. The symptoms of this bug were
2133 either the PCRE error "internal error: code overflow" from pcre_compile(),
2134 or a glibc crash with a message such as "pcretest: free(): invalid next
2135 size (fast)". Examples of patterns that provoked this bug (shown in
2136 pcretest format) are:
2137
2138 /(?-x: )/x
2139 /(?x)(?-x: \s*#\s*)/
2140 /((?i)[\x{c0}])/8
2141 /(?i:[\x{c0}])/8
2142
2143 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2144 is now done differently.
2145
2146 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2147 wrapper classes; (b) implement a new function in the C++ scanner that is
2148 more efficient than the old way of doing things because it avoids levels of
2149 recursion in the regex matching; (c) add a paragraph to the documentation
2150 for the FullMatch() function.
2151
2152 6. The escape sequence \n was being treated as whatever was defined as
2153 "newline". Not only was this contrary to the documentation, which states
2154 that \n is character 10 (hex 0A), but it also went horribly wrong when
2155 "newline" was defined as CRLF. This has been fixed.
2156
2157 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2158 was being set to -1 for the "end of line" case (supposedly a value that no
2159 character can have). Though this value is never used (the check for end of
2160 line is "zero bytes in current character"), it caused compiler complaints.
2161 I've changed it to 0xffffffff.
2162
2163 8. In pcre_version.c, the version string was being built by a sequence of
2164 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2165 string (as it is for production releases) called a macro with an empty
2166 argument. The C standard says the result of this is undefined. The gcc
2167 compiler treats it as an empty string (which was what was wanted) but it is
2168 reported that Visual C gives an error. The source has been hacked around to
2169 avoid this problem.
2170
2171 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2172 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2173 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2174 of them did).
2175
2176 10. Originally, pcretest opened its input and output without "b"; then I was
2177 told that "b" was needed in some environments, so it was added for release
2178 5.0 to both the input and output. (It makes no difference on Unix-like
2179 systems.) Later I was told that it is wrong for the input on Windows. I've
2180 now abstracted the modes into two macros, to make it easier to fiddle with
2181 them, and removed "b" from the input mode under Windows.
2182
2183 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2184
2185 12. Added -help and --help to pcretest as an official way of being reminded
2186 of the options.
2187
2188 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2189 and pcrecpp.cc because they annoy compilers at high warning levels.
2190
2191 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2192
2193 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2194 scripts are not C programs :-) and which was not noticed because it works
2195 on Linux.
2196
2197 16. pcretest is supposed to handle any length of pattern and data line (as one
2198 line or as a continued sequence of lines) by extending its input buffer if
2199 necessary. This feature was broken for very long pattern lines, leading to
2200 a string of junk being passed to pcre_compile() if the pattern was longer
2201 than about 50K.
2202
2203 17. I have done a major re-factoring of the way pcre_compile() computes the
2204 amount of memory needed for a compiled pattern. Previously, there was code
2205 that made a preliminary scan of the pattern in order to do this. That was
2206 OK when PCRE was new, but as the facilities have expanded, it has become
2207 harder and harder to keep it in step with the real compile phase, and there
2208 have been a number of bugs (see for example, 4 above). I have now found a
2209 cunning way of running the real compile function in a "fake" mode that
2210 enables it to compute how much memory it would need, while actually only
2211 ever using a few hundred bytes of working memory and without too many
2212 tests of the mode. This should make future maintenance and development
2213 easier. A side effect of this work is that the limit of 200 on the nesting
2214 depth of parentheses has been removed (though this was never a serious
2215 limitation, I suspect). However, there is a downside: pcre_compile() now
2216 runs more slowly than before (30% or more, depending on the pattern). I
2217 hope this isn't a big issue. There is no effect on runtime performance.
2218
2219 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2220 newline (only possible for the last line of a file) and it was a
2221 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2222
2223 19. Added additional timing features to pcretest. (1) The -tm option now times
2224 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2225 separate command line item, by a number that specifies the number of
2226 repeats to use when timing. The default is 50000; this gives better
2227 precision, but takes uncomfortably long for very large patterns.
2228
2229 20. Extended pcre_study() to be more clever in cases where a branch of a
2230 subpattern has no definite first character. For example, (a*|b*)[cd] would
2231 previously give no result from pcre_study(). Now it recognizes that the
2232 first character must be a, b, c, or d.
2233
2234 21. There was an incorrect error "recursive call could loop indefinitely" if
2235 a subpattern (or the entire pattern) that was being tested for matching an
2236 empty string contained only one non-empty item after a nested subpattern.
2237 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2238 incorrectly, because the \d was being skipped in the check.
2239
2240 22. The pcretest program now has a new pattern option /B and a command line
2241 option -b, which is equivalent to adding /B to every pattern. This causes
2242 it to show the compiled bytecode, without the additional information that
2243 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2244 is the same as /B/I).
2245
2246 23. A new optimization is now able automatically to treat some sequences such
2247 as a*b as a*+b. More specifically, if something simple (such as a character
2248 or a simple class like \d) has an unlimited quantifier, and is followed by
2249 something that cannot possibly match the quantified thing, the quantifier
2250 is automatically "possessified".
2251
2252 24. A recursive reference to a subpattern whose number was greater than 39
2253 went wrong under certain circumstances in UTF-8 mode. This bug could also
2254 have affected the operation of pcre_study().
2255
2256 25. Realized that a little bit of performance could be had by replacing
2257 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2258
2259 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2260
2261 27. Possessive quantifiers such as a++ were previously implemented by turning
2262 them into atomic groups such as ($>a+). Now they have their own opcodes,
2263 which improves performance. This includes the automatically created ones
2264 from 23 above.
2265
2266 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2267 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2268 the first matched character to be a colon. This applied both to named and
2269 numbered groups.
2270
2271 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2272
2273 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2274 building PCRE on Windows easier, so I have included it in the distribution.
2275
2276 31. There is now a check in pcretest against a ridiculously large number being
2277 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2278 loop, the loop is abandoned.
2279
2280 32. Forward references to subpatterns in conditions such as (?(2)...) where
2281 subpattern 2 is defined later cause pcre_compile() to search forwards in
2282 the pattern for the relevant set of parentheses. This search went wrong
2283 when there were unescaped parentheses in a character class, parentheses
2284 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2285
2286 33. "Subroutine" calls and backreferences were previously restricted to
2287 referencing subpatterns earlier in the regex. This restriction has now
2288 been removed.
2289
2290 34. Added a number of extra features that are going to be in Perl 5.10. On the
2291 whole, these are just syntactic alternatives for features that PCRE had
2292 previously implemented using the Python syntax or my own invention. The
2293 other formats are all retained for compatibility.
2294
2295 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2296 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2297 also .NET compatible.
2298
2299 (b) A recursion or subroutine call to a named group can now be defined as
2300 (?&name) as well as (?P>name).
2301
2302 (c) A backreference to a named group can now be defined as \k<name> or
2303 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2304 5.10, are also .NET compatible.
2305
2306 (d) A conditional reference to a named group can now use the syntax
2307 (?(<name>) or (?('name') as well as (?(name).
2308
2309 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2310 groups (named and numbered) that are never evaluated inline, but can be
2311 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2312 is always false. There may be only one alternative in such a group.
2313
2314 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2315 as the simple (?(R). The condition is true only if the most recent
2316 recursion is that of the given number or name. It does not search out
2317 through the entire recursion stack.
2318
2319 (g) The escape \gN or \g{N} has been added, where N is a positive or
2320 negative number, specifying an absolute or relative reference.
2321
2322 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2323 some "unreachable code" warnings.
2324
2325 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2326 things, this adds five new scripts.
2327
2328 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2329 There were also incompatibilities regarding the handling of \Q..\E inside
2330 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2331 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2332
2333 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2334 matches an empty string, and forcibly breaks the loop. There were bugs in
2335 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2336 against aaaa the result was just "a" rather than "aaaa", for example. Two
2337 separate and independent bugs (that affected different cases) have been
2338 fixed.
2339
2340 39. Refactored the code to abolish the use of different opcodes for small
2341 capturing bracket numbers. This is a tidy that I avoided doing when I
2342 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2343 The new approach is not only tidier, it makes it possible to reduce the
2344 memory needed to fix the previous bug (38).
2345
2346 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2347 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2348 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2349 mode.
2350
2351 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2352 report.
2353
2354 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2355 copy construction and assignment in the C++ wrapper.
2356
2357 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2358 couple of bugs that could have given wrong results in the "--newline=crlf"
2359 case.
2360
2361 44. Added a number of casts and did some reorganization of signed/unsigned int
2362 variables following suggestions from Dair Grant. Also renamed the variable
2363 "this" as "item" because it is a C++ keyword.
2364
2365 45. Arranged for dftables to add
2366
2367 #include "pcre_internal.h"
2368
2369 to pcre_chartables.c because without it, gcc 4.x may remove the array
2370 definition from the final binary if PCRE is built into a static library and
2371 dead code stripping is activated.
2372
2373 46. For an unanchored pattern, if a match attempt fails at the start of a
2374 newline sequence, and the newline setting is CRLF or ANY, and the next two
2375 characters are CRLF, advance by two characters instead of one.
2376
2377
2378 Version 6.7 04-Jul-06
2379 ---------------------
2380
2381 1. In order to handle tests when input lines are enormously long, pcretest has
2382 been re-factored so that it automatically extends its buffers when
2383 necessary. The code is crude, but this _is_ just a test program. The
2384 default size has been increased from 32K to 50K.
2385
2386 2. The code in pcre_study() was using the value of the re argument before
2387 testing it for NULL. (Of course, in any sensible call of the function, it
2388 won't be NULL.)
2389
2390 3. The memmove() emulation function in pcre_internal.h, which is used on
2391 systems that lack both memmove() and bcopy() - that is, hardly ever -
2392 was missing a "static" storage class specifier.
2393
2394 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2395 containing an extended class (one that cannot be represented by a bitmap
2396 because it contains high-valued characters or Unicode property items, e.g.
2397 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2398 pattern, but PCRE should not loop if you do not (it no longer does).
2399 [Detail: two cases were found: (a) a repeated subpattern containing an
2400 extended class; (b) a recursive reference to a subpattern that followed a
2401 previous extended class. It wasn't skipping over the extended class
2402 correctly when UTF-8 mode was not set.]
2403
2404 5. A negated single-character class was not being recognized as fixed-length
2405 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2406 compile error "lookbehind assertion is not fixed length".
2407
2408 6. The RunPerlTest auxiliary script was showing an unexpected difference
2409 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2410 write a Perl script that can interpret lines of an input file either as
2411 byte characters or as UTF-8, which is what "perltest" was being required to
2412 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2413 can't do is switch easily at run time between having the "use utf8;" pragma
2414 or not. In the end, I fudged it by using the RunPerlTest script to insert
2415 "use utf8;" explicitly for the UTF-8 tests.
2416
2417 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2418 the end of the subject string, contrary to the documentation and to what
2419 Perl does. This was true of both matching functions. Now it matches only at
2420 the start of the subject and immediately after *internal* newlines.
2421
2422 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2423 a pointer to an int instead of a pointer to an unsigned long int. This
2424 caused problems on 64-bit systems.
2425
2426 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2427 instance of the 'standard' template library not being so standard".
2428
2429 10. There was no check on the number of named subpatterns nor the maximum
2430 length of a subpattern name. The product of these values is used to compute
2431 the size of the memory block for a compiled pattern. By supplying a very
2432 long subpattern name and a large number of named subpatterns, the size
2433 computation could be caused to overflow. This is now prevented by limiting
2434 the length of names to 32 characters, and the number of named subpatterns
2435 to 10,000.
2436
2437 11. Subpatterns that are repeated with specific counts have to be replicated in
2438 the compiled pattern. The size of memory for this was computed from the
2439 length of the subpattern and the repeat count. The latter is limited to
2440 65535, but there was no limit on the former, meaning that integer overflow
2441 could in principle occur. The compiled length of a repeated subpattern is
2442 now limited to 30,000 bytes in order to prevent this.
2443
2444 12. Added the optional facility to have named substrings with the same name.
2445
2446 13. Added the ability to use a named substring as a condition, using the
2447 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2448 are numbers (not recommended). Forward references are permitted.
2449
2450 14. Added forward references in named backreferences (if you see what I mean).
2451
2452 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2453 pattern could run off the end of the subject. For example, the pattern
2454 "(?s)(.{1,5})"8 did this with the subject "ab".
2455
2456 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2457 PCRE_CASELESS was set when matching characters that were quantified with ?
2458 or *.
2459
2460 17. A character class other than a single negated character that had a minimum
2461 but no maximum quantifier - for example [ab]{6,} - was not handled
2462 correctly by pce_dfa_exec(). It would match only one character.
2463
2464 18. A valid (though odd) pattern that looked like a POSIX character
2465 class but used an invalid character after [ (for example [[,abc,]]) caused
2466 pcre_compile() to give the error "Failed: internal error: code overflow" or
2467 in some cases to crash with a glibc free() error. This could even happen if
2468 the pattern terminated after [[ but there just happened to be a sequence of
2469 letters, a binary zero, and a closing ] in the memory that followed.
2470
2471 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2472 over the years. Originally (before any Unicode support), just the bottom 8
2473 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2474 output from "man perlunicode" includes this:
2475
2476 The regular expression compiler produces polymorphic opcodes. That
2477 is, the pattern adapts to the data and automatically switches to
2478 the Unicode character scheme when presented with Unicode data--or
2479 instead uses a traditional byte scheme when presented with byte
2480 data.
2481
2482 Sadly, a wide octal escape does not cause a switch, and in a string with
2483 no other multibyte characters, these octal escapes are treated as before.
2484 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2485 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2486 Unicode string.
2487
2488 I have not perpetrated such confusion in PCRE. Up till now, it took just
2489 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2490 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2491 translate to the appropriate multibyte character.
2492
2493 29. Applied some refactoring to reduce the number of warnings from Microsoft
2494 and Borland compilers. This has included removing the fudge introduced
2495 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2496 a warning about an unused variable.
2497
2498 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2499 characters since release 4.0, because Perl (from release 5.004) does not.
2500 [Or at least, is documented not to: some releases seem to be in conflict
2501 with the documentation.] However, when a pattern was studied with
2502 pcre_study() and all its branches started with \s, PCRE still included VT
2503 as a possible starting character. Of course, this did no harm; it just
2504 caused an unnecessary match attempt.
2505
2506 22. Removed a now-redundant internal flag bit that recorded the fact that case
2507 dependency changed within the pattern. This was once needed for "required
2508 byte" processing, but is no longer used. This recovers a now-scarce options
2509 bit. Also moved the least significant internal flag bit to the most-
2510 significant bit of the word, which was not previously used (hangover from
2511 the days when it was an int rather than a uint) to free up another bit for
2512 the future.
2513
2514 23. Added support for CRLF line endings as well as CR and LF. As well as the
2515 default being selectable at build time, it can now be changed at runtime
2516 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2517 specify that it is scanning data with non-default line endings.
2518
2519 24. Changed the definition of CXXLINK to make it agree with the definition of
2520 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2521
2522 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2523 recursions. This makes a big different to stack usage for some patterns.
2524
2525 26. If a subpattern containing a named recursion or subroutine reference such
2526 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2527 the space required for the compiled pattern went wrong and gave too small a
2528 value. Depending on the environment, this could lead to "Failed: internal
2529 error: code overflow at offset 49" or "glibc detected double free or
2530 corruption" errors.
2531
2532 27. Applied patches from Google (a) to support the new newline modes and (b) to
2533 advance over multibyte UTF-8 characters in GlobalReplace.
2534
2535 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2536 difference for some implementation of PCRE in some Windows version.
2537
2538 29. Added some extra testing facilities to pcretest:
2539
2540 \q<number> in a data line sets the "match limit" value
2541 \Q<number> in a data line sets the "match recursion limt" value
2542 -S <number> sets the stack size, where <number> is in megabytes
2543
2544 The -S option isn't available for Windows.
2545
2546
2547 Version 6.6 06-Feb-06
2548 ---------------------
2549
2550 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2551 in pcreposix.h. I have copied the definition from pcre.h.
2552
2553 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2554 because pcre.h is no longer a built file.
2555
2556 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2557 not normally included in the compiled code.
2558
2559
2560 Version 6.5 01-Feb-06
2561 ---------------------
2562
2563 1. When using the partial match feature with pcre_dfa_exec(), it was not
2564 anchoring the second and subsequent partial matches at the new starting
2565 point. This could lead to incorrect results. For example, with the pattern
2566 /1234/, partially matching against "123" and then "a4" gave a match.
2567
2568 2. Changes to pcregrep:
2569
2570 (a) All non-match returns from pcre_exec() were being treated as failures
2571 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2572 error message is output. Some extra information is given for the
2573 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2574 probably the only errors that are likely to be caused by users (by
2575 specifying a regex that has nested indefinite repeats, for instance).
2576 If there are more than 20 of these errors, pcregrep is abandoned.
2577
2578 (b) A binary zero was treated as data while matching, but terminated the
2579 output line if it was written out. This has been fixed: binary zeroes
2580 are now no different to any other data bytes.
2581
2582 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2583 used to set a locale for matching. The --locale=xxxx long option has
2584 been added (no short equivalent) to specify a locale explicitly on the
2585 pcregrep command, overriding the environment variables.
2586
2587 (d) When -B was used with -n, some line numbers in the output were one less
2588 than they should have been.
2589
2590 (e) Added the -o (--only-matching) option.
2591
2592 (f) If -A or -C was used with -c (count only), some lines of context were
2593 accidentally printed for the final match.
2594
2595 (g) Added the -H (--with-filename) option.
2596
2597 (h) The combination of options -rh failed to suppress file names for files
2598 that were found from directory arguments.
2599
2600 (i) Added the -D (--devices) and -d (--directories) options.
2601
2602 (j) Added the -F (--fixed-strings) option.
2603
2604 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2605
2606 (l) Added the --colo(u)r option.
2607
2608 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2609 is not present by default.
2610
2611 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2612 items such as (?R) or (?1), when the recursion could match a number of
2613 alternatives. If it matched one of the alternatives, but subsequently,
2614 outside the recursion, there was a failure, the code tried to back up into
2615 the recursion. However, because of the way PCRE is implemented, this is not
2616 possible, and the result was an incorrect result from the match.
2617
2618 In order to prevent this happening, the specification of recursion has
2619 been changed so that all such subpatterns are automatically treated as
2620 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2621
2622 4. I had overlooked the fact that, in some locales, there are characters for
2623 which isalpha() is true but neither isupper() nor islower() are true. In
2624 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2625 and ordfeminine) are like this. This affected the treatment of \w and \W
2626 when they appeared in character classes, but not when they appeared outside
2627 a character class. The bit map for "word" characters is now created
2628 separately from the results of isalnum() instead of just taking it from the
2629 upper, lower, and digit maps. (Plus the underscore character, of course.)
2630
2631 5. The above bug also affected the handling of POSIX character classes such as
2632 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2633 permanent tables. Instead, the bit maps for such a class were previously
2634 created as the appropriate unions of the upper, lower, and digit bitmaps.
2635 Now they are created by subtraction from the [[:word:]] class, which has
2636 its own bitmap.
2637
2638 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2639 It is created by subtracting the vertical space characters (\x09, \x0a,
2640 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2641 subtraction was done in the overall bitmap for a character class, meaning
2642 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2643 be recognized. This bug has been fixed.
2644
2645 7. Patches from the folks at Google:
2646
2647 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2648 real life, but is still worth protecting against".
2649
2650 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2651 regular expressions".
2652
2653 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2654 have it.
2655
2656 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2657 "configure" and the latter not, in order to fix a problem somebody had
2658 with compiling the Arg class on HP-UX.
2659
2660 (e) Improve the error-handling of the C++ wrapper a little bit.
2661
2662 (f) New tests for checking recursion limiting.
2663
2664 8. The pcre_memmove() function, which is used only if the environment does not
2665 have a standard memmove() function (and is therefore rarely compiled),
2666 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2667 returning a result (though PCRE never actually uses the result).
2668
2669 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2670 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2671 returned instead of calling malloc() with an overflowing number that would
2672 most likely cause subsequent chaos.
2673
2674 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2675
2676 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2677 with this option is matched, the nmatch and pmatch options of regexec() are
2678 ignored.
2679
2680 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2681 provided in case anyone wants to the the POSIX interface with UTF-8
2682 strings.
2683
2684 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2685 C++ linking (needed for some HP-UX environments).
2686
2687 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2688 (unused parameter) and in the pcre_printint() function (omitted "default"
2689 switch label when the default is to do nothing).
2690
2691 15. Added some code to make it possible, when PCRE is compiled as a C++
2692 library, to replace subject pointers for pcre_exec() with a smart pointer
2693 class, thus making it possible to process discontinuous strings.
2694
2695 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2696 much the same function. They were added by different people who were trying
2697 to make PCRE easy to compile on non-Unix systems. It has been suggested
2698 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2699 for compiling on Windows systems. I have therefore replaced it with
2700 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2701 defaults to "extern" for C or "extern C" for C++, which works fine on
2702 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2703 SCOPE with something explicit in config.h. In addition:
2704
2705 (a) pcreposix.h still had just "extern" instead of either of these macros;
2706 I have replaced it with PCRE_DATA_SCOPE.
2707
2708 (b) Functions such as _pcre_xclass(), which are internal to the library,
2709 but external in the C sense, all had PCRE_EXPORT in their definitions.
2710 This is apparently wrong for the Windows case, so I have removed it.
2711 (It makes no difference on Unix-like systems.)
2712
2713 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2714 of recursive calls to match(). This is different to MATCH_LIMIT because
2715 that limits the total number of calls to match(), not all of which increase
2716 the depth of recursion. Limiting the recursion depth limits the amount of
2717 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2718 when PCRE is compiled, and changed at run time. A patch from Google adds
2719 this functionality to the C++ interface.
2720
2721 18. Changes to the handling of Unicode character properties:
2722
2723 (a) Updated the table to Unicode 4.1.0.
2724
2725 (b) Recognize characters that are not in the table as "Cn" (undefined).
2726
2727 (c) I revised the way the table is implemented to a much improved format
2728 which includes recognition of ranges. It now supports the ranges that
2729 are defined in UnicodeData.txt, and it also amalgamates other
2730 characters into ranges. This has reduced the number of entries in the
2731 table from around 16,000 to around 3,000, thus reducing its size
2732 considerably. I realized I did not need to use a tree structure after
2733 all - a binary chop search is just as efficient. Having reduced the
2734 number of entries, I extended their size from 6 bytes to 8 bytes to
2735 allow for more data.
2736
2737 (d) Added support for Unicode script names via properties such as \p{Han}.
2738
2739 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2740 matching that character.
2741
2742 20. When matching a repeated Unicode property with a minimum greater than zero,
2743 (for example \pL{2,}), PCRE could look past the end of the subject if it
2744 reached it while seeking the minimum number of characters. This could
2745 happen only if some of the characters were more than one byte long, because
2746 there is a check for at least the minimum number of bytes.
2747
2748 21. Refactored the implementation of \p and \P so as to be more general, to
2749 allow for more different types of property in future. This has changed the
2750 compiled form incompatibly. Anybody with saved compiled patterns that use
2751 \p or \P will have to recompile them.
2752
2753 22. Added "Any" and "L&" to the supported property types.
2754
2755 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2756 but give a compile time error if the value is greater than 0xff.
2757
2758 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2759 accidentally not being installed or uninstalled.
2760
2761 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2762 made were to insert the current release number. This seemed silly, because
2763 it made things harder for people building PCRE on systems that don't run
2764 "configure". I have turned pcre.h into a distributed file, no longer built
2765 by "configure", with the version identification directly included. There is
2766 no longer a pcre.h.in file.
2767
2768 However, this change necessitated a change to the pcre-config script as
2769 well. It is built from pcre-config.in, and one of the substitutions was the
2770 release number. I have updated configure.ac so that ./configure now finds
2771 the release number by grepping pcre.h.
2772
2773 26. Added the ability to run the tests under valgrind.
2774
2775
2776 Version 6.4 05-Sep-05
2777 ---------------------
2778
2779 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2780 "--" to be printed when multiple files were scanned, even when none of the
2781 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2782 consider it to be a bug, and have restored the previous behaviour.
2783
2784 2. A couple of code tidies to get rid of compiler warnings.
2785
2786 3. The pcretest program used to cheat by referring to symbols in the library
2787 whose names begin with _pcre_. These are internal symbols that are not
2788 really supposed to be visible externally, and in some environments it is
2789 possible to suppress them. The cheating is now confined to including
2790 certain files from the library's source, which is a bit cleaner.
2791
2792 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2793 file's purpose clearer.
2794
2795 5. Reorganized pcre_ucp_findchar().
2796
2797
2798 Version 6.3 15-Aug-05
2799 ---------------------
2800
2801 1. The file libpcre.pc.in did not have general read permission in the tarball.
2802
2803 2. There were some problems when building without C++ support:
2804
2805 (a) If C++ support was not built, "make install" and "make test" still
2806 tried to test it.
2807
2808 (b) There were problems when the value of CXX was explicitly set. Some
2809 changes have been made to try to fix these, and ...
2810
2811 (c) --disable-cpp can now be used to explicitly disable C++ support.
2812
2813 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2814 backslash in a target when C++ was disabled. This confuses some
2815 versions of "make", apparently. Using an intermediate variable solves
2816 this. (Same for CPP_LOBJ.)
2817
2818 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2819 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2820 necessary on certain architectures.
2821
2822 4. Added a setting of -export-symbols-regex to the link command to remove
2823 those symbols that are exported in the C sense, but actually are local
2824 within the library, and not documented. Their names all begin with
2825 "_pcre_". This is not a perfect job, because (a) we have to except some
2826 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2827 available (and never for static libraries). I have made a note to try to
2828 find a way round (a) in the future.
2829
2830
2831 Version 6.2 01-Aug-05
2832 ---------------------
2833
2834 1. There was no test for integer overflow of quantifier values. A construction
2835 such as {1111111111111111} would give undefined results. What is worse, if
2836 a minimum quantifier for a parenthesized subpattern overflowed and became
2837 negative, the calculation of the memory size went wrong. This could have
2838 led to memory overwriting.
2839
2840 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2841
2842 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2843 operating environments where this matters.
2844
2845 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2846 PCRE options from within the C++ wrapper.
2847
2848 5. Named capturing subpatterns were not being correctly counted when a pattern
2849 was compiled. This caused two problems: (a) If there were more than 100
2850 such subpatterns, the calculation of the memory needed for the whole
2851 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2852 back references of the form \12, where the number was greater than 9, were
2853 not recognized as back references, even though there were sufficient
2854 previous subpatterns.
2855
2856 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2857 versions of gcc, e.g. 2.95.4.
2858
2859
2860 Version 6.1 21-Jun-05
2861 ---------------------
2862
2863 1. There was one reference to the variable "posix" in pcretest.c that was not
2864 surrounded by "#if !defined NOPOSIX".
2865
2866 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2867 the cross-check on the old pcre_info() function, for the benefit of the
2868 cut-down version of PCRE that is currently imported into Exim.
2869
2870 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2871 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2872 patterns that start (?i) but I don't think that matters. The use of (?i) is
2873 just an example; this all applies to the other options as well.
2874
2875 4. Since libtool seems to echo the compile commands it is issuing, the output
2876 from "make" can be reduced a bit by putting "@" in front of each libtool
2877 compile command.
2878
2879 5. Patch from the folks at Google for configure.in to be a bit more thorough
2880 in checking for a suitable C++ installation before trying to compile the
2881 C++ stuff. This should fix a reported problem when a compiler was present,
2882 but no suitable headers.
2883
2884 6. The man pages all had just "PCRE" as their title. I have changed them to
2885 be the relevant file name. I have also arranged that these names are
2886 retained in the file doc/pcre.txt, which is a concatenation in text format
2887 of all the man pages except the little individual ones for each function.
2888
2889 7. The NON-UNIX-USE file had not been updated for the different set of source
2890 files that come with release 6. I also added a few comments about the C++
2891 wrapper.
2892
2893
2894 Version 6.0 07-Jun-05
2895 ---------------------
2896
2897 1. Some minor internal re-organization to help with my DFA experiments.
2898
2899 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2900 didn't matter for the library itself when fully configured, but did matter
2901 when compiling without UCP support, or within Exim, where the ucp files are
2902 not imported.
2903
2904 3. Refactoring of the library code to split up the various functions into
2905 different source modules. The addition of the new DFA matching code (see
2906 below) to a single monolithic source would have made it really too
2907 unwieldy, quite apart from causing all the code to be include in a
2908 statically linked application, when only some functions are used. This is
2909 relevant even without the DFA addition now that patterns can be compiled in
2910 one application and matched in another.
2911
2912 The downside of splitting up is that there have to be some external
2913 functions and data tables that are used internally in different modules of
2914 the library but which are not part of the API. These have all had their
2915 names changed to start with "_pcre_" so that they are unlikely to clash
2916 with other external names.
2917
2918 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2919 a different (DFA) algorithm. Although it is slower than the original
2920 function, it does have some advantages for certain types of matching
2921 problem.
2922
2923 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2924 including restarting after a partial match.
2925
2926 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2927 defined when compiling for Windows was sent to me. I have put it into the
2928 code, though I have no means of testing or verifying it.
2929
2930 7. Added the pcre_refcount() auxiliary function.
2931
2932 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2933 match before or at the first newline in the subject string. In pcretest,
2934 the /f option on a pattern can be used to set this.
2935
2936 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2937 would behave wrongly. This has been present in PCRE since release 4.0.
2938
2939 10. A number of changes to the pcregrep command:
2940
2941 (a) Refactored how -x works; insert ^(...)$ instead of setting
2942 PCRE_ANCHORED and checking the length, in preparation for adding
2943 something similar for -w.
2944
2945 (b) Added the -w (match as a word) option.
2946
2947 (c) Refactored the way lines are read and buffered so as to have more
2948 than one at a time available.
2949
2950 (d) Implemented a pcregrep test script.
2951
2952 (e) Added the -M (multiline match) option. This allows patterns to match
2953 over several lines of the subject. The buffering ensures that at least
2954 8K, or the rest of the document (whichever is the shorter) is available
2955 for matching (and similarly the previous 8K for lookbehind assertions).
2956
2957 (f) Changed the --help output so that it now says
2958
2959 -w, --word-regex(p)
2960
2961 instead of two lines, one with "regex" and the other with "regexp"
2962 because that confused at least one person since the short forms are the
2963 same. (This required a bit of code, as the output is generated
2964 automatically from a table. It wasn't just a text change.)
2965
2966 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2967 option but starts with a hyphen. Could be a pattern or a path name
2968 starting with a hyphen, for instance.
2969
2970 (h) "-" can be given as a file name to represent stdin.
2971
2972 (i) When file names are being printed, "(standard input)" is used for
2973 the standard input, for compatibility with GNU grep. Previously
2974 "<stdin>" was used.
2975
2976 (j) The option --label=xxx can be used to supply a name to be used for
2977 stdin when file names are being printed. There is no short form.
2978
2979 (k) Re-factored the options decoding logic because we are going to add
2980 two more options that take data. Such options can now be given in four
2981 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2982
2983 (l) Added the -A, -B, and -C options for requesting that lines of context
2984 around matches be printed.
2985
2986 (m) Added the -L option to print the names of files that do not contain
2987 any matching lines, that is, the complement of -l.
2988
2989 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2990 continue to scan other files.
2991
2992 (o) The -s option was incorrectly implemented. For compatibility with other
2993 greps, it now suppresses the error message for a non-existent or non-
2994 accessible file (but not the return code). There is a new option called
2995 -q that suppresses the output of matching lines, which was what -s was
2996 previously doing.
2997
2998 (p) Added --include and --exclude options to specify files for inclusion
2999 and exclusion when recursing.
3000
3001 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3002 Hopefully, it now does.
3003
3004 12. Missing cast in pcre_study().
3005
3006 13. Added an "uninstall" target to the makefile.
3007
3008 14. Replaced "extern" in the function prototypes in Makefile.in with
3009 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3010 world, but is set differently for Windows.
3011
3012 15. Added a second compiling function called pcre_compile2(). The only
3013 difference is that it has an extra argument, which is a pointer to an
3014 integer error code. When there is a compile-time failure, this is set
3015 non-zero, in addition to the error test pointer being set to point to an
3016 error message. The new argument may be NULL if no error number is required
3017 (but then you may as well call pcre_compile(), which is now just a
3018 wrapper). This facility is provided because some applications need a
3019 numeric error indication, but it has also enabled me to tidy up the way
3020 compile-time errors are handled in the POSIX wrapper.
3021
3022 16. Added VPATH=.libs to the makefile; this should help when building with one
3023 prefix path and installing with another. (Or so I'm told by someone who
3024 knows more about this stuff than I do.)
3025
3026 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3027 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3028 match everything, including newlines. This is not POSIX-compatible, but
3029 somebody wanted the feature. From pcretest it can be activated by using
3030 both the P and the s flags.
3031
3032 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3033
3034 19. libpcre.pc was being incorrectly installed as executable.
3035
3036 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3037 it now also looks for '\r' so that it will work unmodified on Windows.
3038
3039 21. Added Google's contributed C++ wrapper to the distribution.
3040
3041 22. Added some untidy missing memory free() calls in pcretest, to keep
3042 Electric Fence happy when testing.
3043
3044
3045
3046 Version 5.0 13-Sep-04
3047 ---------------------
3048
3049 1. Internal change: literal characters are no longer packed up into items
3050 containing multiple characters in a single byte-string. Each character
3051 is now matched using a separate opcode. However, there may be more than one
3052 byte in the character in UTF-8 mode.
3053
3054 2. The pcre_callout_block structure has two new fields: pattern_position and
3055 next_item_length. These contain the offset in the pattern to the next match
3056 item, and its length, respectively.
3057
3058 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3059 insertion of callouts before each pattern item. Added the /C option to
3060 pcretest to make use of this.
3061
3062 4. On the advice of a Windows user, the lines
3063
3064 #if defined(_WIN32) || defined(WIN32)
3065 _setmode( _fileno( stdout ), 0x8000 );
3066 #endif /* defined(_WIN32) || defined(WIN32) */
3067
3068 have been added to the source of pcretest. This apparently does useful
3069 magic in relation to line terminators.
3070
3071 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3072 for the benefit of those environments where the "b" makes a difference.
3073
3074 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3075 to know about it. I have put a hack into configure.in that adds in code
3076 to set GCC=yes if CC=icc. This seems to end up at a point in the
3077 generated configure script that is early enough to affect the setting of
3078 compiler options, which is what is needed, but I have no means of testing
3079 whether it really works. (The user who reported this had patched the
3080 generated configure script, which of course I cannot do.)
3081
3082 LATER: After change 22 below (new libtool files), the configure script
3083 seems to know about icc (and also ecc). Therefore, I have commented out
3084 this hack in configure.in.
3085
3086 7. Added support for pkg-config (2 patches were sent in).
3087
3088 8. Negated POSIX character classes that used a combination of internal tables
3089 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3090 [[:^ascii]]. Typically, they would match almost any characters. The other
3091 POSIX classes were not broken in this way.
3092
3093 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3094 to find the match, as PCRE was deluded into thinking that the match had to
3095 start at the start point or following a newline. The same bug applied to
3096 patterns with negative forward assertions or any backward assertions
3097 preceding ".*" at the start, unless the pattern required a fixed first
3098 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3099
3100 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3101 starting at the last subject character, bytes beyond the end of the subject
3102 string were read.
3103
3104 11. Renamed the variable "class" as "classbits" to make life easier for C++
3105 users. (Previously there was a macro definition, but it apparently wasn't
3106 enough.)
3107
3108 12. Added the new field "tables" to the extra data so that tables can be passed
3109 in at exec time, or the internal tables can be re-selected. This allows
3110 a compiled regex to be saved and re-used at a later time by a different
3111 program that might have everything at different addresses.
3112
3113 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3114 -R library as well as a -L library.
3115
3116 14. The debugging options of pcretest (-d on the command line or D on a
3117 pattern) showed incorrect output for anything following an extended class
3118 that contained multibyte characters and which was followed by a quantifier.
3119
3120 15. Added optional support for general category Unicode character properties
3121 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3122 support. It adds about 90K to the size of the library. The meanings of the
3123 inbuilt class escapes such as \d and \s have NOT been changed.
3124
3125 16. Updated pcredemo.c to include calls to free() to release the memory for the
3126 compiled pattern.
3127
3128 17. The generated file chartables.c was being created in the source directory
3129 instead of in the building directory. This caused the build to fail if the
3130 source directory was different from the building directory, and was
3131 read-only.
3132
3133 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3134 file. No doubt somebody will tell me if they don't make sense... Also added
3135 Dan Mooney's comments about building on OpenVMS.
3136
3137 19. Added support for partial matching via the PCRE_PARTIAL option for
3138 pcre_exec() and the \P data escape in pcretest.
3139
3140 20. Extended pcretest with 3 new pattern features:
3141
3142 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3143 write the compiled pattern to the file whose name is "rest-of-line".
3144 This is a straight binary dump of the data, with the saved pointer to
3145 the character tables forced to be NULL. The study data, if any, is
3146 written too. After writing, pcretest reads a new pattern.
3147
3148 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3149 compiled pattern from the given file. There must not be any
3150 occurrences of "<" in the file name (pretty unlikely); if there are,
3151 pcretest will instead treat the initial "<" as a pattern delimiter.
3152 After reading in the pattern, pcretest goes on to read data lines as
3153 usual.
3154
3155 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3156 and 16-bit fields in a compiled pattern, to simulate a pattern that
3157 was compiled on a host of opposite endianness.
3158
3159 21. The pcre-exec() function can now cope with patterns that were compiled on
3160 hosts of opposite endianness, with this restriction:
3161
3162 As for any compiled expression that is saved and used later, the tables
3163 pointer field cannot be preserved; the extra_data field in the arguments
3164 to pcre_exec() should be used to pass in a tables address if a value
3165 other than the default internal tables were used at compile time.
3166
3167 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3168 now diagnosed as an error. Previously, most of the time, a negative number
3169 would have been treated as zero, but if in addition "ovector" was passed as
3170 NULL, a crash could occur.
3171
3172 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3173 new versions from the libtool 1.5 distribution (the last one is a copy of
3174 a file called libtool.m4). This seems to have fixed the need to patch
3175 "configure" to support Darwin 1.3 (which I used to do). However, I still
3176 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3177 workstation).
3178
3179 24. Changed the PCRE licence to be the more standard "BSD" licence.
3180
3181
3182 Version 4.5 01-Dec-03
3183 ---------------------
3184
3185 1. There has been some re-arrangement of the code for the match() function so
3186 that it can be compiled in a version that does not call itself recursively.
3187 Instead, it keeps those local variables that need separate instances for
3188 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3189 needs to "recurse". Keeping track of where control must go is done by means
3190 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3191 hide most of the details from the main code, and operates only if
3192 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3193 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3194 operating.
3195
3196 To make it easier for callers to provide specially tailored get/free
3197 functions for this usage, two new functions, pcre_stack_malloc, and
3198 pcre_stack_free, are used. They are always called in strict stacking order,
3199 and the size of block requested is always the same.
3200
3201 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3202 PCRE has been compiled to use the stack or the heap for recursion. The
3203 -C option of pcretest uses this to show which version is compiled.
3204
3205 A new data escape \S, is added to pcretest; it causes the amounts of store
3206 obtained and freed by both kinds of malloc/free at match time to be added
3207 to the output.
3208
3209 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3210 what's available on my current Linux desktop machine.
3211
3212 3. When matching a UTF-8 string, the test for a valid string at the start has
3213 been extended. If start_offset is not zero, PCRE now checks that it points
3214 to a byte that is the start of a UTF-8 character. If not, it returns
3215 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3216 this is necessary because there may be backward assertions in the pattern.
3217 When matching the same subject several times, it may save resources to use
3218 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3219
3220 4. The code for checking the validity of UTF-8 strings has been tightened so
3221 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3222 containing "overlong sequences".
3223
3224 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3225 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3226 should have been "&", but it just so happened that all the cases this let
3227 through by mistake were picked up later in the function.
3228
3229 6. I had used a variable called "isblank" - this is a C99 function, causing
3230 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3231
3232 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3233 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3234 so the version is shown; (c) stop "make test" echoing "./RunTest".
3235
3236 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3237
3238 9. The prototype for memmove() for systems that don't have it was using
3239 size_t, but the inclusion of the header that defines size_t was later. I've
3240 moved the #includes for the C headers earlier to avoid this.
3241
3242 10. Added some adjustments to the code to make it easier to compiler on certain
3243 special systems:
3244
3245 (a) Some "const" qualifiers were missing.
3246 (b) Added the macro EXPORT before all exported functions; by default this
3247 is defined to be empty.
3248 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3249 that it reads its output file name as an argument instead of writing
3250 to the standard output and assuming this can be redirected.
3251
3252 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3253 class containing characters with values greater than 255, PCRE compilation
3254 went into a loop.
3255
3256 12. A recursive reference to a subpattern that was within another subpattern
3257 that had a minimum quantifier of zero caused PCRE to crash. For example,
3258 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3259 recursion. If the recursively-called subpattern itself had a zero repeat,
3260 that was OK.
3261
3262 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3263 buffer into which it was copied (for escape processing) was still set at
3264 1024, so long lines caused crashes.
3265
3266 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3267 "internal error: code overflow...". This applied to any character class
3268 that was followed by a possessive quantifier.
3269
3270 15. Modified the Makefile to add libpcre.la as a prerequisite for
3271 libpcreposix.la because I was told this is needed for a parallel build to
3272 work.
3273
3274 16. If a pattern that contained .* following optional items at the start was
3275 studied, the wrong optimizing data was generated, leading to matching
3276 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3277 matching string must start with a or b or c. The correct conclusion for
3278 this pattern is that a match can start with any character.
3279
3280
3281 Version 4.4 13-Aug-03
3282 ---------------------
3283
3284 1. In UTF-8 mode, a character class containing characters with values between
3285 127 and 255 was not handled correctly if the compiled pattern was studied.
3286 In fixing this, I have also improved the studying algorithm for such
3287 classes (slightly).
3288
3289 2. Three internal functions had redundant arguments passed to them. Removal
3290 might give a very teeny performance improvement.
3291
3292 3. Documentation bug: the value of the capture_top field in a callout is *one
3293 more than* the number of the hightest numbered captured substring.
3294
3295 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3296 in incorrectly linking with a previously installed version. They now link
3297 explicitly with libpcre.la.
3298
3299 5. configure.in no longer needs to recognize Cygwin specially.
3300
3301 6. A problem in pcre.in for Windows platforms is fixed.
3302
3303 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3304 pcretest, it used to include the size of the study block as part of its
3305 output. Unfortunately, the structure contains a field that has a different
3306 size on different hardware architectures. This meant that the tests that
3307 showed this size failed. As the block is currently always of a fixed size,
3308 this information isn't actually particularly useful in pcretest output, so
3309 I have just removed it.
3310
3311 8. Three pre-processor statements accidentally did not start in column 1.
3312 Sadly, there are *still* compilers around that complain, even though
3313 standard C has not required this for well over a decade. Sigh.
3314
3315 9. In pcretest, the code for checking callouts passed small integers in the
3316 callout_data field, which is a void * field. However, some picky compilers
3317 complained about the casts involved for this on 64-bit systems. Now
3318 pcretest passes the address of the small integer instead, which should get
3319 rid of the warnings.
3320
3321 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3322 both compile and run time, and gives an error if an invalid UTF-8 sequence
3323 is found. There is a option for disabling this check in cases where the
3324 string is known to be correct and/or the maximum performance is wanted.
3325
3326 11. In response to a bug report, I changed one line in Makefile.in from
3327
3328 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3329 to
3330 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3331
3332 to look similar to other lines, but I have no way of telling whether this
3333 is the right thing to do, as I do not use Windows. No doubt I'll get told
3334 if it's wrong...
3335
3336
3337 Version 4.3 21-May-03
3338 ---------------------
3339
3340 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3341 Makefile.
3342
3343 2. Some refactoring to improve the quality of the code:
3344
3345 (i) The utf8_table... variables are now declared "const".
3346
3347 (ii) The code for \cx, which used the "case flipping" table to upper case
3348 lower case letters, now just substracts 32. This is ASCII-specific,
3349 but the whole concept of \cx is ASCII-specific, so it seems
3350 reasonable.
3351
3352 (iii) PCRE was using its character types table to recognize decimal and
3353 hexadecimal digits in the pattern. This is silly, because it handles
3354 only 0-9, a-f, and A-F, but the character types table is locale-
3355 specific, which means strange things might happen. A private
3356 table is now used for this - though it costs 256 bytes, a table is
3357 much faster than multiple explicit tests. Of course, the standard
3358 character types table is still used for matching digits in subject
3359 strings against \d.
3360
3361 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3362 ending in _t are). So I've renamed it as ESC_tee.
3363
3364 3. The first argument for regexec() in the POSIX wrapper should have been
3365 defined as "const".
3366
3367 4. Changed pcretest to use malloc() for its buffers so that they can be
3368 Electric Fenced for debugging.
3369
3370 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3371 to read one or more bytes before the start of the subject string. Often this
3372 had no effect on PCRE's behaviour, but in some circumstances it could
3373 provoke a segmentation fault.
3374
3375 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3376 to try to read one or more bytes before the start of the subject string.
3377
3378 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3379 UTF-8 support could misbehave in various ways if the subject string
3380 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3381 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3382 back over UTF-8 characters.)
3383
3384
3385 Version 4.2 14-Apr-03
3386 ---------------------
3387
3388 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3389
3390 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3391 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3392 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3393 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3394 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3395 and BUILD_EXEEXT
3396 Note: automatic setting of the BUILD variables is not yet working
3397 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3398 compile-time but not at link-time
3399 [LINK]: use for linking executables only
3400 make different versions for Windows and non-Windows
3401 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3402 libraries
3403 [LINK_FOR_BUILD]: new variable
3404 [OBJEXT]: use throughout
3405 [EXEEXT]: use throughout
3406 <winshared>: new target
3407 <wininstall>: new target
3408 <dftables.o>: use native compiler
3409 <dftables>: use native linker
3410 <install>: handle Windows platform correctly
3411 <clean>: ditto
3412 <check>: ditto
3413 copy DLL to top builddir before testing
3414
3415 As part of these changes, -no-undefined was removed again. This was reported
3416 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3417 in any case.
3418
3419 3. Some tidies to get rid of compiler warnings:
3420
3421 . In the match_data structure, match_limit was an unsigned long int, whereas
3422 match_call_count was an int. I've made them both unsigned long ints.
3423
3424 . In pcretest the fact that a const uschar * doesn't automatically cast to
3425 a void * provoked a warning.
3426
3427 . Turning on some more compiler warnings threw up some "shadow" variables
3428 and a few more missing casts.
3429
3430 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3431 option, a class that contained a single character with a value between 128
3432 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3433
3434 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3435 option, a class that contained several characters, but with at least one
3436 whose value was between 128 and 255 caused PCRE to crash.
3437
3438
3439 Version 4.1 12-Mar-03
3440 ---------------------
3441
3442 1. Compiling with gcc -pedantic found a couple of places where casts were
3443 needed, and a string in dftables.c that was longer than standard compilers are
3444 required to support.
3445
3446 2. Compiling with Sun's compiler found a few more places where the code could
3447 be tidied up in order to avoid warnings.
3448
3449 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3450 first of these names is deprecated in the latest Autoconf in favour of the name
3451 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3452 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3453 analogy I have changed it to CFLAGS_FOR_BUILD.
3454
3455 4. Added -no-undefined to the linking command in the Makefile, because this is
3456 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3457 linking step for the pcreposix library.
3458
3459 5. PCRE was failing to diagnose the case of two named groups with the same
3460 name.
3461
3462 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3463 literal character that is needed in the subject for a match, and scans along to
3464 ensure that it is present before embarking on the full matching process. This
3465 saves time in cases of nested unlimited repeats that are never going to match.
3466 Problem: the scan can take a lot of time if the subject is very long (e.g.
3467 megabytes), thus penalizing straightforward matches. It is now done only if the
3468 amount of subject to be scanned is less than 1000 bytes.
3469
3470 7. A lesser problem with the same optimization is that it was recording the
3471 first character of an anchored pattern as "needed", thus provoking a search
3472 right along the subject, even when the first match of the pattern was going to
3473 fail. The "needed" character is now not set for anchored patterns, unless it
3474 follows something in the pattern that is of non-fixed length. Thus, it still
3475 fulfils its original purpose of finding quick non-matches in cases of nested
3476 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3477
3478
3479 Version 4.0 17-Feb-03
3480 ---------------------
3481
3482 1. If a comment in an extended regex that started immediately after a meta-item
3483 extended to the end of string, PCRE compiled incorrect data. This could lead to
3484 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3485
3486 2. Moved to autoconf 2.53 and libtool 1.4.2.
3487
3488 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3489 the special perltest8 script is no longer needed - all the tests can be run
3490 from a single perltest script.
3491
3492 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3493 by \s. It has now been removed in PCRE. This means it isn't recognized as
3494 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3495 class [:space:] *does* include VT, thereby creating a mess.
3496
3497 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3498 space and tab.
3499
3500 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3501 its new features into the main test script, reducing the number of scripts.
3502
3503 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3504 were backward compatible, and made the (?i) apply to the whole pattern, as if
3505 /i were given. Now it behaves more logically, and applies the option setting
3506 only to what follows. PCRE has been changed to follow suit. However, if it
3507 finds options settings right at the start of the pattern, it extracts them into
3508 the global options, as before. Thus, they show up in the info data.
3509
3510 8. Added support for the \Q...\E escape sequence. Characters in between are
3511 treated as literals. This is slightly different from Perl in that $ and @ are
3512 also handled as literals inside the quotes. In Perl, they will cause variable
3513 interpolation. Note the following examples:
3514
3515 Pattern PCRE matches Perl matches
3516
3517 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3518 \Qabc\$xyz\E abc\$xyz abc\$xyz
3519 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3520
3521 For compatibility with Perl, \Q...\E sequences are recognized inside character
3522 classes as well as outside them.
3523
3524 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3525 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3526 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3527 signed/unsigned warnings.
3528
3529 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3530 option for pcretest, so I've replaced it by a simple function that does just
3531 that job.
3532
3533 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3534 "pcregrep -".
3535
3536 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3537 Java package. This provides some syntactic sugar for simple cases of what my
3538 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3539 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3540 item, you can use this simplified notation. Note that only makes sense with
3541 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3542 greediness, whatever the setting of the PCRE_UNGREEDY option.
3543
3544 13. A change of greediness default within a pattern was not taking effect at
3545 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3546 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3547 was abstracted outside.
3548
3549 14. PCRE now supports the \G assertion. It is true when the current matching
3550 position is at the start point of the match. This differs from \A when the
3551 starting offset is non-zero. Used with the /g option of pcretest (or similar
3552 code), it works in the same way as it does for Perl's /g option. If all
3553 alternatives of a regex begin with \G, the expression is anchored to the start
3554 match position, and the "anchored" flag is set in the compiled expression.
3555
3556 15. Some bugs concerning the handling of certain option changes within patterns
3557 have been fixed. These applied to options other than (?ims). For example,
3558 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3559 been the other way round. Some of this was related to change 7 above.
3560
3561 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3562 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3563 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3564 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3565
3566 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3567 mode. Unlike ".", it always matches newline, whatever the setting of
3568 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3569 assertions. Perl allows it, but it doesn't (in general) work because it can't
3570 calculate the length of the lookbehind. At least, that's the case for Perl
3571 5.8.0 - I've been told they are going to document that it doesn't work in
3572 future.
3573
3574 18. Added an error diagnosis for escapes that PCRE does not support: these are
3575 \L, \l, \N, \P, \p, \U, \u, and \X.
3576
3577 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3578 reading past the end of the pattern in cases such as /[abcd/.
3579
3580 20. PCRE was getting more memory than necessary for patterns with classes that
3581 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3582
3583 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3584 compiling PCRE for use with Virtual Pascal.
3585
3586 22. Small fix to the Makefile to make it work properly if the build is done
3587 outside the source tree.
3588
3589 23. Added a new extension: a condition to go with recursion. If a conditional
3590 subpattern starts with (?(R) the "true" branch is used if recursion has
3591 happened, whereas the "false" branch is used only at the top level.
3592
3593 24. When there was a very long string of literal characters (over 255 bytes
3594 without UTF support, over 250 bytes with UTF support), the computation of how
3595 much memory was required could be incorrect, leading to segfaults or other
3596 strange effects.
3597
3598 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3599 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3600 there was a subsequent back reference to those brackets. This meant that, for
3601 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3602 possible to check for precisely this case. All we can do is abandon the
3603 optimization if .* occurs inside capturing brackets when there are any back
3604 references whatsoever. (See below for a better fix that came later.)
3605
3606 26. The handling of the optimization for finding the first character of a
3607 non-anchored pattern, and for finding a character that is required later in the
3608 match were failing in some cases. This didn't break the matching; it just
3609 failed to optimize when it could. The way this is done has been re-implemented.
3610
3611 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3612
3613 28. Added a new feature that provides some of the functionality that Perl
3614 provides with (?{...}). The facility is termed a "callout". The way it is done
3615 in PCRE is for the caller to provide an optional function, by setting
3616 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3617 global variable. By default it is unset, which disables all calling out. To get
3618 the function called, the regex must include (?C) at appropriate points. This
3619 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3620 This provides a means of identifying different callout points. When PCRE
3621 reaches such a point in the regex, if pcre_callout has been set, the external
3622 function is called. It is provided with data in a structure called
3623 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3624 matching continues; if it returns a non-zero value, the match at the current
3625 point fails. However, backtracking will occur if possible. [This was changed
3626 later and other features added - see item 49 below.]
3627
3628 29. pcretest is upgraded to test the callout functionality. It provides a
3629 callout function that displays information. By default, it shows the start of
3630 the match and the current position in the text. There are some new data escapes
3631 to vary what happens:
3632
3633 \C+ in addition, show current contents of captured substrings
3634 \C- do not supply a callout function
3635 \C!n return 1 when callout number n is reached
3636 \C!n!m return 1 when callout number n is reached for the mth time
3637
3638 30. If pcregrep was called with the -l option and just a single file name, it
3639 output "<stdin>" if a match was found, instead of the file name.
3640
3641 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3642 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3643 pcre_exec(). This saves a malloc/free per call. The default value of
3644 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3645 when configuring.
3646
3647 32. The default maximum size of a compiled pattern is 64K. There have been a
3648 few cases of people hitting this limit. The code now uses macros to handle the
3649 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3650 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3651 configuring. Tests 2 and 5 work only with 2-byte links because they output
3652 debugging information about compiled patterns.
3653
3654 33. Internal code re-arrangements:
3655
3656 (a) Moved the debugging function for printing out a compiled regex into
3657 its own source file (printint.c) and used #include to pull it into
3658 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3659 separate copies.
3660
3661 (b) Defined the list of op-code names for debugging as a macro in
3662 internal.h so that it is next to the definition of the opcodes.
3663
3664 (c) Defined a table of op-code lengths for simpler skipping along compiled
3665 code. This is again a macro in internal.h so that it is next to the
3666 definition of the opcodes.
3667
3668 34. Added support for recursive calls to individual subpatterns, along the
3669 lines of Robin Houston's patch (but implemented somewhat differently).
3670
3671 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3672 allow it to read and process whole directories in Win32. This code was
3673 contributed by Lionel Fourquaux; it has not been tested by me.
3674
3675 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3676 used to name a group. Names consist of alphanumerics and underscores, and must
3677 be unique. Back references use the syntax (?P=name) and recursive calls use
3678 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3679 numbers. The function pcre_fullinfo() can be used after compilation to extract
3680 a name/number map. There are three relevant calls:
3681
3682 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3683 PCRE_INFO_NAMECOUNT yields the number of entries
3684 PCRE_INFO_NAMETABLE yields a pointer to the map.
3685
3686 The map is a vector of fixed-size entries. The size of each entry depends on
3687 the length of the longest name used. The first two bytes of each entry are the
3688 group number, most significant byte first. There follows the corresponding
3689 name, zero terminated. The names are in alphabetical order.
3690
3691 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3692 case instead of 255. Making it the same both with and without UTF-8 support
3693 means that the same test output works with both.
3694
3695 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3696 calling malloc() with a zero argument.
3697
3698 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3699 optimization. I've improved things by keeping a bitmap of backreferences with
3700 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3701 fact referenced, the optimization can be applied. It is unlikely that a
3702 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3703 the match to follow \n) will appear inside brackets with a number greater than
3704 31, but if it does, any back reference > 31 suppresses the optimization.
3705
3706 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3707 of disabling numbered capturing parentheses. Any opening parenthesis that is
3708 not followed by ? behaves as if it were followed by ?: but named parentheses
3709 can still be used for capturing (and they will acquire numbers in the usual
3710 way).
3711
3712 41. Redesigned the return codes from the match() function into yes/no/error so
3713 that errors can be passed back from deep inside the nested calls. A malloc
3714 failure while inside a recursive subpattern call now causes the
3715 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3716
3717 42. It is now possible to set a limit on the number of times the match()
3718 function is called in a call to pcre_exec(). This facility makes it possible to
3719 limit the amount of recursion and backtracking, though not in a directly
3720 obvious way, because the match() function is used in a number of different
3721 circumstances. The count starts from zero for each position in the subject
3722 string (for non-anchored patterns). The default limit is, for compatibility, a
3723 large number, namely 10 000 000. You can change this in two ways:
3724
3725 (a) When configuring PCRE before making, you can use --with-match-limit=n
3726 to set a default value for the compiled library.
3727
3728 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3729 a different value is set. See 45 below.
3730
3731 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3732
3733 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3734 of things that can be changed at compile time. The first argument specifies
3735 what is wanted and the second points to where the information is to be placed.
3736 The current list of available information is:
3737
3738 PCRE_CONFIG_UTF8
3739
3740 The output is an integer that is set to one if UTF-8 support is available;
3741 otherwise it is set to zero.
3742
3743 PCRE_CONFIG_NEWLINE
3744
3745 The output is an integer that it set to the value of the code that is used for
3746 newline. It is either LF (10) or CR (13).
3747
3748 PCRE_CONFIG_LINK_SIZE
3749
3750 The output is an integer that contains the number of bytes used for internal
3751 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3752
3753 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3754
3755 The output is an integer that contains the threshold above which the POSIX
3756 interface uses malloc() for output vectors. See item 31 above.
3757
3758 PCRE_CONFIG_MATCH_LIMIT
3759
3760 The output is an unsigned integer that contains the default limit of the number
3761 of match() calls in a pcre_exec() execution. See 42 above.
3762
3763 44. pcretest has been upgraded by the addition of the -C option. This causes it
3764 to extract all the available output from the new pcre_config() function, and to
3765 output it. The program then exits immediately.
3766
3767 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3768 order to support additional features. One way would have been to define
3769 pcre_exec2() (for example) with extra arguments, but this would not have been
3770 extensible, and would also have required all calls to the original function to
3771 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3772 is used for passing in "extra" data from pcre_study().
3773
3774 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3775 contains the following fields:
3776
3777 flags a bitmap indicating which of the following fields are set
3778 study_data opaque data from pcre_study()
3779 match_limit a way of specifying a limit on match() calls for a specific
3780 call to pcre_exec()
3781 callout_data data for callouts (see 49 below)
3782
3783 The flag bits are also defined in pcre.h, and are
3784
3785 PCRE_EXTRA_STUDY_DATA
3786 PCRE_EXTRA_MATCH_LIMIT
3787 PCRE_EXTRA_CALLOUT_DATA
3788
3789 The pcre_study() function now returns one of these new pcre_extra blocks, with
3790 the actual study data pointed to by the study_data field, and the
3791 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3792 before. That is, this change is entirely upwards-compatible and requires no
3793 change to existing code.
3794
3795 If you want to pass in additional data to pcre_exec(), you can either place it
3796 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3797 block.
3798
3799 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3800 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3801 times with different match limits, until it finds the minimum value needed for
3802 pcre_exec() to complete. The value is then output. This can be instructive; for
3803 most simple matches the number is quite small, but for pathological cases it
3804 gets very large very quickly.
3805
3806 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3807 returns the size of the data block pointed to by the study_data field in a
3808 pcre_extra block, that is, the value that was passed as the argument to
3809 pcre_malloc() when PCRE was getting memory in which to place the information
3810 created by pcre_study(). The fourth argument should point to a size_t variable.
3811 pcretest has been extended so that this information is shown after a successful
3812 pcre_study() call when information about the compiled regex is being displayed.
3813
3814 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3815 because what follows is always an absolute path. (Later: it turns out that this
3816 is more than cosmetic for MinGW, because it doesn't like empty path
3817 components.)
3818
3819 49. Some changes have been made to the callout feature (see 28 above):
3820
3821 (i) A callout function now has three choices for what it returns:
3822
3823 0 => success, carry on matching
3824 > 0 => failure at this point, but backtrack if possible
3825 < 0 => serious error, return this value from pcre_exec()
3826
3827 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3828 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3829 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3830 use by callout functions. It will never be used by PCRE itself.
3831
3832 (ii) The pcre_extra structure (see 45 above) has a void * field called
3833 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3834 pcre_callout_block structure has a field of the same name. The contents of
3835 the field passed in the pcre_extra structure are passed to the callout
3836 function in the corresponding field in the callout block. This makes it
3837 easier to use the same callout-containing regex from multiple threads. For
3838 testing, the pcretest program has a new data escape
3839
3840 \C*n pass the number n (may be negative) as callout_data
3841
3842 If the callout function in pcretest receives a non-zero value as
3843 callout_data, it returns that value.
3844
3845 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3846 there were some redundant $(CFLAGS) in commands that are now specified as
3847 $(LINK), which already includes $(CFLAGS).
3848
3849 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3850 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3851 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3852 one-byte characters throughout. Note that case-insensitive matching applies
3853 only to characters whose values are less than 256. PCRE doesn't support the
3854 notion of cases for higher-valued characters.
3855
3856 (i) A character class whose characters are all within 0-255 is handled as
3857 a bit map, and the map is inverted for negative classes. Previously, a
3858 character > 255 always failed to match such a class; however it should
3859 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3860
3861 (ii) A negated character class with a single character < 255 is coded as
3862 "not this character" (OP_NOT). This wasn't working properly when the test
3863 character was multibyte, either singly or repeated.
3864
3865 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3866 mode, for example: \x{100}{2,3}.
3867
3868 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3869 singly or repeated) now correctly test multibyte characters. However,
3870 PCRE doesn't recognize any characters with values greater than 255 as
3871 digits, spaces, or word characters. Such characters always match \D, \S,
3872 and \W, and never match \d, \s, or \w.
3873
3874 (v) Classes may now contain characters and character ranges with values
3875 greater than 255. For example: [ab\x{100}-\x{400}].
3876
3877 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3878 PCRE in UTF-8 mode.
3879
3880 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3881 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3882 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3883 value.)
3884
3885 53. The single man page has become too large. I have therefore split it up into
3886 a number of separate man pages. These also give rise to individual HTML pages;
3887 these are now put in a separate directory, and there is an index.html page that
3888 lists them all. Some hyperlinking between the pages has been installed.
3889
3890 54. Added convenience functions for handling named capturing parentheses.
3891
3892 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3893 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3894 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3895 are faulted.
3896
3897 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3898 calling configure. These values are used when compiling the dftables.c program
3899 which is run to generate the source of the default character tables. They
3900 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3901 you will need to set these values.
3902
3903 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3904
3905
3906 Version 3.9 02-Jan-02
3907 ---------------------
3908
3909 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3910
3911 2. If --disable-static was given, the building process failed when trying to
3912 build pcretest and pcregrep. (For some reason it was using libtool to compile
3913 them, which is not right, as they aren't part of the library.)
3914
3915
3916 Version 3.8 18-Dec-01
3917 ---------------------
3918
3919 1. The experimental UTF-8 code was completely screwed up. It was packing the
3920 bytes in the wrong order. How dumb can you get?
3921
3922
3923 Version 3.7 29-Oct-01
3924 ---------------------
3925
3926 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3927 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3928 this didn't happen under Solaris 8, where I normally test things.
3929
3930 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3931 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3932 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3933 this fix an hour or so after the initial 3.7 release.)
3934
3935
3936 Version 3.6 23-Oct-01
3937 ---------------------
3938
3939 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3940 offsets passed as NULL with zero offset count.
3941
3942 2. The config.guess and config.sub files had not been updated when I moved to
3943 the latest autoconf.
3944
3945
3946 Version 3.5 15-Aug-01
3947 ---------------------
3948
3949 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3950 had been forgotten.
3951
3952 2. By using declared but undefined structures, we can avoid using "void"
3953 definitions in pcre.h while keeping the internal definitions of the structures
3954 private.
3955
3956 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3957 user point of view, this means that both static and shared libraries are built
3958 by default, but this can be individually controlled. More of the work of
3959 handling this static/shared cases is now inside libtool instead of PCRE's make
3960 file.
3961
3962 4. The pcretest utility is now installed along with pcregrep because it is
3963 useful for users (to test regexs) and by doing this, it automatically gets
3964 relinked by libtool. The documentation has been turned into a man page, so
3965 there are now .1, .txt, and .html versions in /doc.
3966
3967 5. Upgrades to pcregrep:
3968 (i) Added long-form option names like gnu grep.
3969 (ii) Added --help to list all options with an explanatory phrase.
3970 (iii) Added -r, --recursive to recurse into sub-directories.
3971 (iv) Added -f, --file to read patterns from a file.
3972
3973 6. pcre_exec() was referring to its "code" argument before testing that
3974 argument for NULL (and giving an error if it was NULL).
3975
3976 7. Upgraded Makefile.in to allow for compiling in a different directory from
3977 the source directory.
3978
3979 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3980 options bits, the pointer it was passed was to an int instead of to an unsigned
3981 long int. This mattered only on 64-bit systems.
3982
3983 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3984 generated) instead of pcre.in, which it its source. Also made the same change
3985 in several of the .c files.
3986
3987 10. A new release of gcc defines printf() as a macro, which broke pcretest
3988 because it had an ifdef in the middle of a string argument for printf(). Fixed
3989 by using separate calls to printf().
3990
3991 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3992 script, to force use of CR or LF instead of \n in the source. On non-Unix
3993 systems, the value can be set in config.h.
3994
3995 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3996 absolute limit. Changed the text of the error message to make this clear, and
3997 likewise updated the man page.
3998
3999 13. The limit of 99 on the number of capturing subpatterns has been removed.
4000 The new limit is 65535, which I hope will not be a "real" limit.
4001
4002
4003 Version 3.4 22-Aug-00
4004 ---------------------
4005
4006 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4007
4008 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4009
4010
4011 Version 3.3 01-Aug-00
4012 ---------------------
4013
4014 1. If an octal character was given, but the value was greater than \377, it
4015 was not getting masked to the least significant bits, as documented. This could
4016 lead to crashes in some systems.
4017
4018 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4019 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4020
4021 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4022 These just pass their arguments on to (pcre_free)(), but they are provided
4023 because some uses of PCRE bind it to non-C systems that can call its functions,
4024 but cannot call free() or pcre_free() directly.
4025
4026 4. Add "make test" as a synonym for "make check". Corrected some comments in
4027 the Makefile.
4028
4029 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4030 Makefile.
4031
4032 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4033 command called pgrep for grepping around the active processes.
4034
4035 7. Added the beginnings of support for UTF-8 character strings.
4036
4037 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4038 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4039 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4040 out for the ar command.)
4041
4042
4043 Version 3.2 12-May-00
4044 ---------------------
4045
4046 This is purely a bug fixing release.
4047
4048 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4049 of ZA. This was just one example of several cases that could provoke this bug,
4050 which was introduced by change 9 of version 2.00. The code for breaking
4051 infinite loops after an iteration that matches an empty string was't working
4052 correctly.
4053
4054 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4055 when matched against abbab (for example). After matching an empty string, it
4056 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4057 caused it to match further down the string than it should.
4058
4059 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4060 was there because it doesn't seem to be needed, and it causes trouble on some
4061 systems, as it is not a Standard C header. It has been removed.
4062
4063 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4064 were reported on the Macintosh. The changes were from
4065
4066 while ((c = *(++ptr)) != 0 && c != '\n');
4067 to
4068 while ((c = *(++ptr)) != 0 && c != '\n') ;
4069
4070 Totally extraordinary, but if that's what it takes...
4071
4072 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4073 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4074 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4075 assumes the way PCRE uses memmove() (always moving upwards).
4076
4077 6. PCRE is being used in one environment where strchr() is not available. There
4078 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4079 faster code anyway.
4080
4081
4082 Version 3.1 09-Feb-00
4083 ---------------------
4084
4085 The only change in this release is the fixing of some bugs in Makefile.in for
4086 the "install" target:
4087
4088 (1) It was failing to install pcreposix.h.
4089
4090 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4091
4092
4093 Version 3.0 01-Feb-00
4094 ---------------------
4095
4096 1. Add support for the /+ modifier to perltest (to output $` like it does in
4097 pcretest).
4098
4099 2. Add support for the /g modifier to perltest.
4100
4101 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4102 matches null strings.
4103
4104 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4105 pattern. Perl treats empty patterns specially - it reuses the most recent
4106 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4107 effect.
4108
4109 5. The POSIX interface was broken in that it was just handing over the POSIX
4110 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4111 required a bigger vector, with some working space on the end. This means that
4112 the POSIX wrapper now has to get and free some memory, and copy the results.
4113
4114 6. Added some simple autoconf support, placing the test data and the
4115 documentation in separate directories, re-organizing some of the
4116 information files, and making it build pcre-config (a GNU standard). Also added
4117 libtool support for building PCRE as a shared library, which is now the
4118 default.
4119
4120 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4121 09 are not valid octal constants. Single digits will be used for minor values
4122 less than 10.
4123
4124 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4125 existing programs that set these in the POSIX interface can use PCRE without
4126 modification.
4127
4128 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4129 return all that pcre_info() returns, plus additional data. The pcre_info()
4130 function is retained for compatibility, but is considered to be obsolete.
4131
4132 10. Added experimental recursion feature (?R) to handle one common case that
4133 Perl 5.6 will be able to do with (?p{...}).
4134
4135 11. Added support for POSIX character classes like [:alpha:], which Perl is
4136 adopting.
4137
4138
4139 Version 2.08 31-Aug-99
4140 ----------------------
4141
4142 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4143 trying to match at the startoffset position, but instead was moving forward to
4144 the next newline as if a previous match had failed.
4145
4146 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4147 and could get into a loop if a null string was matched other than at the start
4148 of the subject.
4149
4150 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4151 be distinguished at compile time, and for completeness also added PCRE_DATE.
4152
4153 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4154 in GnuWin32 environments.
4155
4156
4157 Version 2.07 29-Jul-99
4158 ----------------------
4159
4160 1. The documentation is now supplied in plain text form and HTML as well as in
4161 the form of man page sources.
4162
4163 2. C++ compilers don't like assigning (void *) values to other pointer types.
4164 In particular this affects malloc(). Although there is no problem in Standard
4165 C, I've put in casts to keep C++ compilers happy.
4166
4167 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4168 should be (const char *).
4169
4170 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4171 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4172 However, I haven't made this a standard facility. The documentation doesn't
4173 mention it, and the Makefile doesn't support it.
4174
4175 5. The Makefile now contains an "install" target, with editable destinations at
4176 the top of the file. The pcretest program is not installed.
4177
4178 6. pgrep -V now gives the PCRE version number and date.
4179
4180 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4181 causing the entire string to be ignored, instead of just the last character.
4182
4183 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4184 non-matching string, it can take a very, very long time, even for strings of
4185 quite modest length, because of the nested recursion. PCRE now does better in
4186 some of these cases. It does this by remembering the last required literal
4187 character in the pattern, and pre-searching the subject to ensure it is present
4188 before running the real match. In other words, it applies a heuristic to detect
4189 some types of certain failure quickly, and in the above example, if presented
4190 with a string that has no trailing " it gives "no match" very quickly.
4191
4192 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4193 other alternatives are tried instead.
4194
4195
4196 Version 2.06 09-Jun-99
4197 ----------------------
4198
4199 1. Change pcretest's output for amount of store used to show just the code
4200 space, because the remainder (the data block) varies in size between 32-bit and
4201 64-bit systems.
4202
4203 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4204 start matching at. This allows lookbehinds to work when searching for multiple
4205 occurrences in a string.
4206
4207 3. Added additional options to pcretest for testing multiple occurrences:
4208
4209 /+ outputs the rest of the string that follows a match
4210 /g loops for multiple occurrences, using the new startoffset argument
4211 /G loops for multiple occurrences by passing an incremented pointer
4212
4213 4. PCRE wasn't doing the "first character" optimization for patterns starting
4214 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4215 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4216 the letter 'x'. On long subject strings, this gives a significant speed-up.
4217
4218
4219 Version 2.05 21-Apr-99
4220 ----------------------
4221
4222 1. Changed the type of magic_number from int to long int so that it works
4223 properly on 16-bit systems.
4224
4225 2. Fixed a bug which caused patterns starting with .* not to work correctly
4226 when the subject string contained newline characters. PCRE was assuming
4227 anchoring for such patterns in all cases, which is not correct because .* will
4228 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4229 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4230 must be retried after every newline in the subject.
4231
4232
4233 Version 2.04 18-Feb-99
4234 ----------------------
4235
4236 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4237 computation of the store needed to hold the pattern was incorrect (too large).
4238 If such patterns were nested a few deep, this could multiply and become a real
4239 problem.
4240
4241 2. Added /M option to pcretest to show the memory requirement of a specific
4242 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4243
4244 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4245 compiled in such a way that the backtracking after subsequent failure was
4246 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4247 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4248
4249
4250 Version 2.03 02-Feb-99
4251 ----------------------
4252
4253 1. Fixed typo and small mistake in man page.
4254
4255 2. Added 4th condition (GPL supersedes if conflict) and created separate
4256 LICENCE file containing the conditions.
4257
4258 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4259 Perl, that is the internal \ allows the delimiter to be included in the
4260 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4261 the final delimiter, add \ to the end of the pattern (to test the error).
4262
4263 4. Added the convenience functions for extracting substrings after a successful
4264 match. Updated pcretest to make it able to test these functions.
4265
4266
4267 Version 2.02 14-Jan-99
4268 ----------------------
4269
4270 1. Initialized the working variables associated with each extraction so that
4271 their saving and restoring doesn't refer to uninitialized store.
4272
4273 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4274 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4275 fix the problem.
4276
4277 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4278 calls, and wasn't printing the correct value for compiling calls. Increased the
4279 default value of LOOPREPEAT, and the number of significant figures in the
4280 times.
4281
4282 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4283
4284 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4285 a building problem on Windows NT with a FAT file system.
4286
4287
4288 Version 2.01 21-Oct-98
4289 ----------------------
4290
4291 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4292 to character tables built by pcre_maketables() in the current locale. If NULL
4293 is passed, the default tables are used.
4294
4295
4296 Version 2.00 24-Sep-98
4297 ----------------------
4298
4299 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4300 it any more.
4301
4302 2. Allow quantification of (?>) groups, and make it work correctly.
4303
4304 3. The first character computation wasn't working for (?>) groups.
4305
4306 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4307 end of the subject) and add 5.005's \z, which really does match only at the
4308 very end of the subject.
4309
4310 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4311
4312 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4313 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4314 localized options. All options to pcre_study() were also removed.
4315
4316 7. Add other new features from 5.005:
4317
4318 $(?<= positive lookbehind
4319 $(?<! negative lookbehind
4320 (?imsx-imsx) added the unsetting capability
4321 such a setting is global if at outer level; local otherwise
4322 (?imsx-imsx:) non-capturing groups with option setting
4323 (?(cond)re|re) conditional pattern matching
4324
4325 A backreference to itself in a repeated group matches the previous
4326 captured string.
4327
4328 8. General tidying up of studying (both automatic and via "study")
4329 consequential on the addition of new assertions.
4330
4331 9. As in 5.005, unlimited repeated groups that could match an empty substring
4332 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4333 runtime if any iteration does actually match an empty substring.
4334
4335 10. Include the RunTest script in the distribution.
4336
4337 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4338 discrepancies, some of which were old and were also with respect to 5.004. They
4339 have now been fixed.
4340
4341
4342 Version 1.09 28-Apr-98
4343 ----------------------
4344
4345 1. A negated single character class followed by a quantifier with a minimum
4346 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4347 program crashes, or just wrong answers. This did not apply to negated classes
4348 containing more than one character, or to minima other than one.
4349
4350
4351 Version 1.08 27-Mar-98
4352 ----------------------
4353
4354 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4355
4356 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4357 latter must appear before anything that relies on it in the pattern.
4358
4359
4360 Version 1.07 16-Feb-98
4361 ----------------------
4362
4363 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4364 repeat of a potentially empty string).
4365
4366
4367 Version 1.06 23-Jan-98
4368 ----------------------
4369
4370 1. Added Markus Oberhumer's little patches for C++.
4371
4372 2. Literal strings longer than 255 characters were broken.
4373
4374
4375 Version 1.05 23-Dec-97
4376 ----------------------
4377
4378 1. Negated character classes containing more than one character were failing if
4379 PCRE_CASELESS was set at run time.
4380
4381
4382 Version 1.04 19-Dec-97
4383 ----------------------
4384
4385 1. Corrected the man page, where some "const" qualifiers had been omitted.
4386
4387 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4388 input syntax.
4389
4390 3. Fixed memory leak which occurred when a regex with back references was
4391 matched with an offsets vector that wasn't big enough. The temporary memory
4392 that is used in this case wasn't being freed if the match failed.
4393
4394 4. Tidied pcretest to ensure it frees memory that it gets.
4395
4396 5. Temporary memory was being obtained in the case where the passed offsets
4397 vector was exactly big enough.
4398
4399 6. Corrected definition of offsetof() from change 5 below.
4400
4401 7. I had screwed up change 6 below and broken the rules for the use of
4402 setjmp(). Now fixed.
4403
4404
4405 Version 1.03 18-Dec-97
4406 ----------------------
4407
4408 1. A erroneous regex with a missing opening parenthesis was correctly
4409 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4410 on some systems.
4411
4412 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4413 it was reported that one broken compiler failed on the former because "code" is
4414 also an independent variable.
4415
4416 3. The erroneous regex a[]b caused an array overrun reference.
4417
4418 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4419 fail on data ending with that character. (It was going on too far, and checking
4420 the next character, typically a binary zero.) This was specific to the
4421 optimized code for single-character negative classes.
4422
4423 5. Added a contributed patch from the TIN world which does the following:
4424
4425 + Add an undef for memmove, in case the the system defines a macro for it.
4426
4427 + Add a definition of offsetof(), in case there isn't one. (I don't know
4428 the reason behind this - offsetof() is part of the ANSI standard - but
4429 it does no harm).
4430
4431 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4432 most of the places where whitespace preceded '#'. I have given up and
4433 allowed the remaining 2 cases to be at the margin.
4434
4435 + Rename some variables in pcre to eliminate shadowing. This seems very
4436 pedantic, but does no harm, of course.
4437
4438 6. Moved the call to setjmp() into its own function, to get rid of warnings
4439 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4440
4441 7. Constructs such as \d{8,} were compiling into the equivalent of
4442 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4443 outcome, but in this particular case used more store than had been allocated,
4444 which caused the bug to be discovered because it threw up an internal error.
4445
4446 8. The debugging code in both pcre and pcretest for outputting the compiled
4447 form of a regex was going wrong in the case of back references followed by
4448 curly-bracketed repeats.
4449
4450
4451 Version 1.02 12-Dec-97
4452 ----------------------
4453
4454 1. Typos in pcre.3 and comments in the source fixed.
4455
4456 2. Applied a contributed patch to get rid of places where it used to remove
4457 'const' from variables, and fixed some signed/unsigned and uninitialized
4458 variable warnings.
4459
4460 3. Added the "runtest" target to Makefile.
4461
4462 4. Set default compiler flag to -O2 rather than just -O.
4463
4464
4465 Version 1.01 19-Nov-97
4466 ----------------------
4467
4468 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4469 like /([ab]*)*/, that is, for classes with more than one character in them.
4470
4471 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4472 as /((?>a*))*/ (a PCRE_EXTRA facility).
4473
4474
4475 Version 1.00 18-Nov-97
4476 ----------------------
4477
4478 1. Added compile-time macros to support systems such as SunOS4 which don't have
4479 memmove() or strerror() but have other things that can be used instead.
4480
4481 2. Arranged that "make clean" removes the executables.
4482
4483
4484 Version 0.99 27-Oct-97
4485 ----------------------
4486
4487 1. Fixed bug in code for optimizing classes with only one character. It was
4488 initializing a 32-byte map regardless, which could cause it to run off the end
4489 of the memory it had got.
4490
4491 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4492
4493
4494 Version 0.98 22-Oct-97
4495 ----------------------
4496
4497 1. Fixed bug in code for handling temporary memory usage when there are more
4498 back references than supplied space in the ovector. This could cause segfaults.
4499
4500
4501 Version 0.97 21-Oct-97
4502 ----------------------
4503
4504 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4505
4506 2. Optimized negated single characters not to use a bit map.
4507
4508 3. Brought error texts together as macro definitions; clarified some of them;
4509 fixed one that was wrong - it said "range out of order" when it meant "invalid
4510 escape sequence".
4511
4512 4. Changed some char * arguments to const char *.
4513
4514 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4515
4516 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4517 pcretest.
4518
4519
4520 Version 0.96 16-Oct-97
4521 ----------------------
4522
4523 1. Added a simple "pgrep" utility to the distribution.
4524
4525 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4526 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4527 where "ddd" means "one or more decimal digits".
4528
4529 3. Fixed serious bug. If a pattern had a back reference, but the call to
4530 pcre_exec() didn't supply a large enough ovector to record the related
4531 identifying subpattern, the match always failed. PCRE now remembers the number
4532 of the largest back reference, and gets some temporary memory in which to save
4533 the offsets during matching if necessary, in order to ensure that
4534 backreferences always work.
4535
4536 4. Increased the compatibility with Perl in a number of ways:
4537
4538 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4539 to request this handling. The option can be set at compile or exec time.
4540
4541 (b) $ matches before a terminating newline by default; an option
4542 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4543 mode). The option can be set at compile or exec time.
4544
4545 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4546 the same as Perl's. If the decimal number it represents is less than 10
4547 or there aren't that many previous left capturing parentheses, an octal
4548 escape is read. Inside a character class, it's always an octal escape,
4549 even if it is a single digit.
4550
4551 (d) An escaped but undefined alphabetic character is taken as a literal,
4552 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4553 escapes.
4554
4555 (e) {0} is now permitted. (The previous item is removed from the compiled
4556 pattern).
4557
4558 5. Changed all the names of code files so that the basic parts are no longer
4559 than 10 characters, and abolished the teeny "globals.c" file.
4560
4561 6. Changed the handling of character classes; they are now done with a 32-byte
4562 bit map always.
4563
4564 7. Added the -d and /D options to pcretest to make it possible to look at the
4565 internals of compilation without having to recompile pcre.
4566
4567
4568 Version 0.95 23-Sep-97
4569 ----------------------
4570
4571 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4572 \x20 at the start of a run of normal characters. These were being treated as
4573 real characters, instead of the source characters being re-checked.
4574
4575
4576 Version 0.94 18-Sep-97
4577 ----------------------
4578
4579 1. The functions are now thread-safe, with the caveat that the global variables
4580 containing pointers to malloc() and free() or alternative functions are the
4581 same for all threads.
4582
4583 2. Get pcre_study() to generate a bitmap of initial characters for non-
4584 anchored patterns when this is possible, and use it if passed to pcre_exec().
4585
4586
4587 Version 0.93 15-Sep-97
4588 ----------------------
4589
4590 1. /(b)|(:+)/ was computing an incorrect first character.
4591
4592 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4593 but not actually doing anything yet.
4594
4595 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4596 as Perl does (e.g. [-az] or [az-]).
4597
4598 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4599 all possible positions.
4600
4601 5. Split up into different modules to avoid including unneeded functions in a
4602 compiled binary. However, compile and exec are still in one module. The "study"
4603 function is split off.
4604
4605 6. The character tables are now in a separate module whose source is generated
4606 by an auxiliary program - but can then be edited by hand if required. There are
4607 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4608 toupper() in the code.
4609
4610 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4611 make them global. Abolish the function for setting them, as the caller can now
4612 set them directly.
4613
4614
4615 Version 0.92 11-Sep-97
4616 ----------------------
4617
4618 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4619 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4620
4621 2. Caseless matching was not working in character classes if the characters in
4622 the pattern were in upper case.
4623
4624 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4625
4626 4. Make PCRE_ANCHORED public and accept as a compile option.
4627
4628 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4629 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4630 pass them.
4631
4632 6. Give an error if bad option bits passed at compile or run time.
4633
4634 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4635 pcretest to cause it to pass that flag.
4636
4637 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4638 options, and the first character, if set.
4639
4640 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4641
4642
4643 Version 0.91 10-Sep-97
4644 ----------------------
4645
4646 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4647 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4648
4649 2. PCRE was looping on encountering an indefinitely repeated back reference to
4650 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4651 Perl does - treats the match as successful.
4652
4653 ****

Properties

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

webmaster@exim.org
ViewVC Help
Powered by ViewVC 1.1.12