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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 889 - (show annotations)
Tue Jan 17 14:51:50 2012 UTC (3 years, 6 months ago) by ph10
File size: 225547 byte(s)
Error occurred while calculating annotation data.
More portable way of linking pcretest with -lreadline.
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 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
53 -lreadline.
54
55
56 Version 8.21 12-Dec-2011
57 ------------------------
58
59 1. Updating the JIT compiler.
60
61 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
62 are added as well.
63
64 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
65 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
66 calling _pcre_jit_exec. Some extra comments are added.
67
68 4. (*MARK) settings inside atomic groups that do not contain any capturing
69 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
70 was introduced by change 18 for 8.20.
71
72 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
73 ECMA-262 standard.
74
75 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
76 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
77 This bug was probably introduced by change 9 of 8.13.
78
79 7. While fixing 6 above, I noticed that a number of other items were being
80 incorrectly rejected as "not fixed length". This arose partly because newer
81 opcodes had not been added to the fixed-length checking code. I have (a)
82 corrected the bug and added tests for these items, and (b) arranged for an
83 error to occur if an unknown opcode is encountered while checking for fixed
84 length instead of just assuming "not fixed length". The items that were
85 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
86 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
87 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
88
89 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
90 being incorrectly compiled and would have given unpredicatble results.
91
92 9. A possessively repeated subpattern with minimum repeat count greater than
93 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
94 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
95 the first (A) could occur when it should not.
96
97 10. Add a cast and remove a redundant test from the code.
98
99 11. JIT should use pcre_malloc/pcre_free for allocation.
100
101 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
102 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
103 is anything other than /usr, -L is still shown).
104
105 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
106
107 14. Perl does not support \N without a following name in a [] class; PCRE now
108 also gives an error.
109
110 15. If a forward reference was repeated with an upper limit of around 2000,
111 it caused the error "internal error: overran compiling workspace". The
112 maximum number of forward references (including repeats) was limited by the
113 internal workspace, and dependent on the LINK_SIZE. The code has been
114 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
115 the default depends on LINK_SIZE. There is a new upper limit (for safety)
116 of around 200,000 forward references. While doing this, I also speeded up
117 the filling in of repeated forward references.
118
119 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
120 incorrectly expecting the subject to contain another "a" after the start.
121
122 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
123 in the match, the SKIP should be ignored. This was not happening; instead
124 the SKIP was being treated as NOMATCH. For patterns such as
125 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
126 tested.
127
128 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
129 now much more compatible with Perl, in particular in cases where the result
130 is a non-match for a non-anchored pattern. For example, if
131 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
132 "m", where previously it did not return a name. A side effect of this
133 change is that for partial matches, the last encountered mark name is
134 returned, as for non matches. A number of tests that were previously not
135 Perl-compatible have been moved into the Perl-compatible test files. The
136 refactoring has had the pleasing side effect of removing one argument from
137 the match() function, thus reducing its stack requirements.
138
139 19. If the /S+ option was used in pcretest to study a pattern using JIT,
140 subsequent uses of /S (without +) incorrectly behaved like /S+.
141
142 21. Retrieve executable code size support for the JIT compiler and fixing
143 some warnings.
144
145 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
146 not work when the shorter character appeared right at the end of the
147 subject string.
148
149 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
150 systems.
151
152 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
153 output it when the /M option is used in pcretest.
154
155 25. The CheckMan script was not being included in the distribution. Also, added
156 an explicit "perl" to run Perl scripts from the PrepareRelease script
157 because this is reportedly needed in Windows.
158
159 26. If study data was being save in a file and studying had not found a set of
160 "starts with" bytes for the pattern, the data written to the file (though
161 never used) was taken from uninitialized memory and so caused valgrind to
162 complain.
163
164 27. Updated RunTest.bat as provided by Sheri Pierce.
165
166 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
167
168 29. Computation of memory usage for the table of capturing group names was
169 giving an unnecessarily large value.
170
171
172 Version 8.20 21-Oct-2011
173 ------------------------
174
175 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
176 a POSIX class. After further experiments with Perl, which convinced me that
177 Perl has bugs and confusions, a closing square bracket is no longer allowed
178 in a POSIX name. This bug also affected patterns with classes that started
179 with full stops.
180
181 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
182 captured substring, but while checking the failing first alternative,
183 substring 1 is temporarily captured. If the output vector supplied to
184 pcre_exec() was not big enough for this capture, the yield of the function
185 was still zero ("insufficient space for captured substrings"). This cannot
186 be totally fixed without adding another stack variable, which seems a lot
187 of expense for a edge case. However, I have improved the situation in cases
188 such as /(a)(b)x|abc/ matched against "abc", where the return code
189 indicates that fewer than the maximum number of slots in the ovector have
190 been set.
191
192 3. Related to (2) above: when there are more back references in a pattern than
193 slots in the output vector, pcre_exec() uses temporary memory during
194 matching, and copies in the captures as far as possible afterwards. It was
195 using the entire output vector, but this conflicts with the specification
196 that only 2/3 is used for passing back captured substrings. Now it uses
197 only the first 2/3, for compatibility. This is, of course, another edge
198 case.
199
200 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
201 main code base, and can be used by building with --enable-jit. When this is
202 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
203 runtime --no-jit option is given.
204
205 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
206 ovector, the return from the function was zero, implying that there were
207 other matches that did not fit. The correct "exactly full" value is now
208 returned.
209
210 6. If a subpattern that was called recursively or as a subroutine contained
211 (*PRUNE) or any other control that caused it to give a non-standard return,
212 invalid errors such as "Error -26 (nested recursion at the same subject
213 position)" or even infinite loops could occur.
214
215 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
216 computing the minimum length on reaching *ACCEPT, and so ended up with the
217 wrong value of 1 rather than 0. Further investigation indicates that
218 computing a minimum subject length in the presence of *ACCEPT is difficult
219 (think back references, subroutine calls), and so I have changed the code
220 so that no minimum is registered for a pattern that contains *ACCEPT.
221
222 8. If (*THEN) was present in the first (true) branch of a conditional group,
223 it was not handled as intended. [But see 16 below.]
224
225 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
226 Sheri Pierce.
227
228 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
229 the first byte in a match must be "a".
230
231 11. Change 17 for 8.13 increased the recursion depth for patterns like
232 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
233 pattern contains any instances of (*THEN). If it does not, the old
234 optimizations are restored. It would be nice to do this on a per-group
235 basis, but at the moment that is not feasible.
236
237 12. In some environments, the output of pcretest -C is CRLF terminated. This
238 broke RunTest's code that checks for the link size. A single white space
239 character after the value is now allowed for.
240
241 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
242 For "fr", it uses the Windows-specific input and output files.
243
244 14. If (*THEN) appeared in a group that was called recursively or as a
245 subroutine, it did not work as intended. [But see next item.]
246
247 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
248 pattern fragments (but not containing any | characters). If A and B are
249 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
250 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
251 D. In other words, Perl considers parentheses that do not contain any |
252 characters to be part of a surrounding alternative, whereas PCRE was
253 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
254 differently. PCRE now behaves in the same way as Perl, except in the case
255 of subroutine/recursion calls such as (?1) which have in any case always
256 been different (but PCRE had them first :-).
257
258 16. Related to 15 above: Perl does not treat the | in a conditional group as
259 creating alternatives. Such a group is treated in the same way as an
260 ordinary group without any | characters when processing (*THEN). PCRE has
261 been changed to match Perl's behaviour.
262
263 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
264 RunGrepTest script failed.
265
266 18. Change 22 for version 13 caused atomic groups to use more stack. This is
267 inevitable for groups that contain captures, but it can lead to a lot of
268 stack use in large patterns. The old behaviour has been restored for atomic
269 groups that do not contain any capturing parentheses.
270
271 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
272 suppress the check for a minimum subject length at run time. (If it was
273 given to pcre_exec() or pcre_dfa_exec() it did work.)
274
275 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
276 fail to work when decoding hex characters in data strings in EBCDIC
277 environments.
278
279 21. It appears that in at least one Mac OS environment, the isxdigit() function
280 is implemented as a macro that evaluates to its argument more than once,
281 contravening the C 90 Standard (I haven't checked a later standard). There
282 was an instance in pcretest which caused it to go wrong when processing
283 \x{...} escapes in subject strings. The has been rewritten to avoid using
284 things like p++ in the argument of isxdigit().
285
286
287 Version 8.13 16-Aug-2011
288 ------------------------
289
290 1. The Unicode data tables have been updated to Unicode 6.0.0.
291
292 2. Two minor typos in pcre_internal.h have been fixed.
293
294 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
295 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
296 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
297
298 4. There were a number of related bugs in the code for matching backrefences
299 caselessly in UTF-8 mode when codes for the characters concerned were
300 different numbers of bytes. For example, U+023A and U+2C65 are an upper
301 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
302 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
303 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
304 2-byte code at the end of the subject (it thought there wasn't enough data
305 left).
306
307 5. Comprehensive information about what went wrong is now returned by
308 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
309 as the output vector has at least 2 elements. The offset of the start of
310 the failing character and a reason code are placed in the vector.
311
312 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
313 now returned is for the first byte of the failing character, instead of the
314 last byte inspected. This is an incompatible change, but I hope it is small
315 enough not to be a problem. It makes the returned offset consistent with
316 pcre_exec() and pcre_dfa_exec().
317
318 7. pcretest now gives a text phrase as well as the error number when
319 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
320 failure, the offset and reason code are output.
321
322 8. When \R was used with a maximizing quantifier it failed to skip backwards
323 over a \r\n pair if the subsequent match failed. Instead, it just skipped
324 back over a single character (\n). This seems wrong (because it treated the
325 two characters as a single entity when going forwards), conflicts with the
326 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
327 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
328 has been changed.
329
330 9. Some internal refactoring has changed the processing so that the handling
331 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
332 time (the PCRE_DOTALL option was changed this way some time ago: version
333 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
334 which was always a bit of a fudge. It also means that there is one less
335 argument for the match() function, which reduces its stack requirements
336 slightly. This change also fixes an incompatibility with Perl: the pattern
337 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
338
339 10. More internal refactoring has drastically reduced the number of recursive
340 calls to match() for possessively repeated groups such as (abc)++ when
341 using pcre_exec().
342
343 11. While implementing 10, a number of bugs in the handling of groups were
344 discovered and fixed:
345
346 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
347 (a|)*(?1) gave a compile-time internal error.
348 ((a|)+)+ did not notice that the outer group could match an empty string.
349 (^a|^)+ was not marked as anchored.
350 (.*a|.*)+ was not marked as matching at start or after a newline.
351
352 12. Yet more internal refactoring has removed another argument from the match()
353 function. Special calls to this function are now indicated by setting a
354 value in a variable in the "match data" data block.
355
356 13. Be more explicit in pcre_study() instead of relying on "default" for
357 opcodes that mean there is no starting character; this means that when new
358 ones are added and accidentally left out of pcre_study(), testing should
359 pick them up.
360
361 14. The -s option of pcretest has been documented for ages as being an old
362 synonym of -m (show memory usage). I have changed it to mean "force study
363 for every regex", that is, assume /S for every regex. This is similar to -i
364 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
365 using it. It makes it easier to run collections of tests with and without
366 study enabled, and thereby test pcre_study() more easily. All the standard
367 tests are now run with and without -s (but some patterns can be marked as
368 "never study" - see 20 below).
369
370 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
371 restoration of the capturing data to the outer values was not happening
372 correctly.
373
374 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
375 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
376 pattern had matched an empty string, and so incorrectly returned a no
377 match.
378
379 17. There was optimizing code for the last branch of non-capturing parentheses,
380 and also for the obeyed branch of a conditional subexpression, which used
381 tail recursion to cut down on stack usage. Unfortunately, now that there is
382 the possibility of (*THEN) occurring in these branches, tail recursion is
383 no longer possible because the return has to be checked for (*THEN). These
384 two optimizations have therefore been removed. [But see 8.20/11 above.]
385
386 18. If a pattern containing \R was studied, it was assumed that \R always
387 matched two bytes, thus causing the minimum subject length to be
388 incorrectly computed because \R can also match just one byte.
389
390 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
391 was incorrectly computed.
392
393 20. If /S is present twice on a test pattern in pcretest input, it now
394 *disables* studying, thereby overriding the use of -s on the command line
395 (see 14 above). This is necessary for one or two tests to keep the output
396 identical in both cases.
397
398 21. When (*ACCEPT) was used in an assertion that matched an empty string and
399 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
400
401 22. When an atomic group that contained a capturing parenthesis was
402 successfully matched, but the branch in which it appeared failed, the
403 capturing was not being forgotten if a higher numbered group was later
404 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
405 group 1 to "a", when in fact it should be unset. This applied to multi-
406 branched capturing and non-capturing groups, repeated or not, and also to
407 positive assertions (capturing in negative assertions does not happen
408 in PCRE) and also to nested atomic groups.
409
410 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
411 subject after a captured substring, to make it easier to tell which of a
412 number of identical substrings has been captured.
413
414 24. The way atomic groups are processed by pcre_exec() has been changed so that
415 if they are repeated, backtracking one repetition now resets captured
416 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
417 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
418 "aaa". Previously, it would have been "a". As part of this code
419 refactoring, the way recursive calls are handled has also been changed.
420
421 25. If an assertion condition captured any substrings, they were not passed
422 back unless some other capturing happened later. For example, if
423 (?(?=(a))a) was matched against "a", no capturing was returned.
424
425 26. When studying a pattern that contained subroutine calls or assertions,
426 the code for finding the minimum length of a possible match was handling
427 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
428 group 1 called group 2 while simultaneously a separate group 2 called group
429 1). A stack overflow occurred in this case. I have fixed this by limiting
430 the recursion depth to 10.
431
432 27. Updated RunTest.bat in the distribution to the version supplied by Tom
433 Fortmann. This supports explicit test numbers on the command line, and has
434 argument validation and error reporting.
435
436 28. An instance of \X with an unlimited repeat could fail if at any point the
437 first character it looked at was a mark character.
438
439 29. Some minor code refactoring concerning Unicode properties and scripts
440 should reduce the stack requirement of match() slightly.
441
442 30. Added the '=' option to pcretest to check the setting of unused capturing
443 slots at the end of the pattern, which are documented as being -1, but are
444 not included in the return count.
445
446 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
447 compiled something random. Now it gives a compile-time error (as does
448 Perl).
449
450 32. A *MARK encountered during the processing of a positive assertion is now
451 recorded and passed back (compatible with Perl).
452
453 33. If --only-matching or --colour was set on a pcregrep call whose pattern
454 had alternative anchored branches, the search for a second match in a line
455 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
456 matched the line "0102" twice. The same bug affected patterns that started
457 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
458 twice.
459
460 34. Previously, PCRE did not allow quantification of assertions. However, Perl
461 does, and because of capturing effects, quantifying parenthesized
462 assertions may at times be useful. Quantifiers are now allowed for
463 parenthesized assertions.
464
465 35. A minor code tidy in pcre_compile() when checking options for \R usage.
466
467 36. \g was being checked for fancy things in a character class, when it should
468 just be a literal "g".
469
470 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
471 appearance of a nested POSIX class supersedes an apparent external class.
472 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
473 unescaped square brackets may also appear as part of class names. For
474 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
475 more like Perl. (But see 8.20/1 above.)
476
477 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
478 was because it thought it was \N{name}, which is not supported).
479
480 39. Add minix to OS list not supporting the -S option in pcretest.
481
482 40. PCRE tries to detect cases of infinite recursion at compile time, but it
483 cannot analyze patterns in sufficient detail to catch mutual recursions
484 such as ((?1))((?2)). There is now a runtime test that gives an error if a
485 subgroup is called recursively as a subpattern for a second time at the
486 same position in the subject string. In previous releases this might have
487 been caught by the recursion limit, or it might have run out of stack.
488
489 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
490 happen only once. PCRE was, however incorrectly giving a compile time error
491 "recursive call could loop indefinitely" because it cannot analyze the
492 pattern in sufficient detail. The compile time test no longer happens when
493 PCRE is compiling a conditional subpattern, but actual runaway loops are
494 now caught at runtime (see 40 above).
495
496 42. It seems that Perl allows any characters other than a closing parenthesis
497 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
498 has been changed to be the same.
499
500 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
501 as not to get warnings when autogen.sh is called. Also changed
502 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
503
504 44. To help people who use pcregrep to scan files containing exceedingly long
505 lines, the following changes have been made:
506
507 (a) The default value of the buffer size parameter has been increased from
508 8K to 20K. (The actual buffer used is three times this size.)
509
510 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
511 PCRE is built.
512
513 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
514 to be set at run time.
515
516 (d) Numerical values in pcregrep options can be followed by K or M, for
517 example --buffer-size=50K.
518
519 (e) If a line being scanned overflows pcregrep's buffer, an error is now
520 given and the return code is set to 2.
521
522 45. Add a pointer to the latest mark to the callout data block.
523
524 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
525 partial match of an empty string instead of no match. This was specific to
526 the use of ".".
527
528 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
529 complete match instead of a partial match. This bug was dependent on both
530 the PCRE_UTF8 and PCRE_DOTALL options being set.
531
532 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
533 starting byte set, because \b was not being ignored.
534
535
536 Version 8.12 15-Jan-2011
537 ------------------------
538
539 1. Fixed some typos in the markup of the man pages, and wrote a script that
540 checks for such things as part of the documentation building process.
541
542 2. On a big-endian 64-bit system, pcregrep did not correctly process the
543 --match-limit and --recursion-limit options (added for 8.11). In
544 particular, this made one of the standard tests fail. (The integer value
545 went into the wrong half of a long int.)
546
547 3. If the --colour option was given to pcregrep with -v (invert match), it
548 did strange things, either producing crazy output, or crashing. It should,
549 of course, ignore a request for colour when reporting lines that do not
550 match.
551
552 4. Another pcregrep bug caused similar problems if --colour was specified with
553 -M (multiline) and the pattern match finished with a line ending.
554
555 5. In pcregrep, when a pattern that ended with a literal newline sequence was
556 matched in multiline mode, the following line was shown as part of the
557 match. This seems wrong, so I have changed it.
558
559 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
560 the check for further matches in the same line (so they could be coloured)
561 to overrun the end of the current line. If another match was found, it was
562 incorrectly shown (and then shown again when found in the next line).
563
564 7. If pcregrep was compiled under Windows, there was a reference to the
565 function pcregrep_exit() before it was defined. I am assuming this was
566 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
567 reported by a user. I've moved the definition above the reference.
568
569
570 Version 8.11 10-Dec-2010
571 ------------------------
572
573 1. (*THEN) was not working properly if there were untried alternatives prior
574 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
575 backtracked to try for "b" instead of moving to the next alternative branch
576 at the same level (in this case, to look for "c"). The Perl documentation
577 is clear that when (*THEN) is backtracked onto, it goes to the "next
578 alternative in the innermost enclosing group".
579
580 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
581 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
582 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
583 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
584 (*THEN).
585
586 3. If \s appeared in a character class, it removed the VT character from
587 the class, even if it had been included by some previous item, for example
588 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
589 of \s, but is part of the POSIX "space" class.)
590
591 4. A partial match never returns an empty string (because you can always
592 match an empty string at the end of the subject); however the checking for
593 an empty string was starting at the "start of match" point. This has been
594 changed to the "earliest inspected character" point, because the returned
595 data for a partial match starts at this character. This means that, for
596 example, /(?<=abc)def/ gives a partial match for the subject "abc"
597 (previously it gave "no match").
598
599 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
600 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
601 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
602 has an implication that the given string is incomplete (because a partial
603 match is preferred over a full match). For this reason, these items now
604 give a partial match in this situation. [Aside: previously, the one case
605 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
606 match rather than a full match, which was wrong by the old rules, but is
607 now correct.]
608
609 6. There was a bug in the handling of #-introduced comments, recognized when
610 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
611 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
612 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
613 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
614 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
615 places in pcre_compile().
616
617 7. Related to (6) above, when pcre_compile() was skipping #-introduced
618 comments when looking ahead for named forward references to subpatterns,
619 the only newline sequence it recognized was NL. It now handles newlines
620 according to the set newline convention.
621
622 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
623 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
624 cater for a lack of strerror(). These oversights have been fixed.
625
626 9. Added --match-limit and --recursion-limit to pcregrep.
627
628 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
629
630 11. When the -o option was used, pcregrep was setting a return code of 1, even
631 when matches were found, and --line-buffered was not being honoured.
632
633 12. Added an optional parentheses number to the -o and --only-matching options
634 of pcregrep.
635
636 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
637 can match an empty string. The code to do it in pcretest and pcredemo
638 needed fixing:
639
640 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
641 only one byte after an empty string match just before CRLF (this case
642 just got forgotten; "any" and "anycrlf" were OK).
643
644 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
645 mode when an empty string match preceded an ASCII character followed by
646 a non-ASCII character. (The code for advancing by one character rather
647 than one byte was nonsense.)
648
649 (c) The pcredemo.c sample program did not have any code at all to handle
650 the cases when CRLF is a valid newline sequence.
651
652 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
653 as a starting offset was within the subject string. There is now a new
654 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
655 negative or greater than the length of the string. In order to test this,
656 pcretest is extended to allow the setting of negative starting offsets.
657
658 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
659 starting offset points to the beginning of a UTF-8 character was
660 unnecessarily clumsy. I tidied it up.
661
662 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
663 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
664
665 17. Nobody had reported that the --include_dir option, which was added in
666 release 7.7 should have been called --include-dir (hyphen, not underscore)
667 for compatibility with GNU grep. I have changed it to --include-dir, but
668 left --include_dir as an undocumented synonym, and the same for
669 --exclude-dir, though that is not available in GNU grep, at least as of
670 release 2.5.4.
671
672 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
673 characters from a string of bytes) have been redefined so as not to use
674 loops, in order to improve performance in some environments. At the same
675 time, I abstracted some of the common code into auxiliary macros to save
676 repetition (this should not affect the compiled code).
677
678 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
679 compile-time error is now given if \c is not followed by an ASCII
680 character, that is, a byte less than 128. (In EBCDIC mode, the code is
681 different, and any byte value is allowed.)
682
683 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
684 START_OPTIMIZE option, which is now allowed at compile time - but just
685 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
686 to pcregrep and other applications that have no direct access to PCRE
687 options. The new /Y option in pcretest sets this option when calling
688 pcre_compile().
689
690 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
691 back references. Groups containing recursive back references were forced to
692 be atomic by that change, but in the case of named groups, the amount of
693 memory required was incorrectly computed, leading to "Failed: internal
694 error: code overflow". This has been fixed.
695
696 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
697 pcretest.c, to avoid build problems in some Borland environments.
698
699
700 Version 8.10 25-Jun-2010
701 ------------------------
702
703 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
704 THEN.
705
706 2. (*ACCEPT) was not working when inside an atomic group.
707
708 3. Inside a character class, \B is treated as a literal by default, but
709 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
710 causes the error). The code is unchanged, but I tidied the documentation.
711
712 4. Inside a character class, PCRE always treated \R and \X as literals,
713 whereas Perl faults them if its -w option is set. I have changed PCRE so
714 that it faults them when PCRE_EXTRA is set.
715
716 5. Added support for \N, which always matches any character other than
717 newline. (It is the same as "." when PCRE_DOTALL is not set.)
718
719 6. When compiling pcregrep with newer versions of gcc which may have
720 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
721 declared with attribute warn_unused_result" were given. Just casting the
722 result to (void) does not stop the warnings; a more elaborate fudge is
723 needed. I've used a macro to implement this.
724
725 7. Minor change to pcretest.c to avoid a compiler warning.
726
727 8. Added four artifical Unicode properties to help with an option to make
728 \s etc use properties (see next item). The new properties are: Xan
729 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
730
731 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
732 use Unicode properties. (*UCP) at the start of a pattern can be used to set
733 this option. Modified pcretest to add /W to test this facility. Added
734 REG_UCP to make it available via the POSIX interface.
735
736 10. Added --line-buffered to pcregrep.
737
738 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
739 studied, and the match started with a letter with a code point greater than
740 127 whose first byte was different to the first byte of the other case of
741 the letter, the other case of this starting letter was not recognized
742 (#976).
743
744 12. If a pattern that was studied started with a repeated Unicode property
745 test, for example, \p{Nd}+, there was the theoretical possibility of
746 setting up an incorrect bitmap of starting bytes, but fortunately it could
747 not have actually happened in practice until change 8 above was made (it
748 added property types that matched character-matching opcodes).
749
750 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
751 possible starting bytes for non-anchored patterns.
752
753 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
754 \R, and also a number of cases that involve Unicode properties, both
755 explicit and implicit when PCRE_UCP is set.
756
757 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
758 input, it could crash or give wrong results if characters with values
759 greater than 0xc0 were present in the subject string. (Detail: it assumed
760 UTF-8 input when processing these items.)
761
762 16. Added a lot of (int) casts to avoid compiler warnings in systems where
763 size_t is 64-bit (#991).
764
765 17. Added a check for running out of memory when PCRE is compiled with
766 --disable-stack-for-recursion (#990).
767
768 18. If the last data line in a file for pcretest does not have a newline on
769 the end, a newline was missing in the output.
770
771 19. The default pcre_chartables.c file recognizes only ASCII characters (values
772 less than 128) in its various bitmaps. However, there is a facility for
773 generating tables according to the current locale when PCRE is compiled. It
774 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
775 characters, are recognized by isspace() and therefore were getting set in
776 these tables, and indeed these tables seem to approximate to ISO 8859. This
777 caused a problem in UTF-8 mode when pcre_study() was used to create a list
778 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
779 which of course cannot start UTF-8 characters. I have changed the code so
780 that only real ASCII characters (less than 128) and the correct starting
781 bytes for UTF-8 encodings are set for characters greater than 127 when in
782 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
783 altogether.)
784
785 20. Added the /T option to pcretest so as to be able to run tests with non-
786 standard character tables, thus making it possible to include the tests
787 used for 19 above in the standard set of tests.
788
789 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
790 reference to a subpattern the other side of a comment that contains an
791 opening parenthesis caused either an internal compiling error, or a
792 reference to the wrong subpattern.
793
794
795 Version 8.02 19-Mar-2010
796 ------------------------
797
798 1. The Unicode data tables have been updated to Unicode 5.2.0.
799
800 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
801 configured.
802
803 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
804 original author of that file, following a query about its status.
805
806 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
807 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
808
809 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
810 quantifier applied to a forward-referencing subroutine call, could compile
811 incorrect code or give the error "internal error: previously-checked
812 referenced subpattern not found".
813
814 6. Both MS Visual Studio and Symbian OS have problems with initializing
815 variables to point to external functions. For these systems, therefore,
816 pcre_malloc etc. are now initialized to local functions that call the
817 relevant global functions.
818
819 7. There were two entries missing in the vectors called coptable and poptable
820 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
821 I've fixed the data, and added a kludgy way of testing at compile time that
822 the lengths are correct (equal to the number of opcodes).
823
824 8. Following on from 7, I added a similar kludge to check the length of the
825 eint vector in pcreposix.c.
826
827 9. Error texts for pcre_compile() are held as one long string to avoid too
828 much relocation at load time. To find a text, the string is searched,
829 counting zeros. There was no check for running off the end of the string,
830 which could happen if a new error number was added without updating the
831 string.
832
833 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
834
835 11. \K was not working if it appeared in an atomic group or in a group that
836 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
837 \K is "not well defined" if used in an assertion. PCRE now accepts it if
838 the assertion is positive, but not if it is negative.
839
840 12. Change 11 fortuitously reduced the size of the stack frame used in the
841 "match()" function of pcre_exec.c by one pointer. Forthcoming
842 implementation of support for (*MARK) will need an extra pointer on the
843 stack; I have reserved it now, so that the stack frame size does not
844 decrease.
845
846 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
847 item in branch that calls a recursion is a subroutine call - as in the
848 second branch in the above example - was incorrectly given the compile-
849 time error "recursive call could loop indefinitely" because pcre_compile()
850 was not correctly checking the subroutine for matching a non-empty string.
851
852 14. The checks for overrunning compiling workspace could trigger after an
853 overrun had occurred. This is a "should never occur" error, but it can be
854 triggered by pathological patterns such as hundreds of nested parentheses.
855 The checks now trigger 100 bytes before the end of the workspace.
856
857 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
858
859
860 Version 8.01 19-Jan-2010
861 ------------------------
862
863 1. If a pattern contained a conditional subpattern with only one branch (in
864 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
865 computed the wrong minimum data length (which is of course zero for such
866 subpatterns). This could cause incorrect "no match" results.
867
868 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
869 the pattern is reset in the first branch, pcre_compile() failed with
870 "internal error: code overflow at offset...". This happened only when
871 the reset was to the original external option setting. (An optimization
872 abstracts leading options settings into an external setting, which was the
873 cause of this.)
874
875 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
876 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
877 assertion pattern did not match (meaning that the assertion was true), it
878 was incorrectly treated as false if the SKIP had been reached during the
879 matching. This also applied to assertions used as conditions.
880
881 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
882 assertion subpattern, including such a pattern used as a condition,
883 unpredictable results occurred, instead of the error return
884 PCRE_ERROR_DFA_UITEM.
885
886 5. The C++ GlobalReplace function was not working like Perl for the special
887 situation when an empty string is matched. It now does the fancy magic
888 stuff that is necessary.
889
890 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
891 removed. (These were left over from very, very early versions of PCRE.)
892
893 7. Some cosmetic changes to the code to make life easier when compiling it
894 as part of something else:
895
896 (a) Change DEBUG to PCRE_DEBUG.
897
898 (b) In pcre_compile(), rename the member of the "branch_chain" structure
899 called "current" as "current_branch", to prevent a collision with the
900 Linux macro when compiled as a kernel module.
901
902 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
903 prevent a collision with the Linux macro when compiled as a kernel
904 module.
905
906 8. In pcre_compile() there are some checks for integer overflows that used to
907 cast potentially large values to (double). This has been changed to that
908 when building, a check for int64_t is made, and if it is found, it is used
909 instead, thus avoiding the use of floating point arithmetic. (There is no
910 other use of FP in PCRE.) If int64_t is not found, the fallback is to
911 double.
912
913 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
914 2005 (difference between two addresses compared to an unsigned value).
915
916 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
917 custom one, because of the following reported problem in Windows:
918
919 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
920 under Win32.
921 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
922 therefore missing the function definition.
923 - The compiler thus generates a "C" signature for the test function.
924 - The linker fails to find the "C" function.
925 - PCRE fails to configure if asked to do so against libbz2.
926
927 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
928 messages were output:
929
930 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
931 rerunning libtoolize, to keep the correct libtool macros in-tree.
932 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
933
934 I have done both of these things.
935
936 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
937 most of the time, it *can* run out if it is given a pattern that contains a
938 runaway infinite recursion. I updated the discussion in the pcrestack man
939 page.
940
941 13. Now that we have gone to the x.xx style of version numbers, the minor
942 version may start with zero. Using 08 or 09 is a bad idea because users
943 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
944 interpreted as invalid octal numbers. I've updated the previous comment in
945 configure.ac, and also added a check that gives an error if 08 or 09 are
946 used.
947
948 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
949 causing partial matching to fail when the end of the subject matched \W
950 in a UTF-8 pattern where \W was quantified with a minimum of 3.
951
952 15. There were some discrepancies between the declarations in pcre_internal.h
953 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
954 their definitions. The declarations used "const uschar *" and the
955 definitions used USPTR. Even though USPTR is normally defined as "const
956 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
957 reported that: "This difference in casting confuses some C++ compilers, for
958 example, SunCC recognizes above declarations as different functions and
959 generates broken code for hbpcre." I have changed the declarations to use
960 USPTR.
961
962 16. GNU libtool is named differently on some systems. The autogen.sh script now
963 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
964 (FreeBSD).
965
966 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
967 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
968 comment: "Figure out how to create a longlong from a string: strtoll and
969 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
970 instance, but it only takes 2 args instead of 3!"
971
972 18. A subtle bug concerned with back references has been fixed by a change of
973 specification, with a corresponding code fix. A pattern such as
974 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
975 refers, was giving matches when it shouldn't. For example, xa=xaaa would
976 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
977 same bug. Such groups have to be quantified to be useful, or contained
978 inside another quantified group. (If there's no repetition, the reference
979 can never match.) The problem arises because, having left the group and
980 moved on to the rest of the pattern, a later failure that backtracks into
981 the group uses the captured value from the final iteration of the group
982 rather than the correct earlier one. I have fixed this in PCRE by forcing
983 any group that contains a reference to itself to be an atomic group; that
984 is, there cannot be any backtracking into it once it has completed. This is
985 similar to recursive and subroutine calls.
986
987
988 Version 8.00 19-Oct-09
989 ----------------------
990
991 1. The table for translating pcre_compile() error codes into POSIX error codes
992 was out-of-date, and there was no check on the pcre_compile() error code
993 being within the table. This could lead to an OK return being given in
994 error.
995
996 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
997 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
998 in a Windows environment.
999
1000 3. The pcregrep --count option prints the count for each file even when it is
1001 zero, as does GNU grep. However, pcregrep was also printing all files when
1002 --files-with-matches was added. Now, when both options are given, it prints
1003 counts only for those files that have at least one match. (GNU grep just
1004 prints the file name in this circumstance, but including the count seems
1005 more useful - otherwise, why use --count?) Also ensured that the
1006 combination -clh just lists non-zero counts, with no names.
1007
1008 4. The long form of the pcregrep -F option was incorrectly implemented as
1009 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1010 but it seems right to fix it, and I didn't think it was worth preserving
1011 the old behaviour.
1012
1013 5. The command line items --regex=pattern and --regexp=pattern were not
1014 recognized by pcregrep, which required --regex pattern or --regexp pattern
1015 (with a space rather than an '='). The man page documented the '=' forms,
1016 which are compatible with GNU grep; these now work.
1017
1018 6. No libpcreposix.pc file was created for pkg-config; there was just
1019 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1020
1021 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1022 when UCP support is not needed, by modifying the Python script that
1023 generates it from Unicode data files. This should not matter if the module
1024 is correctly used as a library, but I received one complaint about 50K of
1025 unwanted data. My guess is that the person linked everything into his
1026 program rather than using a library. Anyway, it does no harm.
1027
1028 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1029 was a minimum greater than 1 for a wide character in a possessive
1030 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1031 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1032 character. Chaos in the form of incorrect output or a compiling loop could
1033 result.
1034
1035 9. The restrictions on what a pattern can contain when partial matching is
1036 requested for pcre_exec() have been removed. All patterns can now be
1037 partially matched by this function. In addition, if there are at least two
1038 slots in the offset vector, the offset of the earliest inspected character
1039 for the match and the offset of the end of the subject are set in them when
1040 PCRE_ERROR_PARTIAL is returned.
1041
1042 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1043 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1044 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1045 and may be more useful for multi-segment matching.
1046
1047 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1048 used to be given if ever the end of the subject was reached; now it is
1049 given only if matching could not proceed because another character was
1050 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1051 string "Z", which now yields "no match" instead of "partial match". In the
1052 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1053 final character ended with (*FAIL).
1054
1055 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1056 if the pattern had a "must contain" character that was already found in the
1057 earlier partial match, unless partial matching was again requested. For
1058 example, with the pattern /dog.(body)?/, the "must contain" character is
1059 "g". If the first part-match was for the string "dog", restarting with
1060 "sbody" failed. This bug has been fixed.
1061
1062 13. The string returned by pcre_dfa_exec() after a partial match has been
1063 changed so that it starts at the first inspected character rather than the
1064 first character of the match. This makes a difference only if the pattern
1065 starts with a lookbehind assertion or \b or \B (\K is not supported by
1066 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1067 matching functions compatible, and I think it's the right thing to do.
1068
1069 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1070 so that the demonstration program is easily available in environments where
1071 PCRE has not been installed from source.
1072
1073 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1074 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1075 library.
1076
1077 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1078 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1079 is not the first non-POSIX option to be added. Clearly some people find
1080 these options useful.
1081
1082 17. If a caller to the POSIX matching function regexec() passes a non-zero
1083 value for nmatch with a NULL value for pmatch, the value of
1084 nmatch is forced to zero.
1085
1086 18. RunGrepTest did not have a test for the availability of the -u option of
1087 the diff command, as RunTest does. It now checks in the same way as
1088 RunTest, and also checks for the -b option.
1089
1090 19. If an odd number of negated classes containing just a single character
1091 interposed, within parentheses, between a forward reference to a named
1092 subpattern and the definition of the subpattern, compilation crashed with
1093 an internal error, complaining that it could not find the referenced
1094 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1095 [The bug was that it was starting one character too far in when skipping
1096 over the character class, thus treating the ] as data rather than
1097 terminating the class. This meant it could skip too much.]
1098
1099 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1100 /g option in pcretest when the pattern contains \K, which makes it possible
1101 to have an empty string match not at the start, even when the pattern is
1102 anchored. Updated pcretest and pcredemo to use this option.
1103
1104 21. If the maximum number of capturing subpatterns in a recursion was greater
1105 than the maximum at the outer level, the higher number was returned, but
1106 with unset values at the outer level. The correct (outer level) value is
1107 now given.
1108
1109 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1110 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1111 make it do so. The string so far is captured, making this feature
1112 compatible with Perl.
1113
1114 23. The tests have been re-organized, adding tests 11 and 12, to make it
1115 possible to check the Perl 5.10 features against Perl 5.10.
1116
1117 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1118 pattern matches a fixed length string. PCRE did not allow this; now it
1119 does. Neither allows recursion.
1120
1121 25. I finally figured out how to implement a request to provide the minimum
1122 length of subject string that was needed in order to match a given pattern.
1123 (It was back references and recursion that I had previously got hung up
1124 on.) This code has now been added to pcre_study(); it finds a lower bound
1125 to the length of subject needed. It is not necessarily the greatest lower
1126 bound, but using it to avoid searching strings that are too short does give
1127 some useful speed-ups. The value is available to calling programs via
1128 pcre_fullinfo().
1129
1130 26. While implementing 25, I discovered to my embarrassment that pcretest had
1131 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1132 study optimizations had never been tested with that matching function.
1133 Oops. What is worse, even when it was passed study data, there was a bug in
1134 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1135 were also very few tests of studied patterns with pcre_dfa_exec().
1136
1137 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1138 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1139 on the other side of the coin, they are no longer allowed to have different
1140 names, because these cannot be distinguished in PCRE, and this has caused
1141 confusion. (This is a difference from Perl.)
1142
1143 28. When duplicate subpattern names are present (necessarily with different
1144 numbers, as required by 27 above), and a test is made by name in a
1145 conditional pattern, either for a subpattern having been matched, or for
1146 recursion in such a pattern, all the associated numbered subpatterns are
1147 tested, and the overall condition is true if the condition is true for any
1148 one of them. This is the way Perl works, and is also more like the way
1149 testing by number works.
1150
1151
1152 Version 7.9 11-Apr-09
1153 ---------------------
1154
1155 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1156 (pcretest), all targets were linked against these libraries. This included
1157 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1158 libraries. This caused unwanted dependencies to be created. This problem
1159 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1160 pcretest is linked with readline.
1161
1162 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1163 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1164 moved outside it again, because FALSE and TRUE are already defined in AIX,
1165 but BOOL is not.
1166
1167 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1168 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1169
1170 4. The pcregrep documentation said spaces were inserted as well as colons (or
1171 hyphens) following file names and line numbers when outputting matching
1172 lines. This is not true; no spaces are inserted. I have also clarified the
1173 wording for the --colour (or --color) option.
1174
1175 5. In pcregrep, when --colour was used with -o, the list of matching strings
1176 was not coloured; this is different to GNU grep, so I have changed it to be
1177 the same.
1178
1179 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1180 each matching line was coloured. Now it goes on to look for further matches
1181 of any of the test patterns, which is the same behaviour as GNU grep.
1182
1183 7. A pattern that could match an empty string could cause pcregrep to loop; it
1184 doesn't make sense to accept an empty string match in pcregrep, so I have
1185 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1186 seems to be how GNU grep behaves.
1187
1188 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1189 start or after a newline", because the conditional assertion was not being
1190 correctly handled. The rule now is that both the assertion and what follows
1191 in the first alternative must satisfy the test.
1192
1193 9. If auto-callout was enabled in a pattern with a conditional group whose
1194 condition was an assertion, PCRE could crash during matching, both with
1195 pcre_exec() and pcre_dfa_exec().
1196
1197 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1198 used for matching.
1199
1200 11. Unicode property support in character classes was not working for
1201 characters (bytes) greater than 127 when not in UTF-8 mode.
1202
1203 12. Added the -M command line option to pcretest.
1204
1205 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1206
1207 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1208
1209 16. Added comments and documentation about mis-use of no_arg in the C++
1210 wrapper.
1211
1212 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1213 from Martin Jerabek that uses macro names for all relevant character and
1214 string constants.
1215
1216 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1217 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1218 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1219 these, but not everybody uses configure.
1220
1221 19. A conditional group that had only one branch was not being correctly
1222 recognized as an item that could match an empty string. This meant that an
1223 enclosing group might also not be so recognized, causing infinite looping
1224 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1225 with the subject "ab", where knowledge that the repeated group can match
1226 nothing is needed in order to break the loop.
1227
1228 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1229 exec(), but without supplying a callout function, matching went wrong.
1230
1231 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1232 leak if the size of the offset vector was greater than 30. When the vector
1233 is smaller, the saved offsets during recursion go onto a local stack
1234 vector, but for larger vectors malloc() is used. It was failing to free
1235 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1236 error, in fact).
1237
1238 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1239 heapframe that is used only when UTF-8 support is enabled. This caused no
1240 problem, but was untidy.
1241
1242 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1243 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1244 included within another project.
1245
1246 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1247 slightly modified by me:
1248
1249 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1250 not building pcregrep.
1251
1252 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1253 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1254
1255 25. Forward references, both numeric and by name, in patterns that made use of
1256 duplicate group numbers, could behave incorrectly or give incorrect errors,
1257 because when scanning forward to find the reference group, PCRE was not
1258 taking into account the duplicate group numbers. A pattern such as
1259 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1260
1261 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1262 the feature of a custom pointer more persuasive (as requested by a user).
1263
1264 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1265 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1266 pre-defined.
1267
1268 28. Added support for (*UTF8) at the start of a pattern.
1269
1270 29. Arrange for flags added by the "release type" setting in CMake to be shown
1271 in the configuration summary.
1272
1273
1274 Version 7.8 05-Sep-08
1275 ---------------------
1276
1277 1. Replaced UCP searching code with optimized version as implemented for Ad
1278 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1279 stage table and inline lookup instead of a function, giving speed ups of 2
1280 to 5 times on some simple patterns that I tested. Permission was given to
1281 distribute the MultiStage2.py script that generates the tables (it's not in
1282 the tarball, but is in the Subversion repository).
1283
1284 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1285 scripts.
1286
1287 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1288 a group with a zero qualifier. The result of the study could be incorrect,
1289 or the function might crash, depending on the pattern.
1290
1291 4. Caseless matching was not working for non-ASCII characters in back
1292 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1293 It now works when Unicode Property Support is available.
1294
1295 5. In pcretest, an escape such as \x{de} in the data was always generating
1296 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1297 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1298 truncation.
1299
1300 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1301
1302 7. Added two (int) casts to pcregrep when printing the difference of two
1303 pointers, in case they are 64-bit values.
1304
1305 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1306 test 2 if it fails.
1307
1308 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1309 and a #define of that name to empty if it is not externally set. This is to
1310 allow users of MSVC to set it if necessary.
1311
1312 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1313 the convenience functions in the pcre_get.c source file.
1314
1315 11. An option change at the start of a pattern that had top-level alternatives
1316 could cause overwriting and/or a crash. This command provoked a crash in
1317 some environments:
1318
1319 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1320
1321 This potential security problem was recorded as CVE-2008-2371.
1322
1323 12. For a pattern where the match had to start at the beginning or immediately
1324 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1325 pcre_dfa_exec() could read past the end of the passed subject if there was
1326 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1327 pcretest so that it places the subject at the end of its malloc-ed buffer.
1328
1329 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1330 exec() might read past the end of the data buffer in UTF-8 mode.
1331
1332 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1333 the data contained the byte 0x85 as part of a UTF-8 character within its
1334 first line. This applied both to normal and DFA matching.
1335
1336 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1337 /^[^d]*?$/8 failed to match "abc".
1338
1339 16. Added a missing copyright notice to pcrecpp_internal.h.
1340
1341 17. Make it more clear in the documentation that values returned from
1342 pcre_exec() in ovector are byte offsets, not character counts.
1343
1344 18. Tidied a few places to stop certain compilers from issuing warnings.
1345
1346 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1347 supplied by Stefan Weber. I made a further small update for 7.8 because
1348 there is a change of source arrangements: the pcre_searchfuncs.c module is
1349 replaced by pcre_ucd.c.
1350
1351
1352 Version 7.7 07-May-08
1353 ---------------------
1354
1355 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1356 a string to a long long, pretend we don't even have a long long." This is
1357 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1358
1359 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1360 pre-7.6 versions, which defined a global no_arg variable instead of putting
1361 it in the RE class. (See also #8 below.)
1362
1363 3. Remove a line of dead code, identified by coverity and reported by Nuno
1364 Lopes.
1365
1366 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1367
1368 (1) The include/exclude patterns were being applied to the whole pathnames
1369 of files, instead of just to the final components.
1370
1371 (2) If there was more than one level of directory, the subdirectories were
1372 skipped unless they satisfied the include/exclude conditions. This is
1373 inconsistent with GNU grep (and could even be seen as contrary to the
1374 pcregrep specification - which I improved to make it absolutely clear).
1375 The action now is always to scan all levels of directory, and just
1376 apply the include/exclude patterns to regular files.
1377
1378 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1379 --exclude_dir in the tests to avoid scanning .svn directories.
1380
1381 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1382 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1383 doesn't support NULs in patterns.
1384
1385 7. Added some missing "const"s to declarations of static tables in
1386 pcre_compile.c and pcre_dfa_exec.c.
1387
1388 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1389 caused by fix #2 above. (Subsequently also a second patch to fix the
1390 first patch. And a third patch - this was a messy problem.)
1391
1392 9. Applied Craig's patch to remove the use of push_back().
1393
1394 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1395 matching function regexec().
1396
1397 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1398 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1399 references. PCRE supports relative numbers with this syntax (I don't think
1400 Oniguruma does).
1401
1402 12. Previously, a group with a zero repeat such as (...){0} was completely
1403 omitted from the compiled regex. However, this means that if the group
1404 was called as a subroutine from elsewhere in the pattern, things went wrong
1405 (an internal error was given). Such groups are now left in the compiled
1406 pattern, with a new opcode that causes them to be skipped at execution
1407 time.
1408
1409 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1410 to the way PCRE behaves:
1411
1412 (a) A lone ] character is dis-allowed (Perl treats it as data).
1413
1414 (b) A back reference to an unmatched subpattern matches an empty string
1415 (Perl fails the current match path).
1416
1417 (c) A data ] in a character class must be notated as \] because if the
1418 first data character in a class is ], it defines an empty class. (In
1419 Perl it is not possible to have an empty class.) The empty class []
1420 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1421 The negative empty class [^] matches any one character, independently
1422 of the DOTALL setting.
1423
1424 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1425 non-existent subpattern following a character class starting with ']' and
1426 containing () gave an internal compiling error instead of "reference to
1427 non-existent subpattern". Fortunately, when the pattern did exist, the
1428 compiled code was correct. (When scanning forwards to check for the
1429 existencd of the subpattern, it was treating the data ']' as terminating
1430 the class, so got the count wrong. When actually compiling, the reference
1431 was subsequently set up correctly.)
1432
1433 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1434 it was being rejected as not supported by pcre_dfa_exec(), even though
1435 other assertions are supported. I have made pcre_dfa_exec() support
1436 (*FAIL).
1437
1438 16. The implementation of 13c above involved the invention of a new opcode,
1439 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1440 cannot be changed at match time, I realized I could make a small
1441 improvement to matching performance by compiling OP_ALLANY instead of
1442 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1443 on the OP_ANY path.
1444
1445 17. Compiling pcretest on Windows with readline support failed without the
1446 following two fixes: (1) Make the unistd.h include conditional on
1447 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1448
1449 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1450 ncurses library to be included for pcretest when ReadLine support is
1451 requested, but also to allow for it to be overridden. This patch came from
1452 Daniel Bergström.
1453
1454 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1455 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1456 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1457 spotting this.
1458
1459
1460 Version 7.6 28-Jan-08
1461 ---------------------
1462
1463 1. A character class containing a very large number of characters with
1464 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1465 overflow.
1466
1467 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1468 HAVE_LONG_LONG is not defined.
1469
1470 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1471 bring them up to date and include new features. This patch includes:
1472
1473 - Fixed PH's badly added libz and libbz2 support.
1474 - Fixed a problem with static linking.
1475 - Added pcredemo. [But later removed - see 7 below.]
1476 - Fixed dftables problem and added an option.
1477 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1478 HAVE_LONG_LONG.
1479 - Added readline support for pcretest.
1480 - Added an listing of the option settings after cmake has run.
1481
1482 4. A user submitted a patch to Makefile that makes it easy to create
1483 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1484 Makefile.am that cause it to include this special target, without
1485 affecting anything else. Note that the same mingw target plus all
1486 the other distribution libraries and programs are now supported
1487 when configuring with CMake (see 6 below) instead of with
1488 Configure/Make.
1489
1490 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1491 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1492 exported in the Windows port". It has not yet been confirmed that the patch
1493 solves the problem, but it does no harm.
1494
1495 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1496 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1497 with CMake, and also correct the comment about stack recursion.
1498
1499 7. Remove the automatic building of pcredemo from the ./configure system and
1500 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1501 of a program that users should build themselves after PCRE is installed, so
1502 building it automatically is not really right. What is more, it gave
1503 trouble in some build environments.
1504
1505 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1506
1507
1508 Version 7.5 10-Jan-08
1509 ---------------------
1510
1511 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1512 values in parens when parsing an RE using the C++ wrapper."
1513
1514 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1515 Characters greater than 255 were excluded from the class instead of being
1516 included.
1517
1518 3. The same bug as (2) above applied to negated POSIX classes such as
1519 [:^space:].
1520
1521 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1522 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1523 I have changed it.
1524
1525 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1526 first named subpattern) and a construct such as (?&a) would reference the
1527 first named subpattern whose name started with "a" (in other words, the
1528 length check was missing). Both these problems are fixed. "Subpattern name
1529 expected" is now given for (?&) (a zero-length name), and this patch also
1530 makes it give the same error for \k'' (previously it complained that that
1531 was a reference to a non-existent subpattern).
1532
1533 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1534 this is right because (?- can be followed by option settings as well as by
1535 digits. I have, however, made the messages clearer.
1536
1537 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1538 than the number used in the conditional) now cause a compile-time error.
1539 This is actually not compatible with Perl, which accepts such patterns, but
1540 treats the conditional as always being FALSE (as PCRE used to), but it
1541 seems to me that giving a diagnostic is better.
1542
1543 8. Change "alphameric" to the more common word "alphanumeric" in comments
1544 and messages.
1545
1546 9. Fix two occurrences of "backslash" in comments that should have been
1547 "backspace".
1548
1549 10. Remove two redundant lines of code that can never be obeyed (their function
1550 was moved elsewhere).
1551
1552 11. The program that makes PCRE's Unicode character property table had a bug
1553 which caused it to generate incorrect table entries for sequences of
1554 characters that have the same character type, but are in different scripts.
1555 It amalgamated them into a single range, with the script of the first of
1556 them. In other words, some characters were in the wrong script. There were
1557 thirteen such cases, affecting characters in the following ranges:
1558
1559 U+002b0 - U+002c1
1560 U+0060c - U+0060d
1561 U+0061e - U+00612
1562 U+0064b - U+0065e
1563 U+0074d - U+0076d
1564 U+01800 - U+01805
1565 U+01d00 - U+01d77
1566 U+01d9b - U+01dbf
1567 U+0200b - U+0200f
1568 U+030fc - U+030fe
1569 U+03260 - U+0327f
1570 U+0fb46 - U+0fbb1
1571 U+10450 - U+1049d
1572
1573 12. The -o option (show only the matching part of a line) for pcregrep was not
1574 compatible with GNU grep in that, if there was more than one match in a
1575 line, it showed only the first of them. It now behaves in the same way as
1576 GNU grep.
1577
1578 13. If the -o and -v options were combined for pcregrep, it printed a blank
1579 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1580 does the same. The return code can be used to tell if there were any
1581 non-matching lines.
1582
1583 14. Added --file-offsets and --line-offsets to pcregrep.
1584
1585 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1586 infinitely looping recursion. The bug was that positive lookaheads were not
1587 being skipped when checking for a possible empty match (negative lookaheads
1588 and both kinds of lookbehind were skipped).
1589
1590 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1591 inclusion of <windows.h> to before rather than after the definition of
1592 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1593
1594 17. Specifying a possessive quantifier with a specific limit for a Unicode
1595 character property caused pcre_compile() to compile bad code, which led at
1596 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1597 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1598 caused the error; without that there was no problem.
1599
1600 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1601
1602 19. Added --enable-pcretest-libreadline.
1603
1604 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1605 RE::GlobalReplace(). As a result, the number of replacements returned was
1606 double what it should be. I removed one of the increments, but Craig sent a
1607 later patch that removed the other one (the right fix) and added unit tests
1608 that check the return values (which was not done before).
1609
1610 21. Several CMake things:
1611
1612 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1613 the names libpcre and libpcreposix, not just pcre and pcreposix.
1614
1615 (2) The above change means that pcretest and pcregrep are now correctly
1616 linked with the newly-built libraries, not previously installed ones.
1617
1618 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1619
1620 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1621 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1622 UTF-8 newline character). The key issue is that the pattern starts .*;
1623 this means that the match must be either at the beginning, or after a
1624 newline. The bug was in the code for advancing after a failed match and
1625 checking that the new position followed a newline. It was not taking
1626 account of UTF-8 characters correctly.
1627
1628 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1629 character classes. PCRE was not treating the sequence [:...:] as a
1630 character class unless the ... were all letters. Perl, however, seems to
1631 allow any characters between [: and :], though of course it rejects as
1632 unknown any "names" that contain non-letters, because all the known class
1633 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1634 for example, whereas PCRE did not - it did not recognize a POSIX character
1635 class. This seemed a bit dangerous, so the code has been changed to be
1636 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1637 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1638 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1639 Perl does, and where it didn't before.
1640
1641 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1642 Windows environments %n is disabled by default.
1643
1644
1645 Version 7.4 21-Sep-07
1646 ---------------------
1647
1648 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1649 means that a class such as [\s] counted as "explicit reference to CR or
1650 LF". That isn't really right - the whole point of the change was to try to
1651 help when there was an actual mention of one of the two characters. So now
1652 the change happens only if \r or \n (or a literal CR or LF) character is
1653 encountered.
1654
1655 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1656 of both had grown to the point where there were only 3 bits left.
1657 Fortunately, there was spare space in the data structure, and so I have
1658 moved the internal flags into a new 16-bit field to free up more option
1659 bits.
1660
1661 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1662 but did not set the internal JCHANGED flag - either of these is enough to
1663 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1664 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1665 start sets both bits.
1666
1667 4. Added options (at build time, compile time, exec time) to change \R from
1668 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1669
1670 5. doc/pcresyntax.html was missing from the distribution.
1671
1672 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1673 compatibility, even though it is no longer used.
1674
1675 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1676 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1677 windows.h file is present (where different names are used). [This was
1678 reversed later after testing - see 16 below.]
1679
1680 8. Changed all #include <config.h> to #include "config.h". There were also
1681 some further <pcre.h> cases that I changed to "pcre.h".
1682
1683 9. When pcregrep was used with the --colour option, it missed the line ending
1684 sequence off the lines that it output.
1685
1686 10. It was pointed out to me that arrays of string pointers cause lots of
1687 relocations when a shared library is dynamically loaded. A technique of
1688 using a single long string with a table of offsets can drastically reduce
1689 these. I have refactored PCRE in four places to do this. The result is
1690 dramatic:
1691
1692 Originally: 290
1693 After changing UCP table: 187
1694 After changing error message table: 43
1695 After changing table of "verbs" 36
1696 After changing table of Posix names 22
1697
1698 Thanks to the folks working on Gregex for glib for this insight.
1699
1700 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1701 unicode-properties was also set.
1702
1703 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1704
1705 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1706 checked only for CRLF.
1707
1708 14. Added casts to pcretest.c to avoid compiler warnings.
1709
1710 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1711
1712 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1713 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1714 entirely. This removes changes made in 7 above.
1715
1716 17. The CMake files have been updated, and there is now more information about
1717 building with CMake in the NON-UNIX-USE document.
1718
1719
1720 Version 7.3 28-Aug-07
1721 ---------------------
1722
1723 1. In the rejigging of the build system that eventually resulted in 7.1, the
1724 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1725 brackets there is not right, since it causes compilers to look for an
1726 installed pcre.h, not the version that is in the source that is being
1727 compiled (which of course may be different). I have changed it back to:
1728
1729 #include "pcre.h"
1730
1731 I have a vague recollection that the change was concerned with compiling in
1732 different directories, but in the new build system, that is taken care of
1733 by the VPATH setting the Makefile.
1734
1735 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1736 when the subject happened to end in the byte 0x85 (e.g. if the last
1737 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1738 characters but of course it shouldn't be taken as a newline when it is part
1739 of another character. The bug was that, for an unlimited repeat of . in
1740 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1741 characters when looking for a newline.
1742
1743 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1744
1745 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1746 in debug output.
1747
1748 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1749 long printing in the pcrecpp unittest when running under MinGW.
1750
1751 6. ESC_K was left out of the EBCDIC table.
1752
1753 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1754 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1755 limit also applies to "virtual nesting" when a pattern is recursive, and in
1756 this case 1000 isn't so big. I have been able to remove this limit at the
1757 expense of backing off one optimization in certain circumstances. Normally,
1758 when pcre_exec() would call its internal match() function recursively and
1759 immediately return the result unconditionally, it uses a "tail recursion"
1760 feature to save stack. However, when a subpattern that can match an empty
1761 string has an unlimited repetition quantifier, it no longer makes this
1762 optimization. That gives it a stack frame in which to save the data for
1763 checking that an empty string has been matched. Previously this was taken
1764 from the 1000-entry workspace that had been reserved. So now there is no
1765 explicit limit, but more stack is used.
1766
1767 8. Applied Daniel's patches to solve problems with the import/export magic
1768 syntax that is required for Windows, and which was going wrong for the
1769 pcreposix and pcrecpp parts of the library. These were overlooked when this
1770 problem was solved for the main library.
1771
1772 9. There were some crude static tests to avoid integer overflow when computing
1773 the size of patterns that contain repeated groups with explicit upper
1774 limits. As the maximum quantifier is 65535, the maximum group length was
1775 set at 30,000 so that the product of these two numbers did not overflow a
1776 32-bit integer. However, it turns out that people want to use groups that
1777 are longer than 30,000 bytes (though not repeat them that many times).
1778 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1779 made it possible to implement the integer overflow checks in a much more
1780 dynamic way, which I have now done. The artificial limitation on group
1781 length has been removed - we now have only the limit on the total length of
1782 the compiled pattern, which depends on the LINK_SIZE setting.
1783
1784 10. Fixed a bug in the documentation for get/copy named substring when
1785 duplicate names are permitted. If none of the named substrings are set, the
1786 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1787 empty string.
1788
1789 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1790 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1791 because the ] is interpreted as the first data character and the
1792 terminating ] is not found. PCRE has been made compatible with Perl in this
1793 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1794 cause memory overwriting.
1795
1796 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1797 string has been matched (to stop an infinite loop). It was not recognizing
1798 a conditional subpattern that could match an empty string if that
1799 subpattern was within another subpattern. For example, it looped when
1800 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1801 condition was not nested. This bug has been fixed.
1802
1803 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1804 past the start of the subject in the presence of bytes with the top bit
1805 set, for example "\x8aBCD".
1806
1807 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1808 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1809
1810 14. Optimized (?!) to (*FAIL).
1811
1812 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1813 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1814 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1815 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1816 does: it's just the validity check that is more restrictive.
1817
1818 16. Inserted checks for integer overflows during escape sequence (backslash)
1819 processing, and also fixed erroneous offset values for syntax errors during
1820 backslash processing.
1821
1822 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1823 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1824
1825 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1826 caused an overrun.
1827
1828 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1829 something other than just ASCII characters) inside a group that had an
1830 unlimited repeat caused a loop at compile time (while checking to see
1831 whether the group could match an empty string).
1832
1833 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1834 [\P{Any}] did so. (Error in the code for printing property names.)
1835
1836 21. An orphan \E inside a character class could cause a crash.
1837
1838 22. A repeated capturing bracket such as (A)? could cause a wild memory
1839 reference during compilation.
1840
1841 23. There are several functions in pcre_compile() that scan along a compiled
1842 expression for various reasons (e.g. to see if it's fixed length for look
1843 behind). There were bugs in these functions when a repeated \p or \P was
1844 present in the pattern. These operators have additional parameters compared
1845 with \d, etc, and these were not being taken into account when moving along
1846 the compiled data. Specifically:
1847
1848 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1849 length.
1850
1851 (b) An item such as \pL+ within a repeated group could cause crashes or
1852 loops.
1853
1854 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1855 "reference to non-existent subpattern" error.
1856
1857 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1858
1859 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1860 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1861
1862 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1863
1864 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1865 character were causing crashes (broken optimization).
1866
1867 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1868 \p or \P) caused a compile-time loop.
1869
1870 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1871 break. For example, the unstudied pattern [\r\n]A does not match the string
1872 "\r\nA" because change 7.0/46 below moves the current point on by two
1873 characters after failing to match at the start. However, the pattern \nA
1874 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1875 the same is true. There doesn't seem any very clean way out of this, but
1876 what I have chosen to do makes the common cases work: PCRE now takes note
1877 of whether there can be an explicit match for \r or \n anywhere in the
1878 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1879 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1880 pattern has explicit CR or LF references.
1881
1882 29. Added (*CR) etc for changing newline setting at start of pattern.
1883
1884
1885 Version 7.2 19-Jun-07
1886 ---------------------
1887
1888 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1889 which is apparently normally available under Windows.
1890
1891 2. Re-jig the pcregrep tests with different newline settings in an attempt
1892 to make them independent of the local environment's newline setting.
1893
1894 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1895
1896 4. Some of the "internals" tests were previously cut out when the link size
1897 was not 2, because the output contained actual offsets. The recent new
1898 "Z" feature of pcretest means that these can be cut out, making the tests
1899 usable with all link sizes.
1900
1901 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1902 stack recursion. This gives a massive performance boost under BSD, but just
1903 a small improvement under Linux. However, it saves one field in the frame
1904 in all cases.
1905
1906 6. Added more features from the forthcoming Perl 5.10:
1907
1908 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1909 recursion call. It refers to the nth most recently opened parentheses.
1910
1911 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1912 to be opened parentheses.
1913
1914 (c) Conditions that refer to capturing parentheses can be specified
1915 relatively, for example, (?(-2)... or (?(+3)...
1916
1917 (d) \K resets the start of the current match so that everything before
1918 is not part of it.
1919
1920 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1921
1922 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1923 reference syntax.
1924
1925 (g) (?| introduces a group in which the numbering of parentheses in each
1926 alternative starts with the same number.
1927
1928 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1929
1930 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1931 PCRE_INFO_JCHANGED.
1932
1933 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1934 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1935 for detecting groups that can match an empty string.
1936
1937 9. A pattern with a very large number of alternatives (more than several
1938 hundred) was running out of internal workspace during the pre-compile
1939 phase, where pcre_compile() figures out how much memory will be needed. A
1940 bit of new cunning has reduced the workspace needed for groups with
1941 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1942 workspace instead of running out of the 4096 that are available.
1943
1944 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1945
1946 11. Applied patch from Google to remove an optimization that didn't quite work.
1947 The report of the bug said:
1948
1949 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1950 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1951 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1952
1953 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1954 it matched the wrong number of bytes.
1955
1956
1957 Version 7.1 24-Apr-07
1958 ---------------------
1959
1960 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1961 that is more "standard", making use of automake and other Autotools. There
1962 is some re-arrangement of the files and adjustment of comments consequent
1963 on this.
1964
1965 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1966 for recursive directory scanning broke on some systems because the files
1967 are not scanned in any specific order and on different systems the order
1968 was different. A call to "sort" has been inserted into RunGrepTest for the
1969 approprate test as a short-term fix. In the longer term there may be an
1970 alternative.
1971
1972 3. I had an email from Eric Raymond about problems translating some of PCRE's
1973 man pages to HTML (despite the fact that I distribute HTML pages, some
1974 people do their own conversions for various reasons). The problems
1975 concerned the use of low-level troff macros .br and .in. I have therefore
1976 removed all such uses from the man pages (some were redundant, some could
1977 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1978 HTML has been updated to handle .nf/.fi and to complain if it encounters
1979 .br or .in.
1980
1981 4. Updated comments in configure.ac that get placed in config.h.in and also
1982 arranged for config.h to be included in the distribution, with the name
1983 config.h.generic, for the benefit of those who have to compile without
1984 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1985
1986 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1987 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1988 makevp.bat for the current PCRE, using the additional files
1989 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1990
1991 6. A Windows user reported a minor discrepancy with test 2, which turned out
1992 to be caused by a trailing space on an input line that had got lost in his
1993 copy. The trailing space was an accident, so I've just removed it.
1994
1995 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1996 that is needed.
1997
1998 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1999 as "const" (a) because they are and (b) because it helps the PHP
2000 maintainers who have recently made a script to detect big data structures
2001 in the php code that should be moved to the .rodata section. I remembered
2002 to update Builducptable as well, so it won't revert if ucptable.h is ever
2003 re-created.
2004
2005 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2006 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2007 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2008 support is not required. This saves 1.5-2K of code, which is important in
2009 some applications.
2010
2011 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2012 so as not to refer to the tables, even though these functions will never be
2013 called when UTF-8 support is disabled. Otherwise there are problems with a
2014 shared library.
2015
2016 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2017
2018 (a) It was defining its arguments as char * instead of void *.
2019
2020 (b) It was assuming that all moves were upwards in memory; this was true
2021 a long time ago when I wrote it, but is no longer the case.
2022
2023 The emulated memove() is provided for those environments that have neither
2024 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2025 is clearly not the case, as these two bugs were recently reported.
2026
2027 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2028 and Detrail to create the HTML documentation, the .txt form of the man
2029 pages, and it removes trailing spaces from listed files. It also creates
2030 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2031 case, it wraps all the #defines with #ifndefs. This script should be run
2032 before "make dist".
2033
2034 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2035 with Unicode property support.
2036
2037 (a) For a maximizing quantifier, if the two different cases of the
2038 character were of different lengths in their UTF-8 codings (there are
2039 some cases like this - I found 11), and the matching function had to
2040 back up over a mixture of the two cases, it incorrectly assumed they
2041 were both the same length.
2042
2043 (b) When PCRE was configured to use the heap rather than the stack for
2044 recursion during matching, it was not correctly preserving the data for
2045 the other case of a UTF-8 character when checking ahead for a match
2046 while processing a minimizing repeat. If the check also involved
2047 matching a wide character, but failed, corruption could cause an
2048 erroneous result when trying to check for a repeat of the original
2049 character.
2050
2051 13. Some tidying changes to the testing mechanism:
2052
2053 (a) The RunTest script now detects the internal link size and whether there
2054 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2055 values substituted by "configure". (The RunGrepTest script already did
2056 this for UTF-8.) The configure.ac script no longer substitutes the
2057 relevant variables.
2058
2059 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2060 with length and offset values. This means that the output is different
2061 for different internal link sizes. Test 2 is skipped for link sizes
2062 other than 2 because of this, bypassing the problem. Unfortunately,
2063 there was also a test in test 3 (the locale tests) that used /B and
2064 failed for link sizes other than 2. Rather than cut the whole test out,
2065 I have added a new /Z option to pcretest that replaces the length and
2066 offset values with spaces. This is now used to make test 3 independent
2067 of link size. (Test 2 will be tidied up later.)
2068
2069 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2070 segmentation fault instead of returning the appropriate error message.
2071
2072 15. In multiline mode when the newline sequence was set to "any", the pattern
2073 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2074 This doesn't seem right; it now treats the CRLF combination as the line
2075 ending, and so does not match in that case. It's only a pattern such as ^$
2076 that would hit this one: something like ^ABC$ would have failed after \r
2077 and then tried again after \r\n.
2078
2079 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2080 in an attempt to make files that differ only in their line terminators
2081 compare equal. This works on Linux.
2082
2083 17. Under certain error circumstances pcregrep might try to free random memory
2084 as it exited. This is now fixed, thanks to valgrind.
2085
2086 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2087 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2088 was because its rules for how to advance for /g after matching an empty
2089 string at the end of a line did not allow for this case. They now check for
2090 it specially.
2091
2092 20. pcretest is supposed to handle patterns and data of any length, by
2093 extending its buffers when necessary. It was getting this wrong when the
2094 buffer for a data line had to be extended.
2095
2096 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2097 CRLF as a newline sequence.
2098
2099 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2100 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2101 I have nevertheless tidied it up.
2102
2103 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2104
2105 24. Added a man page for pcre-config.
2106
2107
2108 Version 7.0 19-Dec-06
2109 ---------------------
2110
2111 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2112 moving to gcc 4.1.1.
2113
2114 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2115 sys/time.h, which is documented as needed for this function. It doesn't
2116 seem to matter on Linux, but it showed up on some releases of OS X.
2117
2118 3. It seems that there are systems where bytes whose values are greater than
2119 127 match isprint() in the "C" locale. The "C" locale should be the
2120 default when a C program starts up. In most systems, only ASCII printing
2121 characters match isprint(). This difference caused the output from pcretest
2122 to vary, making some of the tests fail. I have changed pcretest so that:
2123
2124 (a) When it is outputting text in the compiled version of a pattern, bytes
2125 other than 32-126 are always shown as hex escapes.
2126
2127 (b) When it is outputting text that is a matched part of a subject string,
2128 it does the same, unless a different locale has been set for the match
2129 (using the /L modifier). In this case, it uses isprint() to decide.
2130
2131 4. Fixed a major bug that caused incorrect computation of the amount of memory
2132 required for a compiled pattern when options that changed within the
2133 pattern affected the logic of the preliminary scan that determines the
2134 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2135 that the computed length was too small. The symptoms of this bug were
2136 either the PCRE error "internal error: code overflow" from pcre_compile(),
2137 or a glibc crash with a message such as "pcretest: free(): invalid next
2138 size (fast)". Examples of patterns that provoked this bug (shown in
2139 pcretest format) are:
2140
2141 /(?-x: )/x
2142 /(?x)(?-x: \s*#\s*)/
2143 /((?i)[\x{c0}])/8
2144 /(?i:[\x{c0}])/8
2145
2146 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2147 is now done differently.
2148
2149 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2150 wrapper classes; (b) implement a new function in the C++ scanner that is
2151 more efficient than the old way of doing things because it avoids levels of
2152 recursion in the regex matching; (c) add a paragraph to the documentation
2153 for the FullMatch() function.
2154
2155 6. The escape sequence \n was being treated as whatever was defined as
2156 "newline". Not only was this contrary to the documentation, which states
2157 that \n is character 10 (hex 0A), but it also went horribly wrong when
2158 "newline" was defined as CRLF. This has been fixed.
2159
2160 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2161 was being set to -1 for the "end of line" case (supposedly a value that no
2162 character can have). Though this value is never used (the check for end of
2163 line is "zero bytes in current character"), it caused compiler complaints.
2164 I've changed it to 0xffffffff.
2165
2166 8. In pcre_version.c, the version string was being built by a sequence of
2167 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2168 string (as it is for production releases) called a macro with an empty
2169 argument. The C standard says the result of this is undefined. The gcc
2170 compiler treats it as an empty string (which was what was wanted) but it is
2171 reported that Visual C gives an error. The source has been hacked around to
2172 avoid this problem.
2173
2174 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2175 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2176 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2177 of them did).
2178
2179 10. Originally, pcretest opened its input and output without "b"; then I was
2180 told that "b" was needed in some environments, so it was added for release
2181 5.0 to both the input and output. (It makes no difference on Unix-like
2182 systems.) Later I was told that it is wrong for the input on Windows. I've
2183 now abstracted the modes into two macros, to make it easier to fiddle with
2184 them, and removed "b" from the input mode under Windows.
2185
2186 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2187
2188 12. Added -help and --help to pcretest as an official way of being reminded
2189 of the options.
2190
2191 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2192 and pcrecpp.cc because they annoy compilers at high warning levels.
2193
2194 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2195
2196 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2197 scripts are not C programs :-) and which was not noticed because it works
2198 on Linux.
2199
2200 16. pcretest is supposed to handle any length of pattern and data line (as one
2201 line or as a continued sequence of lines) by extending its input buffer if
2202 necessary. This feature was broken for very long pattern lines, leading to
2203 a string of junk being passed to pcre_compile() if the pattern was longer
2204 than about 50K.
2205
2206 17. I have done a major re-factoring of the way pcre_compile() computes the
2207 amount of memory needed for a compiled pattern. Previously, there was code
2208 that made a preliminary scan of the pattern in order to do this. That was
2209 OK when PCRE was new, but as the facilities have expanded, it has become
2210 harder and harder to keep it in step with the real compile phase, and there
2211 have been a number of bugs (see for example, 4 above). I have now found a
2212 cunning way of running the real compile function in a "fake" mode that
2213 enables it to compute how much memory it would need, while actually only
2214 ever using a few hundred bytes of working memory and without too many
2215 tests of the mode. This should make future maintenance and development
2216 easier. A side effect of this work is that the limit of 200 on the nesting
2217 depth of parentheses has been removed (though this was never a serious
2218 limitation, I suspect). However, there is a downside: pcre_compile() now
2219 runs more slowly than before (30% or more, depending on the pattern). I
2220 hope this isn't a big issue. There is no effect on runtime performance.
2221
2222 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2223 newline (only possible for the last line of a file) and it was a
2224 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2225
2226 19. Added additional timing features to pcretest. (1) The -tm option now times
2227 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2228 separate command line item, by a number that specifies the number of
2229 repeats to use when timing. The default is 50000; this gives better
2230 precision, but takes uncomfortably long for very large patterns.
2231
2232 20. Extended pcre_study() to be more clever in cases where a branch of a
2233 subpattern has no definite first character. For example, (a*|b*)[cd] would
2234 previously give no result from pcre_study(). Now it recognizes that the
2235 first character must be a, b, c, or d.
2236
2237 21. There was an incorrect error "recursive call could loop indefinitely" if
2238 a subpattern (or the entire pattern) that was being tested for matching an
2239 empty string contained only one non-empty item after a nested subpattern.
2240 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2241 incorrectly, because the \d was being skipped in the check.
2242
2243 22. The pcretest program now has a new pattern option /B and a command line
2244 option -b, which is equivalent to adding /B to every pattern. This causes
2245 it to show the compiled bytecode, without the additional information that
2246 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2247 is the same as /B/I).
2248
2249 23. A new optimization is now able automatically to treat some sequences such
2250 as a*b as a*+b. More specifically, if something simple (such as a character
2251 or a simple class like \d) has an unlimited quantifier, and is followed by
2252 something that cannot possibly match the quantified thing, the quantifier
2253 is automatically "possessified".
2254
2255 24. A recursive reference to a subpattern whose number was greater than 39
2256 went wrong under certain circumstances in UTF-8 mode. This bug could also
2257 have affected the operation of pcre_study().
2258
2259 25. Realized that a little bit of performance could be had by replacing
2260 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2261
2262 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2263
2264 27. Possessive quantifiers such as a++ were previously implemented by turning
2265 them into atomic groups such as ($>a+). Now they have their own opcodes,
2266 which improves performance. This includes the automatically created ones
2267 from 23 above.
2268
2269 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2270 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2271 the first matched character to be a colon. This applied both to named and
2272 numbered groups.
2273
2274 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2275
2276 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2277 building PCRE on Windows easier, so I have included it in the distribution.
2278
2279 31. There is now a check in pcretest against a ridiculously large number being
2280 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2281 loop, the loop is abandoned.
2282
2283 32. Forward references to subpatterns in conditions such as (?(2)...) where
2284 subpattern 2 is defined later cause pcre_compile() to search forwards in
2285 the pattern for the relevant set of parentheses. This search went wrong
2286 when there were unescaped parentheses in a character class, parentheses
2287 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2288
2289 33. "Subroutine" calls and backreferences were previously restricted to
2290 referencing subpatterns earlier in the regex. This restriction has now
2291 been removed.
2292
2293 34. Added a number of extra features that are going to be in Perl 5.10. On the
2294 whole, these are just syntactic alternatives for features that PCRE had
2295 previously implemented using the Python syntax or my own invention. The
2296 other formats are all retained for compatibility.
2297
2298 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2299 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2300 also .NET compatible.
2301
2302 (b) A recursion or subroutine call to a named group can now be defined as
2303 (?&name) as well as (?P>name).
2304
2305 (c) A backreference to a named group can now be defined as \k<name> or
2306 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2307 5.10, are also .NET compatible.
2308
2309 (d) A conditional reference to a named group can now use the syntax
2310 (?(<name>) or (?('name') as well as (?(name).
2311
2312 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2313 groups (named and numbered) that are never evaluated inline, but can be
2314 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2315 is always false. There may be only one alternative in such a group.
2316
2317 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2318 as the simple (?(R). The condition is true only if the most recent
2319 recursion is that of the given number or name. It does not search out
2320 through the entire recursion stack.
2321
2322 (g) The escape \gN or \g{N} has been added, where N is a positive or
2323 negative number, specifying an absolute or relative reference.
2324
2325 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2326 some "unreachable code" warnings.
2327
2328 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2329 things, this adds five new scripts.
2330
2331 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2332 There were also incompatibilities regarding the handling of \Q..\E inside
2333 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2334 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2335
2336 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2337 matches an empty string, and forcibly breaks the loop. There were bugs in
2338 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2339 against aaaa the result was just "a" rather than "aaaa", for example. Two
2340 separate and independent bugs (that affected different cases) have been
2341 fixed.
2342
2343 39. Refactored the code to abolish the use of different opcodes for small
2344 capturing bracket numbers. This is a tidy that I avoided doing when I
2345 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2346 The new approach is not only tidier, it makes it possible to reduce the
2347 memory needed to fix the previous bug (38).
2348
2349 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2350 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2351 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2352 mode.
2353
2354 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2355 report.
2356
2357 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2358 copy construction and assignment in the C++ wrapper.
2359
2360 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2361 couple of bugs that could have given wrong results in the "--newline=crlf"
2362 case.
2363
2364 44. Added a number of casts and did some reorganization of signed/unsigned int
2365 variables following suggestions from Dair Grant. Also renamed the variable
2366 "this" as "item" because it is a C++ keyword.
2367
2368 45. Arranged for dftables to add
2369
2370 #include "pcre_internal.h"
2371
2372 to pcre_chartables.c because without it, gcc 4.x may remove the array
2373 definition from the final binary if PCRE is built into a static library and
2374 dead code stripping is activated.
2375
2376 46. For an unanchored pattern, if a match attempt fails at the start of a
2377 newline sequence, and the newline setting is CRLF or ANY, and the next two
2378 characters are CRLF, advance by two characters instead of one.
2379
2380
2381 Version 6.7 04-Jul-06
2382 ---------------------
2383
2384 1. In order to handle tests when input lines are enormously long, pcretest has
2385 been re-factored so that it automatically extends its buffers when
2386 necessary. The code is crude, but this _is_ just a test program. The
2387 default size has been increased from 32K to 50K.
2388
2389 2. The code in pcre_study() was using the value of the re argument before
2390 testing it for NULL. (Of course, in any sensible call of the function, it
2391 won't be NULL.)
2392
2393 3. The memmove() emulation function in pcre_internal.h, which is used on
2394 systems that lack both memmove() and bcopy() - that is, hardly ever -
2395 was missing a "static" storage class specifier.
2396
2397 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2398 containing an extended class (one that cannot be represented by a bitmap
2399 because it contains high-valued characters or Unicode property items, e.g.
2400 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2401 pattern, but PCRE should not loop if you do not (it no longer does).
2402 [Detail: two cases were found: (a) a repeated subpattern containing an
2403 extended class; (b) a recursive reference to a subpattern that followed a
2404 previous extended class. It wasn't skipping over the extended class
2405 correctly when UTF-8 mode was not set.]
2406
2407 5. A negated single-character class was not being recognized as fixed-length
2408 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2409 compile error "lookbehind assertion is not fixed length".
2410
2411 6. The RunPerlTest auxiliary script was showing an unexpected difference
2412 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2413 write a Perl script that can interpret lines of an input file either as
2414 byte characters or as UTF-8, which is what "perltest" was being required to
2415 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2416 can't do is switch easily at run time between having the "use utf8;" pragma
2417 or not. In the end, I fudged it by using the RunPerlTest script to insert
2418 "use utf8;" explicitly for the UTF-8 tests.
2419
2420 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2421 the end of the subject string, contrary to the documentation and to what
2422 Perl does. This was true of both matching functions. Now it matches only at
2423 the start of the subject and immediately after *internal* newlines.
2424
2425 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2426 a pointer to an int instead of a pointer to an unsigned long int. This
2427 caused problems on 64-bit systems.
2428
2429 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2430 instance of the 'standard' template library not being so standard".
2431
2432 10. There was no check on the number of named subpatterns nor the maximum
2433 length of a subpattern name. The product of these values is used to compute
2434 the size of the memory block for a compiled pattern. By supplying a very
2435 long subpattern name and a large number of named subpatterns, the size
2436 computation could be caused to overflow. This is now prevented by limiting
2437 the length of names to 32 characters, and the number of named subpatterns
2438 to 10,000.
2439
2440 11. Subpatterns that are repeated with specific counts have to be replicated in
2441 the compiled pattern. The size of memory for this was computed from the
2442 length of the subpattern and the repeat count. The latter is limited to
2443 65535, but there was no limit on the former, meaning that integer overflow
2444 could in principle occur. The compiled length of a repeated subpattern is
2445 now limited to 30,000 bytes in order to prevent this.
2446
2447 12. Added the optional facility to have named substrings with the same name.
2448
2449 13. Added the ability to use a named substring as a condition, using the
2450 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2451 are numbers (not recommended). Forward references are permitted.
2452
2453 14. Added forward references in named backreferences (if you see what I mean).
2454
2455 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2456 pattern could run off the end of the subject. For example, the pattern
2457 "(?s)(.{1,5})"8 did this with the subject "ab".
2458
2459 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2460 PCRE_CASELESS was set when matching characters that were quantified with ?
2461 or *.
2462
2463 17. A character class other than a single negated character that had a minimum
2464 but no maximum quantifier - for example [ab]{6,} - was not handled
2465 correctly by pce_dfa_exec(). It would match only one character.
2466
2467 18. A valid (though odd) pattern that looked like a POSIX character
2468 class but used an invalid character after [ (for example [[,abc,]]) caused
2469 pcre_compile() to give the error "Failed: internal error: code overflow" or
2470 in some cases to crash with a glibc free() error. This could even happen if
2471 the pattern terminated after [[ but there just happened to be a sequence of
2472 letters, a binary zero, and a closing ] in the memory that followed.
2473
2474 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2475 over the years. Originally (before any Unicode support), just the bottom 8
2476 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2477 output from "man perlunicode" includes this:
2478
2479 The regular expression compiler produces polymorphic opcodes. That
2480 is, the pattern adapts to the data and automatically switches to
2481 the Unicode character scheme when presented with Unicode data--or
2482 instead uses a traditional byte scheme when presented with byte
2483 data.
2484
2485 Sadly, a wide octal escape does not cause a switch, and in a string with
2486 no other multibyte characters, these octal escapes are treated as before.
2487 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2488 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2489 Unicode string.
2490
2491 I have not perpetrated such confusion in PCRE. Up till now, it took just
2492 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2493 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2494 translate to the appropriate multibyte character.
2495
2496 29. Applied some refactoring to reduce the number of warnings from Microsoft
2497 and Borland compilers. This has included removing the fudge introduced
2498 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2499 a warning about an unused variable.
2500
2501 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2502 characters since release 4.0, because Perl (from release 5.004) does not.
2503 [Or at least, is documented not to: some releases seem to be in conflict
2504 with the documentation.] However, when a pattern was studied with
2505 pcre_study() and all its branches started with \s, PCRE still included VT
2506 as a possible starting character. Of course, this did no harm; it just
2507 caused an unnecessary match attempt.
2508
2509 22. Removed a now-redundant internal flag bit that recorded the fact that case
2510 dependency changed within the pattern. This was once needed for "required
2511 byte" processing, but is no longer used. This recovers a now-scarce options
2512 bit. Also moved the least significant internal flag bit to the most-
2513 significant bit of the word, which was not previously used (hangover from
2514 the days when it was an int rather than a uint) to free up another bit for
2515 the future.
2516
2517 23. Added support for CRLF line endings as well as CR and LF. As well as the
2518 default being selectable at build time, it can now be changed at runtime
2519 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2520 specify that it is scanning data with non-default line endings.
2521
2522 24. Changed the definition of CXXLINK to make it agree with the definition of
2523 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2524
2525 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2526 recursions. This makes a big different to stack usage for some patterns.
2527
2528 26. If a subpattern containing a named recursion or subroutine reference such
2529 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2530 the space required for the compiled pattern went wrong and gave too small a
2531 value. Depending on the environment, this could lead to "Failed: internal
2532 error: code overflow at offset 49" or "glibc detected double free or
2533 corruption" errors.
2534
2535 27. Applied patches from Google (a) to support the new newline modes and (b) to
2536 advance over multibyte UTF-8 characters in GlobalReplace.
2537
2538 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2539 difference for some implementation of PCRE in some Windows version.
2540
2541 29. Added some extra testing facilities to pcretest:
2542
2543 \q<number> in a data line sets the "match limit" value
2544 \Q<number> in a data line sets the "match recursion limt" value
2545 -S <number> sets the stack size, where <number> is in megabytes
2546
2547 The -S option isn't available for Windows.
2548
2549
2550 Version 6.6 06-Feb-06
2551 ---------------------
2552
2553 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2554 in pcreposix.h. I have copied the definition from pcre.h.
2555
2556 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2557 because pcre.h is no longer a built file.
2558
2559 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2560 not normally included in the compiled code.
2561
2562
2563 Version 6.5 01-Feb-06
2564 ---------------------
2565
2566 1. When using the partial match feature with pcre_dfa_exec(), it was not
2567 anchoring the second and subsequent partial matches at the new starting
2568 point. This could lead to incorrect results. For example, with the pattern
2569 /1234/, partially matching against "123" and then "a4" gave a match.
2570
2571 2. Changes to pcregrep:
2572
2573 (a) All non-match returns from pcre_exec() were being treated as failures
2574 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2575 error message is output. Some extra information is given for the
2576 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2577 probably the only errors that are likely to be caused by users (by
2578 specifying a regex that has nested indefinite repeats, for instance).
2579 If there are more than 20 of these errors, pcregrep is abandoned.
2580
2581 (b) A binary zero was treated as data while matching, but terminated the
2582 output line if it was written out. This has been fixed: binary zeroes
2583 are now no different to any other data bytes.
2584
2585 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2586 used to set a locale for matching. The --locale=xxxx long option has
2587 been added (no short equivalent) to specify a locale explicitly on the
2588 pcregrep command, overriding the environment variables.
2589
2590 (d) When -B was used with -n, some line numbers in the output were one less
2591 than they should have been.
2592
2593 (e) Added the -o (--only-matching) option.
2594
2595 (f) If -A or -C was used with -c (count only), some lines of context were
2596 accidentally printed for the final match.
2597
2598 (g) Added the -H (--with-filename) option.
2599
2600 (h) The combination of options -rh failed to suppress file names for files
2601 that were found from directory arguments.
2602
2603 (i) Added the -D (--devices) and -d (--directories) options.
2604
2605 (j) Added the -F (--fixed-strings) option.
2606
2607 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2608
2609 (l) Added the --colo(u)r option.
2610
2611 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2612 is not present by default.
2613
2614 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2615 items such as (?R) or (?1), when the recursion could match a number of
2616 alternatives. If it matched one of the alternatives, but subsequently,
2617 outside the recursion, there was a failure, the code tried to back up into
2618 the recursion. However, because of the way PCRE is implemented, this is not
2619 possible, and the result was an incorrect result from the match.
2620
2621 In order to prevent this happening, the specification of recursion has
2622 been changed so that all such subpatterns are automatically treated as
2623 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2624
2625 4. I had overlooked the fact that, in some locales, there are characters for
2626 which isalpha() is true but neither isupper() nor islower() are true. In
2627 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2628 and ordfeminine) are like this. This affected the treatment of \w and \W
2629 when they appeared in character classes, but not when they appeared outside
2630 a character class. The bit map for "word" characters is now created
2631 separately from the results of isalnum() instead of just taking it from the
2632 upper, lower, and digit maps. (Plus the underscore character, of course.)
2633
2634 5. The above bug also affected the handling of POSIX character classes such as
2635 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2636 permanent tables. Instead, the bit maps for such a class were previously
2637 created as the appropriate unions of the upper, lower, and digit bitmaps.
2638 Now they are created by subtraction from the [[:word:]] class, which has
2639 its own bitmap.
2640
2641 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2642 It is created by subtracting the vertical space characters (\x09, \x0a,
2643 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2644 subtraction was done in the overall bitmap for a character class, meaning
2645 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2646 be recognized. This bug has been fixed.
2647
2648 7. Patches from the folks at Google:
2649
2650 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2651 real life, but is still worth protecting against".
2652
2653 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2654 regular expressions".
2655
2656 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2657 have it.
2658
2659 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2660 "configure" and the latter not, in order to fix a problem somebody had
2661 with compiling the Arg class on HP-UX.
2662
2663 (e) Improve the error-handling of the C++ wrapper a little bit.
2664
2665 (f) New tests for checking recursion limiting.
2666
2667 8. The pcre_memmove() function, which is used only if the environment does not
2668 have a standard memmove() function (and is therefore rarely compiled),
2669 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2670 returning a result (though PCRE never actually uses the result).
2671
2672 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2673 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2674 returned instead of calling malloc() with an overflowing number that would
2675 most likely cause subsequent chaos.
2676
2677 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2678
2679 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2680 with this option is matched, the nmatch and pmatch options of regexec() are
2681 ignored.
2682
2683 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2684 provided in case anyone wants to the the POSIX interface with UTF-8
2685 strings.
2686
2687 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2688 C++ linking (needed for some HP-UX environments).
2689
2690 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2691 (unused parameter) and in the pcre_printint() function (omitted "default"
2692 switch label when the default is to do nothing).
2693
2694 15. Added some code to make it possible, when PCRE is compiled as a C++
2695 library, to replace subject pointers for pcre_exec() with a smart pointer
2696 class, thus making it possible to process discontinuous strings.
2697
2698 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2699 much the same function. They were added by different people who were trying
2700 to make PCRE easy to compile on non-Unix systems. It has been suggested
2701 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2702 for compiling on Windows systems. I have therefore replaced it with
2703 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2704 defaults to "extern" for C or "extern C" for C++, which works fine on
2705 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2706 SCOPE with something explicit in config.h. In addition:
2707
2708 (a) pcreposix.h still had just "extern" instead of either of these macros;
2709 I have replaced it with PCRE_DATA_SCOPE.
2710
2711 (b) Functions such as _pcre_xclass(), which are internal to the library,
2712 but external in the C sense, all had PCRE_EXPORT in their definitions.
2713 This is apparently wrong for the Windows case, so I have removed it.
2714 (It makes no difference on Unix-like systems.)
2715
2716 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2717 of recursive calls to match(). This is different to MATCH_LIMIT because
2718 that limits the total number of calls to match(), not all of which increase
2719 the depth of recursion. Limiting the recursion depth limits the amount of
2720 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2721 when PCRE is compiled, and changed at run time. A patch from Google adds
2722 this functionality to the C++ interface.
2723
2724 18. Changes to the handling of Unicode character properties:
2725
2726 (a) Updated the table to Unicode 4.1.0.
2727
2728 (b) Recognize characters that are not in the table as "Cn" (undefined).
2729
2730 (c) I revised the way the table is implemented to a much improved format
2731 which includes recognition of ranges. It now supports the ranges that
2732 are defined in UnicodeData.txt, and it also amalgamates other
2733 characters into ranges. This has reduced the number of entries in the
2734 table from around 16,000 to around 3,000, thus reducing its size
2735 considerably. I realized I did not need to use a tree structure after
2736 all - a binary chop search is just as efficient. Having reduced the
2737 number of entries, I extended their size from 6 bytes to 8 bytes to
2738 allow for more data.
2739
2740 (d) Added support for Unicode script names via properties such as \p{Han}.
2741
2742 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2743 matching that character.
2744
2745 20. When matching a repeated Unicode property with a minimum greater than zero,
2746 (for example \pL{2,}), PCRE could look past the end of the subject if it
2747 reached it while seeking the minimum number of characters. This could
2748 happen only if some of the characters were more than one byte long, because
2749 there is a check for at least the minimum number of bytes.
2750
2751 21. Refactored the implementation of \p and \P so as to be more general, to
2752 allow for more different types of property in future. This has changed the
2753 compiled form incompatibly. Anybody with saved compiled patterns that use
2754 \p or \P will have to recompile them.
2755
2756 22. Added "Any" and "L&" to the supported property types.
2757
2758 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2759 but give a compile time error if the value is greater than 0xff.
2760
2761 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2762 accidentally not being installed or uninstalled.
2763
2764 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2765 made were to insert the current release number. This seemed silly, because
2766 it made things harder for people building PCRE on systems that don't run
2767 "configure". I have turned pcre.h into a distributed file, no longer built
2768 by "configure", with the version identification directly included. There is
2769 no longer a pcre.h.in file.
2770
2771 However, this change necessitated a change to the pcre-config script as
2772 well. It is built from pcre-config.in, and one of the substitutions was the
2773 release number. I have updated configure.ac so that ./configure now finds
2774 the release number by grepping pcre.h.
2775
2776 26. Added the ability to run the tests under valgrind.
2777
2778
2779 Version 6.4 05-Sep-05
2780 ---------------------
2781
2782 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2783 "--" to be printed when multiple files were scanned, even when none of the
2784 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2785 consider it to be a bug, and have restored the previous behaviour.
2786
2787 2. A couple of code tidies to get rid of compiler warnings.
2788
2789 3. The pcretest program used to cheat by referring to symbols in the library
2790 whose names begin with _pcre_. These are internal symbols that are not
2791 really supposed to be visible externally, and in some environments it is
2792 possible to suppress them. The cheating is now confined to including
2793 certain files from the library's source, which is a bit cleaner.
2794
2795 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2796 file's purpose clearer.
2797
2798 5. Reorganized pcre_ucp_findchar().
2799
2800
2801 Version 6.3 15-Aug-05
2802 ---------------------
2803
2804 1. The file libpcre.pc.in did not have general read permission in the tarball.
2805
2806 2. There were some problems when building without C++ support:
2807
2808 (a) If C++ support was not built, "make install" and "make test" still
2809 tried to test it.
2810
2811 (b) There were problems when the value of CXX was explicitly set. Some
2812 changes have been made to try to fix these, and ...
2813
2814 (c) --disable-cpp can now be used to explicitly disable C++ support.
2815
2816 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2817 backslash in a target when C++ was disabled. This confuses some
2818 versions of "make", apparently. Using an intermediate variable solves
2819 this. (Same for CPP_LOBJ.)
2820
2821 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2822 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2823 necessary on certain architectures.
2824
2825 4. Added a setting of -export-symbols-regex to the link command to remove
2826 those symbols that are exported in the C sense, but actually are local
2827 within the library, and not documented. Their names all begin with
2828 "_pcre_". This is not a perfect job, because (a) we have to except some
2829 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2830 available (and never for static libraries). I have made a note to try to
2831 find a way round (a) in the future.
2832
2833
2834 Version 6.2 01-Aug-05
2835 ---------------------
2836
2837 1. There was no test for integer overflow of quantifier values. A construction
2838 such as {1111111111111111} would give undefined results. What is worse, if
2839 a minimum quantifier for a parenthesized subpattern overflowed and became
2840 negative, the calculation of the memory size went wrong. This could have
2841 led to memory overwriting.
2842
2843 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2844
2845 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2846 operating environments where this matters.
2847
2848 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2849 PCRE options from within the C++ wrapper.
2850
2851 5. Named capturing subpatterns were not being correctly counted when a pattern
2852 was compiled. This caused two problems: (a) If there were more than 100
2853 such subpatterns, the calculation of the memory needed for the whole
2854 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2855 back references of the form \12, where the number was greater than 9, were
2856 not recognized as back references, even though there were sufficient
2857 previous subpatterns.
2858
2859 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2860 versions of gcc, e.g. 2.95.4.
2861
2862
2863 Version 6.1 21-Jun-05
2864 ---------------------
2865
2866 1. There was one reference to the variable "posix" in pcretest.c that was not
2867 surrounded by "#if !defined NOPOSIX".
2868
2869 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2870 the cross-check on the old pcre_info() function, for the benefit of the
2871 cut-down version of PCRE that is currently imported into Exim.
2872
2873 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2874 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2875 patterns that start (?i) but I don't think that matters. The use of (?i) is
2876 just an example; this all applies to the other options as well.
2877
2878 4. Since libtool seems to echo the compile commands it is issuing, the output
2879 from "make" can be reduced a bit by putting "@" in front of each libtool
2880 compile command.
2881
2882 5. Patch from the folks at Google for configure.in to be a bit more thorough
2883 in checking for a suitable C++ installation before trying to compile the
2884 C++ stuff. This should fix a reported problem when a compiler was present,
2885 but no suitable headers.
2886
2887 6. The man pages all had just "PCRE" as their title. I have changed them to
2888 be the relevant file name. I have also arranged that these names are
2889 retained in the file doc/pcre.txt, which is a concatenation in text format
2890 of all the man pages except the little individual ones for each function.
2891
2892 7. The NON-UNIX-USE file had not been updated for the different set of source
2893 files that come with release 6. I also added a few comments about the C++
2894 wrapper.
2895
2896
2897 Version 6.0 07-Jun-05
2898 ---------------------
2899
2900 1. Some minor internal re-organization to help with my DFA experiments.
2901
2902 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2903 didn't matter for the library itself when fully configured, but did matter
2904 when compiling without UCP support, or within Exim, where the ucp files are
2905 not imported.
2906
2907 3. Refactoring of the library code to split up the various functions into
2908 different source modules. The addition of the new DFA matching code (see
2909 below) to a single monolithic source would have made it really too
2910 unwieldy, quite apart from causing all the code to be include in a
2911 statically linked application, when only some functions are used. This is
2912 relevant even without the DFA addition now that patterns can be compiled in
2913 one application and matched in another.
2914
2915 The downside of splitting up is that there have to be some external
2916 functions and data tables that are used internally in different modules of
2917 the library but which are not part of the API. These have all had their
2918 names changed to start with "_pcre_" so that they are unlikely to clash
2919 with other external names.
2920
2921 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2922 a different (DFA) algorithm. Although it is slower than the original
2923 function, it does have some advantages for certain types of matching
2924 problem.
2925
2926 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2927 including restarting after a partial match.
2928
2929 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2930 defined when compiling for Windows was sent to me. I have put it into the
2931 code, though I have no means of testing or verifying it.
2932
2933 7. Added the pcre_refcount() auxiliary function.
2934
2935 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2936 match before or at the first newline in the subject string. In pcretest,
2937 the /f option on a pattern can be used to set this.
2938
2939 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2940 would behave wrongly. This has been present in PCRE since release 4.0.
2941
2942 10. A number of changes to the pcregrep command:
2943
2944 (a) Refactored how -x works; insert ^(...)$ instead of setting
2945 PCRE_ANCHORED and checking the length, in preparation for adding
2946 something similar for -w.
2947
2948 (b) Added the -w (match as a word) option.
2949
2950 (c) Refactored the way lines are read and buffered so as to have more
2951 than one at a time available.
2952
2953 (d) Implemented a pcregrep test script.
2954
2955 (e) Added the -M (multiline match) option. This allows patterns to match
2956 over several lines of the subject. The buffering ensures that at least
2957 8K, or the rest of the document (whichever is the shorter) is available
2958 for matching (and similarly the previous 8K for lookbehind assertions).
2959
2960 (f) Changed the --help output so that it now says
2961
2962 -w, --word-regex(p)
2963
2964 instead of two lines, one with "regex" and the other with "regexp"
2965 because that confused at least one person since the short forms are the
2966 same. (This required a bit of code, as the output is generated
2967 automatically from a table. It wasn't just a text change.)
2968
2969 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2970 option but starts with a hyphen. Could be a pattern or a path name
2971 starting with a hyphen, for instance.
2972
2973 (h) "-" can be given as a file name to represent stdin.
2974
2975 (i) When file names are being printed, "(standard input)" is used for
2976 the standard input, for compatibility with GNU grep. Previously
2977 "<stdin>" was used.
2978
2979 (j) The option --label=xxx can be used to supply a name to be used for
2980 stdin when file names are being printed. There is no short form.
2981
2982 (k) Re-factored the options decoding logic because we are going to add
2983 two more options that take data. Such options can now be given in four
2984 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2985
2986 (l) Added the -A, -B, and -C options for requesting that lines of context
2987 around matches be printed.
2988
2989 (m) Added the -L option to print the names of files that do not contain
2990 any matching lines, that is, the complement of -l.
2991
2992 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2993 continue to scan other files.
2994
2995 (o) The -s option was incorrectly implemented. For compatibility with other
2996 greps, it now suppresses the error message for a non-existent or non-
2997 accessible file (but not the return code). There is a new option called
2998 -q that suppresses the output of matching lines, which was what -s was
2999 previously doing.
3000
3001 (p) Added --include and --exclude options to specify files for inclusion
3002 and exclusion when recursing.
3003
3004 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3005 Hopefully, it now does.
3006
3007 12. Missing cast in pcre_study().
3008
3009 13. Added an "uninstall" target to the makefile.
3010
3011 14. Replaced "extern" in the function prototypes in Makefile.in with
3012 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3013 world, but is set differently for Windows.
3014
3015 15. Added a second compiling function called pcre_compile2(). The only
3016 difference is that it has an extra argument, which is a pointer to an
3017 integer error code. When there is a compile-time failure, this is set
3018 non-zero, in addition to the error test pointer being set to point to an
3019 error message. The new argument may be NULL if no error number is required
3020 (but then you may as well call pcre_compile(), which is now just a
3021 wrapper). This facility is provided because some applications need a
3022 numeric error indication, but it has also enabled me to tidy up the way
3023 compile-time errors are handled in the POSIX wrapper.
3024
3025 16. Added VPATH=.libs to the makefile; this should help when building with one
3026 prefix path and installing with another. (Or so I'm told by someone who
3027 knows more about this stuff than I do.)
3028
3029 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3030 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3031 match everything, including newlines. This is not POSIX-compatible, but
3032 somebody wanted the feature. From pcretest it can be activated by using
3033 both the P and the s flags.
3034
3035 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3036
3037 19. libpcre.pc was being incorrectly installed as executable.
3038
3039 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3040 it now also looks for '\r' so that it will work unmodified on Windows.
3041
3042 21. Added Google's contributed C++ wrapper to the distribution.
3043
3044 22. Added some untidy missing memory free() calls in pcretest, to keep
3045 Electric Fence happy when testing.
3046
3047
3048
3049 Version 5.0 13-Sep-04
3050 ---------------------
3051
3052 1. Internal change: literal characters are no longer packed up into items
3053 containing multiple characters in a single byte-string. Each character
3054 is now matched using a separate opcode. However, there may be more than one
3055 byte in the character in UTF-8 mode.
3056
3057 2. The pcre_callout_block structure has two new fields: pattern_position and
3058 next_item_length. These contain the offset in the pattern to the next match
3059 item, and its length, respectively.
3060
3061 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3062 insertion of callouts before each pattern item. Added the /C option to
3063 pcretest to make use of this.
3064
3065 4. On the advice of a Windows user, the lines
3066
3067 #if defined(_WIN32) || defined(WIN32)
3068 _setmode( _fileno( stdout ), 0x8000 );
3069 #endif /* defined(_WIN32) || defined(WIN32) */
3070
3071 have been added to the source of pcretest. This apparently does useful
3072 magic in relation to line terminators.
3073
3074 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3075 for the benefit of those environments where the "b" makes a difference.
3076
3077 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3078 to know about it. I have put a hack into configure.in that adds in code
3079 to set GCC=yes if CC=icc. This seems to end up at a point in the
3080 generated configure script that is early enough to affect the setting of
3081 compiler options, which is what is needed, but I have no means of testing
3082 whether it really works. (The user who reported this had patched the
3083 generated configure script, which of course I cannot do.)
3084
3085 LATER: After change 22 below (new libtool files), the configure script
3086 seems to know about icc (and also ecc). Therefore, I have commented out
3087 this hack in configure.in.
3088
3089 7. Added support for pkg-config (2 patches were sent in).
3090
3091 8. Negated POSIX character classes that used a combination of internal tables
3092 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3093 [[:^ascii]]. Typically, they would match almost any characters. The other
3094 POSIX classes were not broken in this way.
3095
3096 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3097 to find the match, as PCRE was deluded into thinking that the match had to
3098 start at the start point or following a newline. The same bug applied to
3099 patterns with negative forward assertions or any backward assertions
3100 preceding ".*" at the start, unless the pattern required a fixed first
3101 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3102
3103 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3104 starting at the last subject character, bytes beyond the end of the subject
3105 string were read.
3106
3107 11. Renamed the variable "class" as "classbits" to make life easier for C++
3108 users. (Previously there was a macro definition, but it apparently wasn't
3109 enough.)
3110
3111 12. Added the new field "tables" to the extra data so that tables can be passed
3112 in at exec time, or the internal tables can be re-selected. This allows
3113 a compiled regex to be saved and re-used at a later time by a different
3114 program that might have everything at different addresses.
3115
3116 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3117 -R library as well as a -L library.
3118
3119 14. The debugging options of pcretest (-d on the command line or D on a
3120 pattern) showed incorrect output for anything following an extended class
3121 that contained multibyte characters and which was followed by a quantifier.
3122
3123 15. Added optional support for general category Unicode character properties
3124 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3125 support. It adds about 90K to the size of the library. The meanings of the
3126 inbuilt class escapes such as \d and \s have NOT been changed.
3127
3128 16. Updated pcredemo.c to include calls to free() to release the memory for the
3129 compiled pattern.
3130
3131 17. The generated file chartables.c was being created in the source directory
3132 instead of in the building directory. This caused the build to fail if the
3133 source directory was different from the building directory, and was
3134 read-only.
3135
3136 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3137 file. No doubt somebody will tell me if they don't make sense... Also added
3138 Dan Mooney's comments about building on OpenVMS.
3139
3140 19. Added support for partial matching via the PCRE_PARTIAL option for
3141 pcre_exec() and the \P data escape in pcretest.
3142
3143 20. Extended pcretest with 3 new pattern features:
3144
3145 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3146 write the compiled pattern to the file whose name is "rest-of-line".
3147 This is a straight binary dump of the data, with the saved pointer to
3148 the character tables forced to be NULL. The study data, if any, is
3149 written too. After writing, pcretest reads a new pattern.
3150
3151 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3152 compiled pattern from the given file. There must not be any
3153 occurrences of "<" in the file name (pretty unlikely); if there are,
3154 pcretest will instead treat the initial "<" as a pattern delimiter.
3155 After reading in the pattern, pcretest goes on to read data lines as
3156 usual.
3157
3158 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3159 and 16-bit fields in a compiled pattern, to simulate a pattern that
3160 was compiled on a host of opposite endianness.
3161
3162 21. The pcre-exec() function can now cope with patterns that were compiled on
3163 hosts of opposite endianness, with this restriction:
3164
3165 As for any compiled expression that is saved and used later, the tables
3166 pointer field cannot be preserved; the extra_data field in the arguments
3167 to pcre_exec() should be used to pass in a tables address if a value
3168 other than the default internal tables were used at compile time.
3169
3170 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3171 now diagnosed as an error. Previously, most of the time, a negative number
3172 would have been treated as zero, but if in addition "ovector" was passed as
3173 NULL, a crash could occur.
3174
3175 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3176 new versions from the libtool 1.5 distribution (the last one is a copy of
3177 a file called libtool.m4). This seems to have fixed the need to patch
3178 "configure" to support Darwin 1.3 (which I used to do). However, I still
3179 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3180 workstation).
3181
3182 24. Changed the PCRE licence to be the more standard "BSD" licence.
3183
3184
3185 Version 4.5 01-Dec-03
3186 ---------------------
3187
3188 1. There has been some re-arrangement of the code for the match() function so
3189 that it can be compiled in a version that does not call itself recursively.
3190 Instead, it keeps those local variables that need separate instances for
3191 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3192 needs to "recurse". Keeping track of where control must go is done by means
3193 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3194 hide most of the details from the main code, and operates only if
3195 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3196 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3197 operating.
3198
3199 To make it easier for callers to provide specially tailored get/free
3200 functions for this usage, two new functions, pcre_stack_malloc, and
3201 pcre_stack_free, are used. They are always called in strict stacking order,
3202 and the size of block requested is always the same.
3203
3204 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3205 PCRE has been compiled to use the stack or the heap for recursion. The
3206 -C option of pcretest uses this to show which version is compiled.
3207
3208 A new data escape \S, is added to pcretest; it causes the amounts of store
3209 obtained and freed by both kinds of malloc/free at match time to be added
3210 to the output.
3211
3212 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3213 what's available on my current Linux desktop machine.
3214
3215 3. When matching a UTF-8 string, the test for a valid string at the start has
3216 been extended. If start_offset is not zero, PCRE now checks that it points
3217 to a byte that is the start of a UTF-8 character. If not, it returns
3218 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3219 this is necessary because there may be backward assertions in the pattern.
3220 When matching the same subject several times, it may save resources to use
3221 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3222
3223 4. The code for checking the validity of UTF-8 strings has been tightened so
3224 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3225 containing "overlong sequences".
3226
3227 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3228 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3229 should have been "&", but it just so happened that all the cases this let
3230 through by mistake were picked up later in the function.
3231
3232 6. I had used a variable called "isblank" - this is a C99 function, causing
3233 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3234
3235 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3236 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3237 so the version is shown; (c) stop "make test" echoing "./RunTest".
3238
3239 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3240
3241 9. The prototype for memmove() for systems that don't have it was using
3242 size_t, but the inclusion of the header that defines size_t was later. I've
3243 moved the #includes for the C headers earlier to avoid this.
3244
3245 10. Added some adjustments to the code to make it easier to compiler on certain
3246 special systems:
3247
3248 (a) Some "const" qualifiers were missing.
3249 (b) Added the macro EXPORT before all exported functions; by default this
3250 is defined to be empty.
3251 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3252 that it reads its output file name as an argument instead of writing
3253 to the standard output and assuming this can be redirected.
3254
3255 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3256 class containing characters with values greater than 255, PCRE compilation
3257 went into a loop.
3258
3259 12. A recursive reference to a subpattern that was within another subpattern
3260 that had a minimum quantifier of zero caused PCRE to crash. For example,
3261 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3262 recursion. If the recursively-called subpattern itself had a zero repeat,
3263 that was OK.
3264
3265 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3266 buffer into which it was copied (for escape processing) was still set at
3267 1024, so long lines caused crashes.
3268
3269 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3270 "internal error: code overflow...". This applied to any character class
3271 that was followed by a possessive quantifier.
3272
3273 15. Modified the Makefile to add libpcre.la as a prerequisite for
3274 libpcreposix.la because I was told this is needed for a parallel build to
3275 work.
3276
3277 16. If a pattern that contained .* following optional items at the start was
3278 studied, the wrong optimizing data was generated, leading to matching
3279 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3280 matching string must start with a or b or c. The correct conclusion for
3281 this pattern is that a match can start with any character.
3282
3283
3284 Version 4.4 13-Aug-03
3285 ---------------------
3286
3287 1. In UTF-8 mode, a character class containing characters with values between
3288 127 and 255 was not handled correctly if the compiled pattern was studied.
3289 In fixing this, I have also improved the studying algorithm for such
3290 classes (slightly).
3291
3292 2. Three internal functions had redundant arguments passed to them. Removal
3293 might give a very teeny performance improvement.
3294
3295 3. Documentation bug: the value of the capture_top field in a callout is *one
3296 more than* the number of the hightest numbered captured substring.
3297
3298 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3299 in incorrectly linking with a previously installed version. They now link
3300 explicitly with libpcre.la.
3301
3302 5. configure.in no longer needs to recognize Cygwin specially.
3303
3304 6. A problem in pcre.in for Windows platforms is fixed.
3305
3306 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3307 pcretest, it used to include the size of the study block as part of its
3308 output. Unfortunately, the structure contains a field that has a different
3309 size on different hardware architectures. This meant that the tests that
3310 showed this size failed. As the block is currently always of a fixed size,
3311 this information isn't actually particularly useful in pcretest output, so
3312 I have just removed it.
3313
3314 8. Three pre-processor statements accidentally did not start in column 1.
3315 Sadly, there are *still* compilers around that complain, even though
3316 standard C has not required this for well over a decade. Sigh.
3317
3318 9. In pcretest, the code for checking callouts passed small integers in the
3319 callout_data field, which is a void * field. However, some picky compilers
3320 complained about the casts involved for this on 64-bit systems. Now
3321 pcretest passes the address of the small integer instead, which should get
3322 rid of the warnings.
3323
3324 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3325 both compile and run time, and gives an error if an invalid UTF-8 sequence
3326 is found. There is a option for disabling this check in cases where the
3327 string is known to be correct and/or the maximum performance is wanted.
3328
3329 11. In response to a bug report, I changed one line in Makefile.in from
3330
3331 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3332 to
3333 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3334
3335 to look similar to other lines, but I have no way of telling whether this
3336 is the right thing to do, as I do not use Windows. No doubt I'll get told
3337 if it's wrong...
3338
3339
3340 Version 4.3 21-May-03
3341 ---------------------
3342
3343 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3344 Makefile.
3345
3346 2. Some refactoring to improve the quality of the code:
3347
3348 (i) The utf8_table... variables are now declared "const".
3349
3350 (ii) The code for \cx, which used the "case flipping" table to upper case
3351 lower case letters, now just substracts 32. This is ASCII-specific,
3352 but the whole concept of \cx is ASCII-specific, so it seems
3353 reasonable.
3354
3355 (iii) PCRE was using its character types table to recognize decimal and
3356 hexadecimal digits in the pattern. This is silly, because it handles
3357 only 0-9, a-f, and A-F, but the character types table is locale-
3358 specific, which means strange things might happen. A private
3359 table is now used for this - though it costs 256 bytes, a table is
3360 much faster than multiple explicit tests. Of course, the standard
3361 character types table is still used for matching digits in subject
3362 strings against \d.
3363
3364 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3365 ending in _t are). So I've renamed it as ESC_tee.
3366
3367 3. The first argument for regexec() in the POSIX wrapper should have been
3368 defined as "const".
3369
3370 4. Changed pcretest to use malloc() for its buffers so that they can be
3371 Electric Fenced for debugging.
3372
3373 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3374 to read one or more bytes before the start of the subject string. Often this
3375 had no effect on PCRE's behaviour, but in some circumstances it could
3376 provoke a segmentation fault.
3377
3378 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3379 to try to read one or more bytes before the start of the subject string.
3380
3381 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3382 UTF-8 support could misbehave in various ways if the subject string
3383 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3384 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3385 back over UTF-8 characters.)
3386
3387
3388 Version 4.2 14-Apr-03
3389 ---------------------
3390
3391 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3392
3393 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3394 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3395 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3396 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3397 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3398 and BUILD_EXEEXT
3399 Note: automatic setting of the BUILD variables is not yet working
3400 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3401 compile-time but not at link-time
3402 [LINK]: use for linking executables only
3403 make different versions for Windows and non-Windows
3404 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3405 libraries
3406 [LINK_FOR_BUILD]: new variable
3407 [OBJEXT]: use throughout
3408 [EXEEXT]: use throughout
3409 <winshared>: new target
3410 <wininstall>: new target
3411 <dftables.o>: use native compiler
3412 <dftables>: use native linker
3413 <install>: handle Windows platform correctly
3414 <clean>: ditto
3415 <check>: ditto
3416 copy DLL to top builddir before testing
3417
3418 As part of these changes, -no-undefined was removed again. This was reported
3419 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3420 in any case.
3421
3422 3. Some tidies to get rid of compiler warnings:
3423
3424 . In the match_data structure, match_limit was an unsigned long int, whereas
3425 match_call_count was an int. I've made them both unsigned long ints.
3426
3427 . In pcretest the fact that a const uschar * doesn't automatically cast to
3428 a void * provoked a warning.
3429
3430 . Turning on some more compiler warnings threw up some "shadow" variables
3431 and a few more missing casts.
3432
3433 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3434 option, a class that contained a single character with a value between 128
3435 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3436
3437 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3438 option, a class that contained several characters, but with at least one
3439 whose value was between 128 and 255 caused PCRE to crash.
3440
3441
3442 Version 4.1 12-Mar-03
3443 ---------------------
3444
3445 1. Compiling with gcc -pedantic found a couple of places where casts were
3446 needed, and a string in dftables.c that was longer than standard compilers are
3447 required to support.
3448
3449 2. Compiling with Sun's compiler found a few more places where the code could
3450 be tidied up in order to avoid warnings.
3451
3452 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3453 first of these names is deprecated in the latest Autoconf in favour of the name
3454 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3455 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3456 analogy I have changed it to CFLAGS_FOR_BUILD.
3457
3458 4. Added -no-undefined to the linking command in the Makefile, because this is
3459 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3460 linking step for the pcreposix library.
3461
3462 5. PCRE was failing to diagnose the case of two named groups with the same
3463 name.
3464
3465 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3466 literal character that is needed in the subject for a match, and scans along to
3467 ensure that it is present before embarking on the full matching process. This
3468 saves time in cases of nested unlimited repeats that are never going to match.
3469 Problem: the scan can take a lot of time if the subject is very long (e.g.
3470 megabytes), thus penalizing straightforward matches. It is now done only if the
3471 amount of subject to be scanned is less than 1000 bytes.
3472
3473 7. A lesser problem with the same optimization is that it was recording the
3474 first character of an anchored pattern as "needed", thus provoking a search
3475 right along the subject, even when the first match of the pattern was going to
3476 fail. The "needed" character is now not set for anchored patterns, unless it
3477 follows something in the pattern that is of non-fixed length. Thus, it still
3478 fulfils its original purpose of finding quick non-matches in cases of nested
3479 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3480
3481
3482 Version 4.0 17-Feb-03
3483 ---------------------
3484
3485 1. If a comment in an extended regex that started immediately after a meta-item
3486 extended to the end of string, PCRE compiled incorrect data. This could lead to
3487 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3488
3489 2. Moved to autoconf 2.53 and libtool 1.4.2.
3490
3491 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3492 the special perltest8 script is no longer needed - all the tests can be run
3493 from a single perltest script.
3494
3495 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3496 by \s. It has now been removed in PCRE. This means it isn't recognized as
3497 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3498 class [:space:] *does* include VT, thereby creating a mess.
3499
3500 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3501 space and tab.
3502
3503 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3504 its new features into the main test script, reducing the number of scripts.
3505
3506 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3507 were backward compatible, and made the (?i) apply to the whole pattern, as if
3508 /i were given. Now it behaves more logically, and applies the option setting
3509 only to what follows. PCRE has been changed to follow suit. However, if it
3510 finds options settings right at the start of the pattern, it extracts them into
3511 the global options, as before. Thus, they show up in the info data.
3512
3513 8. Added support for the \Q...\E escape sequence. Characters in between are
3514 treated as literals. This is slightly different from Perl in that $ and @ are
3515 also handled as literals inside the quotes. In Perl, they will cause variable
3516 interpolation. Note the following examples:
3517
3518 Pattern PCRE matches Perl matches
3519
3520 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3521 \Qabc\$xyz\E abc\$xyz abc\$xyz
3522 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3523
3524 For compatibility with Perl, \Q...\E sequences are recognized inside character
3525 classes as well as outside them.
3526
3527 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3528 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3529 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3530 signed/unsigned warnings.
3531
3532 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3533 option for pcretest, so I've replaced it by a simple function that does just
3534 that job.
3535
3536 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3537 "pcregrep -".
3538
3539 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3540 Java package. This provides some syntactic sugar for simple cases of what my
3541 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3542 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3543 item, you can use this simplified notation. Note that only makes sense with
3544 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3545 greediness, whatever the setting of the PCRE_UNGREEDY option.
3546
3547 13. A change of greediness default within a pattern was not taking effect at
3548 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3549 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3550 was abstracted outside.
3551
3552 14. PCRE now supports the \G assertion. It is true when the current matching
3553 position is at the start point of the match. This differs from \A when the
3554 starting offset is non-zero. Used with the /g option of pcretest (or similar
3555 code), it works in the same way as it does for Perl's /g option. If all
3556 alternatives of a regex begin with \G, the expression is anchored to the start
3557 match position, and the "anchored" flag is set in the compiled expression.
3558
3559 15. Some bugs concerning the handling of certain option changes within patterns
3560 have been fixed. These applied to options other than (?ims). For example,
3561 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3562 been the other way round. Some of this was related to change 7 above.
3563
3564 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3565 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3566 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3567 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3568
3569 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3570 mode. Unlike ".", it always matches newline, whatever the setting of
3571 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3572 assertions. Perl allows it, but it doesn't (in general) work because it can't
3573 calculate the length of the lookbehind. At least, that's the case for Perl
3574 5.8.0 - I've been told they are going to document that it doesn't work in
3575 future.
3576
3577 18. Added an error diagnosis for escapes that PCRE does not support: these are
3578 \L, \l, \N, \P, \p, \U, \u, and \X.
3579
3580 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3581 reading past the end of the pattern in cases such as /[abcd/.
3582
3583 20. PCRE was getting more memory than necessary for patterns with classes that
3584 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3585
3586 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3587 compiling PCRE for use with Virtual Pascal.
3588
3589 22. Small fix to the Makefile to make it work properly if the build is done
3590 outside the source tree.
3591
3592 23. Added a new extension: a condition to go with recursion. If a conditional
3593 subpattern starts with (?(R) the "true" branch is used if recursion has
3594 happened, whereas the "false" branch is used only at the top level.
3595
3596 24. When there was a very long string of literal characters (over 255 bytes
3597 without UTF support, over 250 bytes with UTF support), the computation of how
3598 much memory was required could be incorrect, leading to segfaults or other
3599 strange effects.
3600
3601 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3602 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3603 there was a subsequent back reference to those brackets. This meant that, for
3604 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3605 possible to check for precisely this case. All we can do is abandon the
3606 optimization if .* occurs inside capturing brackets when there are any back
3607 references whatsoever. (See below for a better fix that came later.)
3608
3609 26. The handling of the optimization for finding the first character of a
3610 non-anchored pattern, and for finding a character that is required later in the
3611 match were failing in some cases. This didn't break the matching; it just
3612 failed to optimize when it could. The way this is done has been re-implemented.
3613
3614 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3615
3616 28. Added a new feature that provides some of the functionality that Perl
3617 provides with (?{...}). The facility is termed a "callout". The way it is done
3618 in PCRE is for the caller to provide an optional function, by setting
3619 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3620 global variable. By default it is unset, which disables all calling out. To get
3621 the function called, the regex must include (?C) at appropriate points. This
3622 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3623 This provides a means of identifying different callout points. When PCRE
3624 reaches such a point in the regex, if pcre_callout has been set, the external
3625 function is called. It is provided with data in a structure called
3626 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3627 matching continues; if it returns a non-zero value, the match at the current
3628 point fails. However, backtracking will occur if possible. [This was changed
3629 later and other features added - see item 49 below.]
3630
3631 29. pcretest is upgraded to test the callout functionality. It provides a
3632 callout function that displays information. By default, it shows the start of
3633 the match and the current position in the text. There are some new data escapes
3634 to vary what happens:
3635
3636 \C+ in addition, show current contents of captured substrings
3637 \C- do not supply a callout function
3638 \C!n return 1 when callout number n is reached
3639 \C!n!m return 1 when callout number n is reached for the mth time
3640
3641 30. If pcregrep was called with the -l option and just a single file name, it
3642 output "<stdin>" if a match was found, instead of the file name.
3643
3644 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3645 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3646 pcre_exec(). This saves a malloc/free per call. The default value of
3647 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3648 when configuring.
3649
3650 32. The default maximum size of a compiled pattern is 64K. There have been a
3651 few cases of people hitting this limit. The code now uses macros to handle the
3652 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3653 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3654 configuring. Tests 2 and 5 work only with 2-byte links because they output
3655 debugging information about compiled patterns.
3656
3657 33. Internal code re-arrangements:
3658
3659 (a) Moved the debugging function for printing out a compiled regex into
3660 its own source file (printint.c) and used #include to pull it into
3661 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3662 separate copies.
3663
3664 (b) Defined the list of op-code names for debugging as a macro in
3665 internal.h so that it is next to the definition of the opcodes.
3666
3667 (c) Defined a table of op-code lengths for simpler skipping along compiled
3668 code. This is again a macro in internal.h so that it is next to the
3669 definition of the opcodes.
3670
3671 34. Added support for recursive calls to individual subpatterns, along the
3672 lines of Robin Houston's patch (but implemented somewhat differently).
3673
3674 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3675 allow it to read and process whole directories in Win32. This code was
3676 contributed by Lionel Fourquaux; it has not been tested by me.
3677
3678 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3679 used to name a group. Names consist of alphanumerics and underscores, and must
3680 be unique. Back references use the syntax (?P=name) and recursive calls use
3681 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3682 numbers. The function pcre_fullinfo() can be used after compilation to extract
3683 a name/number map. There are three relevant calls:
3684
3685 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3686 PCRE_INFO_NAMECOUNT yields the number of entries
3687 PCRE_INFO_NAMETABLE yields a pointer to the map.
3688
3689 The map is a vector of fixed-size entries. The size of each entry depends on
3690 the length of the longest name used. The first two bytes of each entry are the
3691 group number, most significant byte first. There follows the corresponding
3692 name, zero terminated. The names are in alphabetical order.
3693
3694 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3695 case instead of 255. Making it the same both with and without UTF-8 support
3696 means that the same test output works with both.
3697
3698 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3699 calling malloc() with a zero argument.
3700
3701 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3702 optimization. I've improved things by keeping a bitmap of backreferences with
3703 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3704 fact referenced, the optimization can be applied. It is unlikely that a
3705 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3706 the match to follow \n) will appear inside brackets with a number greater than
3707 31, but if it does, any back reference > 31 suppresses the optimization.
3708
3709 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3710 of disabling numbered capturing parentheses. Any opening parenthesis that is
3711 not followed by ? behaves as if it were followed by ?: but named parentheses
3712 can still be used for capturing (and they will acquire numbers in the usual
3713 way).
3714
3715 41. Redesigned the return codes from the match() function into yes/no/error so
3716 that errors can be passed back from deep inside the nested calls. A malloc
3717 failure while inside a recursive subpattern call now causes the
3718 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3719
3720 42. It is now possible to set a limit on the number of times the match()
3721 function is called in a call to pcre_exec(). This facility makes it possible to
3722 limit the amount of recursion and backtracking, though not in a directly
3723 obvious way, because the match() function is used in a number of different
3724 circumstances. The count starts from zero for each position in the subject
3725 string (for non-anchored patterns). The default limit is, for compatibility, a
3726 large number, namely 10 000 000. You can change this in two ways:
3727
3728 (a) When configuring PCRE before making, you can use --with-match-limit=n
3729 to set a default value for the compiled library.
3730
3731 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3732 a different value is set. See 45 below.
3733
3734 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3735
3736 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3737 of things that can be changed at compile time. The first argument specifies
3738 what is wanted and the second points to where the information is to be placed.
3739 The current list of available information is:
3740
3741 PCRE_CONFIG_UTF8
3742
3743 The output is an integer that is set to one if UTF-8 support is available;
3744 otherwise it is set to zero.
3745
3746 PCRE_CONFIG_NEWLINE
3747
3748 The output is an integer that it set to the value of the code that is used for
3749 newline. It is either LF (10) or CR (13).
3750
3751 PCRE_CONFIG_LINK_SIZE
3752
3753 The output is an integer that contains the number of bytes used for internal
3754 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3755
3756 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3757
3758 The output is an integer that contains the threshold above which the POSIX
3759 interface uses malloc() for output vectors. See item 31 above.
3760
3761 PCRE_CONFIG_MATCH_LIMIT
3762
3763 The output is an unsigned integer that contains the default limit of the number
3764 of match() calls in a pcre_exec() execution. See 42 above.
3765
3766 44. pcretest has been upgraded by the addition of the -C option. This causes it
3767 to extract all the available output from the new pcre_config() function, and to
3768 output it. The program then exits immediately.
3769
3770 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3771 order to support additional features. One way would have been to define
3772 pcre_exec2() (for example) with extra arguments, but this would not have been
3773 extensible, and would also have required all calls to the original function to
3774 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3775 is used for passing in "extra" data from pcre_study().
3776
3777 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3778 contains the following fields:
3779
3780 flags a bitmap indicating which of the following fields are set
3781 study_data opaque data from pcre_study()
3782 match_limit a way of specifying a limit on match() calls for a specific
3783 call to pcre_exec()
3784 callout_data data for callouts (see 49 below)
3785
3786 The flag bits are also defined in pcre.h, and are
3787
3788 PCRE_EXTRA_STUDY_DATA
3789 PCRE_EXTRA_MATCH_LIMIT
3790 PCRE_EXTRA_CALLOUT_DATA
3791
3792 The pcre_study() function now returns one of these new pcre_extra blocks, with
3793 the actual study data pointed to by the study_data field, and the
3794 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3795 before. That is, this change is entirely upwards-compatible and requires no
3796 change to existing code.
3797
3798 If you want to pass in additional data to pcre_exec(), you can either place it
3799 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3800 block.
3801
3802 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3803 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3804 times with different match limits, until it finds the minimum value needed for
3805 pcre_exec() to complete. The value is then output. This can be instructive; for
3806 most simple matches the number is quite small, but for pathological cases it
3807 gets very large very quickly.
3808
3809 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3810 returns the size of the data block pointed to by the study_data field in a
3811 pcre_extra block, that is, the value that was passed as the argument to
3812 pcre_malloc() when PCRE was getting memory in which to place the information
3813 created by pcre_study(). The fourth argument should point to a size_t variable.
3814 pcretest has been extended so that this information is shown after a successful
3815 pcre_study() call when information about the compiled regex is being displayed.
3816
3817 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3818 because what follows is always an absolute path. (Later: it turns out that this
3819 is more than cosmetic for MinGW, because it doesn't like empty path
3820 components.)
3821
3822 49. Some changes have been made to the callout feature (see 28 above):
3823
3824 (i) A callout function now has three choices for what it returns:
3825
3826 0 => success, carry on matching
3827 > 0 => failure at this point, but backtrack if possible
3828 < 0 => serious error, return this value from pcre_exec()
3829
3830 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3831 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3832 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3833 use by callout functions. It will never be used by PCRE itself.
3834
3835 (ii) The pcre_extra structure (see 45 above) has a void * field called
3836 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3837 pcre_callout_block structure has a field of the same name. The contents of
3838 the field passed in the pcre_extra structure are passed to the callout
3839 function in the corresponding field in the callout block. This makes it
3840 easier to use the same callout-containing regex from multiple threads. For
3841 testing, the pcretest program has a new data escape
3842
3843 \C*n pass the number n (may be negative) as callout_data
3844
3845 If the callout function in pcretest receives a non-zero value as
3846 callout_data, it returns that value.
3847
3848 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3849 there were some redundant $(CFLAGS) in commands that are now specified as
3850 $(LINK), which already includes $(CFLAGS).
3851
3852 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3853 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3854 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3855 one-byte characters throughout. Note that case-insensitive matching applies
3856 only to characters whose values are less than 256. PCRE doesn't support the
3857 notion of cases for higher-valued characters.
3858
3859 (i) A character class whose characters are all within 0-255 is handled as
3860 a bit map, and the map is inverted for negative classes. Previously, a
3861 character > 255 always failed to match such a class; however it should
3862 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3863
3864 (ii) A negated character class with a single character < 255 is coded as
3865 "not this character" (OP_NOT). This wasn't working properly when the test
3866 character was multibyte, either singly or repeated.
3867
3868 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3869 mode, for example: \x{100}{2,3}.
3870
3871 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3872 singly or repeated) now correctly test multibyte characters. However,
3873 PCRE doesn't recognize any characters with values greater than 255 as
3874 digits, spaces, or word characters. Such characters always match \D, \S,
3875 and \W, and never match \d, \s, or \w.
3876
3877 (v) Classes may now contain characters and character ranges with values
3878 greater than 255. For example: [ab\x{100}-\x{400}].
3879
3880 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3881 PCRE in UTF-8 mode.
3882
3883 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3884 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3885 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3886 value.)
3887
3888 53. The single man page has become too large. I have therefore split it up into
3889 a number of separate man pages. These also give rise to individual HTML pages;
3890 these are now put in a separate directory, and there is an index.html page that
3891 lists them all. Some hyperlinking between the pages has been installed.
3892
3893 54. Added convenience functions for handling named capturing parentheses.
3894
3895 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3896 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3897 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3898 are faulted.
3899
3900 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3901 calling configure. These values are used when compiling the dftables.c program
3902 which is run to generate the source of the default character tables. They
3903 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3904 you will need to set these values.
3905
3906 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3907
3908
3909 Version 3.9 02-Jan-02
3910 ---------------------
3911
3912 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3913
3914 2. If --disable-static was given, the building process failed when trying to
3915 build pcretest and pcregrep. (For some reason it was using libtool to compile
3916 them, which is not right, as they aren't part of the library.)
3917
3918
3919 Version 3.8 18-Dec-01
3920 ---------------------
3921
3922 1. The experimental UTF-8 code was completely screwed up. It was packing the
3923 bytes in the wrong order. How dumb can you get?
3924
3925
3926 Version 3.7 29-Oct-01
3927 ---------------------
3928
3929 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3930 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3931 this didn't happen under Solaris 8, where I normally test things.
3932
3933 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3934 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3935 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3936 this fix an hour or so after the initial 3.7 release.)
3937
3938
3939 Version 3.6 23-Oct-01
3940 ---------------------
3941
3942 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3943 offsets passed as NULL with zero offset count.
3944
3945 2. The config.guess and config.sub files had not been updated when I moved to
3946 the latest autoconf.
3947
3948
3949 Version 3.5 15-Aug-01
3950 ---------------------
3951
3952 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3953 had been forgotten.
3954
3955 2. By using declared but undefined structures, we can avoid using "void"
3956 definitions in pcre.h while keeping the internal definitions of the structures
3957 private.
3958
3959 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3960 user point of view, this means that both static and shared libraries are built
3961 by default, but this can be individually controlled. More of the work of
3962 handling this static/shared cases is now inside libtool instead of PCRE's make
3963 file.
3964
3965 4. The pcretest utility is now installed along with pcregrep because it is
3966 useful for users (to test regexs) and by doing this, it automatically gets
3967 relinked by libtool. The documentation has been turned into a man page, so
3968 there are now .1, .txt, and .html versions in /doc.
3969
3970 5. Upgrades to pcregrep:
3971 (i) Added long-form option names like gnu grep.
3972 (ii) Added --help to list all options with an explanatory phrase.
3973 (iii) Added -r, --recursive to recurse into sub-directories.
3974 (iv) Added -f, --file to read patterns from a file.
3975
3976 6. pcre_exec() was referring to its "code" argument before testing that
3977 argument for NULL (and giving an error if it was NULL).
3978
3979 7. Upgraded Makefile.in to allow for compiling in a different directory from
3980 the source directory.
3981
3982 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3983 options bits, the pointer it was passed was to an int instead of to an unsigned
3984 long int. This mattered only on 64-bit systems.
3985
3986 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3987 generated) instead of pcre.in, which it its source. Also made the same change
3988 in several of the .c files.
3989
3990 10. A new release of gcc defines printf() as a macro, which broke pcretest
3991 because it had an ifdef in the middle of a string argument for printf(). Fixed
3992 by using separate calls to printf().
3993
3994 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3995 script, to force use of CR or LF instead of \n in the source. On non-Unix
3996 systems, the value can be set in config.h.
3997
3998 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3999 absolute limit. Changed the text of the error message to make this clear, and
4000 likewise updated the man page.
4001
4002 13. The limit of 99 on the number of capturing subpatterns has been removed.
4003 The new limit is 65535, which I hope will not be a "real" limit.
4004
4005
4006 Version 3.4 22-Aug-00
4007 ---------------------
4008
4009 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4010
4011 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4012
4013
4014 Version 3.3 01-Aug-00
4015 ---------------------
4016
4017 1. If an octal character was given, but the value was greater than \377, it
4018 was not getting masked to the least significant bits, as documented. This could
4019 lead to crashes in some systems.
4020
4021 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4022 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4023
4024 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4025 These just pass their arguments on to (pcre_free)(), but they are provided
4026 because some uses of PCRE bind it to non-C systems that can call its functions,
4027 but cannot call free() or pcre_free() directly.
4028
4029 4. Add "make test" as a synonym for "make check". Corrected some comments in
4030 the Makefile.
4031
4032 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4033 Makefile.
4034
4035 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4036 command called pgrep for grepping around the active processes.
4037
4038 7. Added the beginnings of support for UTF-8 character strings.
4039
4040 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4041 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4042 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4043 out for the ar command.)
4044
4045
4046 Version 3.2 12-May-00
4047 ---------------------
4048
4049 This is purely a bug fixing release.
4050
4051 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4052 of ZA. This was just one example of several cases that could provoke this bug,
4053 which was introduced by change 9 of version 2.00. The code for breaking
4054 infinite loops after an iteration that matches an empty string was't working
4055 correctly.
4056
4057 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4058 when matched against abbab (for example). After matching an empty string, it
4059 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4060 caused it to match further down the string than it should.
4061
4062 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4063 was there because it doesn't seem to be needed, and it causes trouble on some
4064 systems, as it is not a Standard C header. It has been removed.
4065
4066 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4067 were reported on the Macintosh. The changes were from
4068
4069 while ((c = *(++ptr)) != 0 && c != '\n');
4070 to
4071 while ((c = *(++ptr)) != 0 && c != '\n') ;
4072
4073 Totally extraordinary, but if that's what it takes...
4074
4075 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4076 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4077 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4078 assumes the way PCRE uses memmove() (always moving upwards).
4079
4080 6. PCRE is being used in one environment where strchr() is not available. There
4081 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4082 faster code anyway.
4083
4084
4085 Version 3.1 09-Feb-00
4086 ---------------------
4087
4088 The only change in this release is the fixing of some bugs in Makefile.in for
4089 the "install" target:
4090
4091 (1) It was failing to install pcreposix.h.
4092
4093 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4094
4095
4096 Version 3.0 01-Feb-00
4097 ---------------------
4098
4099 1. Add support for the /+ modifier to perltest (to output $` like it does in
4100 pcretest).
4101
4102 2. Add support for the /g modifier to perltest.
4103
4104 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4105 matches null strings.
4106
4107 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4108 pattern. Perl treats empty patterns specially - it reuses the most recent
4109 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4110 effect.
4111
4112 5. The POSIX interface was broken in that it was just handing over the POSIX
4113 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4114 required a bigger vector, with some working space on the end. This means that
4115 the POSIX wrapper now has to get and free some memory, and copy the results.
4116
4117 6. Added some simple autoconf support, placing the test data and the
4118 documentation in separate directories, re-organizing some of the
4119 information files, and making it build pcre-config (a GNU standard). Also added
4120 libtool support for building PCRE as a shared library, which is now the
4121 default.
4122
4123 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4124 09 are not valid octal constants. Single digits will be used for minor values
4125 less than 10.
4126
4127 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4128 existing programs that set these in the POSIX interface can use PCRE without
4129 modification.
4130
4131 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4132 return all that pcre_info() returns, plus additional data. The pcre_info()
4133 function is retained for compatibility, but is considered to be obsolete.
4134
4135 10. Added experimental recursion feature (?R) to handle one common case that
4136 Perl 5.6 will be able to do with (?p{...}).
4137
4138 11. Added support for POSIX character classes like [:alpha:], which Perl is
4139 adopting.
4140
4141
4142 Version 2.08 31-Aug-99
4143 ----------------------
4144
4145 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4146 trying to match at the startoffset position, but instead was moving forward to
4147 the next newline as if a previous match had failed.
4148
4149 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4150 and could get into a loop if a null string was matched other than at the start
4151 of the subject.
4152
4153 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4154 be distinguished at compile time, and for completeness also added PCRE_DATE.
4155
4156 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4157 in GnuWin32 environments.
4158
4159
4160 Version 2.07 29-Jul-99
4161 ----------------------
4162
4163 1. The documentation is now supplied in plain text form and HTML as well as in
4164 the form of man page sources.
4165
4166 2. C++ compilers don't like assigning (void *) values to other pointer types.
4167 In particular this affects malloc(). Although there is no problem in Standard
4168 C, I've put in casts to keep C++ compilers happy.
4169
4170 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4171 should be (const char *).
4172
4173 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4174 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4175 However, I haven't made this a standard facility. The documentation doesn't
4176 mention it, and the Makefile doesn't support it.
4177
4178 5. The Makefile now contains an "install" target, with editable destinations at
4179 the top of the file. The pcretest program is not installed.
4180
4181 6. pgrep -V now gives the PCRE version number and date.
4182
4183 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4184 causing the entire string to be ignored, instead of just the last character.
4185
4186 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4187 non-matching string, it can take a very, very long time, even for strings of
4188 quite modest length, because of the nested recursion. PCRE now does better in
4189 some of these cases. It does this by remembering the last required literal
4190 character in the pattern, and pre-searching the subject to ensure it is present
4191 before running the real match. In other words, it applies a heuristic to detect
4192 some types of certain failure quickly, and in the above example, if presented
4193 with a string that has no trailing " it gives "no match" very quickly.
4194
4195 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4196 other alternatives are tried instead.
4197
4198
4199 Version 2.06 09-Jun-99
4200 ----------------------
4201
4202 1. Change pcretest's output for amount of store used to show just the code
4203 space, because the remainder (the data block) varies in size between 32-bit and
4204 64-bit systems.
4205
4206 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4207 start matching at. This allows lookbehinds to work when searching for multiple
4208 occurrences in a string.
4209
4210 3. Added additional options to pcretest for testing multiple occurrences:
4211
4212 /+ outputs the rest of the string that follows a match
4213 /g loops for multiple occurrences, using the new startoffset argument
4214 /G loops for multiple occurrences by passing an incremented pointer
4215
4216 4. PCRE wasn't doing the "first character" optimization for patterns starting
4217 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4218 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4219 the letter 'x'. On long subject strings, this gives a significant speed-up.
4220
4221
4222 Version 2.05 21-Apr-99
4223 ----------------------
4224
4225 1. Changed the type of magic_number from int to long int so that it works
4226 properly on 16-bit systems.
4227
4228 2. Fixed a bug which caused patterns starting with .* not to work correctly
4229 when the subject string contained newline characters. PCRE was assuming
4230 anchoring for such patterns in all cases, which is not correct because .* will
4231 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4232 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4233 must be retried after every newline in the subject.
4234
4235
4236 Version 2.04 18-Feb-99
4237 ----------------------
4238
4239 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4240 computation of the store needed to hold the pattern was incorrect (too large).
4241 If such patterns were nested a few deep, this could multiply and become a real
4242 problem.
4243
4244 2. Added /M option to pcretest to show the memory requirement of a specific
4245 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4246
4247 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4248 compiled in such a way that the backtracking after subsequent failure was
4249 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4250 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4251
4252
4253 Version 2.03 02-Feb-99
4254 ----------------------
4255
4256 1. Fixed typo and small mistake in man page.
4257
4258 2. Added 4th condition (GPL supersedes if conflict) and created separate
4259 LICENCE file containing the conditions.
4260
4261 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4262 Perl, that is the internal \ allows the delimiter to be included in the
4263 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4264 the final delimiter, add \ to the end of the pattern (to test the error).
4265
4266 4. Added the convenience functions for extracting substrings after a successful
4267 match. Updated pcretest to make it able to test these functions.
4268
4269
4270 Version 2.02 14-Jan-99
4271 ----------------------
4272
4273 1. Initialized the working variables associated with each extraction so that
4274 their saving and restoring doesn't refer to uninitialized store.
4275
4276 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4277 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4278 fix the problem.
4279
4280 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4281 calls, and wasn't printing the correct value for compiling calls. Increased the
4282 default value of LOOPREPEAT, and the number of significant figures in the
4283 times.
4284
4285 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4286
4287 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4288 a building problem on Windows NT with a FAT file system.
4289
4290
4291 Version 2.01 21-Oct-98
4292 ----------------------
4293
4294 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4295 to character tables built by pcre_maketables() in the current locale. If NULL
4296 is passed, the default tables are used.
4297
4298
4299 Version 2.00 24-Sep-98
4300 ----------------------
4301
4302 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4303 it any more.
4304
4305 2. Allow quantification of (?>) groups, and make it work correctly.
4306
4307 3. The first character computation wasn't working for (?>) groups.
4308
4309 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4310 end of the subject) and add 5.005's \z, which really does match only at the
4311 very end of the subject.
4312
4313 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4314
4315 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4316 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4317 localized options. All options to pcre_study() were also removed.
4318
4319 7. Add other new features from 5.005:
4320
4321 $(?<= positive lookbehind
4322 $(?<! negative lookbehind
4323 (?imsx-imsx) added the unsetting capability
4324 such a setting is global if at outer level; local otherwise
4325 (?imsx-imsx:) non-capturing groups with option setting
4326 (?(cond)re|re) conditional pattern matching
4327
4328 A backreference to itself in a repeated group matches the previous
4329 captured string.
4330
4331 8. General tidying up of studying (both automatic and via "study")
4332 consequential on the addition of new assertions.
4333
4334 9. As in 5.005, unlimited repeated groups that could match an empty substring
4335 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4336 runtime if any iteration does actually match an empty substring.
4337
4338 10. Include the RunTest script in the distribution.
4339
4340 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4341 discrepancies, some of which were old and were also with respect to 5.004. They
4342 have now been fixed.
4343
4344
4345 Version 1.09 28-Apr-98
4346 ----------------------
4347
4348 1. A negated single character class followed by a quantifier with a minimum
4349 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4350 program crashes, or just wrong answers. This did not apply to negated classes
4351 containing more than one character, or to minima other than one.
4352
4353
4354 Version 1.08 27-Mar-98
4355 ----------------------
4356
4357 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4358
4359 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4360 latter must appear before anything that relies on it in the pattern.
4361
4362
4363 Version 1.07 16-Feb-98
4364 ----------------------
4365
4366 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4367 repeat of a potentially empty string).
4368
4369
4370 Version 1.06 23-Jan-98
4371 ----------------------
4372
4373 1. Added Markus Oberhumer's little patches for C++.
4374
4375 2. Literal strings longer than 255 characters were broken.
4376
4377
4378 Version 1.05 23-Dec-97
4379 ----------------------
4380
4381 1. Negated character classes containing more than one character were failing if
4382 PCRE_CASELESS was set at run time.
4383
4384
4385 Version 1.04 19-Dec-97
4386 ----------------------
4387
4388 1. Corrected the man page, where some "const" qualifiers had been omitted.
4389
4390 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4391 input syntax.
4392
4393 3. Fixed memory leak which occurred when a regex with back references was
4394 matched with an offsets vector that wasn't big enough. The temporary memory
4395 that is used in this case wasn't being freed if the match failed.
4396
4397 4. Tidied pcretest to ensure it frees memory that it gets.
4398
4399 5. Temporary memory was being obtained in the case where the passed offsets
4400 vector was exactly big enough.
4401
4402 6. Corrected definition of offsetof() from change 5 below.
4403
4404 7. I had screwed up change 6 below and broken the rules for the use of
4405 setjmp(). Now fixed.
4406
4407
4408 Version 1.03 18-Dec-97
4409 ----------------------
4410
4411 1. A erroneous regex with a missing opening parenthesis was correctly
4412 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4413 on some systems.
4414
4415 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4416 it was reported that one broken compiler failed on the former because "code" is
4417 also an independent variable.
4418
4419 3. The erroneous regex a[]b caused an array overrun reference.
4420
4421 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4422 fail on data ending with that character. (It was going on too far, and checking
4423 the next character, typically a binary zero.) This was specific to the
4424 optimized code for single-character negative classes.
4425
4426 5. Added a contributed patch from the TIN world which does the following:
4427
4428 + Add an undef for memmove, in case the the system defines a macro for it.
4429
4430 + Add a definition of offsetof(), in case there isn't one. (I don't know
4431 the reason behind this - offsetof() is part of the ANSI standard - but
4432 it does no harm).
4433
4434 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4435 most of the places where whitespace preceded '#'. I have given up and
4436 allowed the remaining 2 cases to be at the margin.
4437
4438 + Rename some variables in pcre to eliminate shadowing. This seems very
4439 pedantic, but does no harm, of course.
4440
4441 6. Moved the call to setjmp() into its own function, to get rid of warnings
4442 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4443
4444 7. Constructs such as \d{8,} were compiling into the equivalent of
4445 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4446 outcome, but in this particular case used more store than had been allocated,
4447 which caused the bug to be discovered because it threw up an internal error.
4448
4449 8. The debugging code in both pcre and pcretest for outputting the compiled
4450 form of a regex was going wrong in the case of back references followed by
4451 curly-bracketed repeats.
4452
4453
4454 Version 1.02 12-Dec-97
4455 ----------------------
4456
4457 1. Typos in pcre.3 and comments in the source fixed.
4458
4459 2. Applied a contributed patch to get rid of places where it used to remove
4460 'const' from variables, and fixed some signed/unsigned and uninitialized
4461 variable warnings.
4462
4463 3. Added the "runtest" target to Makefile.
4464
4465 4. Set default compiler flag to -O2 rather than just -O.
4466
4467
4468 Version 1.01 19-Nov-97
4469 ----------------------
4470
4471 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4472 like /([ab]*)*/, that is, for classes with more than one character in them.
4473
4474 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4475 as /((?>a*))*/ (a PCRE_EXTRA facility).
4476
4477
4478 Version 1.00 18-Nov-97
4479 ----------------------
4480
4481 1. Added compile-time macros to support systems such as SunOS4 which don't have
4482 memmove() or strerror() but have other things that can be used instead.
4483
4484 2. Arranged that "make clean" removes the executables.
4485
4486
4487 Version 0.99 27-Oct-97
4488 ----------------------
4489
4490 1. Fixed bug in code for optimizing classes with only one character. It was
4491 initializing a 32-byte map regardless, which could cause it to run off the end
4492 of the memory it had got.
4493
4494 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4495
4496
4497 Version 0.98 22-Oct-97
4498 ----------------------
4499
4500 1. Fixed bug in code for handling temporary memory usage when there are more
4501 back references than supplied space in the ovector. This could cause segfaults.
4502
4503
4504 Version 0.97 21-Oct-97
4505 ----------------------
4506
4507 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4508
4509 2. Optimized negated single characters not to use a bit map.
4510
4511 3. Brought error texts together as macro definitions; clarified some of them;
4512 fixed one that was wrong - it said "range out of order" when it meant "invalid
4513 escape sequence".
4514
4515 4. Changed some char * arguments to const char *.
4516
4517 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4518
4519 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4520 pcretest.
4521
4522
4523 Version 0.96 16-Oct-97
4524 ----------------------
4525
4526 1. Added a simple "pgrep" utility to the distribution.
4527
4528 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4529 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4530 where "ddd" means "one or more decimal digits".
4531
4532 3. Fixed serious bug. If a pattern had a back reference, but the call to
4533 pcre_exec() didn't supply a large enough ovector to record the related
4534 identifying subpattern, the match always failed. PCRE now remembers the number
4535 of the largest back reference, and gets some temporary memory in which to save
4536 the offsets during matching if necessary, in order to ensure that
4537 backreferences always work.
4538
4539 4. Increased the compatibility with Perl in a number of ways:
4540
4541 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4542 to request this handling. The option can be set at compile or exec time.
4543
4544 (b) $ matches before a terminating newline by default; an option
4545 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4546 mode). The option can be set at compile or exec time.
4547
4548 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4549 the same as Perl's. If the decimal number it represents is less than 10
4550 or there aren't that many previous left capturing parentheses, an octal
4551 escape is read. Inside a character class, it's always an octal escape,
4552 even if it is a single digit.
4553
4554 (d) An escaped but undefined alphabetic character is taken as a literal,
4555 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4556 escapes.
4557
4558 (e) {0} is now permitted. (The previous item is removed from the compiled
4559 pattern).
4560
4561 5. Changed all the names of code files so that the basic parts are no longer
4562 than 10 characters, and abolished the teeny "globals.c" file.
4563
4564 6. Changed the handling of character classes; they are now done with a 32-byte
4565 bit map always.
4566
4567 7. Added the -d and /D options to pcretest to make it possible to look at the
4568 internals of compilation without having to recompile pcre.
4569
4570
4571 Version 0.95 23-Sep-97
4572 ----------------------
4573
4574 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4575 \x20 at the start of a run of normal characters. These were being treated as
4576 real characters, instead of the source characters being re-checked.
4577
4578
4579 Version 0.94 18-Sep-97
4580 ----------------------
4581
4582 1. The functions are now thread-safe, with the caveat that the global variables
4583 containing pointers to malloc() and free() or alternative functions are the
4584 same for all threads.
4585
4586 2. Get pcre_study() to generate a bitmap of initial characters for non-
4587 anchored patterns when this is possible, and use it if passed to pcre_exec().
4588
4589
4590 Version 0.93 15-Sep-97
4591 ----------------------
4592
4593 1. /(b)|(:+)/ was computing an incorrect first character.
4594
4595 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4596 but not actually doing anything yet.
4597
4598 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4599 as Perl does (e.g. [-az] or [az-]).
4600
4601 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4602 all possible positions.
4603
4604 5. Split up into different modules to avoid including unneeded functions in a
4605 compiled binary. However, compile and exec are still in one module. The "study"
4606 function is split off.
4607
4608 6. The character tables are now in a separate module whose source is generated
4609 by an auxiliary program - but can then be edited by hand if required. There are
4610 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4611 toupper() in the code.
4612
4613 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4614 make them global. Abolish the function for setting them, as the caller can now
4615 set them directly.
4616
4617
4618 Version 0.92 11-Sep-97
4619 ----------------------
4620
4621 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4622 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4623
4624 2. Caseless matching was not working in character classes if the characters in
4625 the pattern were in upper case.
4626
4627 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4628
4629 4. Make PCRE_ANCHORED public and accept as a compile option.
4630
4631 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4632 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4633 pass them.
4634
4635 6. Give an error if bad option bits passed at compile or run time.
4636
4637 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4638 pcretest to cause it to pass that flag.
4639
4640 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4641 options, and the first character, if set.
4642
4643 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4644
4645
4646 Version 0.91 10-Sep-97
4647 ----------------------
4648
4649 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4650 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4651
4652 2. PCRE was looping on encountering an indefinitely repeated back reference to
4653 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4654 Perl does - treats the match as successful.
4655
4656 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5