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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 912 - (show annotations)
Sun Feb 12 16:53:39 2012 UTC (7 years, 9 months ago) by ph10
File size: 226078 byte(s)
Remove bashism from the RunTest script.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.31
5 -----------------------------
6
7 1. Fixing a wrong JIT test case and some compiler warnings.
8
9 2. Removed a bashism from the RunTest script.
10
11
12 Version 8.30 04-February-2012
13 -----------------------------
14
15 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
16 name is defined in ctype.h.
17
18 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
19 only in quite long subpatterns.
20
21 3. Removed the function pcre_info(), which has been obsolete and deprecated
22 since it was replaced by pcre_fullinfo() in February 2000.
23
24 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
25 match a (*MARK), and the match failed at the start of the subject, a
26 reference to memory before the start of the subject could occur. This bug
27 was introduced by fix 17 of release 8.21.
28
29 5. A reference to an unset group with zero minimum repetition was giving
30 totally wrong answers (in non-JavaScript-compatibility mode). For example,
31 /(another)?(\1?)test/ matched against "hello world test". This bug was
32 introduced in release 8.13.
33
34 6. Add support for 16-bit character strings (a large amount of work involving
35 many changes and refactorings).
36
37 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
38 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
39 from a file.
40
41 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
42 rounding is not applied in this particular case).
43
44 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
45 if they appear, or are escaped, in patterns.
46
47 10. Get rid of a number of -Wunused-but-set-variable warnings.
48
49 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
50 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
51 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
52 also returns the mark "x". This bug applied to capturing parentheses,
53 non-capturing parentheses, and atomic parentheses. It also applied to some
54 assertions.
55
56 12. Stephen Kelly's patch to CMakeLists.txt allows it to parse the version
57 information out of configure.ac instead of relying on pcre.h.generic, which
58 is not stored in the repository.
59
60 13. Applied Dmitry V. Levin's patch for a more portable method for linking with
61 -lreadline.
62
63 14. ZH added PCRE_CONFIG_JITTARGET; added its output to pcretest -C.
64
65 15. Applied Graycode's patch to put the top-level frame on the stack rather
66 than the heap when not using the stack for recursion. This gives a
67 performance improvement in many cases when recursion is not deep.
68
69 16. Experimental code added to "pcretest -C" to output the stack frame size.
70
71
72 Version 8.21 12-Dec-2011
73 ------------------------
74
75 1. Updating the JIT compiler.
76
77 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
78 are added as well.
79
80 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
81 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
82 calling _pcre_jit_exec. Some extra comments are added.
83
84 4. (*MARK) settings inside atomic groups that do not contain any capturing
85 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
86 was introduced by change 18 for 8.20.
87
88 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
89 ECMA-262 standard.
90
91 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
92 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
93 This bug was probably introduced by change 9 of 8.13.
94
95 7. While fixing 6 above, I noticed that a number of other items were being
96 incorrectly rejected as "not fixed length". This arose partly because newer
97 opcodes had not been added to the fixed-length checking code. I have (a)
98 corrected the bug and added tests for these items, and (b) arranged for an
99 error to occur if an unknown opcode is encountered while checking for fixed
100 length instead of just assuming "not fixed length". The items that were
101 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
102 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
103 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
104
105 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
106 being incorrectly compiled and would have given unpredicatble results.
107
108 9. A possessively repeated subpattern with minimum repeat count greater than
109 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
110 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
111 the first (A) could occur when it should not.
112
113 10. Add a cast and remove a redundant test from the code.
114
115 11. JIT should use pcre_malloc/pcre_free for allocation.
116
117 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
118 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
119 is anything other than /usr, -L is still shown).
120
121 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
122
123 14. Perl does not support \N without a following name in a [] class; PCRE now
124 also gives an error.
125
126 15. If a forward reference was repeated with an upper limit of around 2000,
127 it caused the error "internal error: overran compiling workspace". The
128 maximum number of forward references (including repeats) was limited by the
129 internal workspace, and dependent on the LINK_SIZE. The code has been
130 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
131 the default depends on LINK_SIZE. There is a new upper limit (for safety)
132 of around 200,000 forward references. While doing this, I also speeded up
133 the filling in of repeated forward references.
134
135 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
136 incorrectly expecting the subject to contain another "a" after the start.
137
138 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
139 in the match, the SKIP should be ignored. This was not happening; instead
140 the SKIP was being treated as NOMATCH. For patterns such as
141 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
142 tested.
143
144 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
145 now much more compatible with Perl, in particular in cases where the result
146 is a non-match for a non-anchored pattern. For example, if
147 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
148 "m", where previously it did not return a name. A side effect of this
149 change is that for partial matches, the last encountered mark name is
150 returned, as for non matches. A number of tests that were previously not
151 Perl-compatible have been moved into the Perl-compatible test files. The
152 refactoring has had the pleasing side effect of removing one argument from
153 the match() function, thus reducing its stack requirements.
154
155 19. If the /S+ option was used in pcretest to study a pattern using JIT,
156 subsequent uses of /S (without +) incorrectly behaved like /S+.
157
158 21. Retrieve executable code size support for the JIT compiler and fixing
159 some warnings.
160
161 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
162 not work when the shorter character appeared right at the end of the
163 subject string.
164
165 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
166 systems.
167
168 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
169 output it when the /M option is used in pcretest.
170
171 25. The CheckMan script was not being included in the distribution. Also, added
172 an explicit "perl" to run Perl scripts from the PrepareRelease script
173 because this is reportedly needed in Windows.
174
175 26. If study data was being save in a file and studying had not found a set of
176 "starts with" bytes for the pattern, the data written to the file (though
177 never used) was taken from uninitialized memory and so caused valgrind to
178 complain.
179
180 27. Updated RunTest.bat as provided by Sheri Pierce.
181
182 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
183
184 29. Computation of memory usage for the table of capturing group names was
185 giving an unnecessarily large value.
186
187
188 Version 8.20 21-Oct-2011
189 ------------------------
190
191 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
192 a POSIX class. After further experiments with Perl, which convinced me that
193 Perl has bugs and confusions, a closing square bracket is no longer allowed
194 in a POSIX name. This bug also affected patterns with classes that started
195 with full stops.
196
197 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
198 captured substring, but while checking the failing first alternative,
199 substring 1 is temporarily captured. If the output vector supplied to
200 pcre_exec() was not big enough for this capture, the yield of the function
201 was still zero ("insufficient space for captured substrings"). This cannot
202 be totally fixed without adding another stack variable, which seems a lot
203 of expense for a edge case. However, I have improved the situation in cases
204 such as /(a)(b)x|abc/ matched against "abc", where the return code
205 indicates that fewer than the maximum number of slots in the ovector have
206 been set.
207
208 3. Related to (2) above: when there are more back references in a pattern than
209 slots in the output vector, pcre_exec() uses temporary memory during
210 matching, and copies in the captures as far as possible afterwards. It was
211 using the entire output vector, but this conflicts with the specification
212 that only 2/3 is used for passing back captured substrings. Now it uses
213 only the first 2/3, for compatibility. This is, of course, another edge
214 case.
215
216 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
217 main code base, and can be used by building with --enable-jit. When this is
218 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
219 runtime --no-jit option is given.
220
221 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
222 ovector, the return from the function was zero, implying that there were
223 other matches that did not fit. The correct "exactly full" value is now
224 returned.
225
226 6. If a subpattern that was called recursively or as a subroutine contained
227 (*PRUNE) or any other control that caused it to give a non-standard return,
228 invalid errors such as "Error -26 (nested recursion at the same subject
229 position)" or even infinite loops could occur.
230
231 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
232 computing the minimum length on reaching *ACCEPT, and so ended up with the
233 wrong value of 1 rather than 0. Further investigation indicates that
234 computing a minimum subject length in the presence of *ACCEPT is difficult
235 (think back references, subroutine calls), and so I have changed the code
236 so that no minimum is registered for a pattern that contains *ACCEPT.
237
238 8. If (*THEN) was present in the first (true) branch of a conditional group,
239 it was not handled as intended. [But see 16 below.]
240
241 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
242 Sheri Pierce.
243
244 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
245 the first byte in a match must be "a".
246
247 11. Change 17 for 8.13 increased the recursion depth for patterns like
248 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
249 pattern contains any instances of (*THEN). If it does not, the old
250 optimizations are restored. It would be nice to do this on a per-group
251 basis, but at the moment that is not feasible.
252
253 12. In some environments, the output of pcretest -C is CRLF terminated. This
254 broke RunTest's code that checks for the link size. A single white space
255 character after the value is now allowed for.
256
257 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
258 For "fr", it uses the Windows-specific input and output files.
259
260 14. If (*THEN) appeared in a group that was called recursively or as a
261 subroutine, it did not work as intended. [But see next item.]
262
263 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
264 pattern fragments (but not containing any | characters). If A and B are
265 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
266 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
267 D. In other words, Perl considers parentheses that do not contain any |
268 characters to be part of a surrounding alternative, whereas PCRE was
269 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
270 differently. PCRE now behaves in the same way as Perl, except in the case
271 of subroutine/recursion calls such as (?1) which have in any case always
272 been different (but PCRE had them first :-).
273
274 16. Related to 15 above: Perl does not treat the | in a conditional group as
275 creating alternatives. Such a group is treated in the same way as an
276 ordinary group without any | characters when processing (*THEN). PCRE has
277 been changed to match Perl's behaviour.
278
279 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
280 RunGrepTest script failed.
281
282 18. Change 22 for version 13 caused atomic groups to use more stack. This is
283 inevitable for groups that contain captures, but it can lead to a lot of
284 stack use in large patterns. The old behaviour has been restored for atomic
285 groups that do not contain any capturing parentheses.
286
287 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
288 suppress the check for a minimum subject length at run time. (If it was
289 given to pcre_exec() or pcre_dfa_exec() it did work.)
290
291 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
292 fail to work when decoding hex characters in data strings in EBCDIC
293 environments.
294
295 21. It appears that in at least one Mac OS environment, the isxdigit() function
296 is implemented as a macro that evaluates to its argument more than once,
297 contravening the C 90 Standard (I haven't checked a later standard). There
298 was an instance in pcretest which caused it to go wrong when processing
299 \x{...} escapes in subject strings. The has been rewritten to avoid using
300 things like p++ in the argument of isxdigit().
301
302
303 Version 8.13 16-Aug-2011
304 ------------------------
305
306 1. The Unicode data tables have been updated to Unicode 6.0.0.
307
308 2. Two minor typos in pcre_internal.h have been fixed.
309
310 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
311 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
312 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
313
314 4. There were a number of related bugs in the code for matching backrefences
315 caselessly in UTF-8 mode when codes for the characters concerned were
316 different numbers of bytes. For example, U+023A and U+2C65 are an upper
317 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
318 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
319 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
320 2-byte code at the end of the subject (it thought there wasn't enough data
321 left).
322
323 5. Comprehensive information about what went wrong is now returned by
324 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
325 as the output vector has at least 2 elements. The offset of the start of
326 the failing character and a reason code are placed in the vector.
327
328 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
329 now returned is for the first byte of the failing character, instead of the
330 last byte inspected. This is an incompatible change, but I hope it is small
331 enough not to be a problem. It makes the returned offset consistent with
332 pcre_exec() and pcre_dfa_exec().
333
334 7. pcretest now gives a text phrase as well as the error number when
335 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
336 failure, the offset and reason code are output.
337
338 8. When \R was used with a maximizing quantifier it failed to skip backwards
339 over a \r\n pair if the subsequent match failed. Instead, it just skipped
340 back over a single character (\n). This seems wrong (because it treated the
341 two characters as a single entity when going forwards), conflicts with the
342 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
343 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
344 has been changed.
345
346 9. Some internal refactoring has changed the processing so that the handling
347 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
348 time (the PCRE_DOTALL option was changed this way some time ago: version
349 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
350 which was always a bit of a fudge. It also means that there is one less
351 argument for the match() function, which reduces its stack requirements
352 slightly. This change also fixes an incompatibility with Perl: the pattern
353 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
354
355 10. More internal refactoring has drastically reduced the number of recursive
356 calls to match() for possessively repeated groups such as (abc)++ when
357 using pcre_exec().
358
359 11. While implementing 10, a number of bugs in the handling of groups were
360 discovered and fixed:
361
362 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
363 (a|)*(?1) gave a compile-time internal error.
364 ((a|)+)+ did not notice that the outer group could match an empty string.
365 (^a|^)+ was not marked as anchored.
366 (.*a|.*)+ was not marked as matching at start or after a newline.
367
368 12. Yet more internal refactoring has removed another argument from the match()
369 function. Special calls to this function are now indicated by setting a
370 value in a variable in the "match data" data block.
371
372 13. Be more explicit in pcre_study() instead of relying on "default" for
373 opcodes that mean there is no starting character; this means that when new
374 ones are added and accidentally left out of pcre_study(), testing should
375 pick them up.
376
377 14. The -s option of pcretest has been documented for ages as being an old
378 synonym of -m (show memory usage). I have changed it to mean "force study
379 for every regex", that is, assume /S for every regex. This is similar to -i
380 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
381 using it. It makes it easier to run collections of tests with and without
382 study enabled, and thereby test pcre_study() more easily. All the standard
383 tests are now run with and without -s (but some patterns can be marked as
384 "never study" - see 20 below).
385
386 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
387 restoration of the capturing data to the outer values was not happening
388 correctly.
389
390 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
391 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
392 pattern had matched an empty string, and so incorrectly returned a no
393 match.
394
395 17. There was optimizing code for the last branch of non-capturing parentheses,
396 and also for the obeyed branch of a conditional subexpression, which used
397 tail recursion to cut down on stack usage. Unfortunately, now that there is
398 the possibility of (*THEN) occurring in these branches, tail recursion is
399 no longer possible because the return has to be checked for (*THEN). These
400 two optimizations have therefore been removed. [But see 8.20/11 above.]
401
402 18. If a pattern containing \R was studied, it was assumed that \R always
403 matched two bytes, thus causing the minimum subject length to be
404 incorrectly computed because \R can also match just one byte.
405
406 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
407 was incorrectly computed.
408
409 20. If /S is present twice on a test pattern in pcretest input, it now
410 *disables* studying, thereby overriding the use of -s on the command line
411 (see 14 above). This is necessary for one or two tests to keep the output
412 identical in both cases.
413
414 21. When (*ACCEPT) was used in an assertion that matched an empty string and
415 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
416
417 22. When an atomic group that contained a capturing parenthesis was
418 successfully matched, but the branch in which it appeared failed, the
419 capturing was not being forgotten if a higher numbered group was later
420 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
421 group 1 to "a", when in fact it should be unset. This applied to multi-
422 branched capturing and non-capturing groups, repeated or not, and also to
423 positive assertions (capturing in negative assertions does not happen
424 in PCRE) and also to nested atomic groups.
425
426 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
427 subject after a captured substring, to make it easier to tell which of a
428 number of identical substrings has been captured.
429
430 24. The way atomic groups are processed by pcre_exec() has been changed so that
431 if they are repeated, backtracking one repetition now resets captured
432 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
433 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
434 "aaa". Previously, it would have been "a". As part of this code
435 refactoring, the way recursive calls are handled has also been changed.
436
437 25. If an assertion condition captured any substrings, they were not passed
438 back unless some other capturing happened later. For example, if
439 (?(?=(a))a) was matched against "a", no capturing was returned.
440
441 26. When studying a pattern that contained subroutine calls or assertions,
442 the code for finding the minimum length of a possible match was handling
443 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
444 group 1 called group 2 while simultaneously a separate group 2 called group
445 1). A stack overflow occurred in this case. I have fixed this by limiting
446 the recursion depth to 10.
447
448 27. Updated RunTest.bat in the distribution to the version supplied by Tom
449 Fortmann. This supports explicit test numbers on the command line, and has
450 argument validation and error reporting.
451
452 28. An instance of \X with an unlimited repeat could fail if at any point the
453 first character it looked at was a mark character.
454
455 29. Some minor code refactoring concerning Unicode properties and scripts
456 should reduce the stack requirement of match() slightly.
457
458 30. Added the '=' option to pcretest to check the setting of unused capturing
459 slots at the end of the pattern, which are documented as being -1, but are
460 not included in the return count.
461
462 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
463 compiled something random. Now it gives a compile-time error (as does
464 Perl).
465
466 32. A *MARK encountered during the processing of a positive assertion is now
467 recorded and passed back (compatible with Perl).
468
469 33. If --only-matching or --colour was set on a pcregrep call whose pattern
470 had alternative anchored branches, the search for a second match in a line
471 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
472 matched the line "0102" twice. The same bug affected patterns that started
473 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
474 twice.
475
476 34. Previously, PCRE did not allow quantification of assertions. However, Perl
477 does, and because of capturing effects, quantifying parenthesized
478 assertions may at times be useful. Quantifiers are now allowed for
479 parenthesized assertions.
480
481 35. A minor code tidy in pcre_compile() when checking options for \R usage.
482
483 36. \g was being checked for fancy things in a character class, when it should
484 just be a literal "g".
485
486 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
487 appearance of a nested POSIX class supersedes an apparent external class.
488 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
489 unescaped square brackets may also appear as part of class names. For
490 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
491 more like Perl. (But see 8.20/1 above.)
492
493 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
494 was because it thought it was \N{name}, which is not supported).
495
496 39. Add minix to OS list not supporting the -S option in pcretest.
497
498 40. PCRE tries to detect cases of infinite recursion at compile time, but it
499 cannot analyze patterns in sufficient detail to catch mutual recursions
500 such as ((?1))((?2)). There is now a runtime test that gives an error if a
501 subgroup is called recursively as a subpattern for a second time at the
502 same position in the subject string. In previous releases this might have
503 been caught by the recursion limit, or it might have run out of stack.
504
505 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
506 happen only once. PCRE was, however incorrectly giving a compile time error
507 "recursive call could loop indefinitely" because it cannot analyze the
508 pattern in sufficient detail. The compile time test no longer happens when
509 PCRE is compiling a conditional subpattern, but actual runaway loops are
510 now caught at runtime (see 40 above).
511
512 42. It seems that Perl allows any characters other than a closing parenthesis
513 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
514 has been changed to be the same.
515
516 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
517 as not to get warnings when autogen.sh is called. Also changed
518 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
519
520 44. To help people who use pcregrep to scan files containing exceedingly long
521 lines, the following changes have been made:
522
523 (a) The default value of the buffer size parameter has been increased from
524 8K to 20K. (The actual buffer used is three times this size.)
525
526 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
527 PCRE is built.
528
529 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
530 to be set at run time.
531
532 (d) Numerical values in pcregrep options can be followed by K or M, for
533 example --buffer-size=50K.
534
535 (e) If a line being scanned overflows pcregrep's buffer, an error is now
536 given and the return code is set to 2.
537
538 45. Add a pointer to the latest mark to the callout data block.
539
540 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
541 partial match of an empty string instead of no match. This was specific to
542 the use of ".".
543
544 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
545 complete match instead of a partial match. This bug was dependent on both
546 the PCRE_UTF8 and PCRE_DOTALL options being set.
547
548 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
549 starting byte set, because \b was not being ignored.
550
551
552 Version 8.12 15-Jan-2011
553 ------------------------
554
555 1. Fixed some typos in the markup of the man pages, and wrote a script that
556 checks for such things as part of the documentation building process.
557
558 2. On a big-endian 64-bit system, pcregrep did not correctly process the
559 --match-limit and --recursion-limit options (added for 8.11). In
560 particular, this made one of the standard tests fail. (The integer value
561 went into the wrong half of a long int.)
562
563 3. If the --colour option was given to pcregrep with -v (invert match), it
564 did strange things, either producing crazy output, or crashing. It should,
565 of course, ignore a request for colour when reporting lines that do not
566 match.
567
568 4. Another pcregrep bug caused similar problems if --colour was specified with
569 -M (multiline) and the pattern match finished with a line ending.
570
571 5. In pcregrep, when a pattern that ended with a literal newline sequence was
572 matched in multiline mode, the following line was shown as part of the
573 match. This seems wrong, so I have changed it.
574
575 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
576 the check for further matches in the same line (so they could be coloured)
577 to overrun the end of the current line. If another match was found, it was
578 incorrectly shown (and then shown again when found in the next line).
579
580 7. If pcregrep was compiled under Windows, there was a reference to the
581 function pcregrep_exit() before it was defined. I am assuming this was
582 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
583 reported by a user. I've moved the definition above the reference.
584
585
586 Version 8.11 10-Dec-2010
587 ------------------------
588
589 1. (*THEN) was not working properly if there were untried alternatives prior
590 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
591 backtracked to try for "b" instead of moving to the next alternative branch
592 at the same level (in this case, to look for "c"). The Perl documentation
593 is clear that when (*THEN) is backtracked onto, it goes to the "next
594 alternative in the innermost enclosing group".
595
596 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
597 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
598 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
599 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
600 (*THEN).
601
602 3. If \s appeared in a character class, it removed the VT character from
603 the class, even if it had been included by some previous item, for example
604 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
605 of \s, but is part of the POSIX "space" class.)
606
607 4. A partial match never returns an empty string (because you can always
608 match an empty string at the end of the subject); however the checking for
609 an empty string was starting at the "start of match" point. This has been
610 changed to the "earliest inspected character" point, because the returned
611 data for a partial match starts at this character. This means that, for
612 example, /(?<=abc)def/ gives a partial match for the subject "abc"
613 (previously it gave "no match").
614
615 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
616 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
617 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
618 has an implication that the given string is incomplete (because a partial
619 match is preferred over a full match). For this reason, these items now
620 give a partial match in this situation. [Aside: previously, the one case
621 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
622 match rather than a full match, which was wrong by the old rules, but is
623 now correct.]
624
625 6. There was a bug in the handling of #-introduced comments, recognized when
626 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
627 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
628 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
629 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
630 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
631 places in pcre_compile().
632
633 7. Related to (6) above, when pcre_compile() was skipping #-introduced
634 comments when looking ahead for named forward references to subpatterns,
635 the only newline sequence it recognized was NL. It now handles newlines
636 according to the set newline convention.
637
638 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
639 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
640 cater for a lack of strerror(). These oversights have been fixed.
641
642 9. Added --match-limit and --recursion-limit to pcregrep.
643
644 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
645
646 11. When the -o option was used, pcregrep was setting a return code of 1, even
647 when matches were found, and --line-buffered was not being honoured.
648
649 12. Added an optional parentheses number to the -o and --only-matching options
650 of pcregrep.
651
652 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
653 can match an empty string. The code to do it in pcretest and pcredemo
654 needed fixing:
655
656 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
657 only one byte after an empty string match just before CRLF (this case
658 just got forgotten; "any" and "anycrlf" were OK).
659
660 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
661 mode when an empty string match preceded an ASCII character followed by
662 a non-ASCII character. (The code for advancing by one character rather
663 than one byte was nonsense.)
664
665 (c) The pcredemo.c sample program did not have any code at all to handle
666 the cases when CRLF is a valid newline sequence.
667
668 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
669 as a starting offset was within the subject string. There is now a new
670 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
671 negative or greater than the length of the string. In order to test this,
672 pcretest is extended to allow the setting of negative starting offsets.
673
674 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
675 starting offset points to the beginning of a UTF-8 character was
676 unnecessarily clumsy. I tidied it up.
677
678 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
679 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
680
681 17. Nobody had reported that the --include_dir option, which was added in
682 release 7.7 should have been called --include-dir (hyphen, not underscore)
683 for compatibility with GNU grep. I have changed it to --include-dir, but
684 left --include_dir as an undocumented synonym, and the same for
685 --exclude-dir, though that is not available in GNU grep, at least as of
686 release 2.5.4.
687
688 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
689 characters from a string of bytes) have been redefined so as not to use
690 loops, in order to improve performance in some environments. At the same
691 time, I abstracted some of the common code into auxiliary macros to save
692 repetition (this should not affect the compiled code).
693
694 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
695 compile-time error is now given if \c is not followed by an ASCII
696 character, that is, a byte less than 128. (In EBCDIC mode, the code is
697 different, and any byte value is allowed.)
698
699 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
700 START_OPTIMIZE option, which is now allowed at compile time - but just
701 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
702 to pcregrep and other applications that have no direct access to PCRE
703 options. The new /Y option in pcretest sets this option when calling
704 pcre_compile().
705
706 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
707 back references. Groups containing recursive back references were forced to
708 be atomic by that change, but in the case of named groups, the amount of
709 memory required was incorrectly computed, leading to "Failed: internal
710 error: code overflow". This has been fixed.
711
712 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
713 pcretest.c, to avoid build problems in some Borland environments.
714
715
716 Version 8.10 25-Jun-2010
717 ------------------------
718
719 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
720 THEN.
721
722 2. (*ACCEPT) was not working when inside an atomic group.
723
724 3. Inside a character class, \B is treated as a literal by default, but
725 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
726 causes the error). The code is unchanged, but I tidied the documentation.
727
728 4. Inside a character class, PCRE always treated \R and \X as literals,
729 whereas Perl faults them if its -w option is set. I have changed PCRE so
730 that it faults them when PCRE_EXTRA is set.
731
732 5. Added support for \N, which always matches any character other than
733 newline. (It is the same as "." when PCRE_DOTALL is not set.)
734
735 6. When compiling pcregrep with newer versions of gcc which may have
736 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
737 declared with attribute warn_unused_result" were given. Just casting the
738 result to (void) does not stop the warnings; a more elaborate fudge is
739 needed. I've used a macro to implement this.
740
741 7. Minor change to pcretest.c to avoid a compiler warning.
742
743 8. Added four artifical Unicode properties to help with an option to make
744 \s etc use properties (see next item). The new properties are: Xan
745 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
746
747 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
748 use Unicode properties. (*UCP) at the start of a pattern can be used to set
749 this option. Modified pcretest to add /W to test this facility. Added
750 REG_UCP to make it available via the POSIX interface.
751
752 10. Added --line-buffered to pcregrep.
753
754 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
755 studied, and the match started with a letter with a code point greater than
756 127 whose first byte was different to the first byte of the other case of
757 the letter, the other case of this starting letter was not recognized
758 (#976).
759
760 12. If a pattern that was studied started with a repeated Unicode property
761 test, for example, \p{Nd}+, there was the theoretical possibility of
762 setting up an incorrect bitmap of starting bytes, but fortunately it could
763 not have actually happened in practice until change 8 above was made (it
764 added property types that matched character-matching opcodes).
765
766 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
767 possible starting bytes for non-anchored patterns.
768
769 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
770 \R, and also a number of cases that involve Unicode properties, both
771 explicit and implicit when PCRE_UCP is set.
772
773 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
774 input, it could crash or give wrong results if characters with values
775 greater than 0xc0 were present in the subject string. (Detail: it assumed
776 UTF-8 input when processing these items.)
777
778 16. Added a lot of (int) casts to avoid compiler warnings in systems where
779 size_t is 64-bit (#991).
780
781 17. Added a check for running out of memory when PCRE is compiled with
782 --disable-stack-for-recursion (#990).
783
784 18. If the last data line in a file for pcretest does not have a newline on
785 the end, a newline was missing in the output.
786
787 19. The default pcre_chartables.c file recognizes only ASCII characters (values
788 less than 128) in its various bitmaps. However, there is a facility for
789 generating tables according to the current locale when PCRE is compiled. It
790 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
791 characters, are recognized by isspace() and therefore were getting set in
792 these tables, and indeed these tables seem to approximate to ISO 8859. This
793 caused a problem in UTF-8 mode when pcre_study() was used to create a list
794 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
795 which of course cannot start UTF-8 characters. I have changed the code so
796 that only real ASCII characters (less than 128) and the correct starting
797 bytes for UTF-8 encodings are set for characters greater than 127 when in
798 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
799 altogether.)
800
801 20. Added the /T option to pcretest so as to be able to run tests with non-
802 standard character tables, thus making it possible to include the tests
803 used for 19 above in the standard set of tests.
804
805 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
806 reference to a subpattern the other side of a comment that contains an
807 opening parenthesis caused either an internal compiling error, or a
808 reference to the wrong subpattern.
809
810
811 Version 8.02 19-Mar-2010
812 ------------------------
813
814 1. The Unicode data tables have been updated to Unicode 5.2.0.
815
816 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
817 configured.
818
819 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
820 original author of that file, following a query about its status.
821
822 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
823 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
824
825 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
826 quantifier applied to a forward-referencing subroutine call, could compile
827 incorrect code or give the error "internal error: previously-checked
828 referenced subpattern not found".
829
830 6. Both MS Visual Studio and Symbian OS have problems with initializing
831 variables to point to external functions. For these systems, therefore,
832 pcre_malloc etc. are now initialized to local functions that call the
833 relevant global functions.
834
835 7. There were two entries missing in the vectors called coptable and poptable
836 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
837 I've fixed the data, and added a kludgy way of testing at compile time that
838 the lengths are correct (equal to the number of opcodes).
839
840 8. Following on from 7, I added a similar kludge to check the length of the
841 eint vector in pcreposix.c.
842
843 9. Error texts for pcre_compile() are held as one long string to avoid too
844 much relocation at load time. To find a text, the string is searched,
845 counting zeros. There was no check for running off the end of the string,
846 which could happen if a new error number was added without updating the
847 string.
848
849 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
850
851 11. \K was not working if it appeared in an atomic group or in a group that
852 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
853 \K is "not well defined" if used in an assertion. PCRE now accepts it if
854 the assertion is positive, but not if it is negative.
855
856 12. Change 11 fortuitously reduced the size of the stack frame used in the
857 "match()" function of pcre_exec.c by one pointer. Forthcoming
858 implementation of support for (*MARK) will need an extra pointer on the
859 stack; I have reserved it now, so that the stack frame size does not
860 decrease.
861
862 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
863 item in branch that calls a recursion is a subroutine call - as in the
864 second branch in the above example - was incorrectly given the compile-
865 time error "recursive call could loop indefinitely" because pcre_compile()
866 was not correctly checking the subroutine for matching a non-empty string.
867
868 14. The checks for overrunning compiling workspace could trigger after an
869 overrun had occurred. This is a "should never occur" error, but it can be
870 triggered by pathological patterns such as hundreds of nested parentheses.
871 The checks now trigger 100 bytes before the end of the workspace.
872
873 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
874
875
876 Version 8.01 19-Jan-2010
877 ------------------------
878
879 1. If a pattern contained a conditional subpattern with only one branch (in
880 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
881 computed the wrong minimum data length (which is of course zero for such
882 subpatterns). This could cause incorrect "no match" results.
883
884 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
885 the pattern is reset in the first branch, pcre_compile() failed with
886 "internal error: code overflow at offset...". This happened only when
887 the reset was to the original external option setting. (An optimization
888 abstracts leading options settings into an external setting, which was the
889 cause of this.)
890
891 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
892 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
893 assertion pattern did not match (meaning that the assertion was true), it
894 was incorrectly treated as false if the SKIP had been reached during the
895 matching. This also applied to assertions used as conditions.
896
897 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
898 assertion subpattern, including such a pattern used as a condition,
899 unpredictable results occurred, instead of the error return
900 PCRE_ERROR_DFA_UITEM.
901
902 5. The C++ GlobalReplace function was not working like Perl for the special
903 situation when an empty string is matched. It now does the fancy magic
904 stuff that is necessary.
905
906 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
907 removed. (These were left over from very, very early versions of PCRE.)
908
909 7. Some cosmetic changes to the code to make life easier when compiling it
910 as part of something else:
911
912 (a) Change DEBUG to PCRE_DEBUG.
913
914 (b) In pcre_compile(), rename the member of the "branch_chain" structure
915 called "current" as "current_branch", to prevent a collision with the
916 Linux macro when compiled as a kernel module.
917
918 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
919 prevent a collision with the Linux macro when compiled as a kernel
920 module.
921
922 8. In pcre_compile() there are some checks for integer overflows that used to
923 cast potentially large values to (double). This has been changed to that
924 when building, a check for int64_t is made, and if it is found, it is used
925 instead, thus avoiding the use of floating point arithmetic. (There is no
926 other use of FP in PCRE.) If int64_t is not found, the fallback is to
927 double.
928
929 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
930 2005 (difference between two addresses compared to an unsigned value).
931
932 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
933 custom one, because of the following reported problem in Windows:
934
935 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
936 under Win32.
937 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
938 therefore missing the function definition.
939 - The compiler thus generates a "C" signature for the test function.
940 - The linker fails to find the "C" function.
941 - PCRE fails to configure if asked to do so against libbz2.
942
943 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
944 messages were output:
945
946 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
947 rerunning libtoolize, to keep the correct libtool macros in-tree.
948 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
949
950 I have done both of these things.
951
952 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
953 most of the time, it *can* run out if it is given a pattern that contains a
954 runaway infinite recursion. I updated the discussion in the pcrestack man
955 page.
956
957 13. Now that we have gone to the x.xx style of version numbers, the minor
958 version may start with zero. Using 08 or 09 is a bad idea because users
959 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
960 interpreted as invalid octal numbers. I've updated the previous comment in
961 configure.ac, and also added a check that gives an error if 08 or 09 are
962 used.
963
964 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
965 causing partial matching to fail when the end of the subject matched \W
966 in a UTF-8 pattern where \W was quantified with a minimum of 3.
967
968 15. There were some discrepancies between the declarations in pcre_internal.h
969 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
970 their definitions. The declarations used "const uschar *" and the
971 definitions used USPTR. Even though USPTR is normally defined as "const
972 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
973 reported that: "This difference in casting confuses some C++ compilers, for
974 example, SunCC recognizes above declarations as different functions and
975 generates broken code for hbpcre." I have changed the declarations to use
976 USPTR.
977
978 16. GNU libtool is named differently on some systems. The autogen.sh script now
979 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
980 (FreeBSD).
981
982 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
983 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
984 comment: "Figure out how to create a longlong from a string: strtoll and
985 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
986 instance, but it only takes 2 args instead of 3!"
987
988 18. A subtle bug concerned with back references has been fixed by a change of
989 specification, with a corresponding code fix. A pattern such as
990 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
991 refers, was giving matches when it shouldn't. For example, xa=xaaa would
992 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
993 same bug. Such groups have to be quantified to be useful, or contained
994 inside another quantified group. (If there's no repetition, the reference
995 can never match.) The problem arises because, having left the group and
996 moved on to the rest of the pattern, a later failure that backtracks into
997 the group uses the captured value from the final iteration of the group
998 rather than the correct earlier one. I have fixed this in PCRE by forcing
999 any group that contains a reference to itself to be an atomic group; that
1000 is, there cannot be any backtracking into it once it has completed. This is
1001 similar to recursive and subroutine calls.
1002
1003
1004 Version 8.00 19-Oct-09
1005 ----------------------
1006
1007 1. The table for translating pcre_compile() error codes into POSIX error codes
1008 was out-of-date, and there was no check on the pcre_compile() error code
1009 being within the table. This could lead to an OK return being given in
1010 error.
1011
1012 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
1013 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
1014 in a Windows environment.
1015
1016 3. The pcregrep --count option prints the count for each file even when it is
1017 zero, as does GNU grep. However, pcregrep was also printing all files when
1018 --files-with-matches was added. Now, when both options are given, it prints
1019 counts only for those files that have at least one match. (GNU grep just
1020 prints the file name in this circumstance, but including the count seems
1021 more useful - otherwise, why use --count?) Also ensured that the
1022 combination -clh just lists non-zero counts, with no names.
1023
1024 4. The long form of the pcregrep -F option was incorrectly implemented as
1025 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1026 but it seems right to fix it, and I didn't think it was worth preserving
1027 the old behaviour.
1028
1029 5. The command line items --regex=pattern and --regexp=pattern were not
1030 recognized by pcregrep, which required --regex pattern or --regexp pattern
1031 (with a space rather than an '='). The man page documented the '=' forms,
1032 which are compatible with GNU grep; these now work.
1033
1034 6. No libpcreposix.pc file was created for pkg-config; there was just
1035 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1036
1037 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1038 when UCP support is not needed, by modifying the Python script that
1039 generates it from Unicode data files. This should not matter if the module
1040 is correctly used as a library, but I received one complaint about 50K of
1041 unwanted data. My guess is that the person linked everything into his
1042 program rather than using a library. Anyway, it does no harm.
1043
1044 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1045 was a minimum greater than 1 for a wide character in a possessive
1046 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1047 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1048 character. Chaos in the form of incorrect output or a compiling loop could
1049 result.
1050
1051 9. The restrictions on what a pattern can contain when partial matching is
1052 requested for pcre_exec() have been removed. All patterns can now be
1053 partially matched by this function. In addition, if there are at least two
1054 slots in the offset vector, the offset of the earliest inspected character
1055 for the match and the offset of the end of the subject are set in them when
1056 PCRE_ERROR_PARTIAL is returned.
1057
1058 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1059 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1060 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1061 and may be more useful for multi-segment matching.
1062
1063 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1064 used to be given if ever the end of the subject was reached; now it is
1065 given only if matching could not proceed because another character was
1066 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1067 string "Z", which now yields "no match" instead of "partial match". In the
1068 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1069 final character ended with (*FAIL).
1070
1071 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1072 if the pattern had a "must contain" character that was already found in the
1073 earlier partial match, unless partial matching was again requested. For
1074 example, with the pattern /dog.(body)?/, the "must contain" character is
1075 "g". If the first part-match was for the string "dog", restarting with
1076 "sbody" failed. This bug has been fixed.
1077
1078 13. The string returned by pcre_dfa_exec() after a partial match has been
1079 changed so that it starts at the first inspected character rather than the
1080 first character of the match. This makes a difference only if the pattern
1081 starts with a lookbehind assertion or \b or \B (\K is not supported by
1082 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1083 matching functions compatible, and I think it's the right thing to do.
1084
1085 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1086 so that the demonstration program is easily available in environments where
1087 PCRE has not been installed from source.
1088
1089 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1090 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1091 library.
1092
1093 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1094 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1095 is not the first non-POSIX option to be added. Clearly some people find
1096 these options useful.
1097
1098 17. If a caller to the POSIX matching function regexec() passes a non-zero
1099 value for nmatch with a NULL value for pmatch, the value of
1100 nmatch is forced to zero.
1101
1102 18. RunGrepTest did not have a test for the availability of the -u option of
1103 the diff command, as RunTest does. It now checks in the same way as
1104 RunTest, and also checks for the -b option.
1105
1106 19. If an odd number of negated classes containing just a single character
1107 interposed, within parentheses, between a forward reference to a named
1108 subpattern and the definition of the subpattern, compilation crashed with
1109 an internal error, complaining that it could not find the referenced
1110 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1111 [The bug was that it was starting one character too far in when skipping
1112 over the character class, thus treating the ] as data rather than
1113 terminating the class. This meant it could skip too much.]
1114
1115 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1116 /g option in pcretest when the pattern contains \K, which makes it possible
1117 to have an empty string match not at the start, even when the pattern is
1118 anchored. Updated pcretest and pcredemo to use this option.
1119
1120 21. If the maximum number of capturing subpatterns in a recursion was greater
1121 than the maximum at the outer level, the higher number was returned, but
1122 with unset values at the outer level. The correct (outer level) value is
1123 now given.
1124
1125 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1126 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1127 make it do so. The string so far is captured, making this feature
1128 compatible with Perl.
1129
1130 23. The tests have been re-organized, adding tests 11 and 12, to make it
1131 possible to check the Perl 5.10 features against Perl 5.10.
1132
1133 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1134 pattern matches a fixed length string. PCRE did not allow this; now it
1135 does. Neither allows recursion.
1136
1137 25. I finally figured out how to implement a request to provide the minimum
1138 length of subject string that was needed in order to match a given pattern.
1139 (It was back references and recursion that I had previously got hung up
1140 on.) This code has now been added to pcre_study(); it finds a lower bound
1141 to the length of subject needed. It is not necessarily the greatest lower
1142 bound, but using it to avoid searching strings that are too short does give
1143 some useful speed-ups. The value is available to calling programs via
1144 pcre_fullinfo().
1145
1146 26. While implementing 25, I discovered to my embarrassment that pcretest had
1147 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1148 study optimizations had never been tested with that matching function.
1149 Oops. What is worse, even when it was passed study data, there was a bug in
1150 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1151 were also very few tests of studied patterns with pcre_dfa_exec().
1152
1153 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1154 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1155 on the other side of the coin, they are no longer allowed to have different
1156 names, because these cannot be distinguished in PCRE, and this has caused
1157 confusion. (This is a difference from Perl.)
1158
1159 28. When duplicate subpattern names are present (necessarily with different
1160 numbers, as required by 27 above), and a test is made by name in a
1161 conditional pattern, either for a subpattern having been matched, or for
1162 recursion in such a pattern, all the associated numbered subpatterns are
1163 tested, and the overall condition is true if the condition is true for any
1164 one of them. This is the way Perl works, and is also more like the way
1165 testing by number works.
1166
1167
1168 Version 7.9 11-Apr-09
1169 ---------------------
1170
1171 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1172 (pcretest), all targets were linked against these libraries. This included
1173 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1174 libraries. This caused unwanted dependencies to be created. This problem
1175 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1176 pcretest is linked with readline.
1177
1178 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1179 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1180 moved outside it again, because FALSE and TRUE are already defined in AIX,
1181 but BOOL is not.
1182
1183 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1184 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1185
1186 4. The pcregrep documentation said spaces were inserted as well as colons (or
1187 hyphens) following file names and line numbers when outputting matching
1188 lines. This is not true; no spaces are inserted. I have also clarified the
1189 wording for the --colour (or --color) option.
1190
1191 5. In pcregrep, when --colour was used with -o, the list of matching strings
1192 was not coloured; this is different to GNU grep, so I have changed it to be
1193 the same.
1194
1195 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1196 each matching line was coloured. Now it goes on to look for further matches
1197 of any of the test patterns, which is the same behaviour as GNU grep.
1198
1199 7. A pattern that could match an empty string could cause pcregrep to loop; it
1200 doesn't make sense to accept an empty string match in pcregrep, so I have
1201 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1202 seems to be how GNU grep behaves.
1203
1204 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1205 start or after a newline", because the conditional assertion was not being
1206 correctly handled. The rule now is that both the assertion and what follows
1207 in the first alternative must satisfy the test.
1208
1209 9. If auto-callout was enabled in a pattern with a conditional group whose
1210 condition was an assertion, PCRE could crash during matching, both with
1211 pcre_exec() and pcre_dfa_exec().
1212
1213 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1214 used for matching.
1215
1216 11. Unicode property support in character classes was not working for
1217 characters (bytes) greater than 127 when not in UTF-8 mode.
1218
1219 12. Added the -M command line option to pcretest.
1220
1221 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1222
1223 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1224
1225 16. Added comments and documentation about mis-use of no_arg in the C++
1226 wrapper.
1227
1228 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1229 from Martin Jerabek that uses macro names for all relevant character and
1230 string constants.
1231
1232 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1233 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1234 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1235 these, but not everybody uses configure.
1236
1237 19. A conditional group that had only one branch was not being correctly
1238 recognized as an item that could match an empty string. This meant that an
1239 enclosing group might also not be so recognized, causing infinite looping
1240 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1241 with the subject "ab", where knowledge that the repeated group can match
1242 nothing is needed in order to break the loop.
1243
1244 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1245 exec(), but without supplying a callout function, matching went wrong.
1246
1247 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1248 leak if the size of the offset vector was greater than 30. When the vector
1249 is smaller, the saved offsets during recursion go onto a local stack
1250 vector, but for larger vectors malloc() is used. It was failing to free
1251 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1252 error, in fact).
1253
1254 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1255 heapframe that is used only when UTF-8 support is enabled. This caused no
1256 problem, but was untidy.
1257
1258 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1259 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1260 included within another project.
1261
1262 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1263 slightly modified by me:
1264
1265 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1266 not building pcregrep.
1267
1268 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1269 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1270
1271 25. Forward references, both numeric and by name, in patterns that made use of
1272 duplicate group numbers, could behave incorrectly or give incorrect errors,
1273 because when scanning forward to find the reference group, PCRE was not
1274 taking into account the duplicate group numbers. A pattern such as
1275 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1276
1277 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1278 the feature of a custom pointer more persuasive (as requested by a user).
1279
1280 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1281 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1282 pre-defined.
1283
1284 28. Added support for (*UTF8) at the start of a pattern.
1285
1286 29. Arrange for flags added by the "release type" setting in CMake to be shown
1287 in the configuration summary.
1288
1289
1290 Version 7.8 05-Sep-08
1291 ---------------------
1292
1293 1. Replaced UCP searching code with optimized version as implemented for Ad
1294 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1295 stage table and inline lookup instead of a function, giving speed ups of 2
1296 to 5 times on some simple patterns that I tested. Permission was given to
1297 distribute the MultiStage2.py script that generates the tables (it's not in
1298 the tarball, but is in the Subversion repository).
1299
1300 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1301 scripts.
1302
1303 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1304 a group with a zero qualifier. The result of the study could be incorrect,
1305 or the function might crash, depending on the pattern.
1306
1307 4. Caseless matching was not working for non-ASCII characters in back
1308 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1309 It now works when Unicode Property Support is available.
1310
1311 5. In pcretest, an escape such as \x{de} in the data was always generating
1312 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1313 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1314 truncation.
1315
1316 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1317
1318 7. Added two (int) casts to pcregrep when printing the difference of two
1319 pointers, in case they are 64-bit values.
1320
1321 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1322 test 2 if it fails.
1323
1324 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1325 and a #define of that name to empty if it is not externally set. This is to
1326 allow users of MSVC to set it if necessary.
1327
1328 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1329 the convenience functions in the pcre_get.c source file.
1330
1331 11. An option change at the start of a pattern that had top-level alternatives
1332 could cause overwriting and/or a crash. This command provoked a crash in
1333 some environments:
1334
1335 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1336
1337 This potential security problem was recorded as CVE-2008-2371.
1338
1339 12. For a pattern where the match had to start at the beginning or immediately
1340 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1341 pcre_dfa_exec() could read past the end of the passed subject if there was
1342 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1343 pcretest so that it places the subject at the end of its malloc-ed buffer.
1344
1345 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1346 exec() might read past the end of the data buffer in UTF-8 mode.
1347
1348 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1349 the data contained the byte 0x85 as part of a UTF-8 character within its
1350 first line. This applied both to normal and DFA matching.
1351
1352 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1353 /^[^d]*?$/8 failed to match "abc".
1354
1355 16. Added a missing copyright notice to pcrecpp_internal.h.
1356
1357 17. Make it more clear in the documentation that values returned from
1358 pcre_exec() in ovector are byte offsets, not character counts.
1359
1360 18. Tidied a few places to stop certain compilers from issuing warnings.
1361
1362 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1363 supplied by Stefan Weber. I made a further small update for 7.8 because
1364 there is a change of source arrangements: the pcre_searchfuncs.c module is
1365 replaced by pcre_ucd.c.
1366
1367
1368 Version 7.7 07-May-08
1369 ---------------------
1370
1371 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1372 a string to a long long, pretend we don't even have a long long." This is
1373 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1374
1375 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1376 pre-7.6 versions, which defined a global no_arg variable instead of putting
1377 it in the RE class. (See also #8 below.)
1378
1379 3. Remove a line of dead code, identified by coverity and reported by Nuno
1380 Lopes.
1381
1382 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1383
1384 (1) The include/exclude patterns were being applied to the whole pathnames
1385 of files, instead of just to the final components.
1386
1387 (2) If there was more than one level of directory, the subdirectories were
1388 skipped unless they satisfied the include/exclude conditions. This is
1389 inconsistent with GNU grep (and could even be seen as contrary to the
1390 pcregrep specification - which I improved to make it absolutely clear).
1391 The action now is always to scan all levels of directory, and just
1392 apply the include/exclude patterns to regular files.
1393
1394 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1395 --exclude_dir in the tests to avoid scanning .svn directories.
1396
1397 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1398 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1399 doesn't support NULs in patterns.
1400
1401 7. Added some missing "const"s to declarations of static tables in
1402 pcre_compile.c and pcre_dfa_exec.c.
1403
1404 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1405 caused by fix #2 above. (Subsequently also a second patch to fix the
1406 first patch. And a third patch - this was a messy problem.)
1407
1408 9. Applied Craig's patch to remove the use of push_back().
1409
1410 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1411 matching function regexec().
1412
1413 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1414 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1415 references. PCRE supports relative numbers with this syntax (I don't think
1416 Oniguruma does).
1417
1418 12. Previously, a group with a zero repeat such as (...){0} was completely
1419 omitted from the compiled regex. However, this means that if the group
1420 was called as a subroutine from elsewhere in the pattern, things went wrong
1421 (an internal error was given). Such groups are now left in the compiled
1422 pattern, with a new opcode that causes them to be skipped at execution
1423 time.
1424
1425 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1426 to the way PCRE behaves:
1427
1428 (a) A lone ] character is dis-allowed (Perl treats it as data).
1429
1430 (b) A back reference to an unmatched subpattern matches an empty string
1431 (Perl fails the current match path).
1432
1433 (c) A data ] in a character class must be notated as \] because if the
1434 first data character in a class is ], it defines an empty class. (In
1435 Perl it is not possible to have an empty class.) The empty class []
1436 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1437 The negative empty class [^] matches any one character, independently
1438 of the DOTALL setting.
1439
1440 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1441 non-existent subpattern following a character class starting with ']' and
1442 containing () gave an internal compiling error instead of "reference to
1443 non-existent subpattern". Fortunately, when the pattern did exist, the
1444 compiled code was correct. (When scanning forwards to check for the
1445 existencd of the subpattern, it was treating the data ']' as terminating
1446 the class, so got the count wrong. When actually compiling, the reference
1447 was subsequently set up correctly.)
1448
1449 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1450 it was being rejected as not supported by pcre_dfa_exec(), even though
1451 other assertions are supported. I have made pcre_dfa_exec() support
1452 (*FAIL).
1453
1454 16. The implementation of 13c above involved the invention of a new opcode,
1455 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1456 cannot be changed at match time, I realized I could make a small
1457 improvement to matching performance by compiling OP_ALLANY instead of
1458 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1459 on the OP_ANY path.
1460
1461 17. Compiling pcretest on Windows with readline support failed without the
1462 following two fixes: (1) Make the unistd.h include conditional on
1463 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1464
1465 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1466 ncurses library to be included for pcretest when ReadLine support is
1467 requested, but also to allow for it to be overridden. This patch came from
1468 Daniel Bergström.
1469
1470 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1471 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1472 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1473 spotting this.
1474
1475
1476 Version 7.6 28-Jan-08
1477 ---------------------
1478
1479 1. A character class containing a very large number of characters with
1480 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1481 overflow.
1482
1483 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1484 HAVE_LONG_LONG is not defined.
1485
1486 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1487 bring them up to date and include new features. This patch includes:
1488
1489 - Fixed PH's badly added libz and libbz2 support.
1490 - Fixed a problem with static linking.
1491 - Added pcredemo. [But later removed - see 7 below.]
1492 - Fixed dftables problem and added an option.
1493 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1494 HAVE_LONG_LONG.
1495 - Added readline support for pcretest.
1496 - Added an listing of the option settings after cmake has run.
1497
1498 4. A user submitted a patch to Makefile that makes it easy to create
1499 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1500 Makefile.am that cause it to include this special target, without
1501 affecting anything else. Note that the same mingw target plus all
1502 the other distribution libraries and programs are now supported
1503 when configuring with CMake (see 6 below) instead of with
1504 Configure/Make.
1505
1506 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1507 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1508 exported in the Windows port". It has not yet been confirmed that the patch
1509 solves the problem, but it does no harm.
1510
1511 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1512 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1513 with CMake, and also correct the comment about stack recursion.
1514
1515 7. Remove the automatic building of pcredemo from the ./configure system and
1516 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1517 of a program that users should build themselves after PCRE is installed, so
1518 building it automatically is not really right. What is more, it gave
1519 trouble in some build environments.
1520
1521 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1522
1523
1524 Version 7.5 10-Jan-08
1525 ---------------------
1526
1527 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1528 values in parens when parsing an RE using the C++ wrapper."
1529
1530 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1531 Characters greater than 255 were excluded from the class instead of being
1532 included.
1533
1534 3. The same bug as (2) above applied to negated POSIX classes such as
1535 [:^space:].
1536
1537 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1538 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1539 I have changed it.
1540
1541 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1542 first named subpattern) and a construct such as (?&a) would reference the
1543 first named subpattern whose name started with "a" (in other words, the
1544 length check was missing). Both these problems are fixed. "Subpattern name
1545 expected" is now given for (?&) (a zero-length name), and this patch also
1546 makes it give the same error for \k'' (previously it complained that that
1547 was a reference to a non-existent subpattern).
1548
1549 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1550 this is right because (?- can be followed by option settings as well as by
1551 digits. I have, however, made the messages clearer.
1552
1553 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1554 than the number used in the conditional) now cause a compile-time error.
1555 This is actually not compatible with Perl, which accepts such patterns, but
1556 treats the conditional as always being FALSE (as PCRE used to), but it
1557 seems to me that giving a diagnostic is better.
1558
1559 8. Change "alphameric" to the more common word "alphanumeric" in comments
1560 and messages.
1561
1562 9. Fix two occurrences of "backslash" in comments that should have been
1563 "backspace".
1564
1565 10. Remove two redundant lines of code that can never be obeyed (their function
1566 was moved elsewhere).
1567
1568 11. The program that makes PCRE's Unicode character property table had a bug
1569 which caused it to generate incorrect table entries for sequences of
1570 characters that have the same character type, but are in different scripts.
1571 It amalgamated them into a single range, with the script of the first of
1572 them. In other words, some characters were in the wrong script. There were
1573 thirteen such cases, affecting characters in the following ranges:
1574
1575 U+002b0 - U+002c1
1576 U+0060c - U+0060d
1577 U+0061e - U+00612
1578 U+0064b - U+0065e
1579 U+0074d - U+0076d
1580 U+01800 - U+01805
1581 U+01d00 - U+01d77
1582 U+01d9b - U+01dbf
1583 U+0200b - U+0200f
1584 U+030fc - U+030fe
1585 U+03260 - U+0327f
1586 U+0fb46 - U+0fbb1
1587 U+10450 - U+1049d
1588
1589 12. The -o option (show only the matching part of a line) for pcregrep was not
1590 compatible with GNU grep in that, if there was more than one match in a
1591 line, it showed only the first of them. It now behaves in the same way as
1592 GNU grep.
1593
1594 13. If the -o and -v options were combined for pcregrep, it printed a blank
1595 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1596 does the same. The return code can be used to tell if there were any
1597 non-matching lines.
1598
1599 14. Added --file-offsets and --line-offsets to pcregrep.
1600
1601 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1602 infinitely looping recursion. The bug was that positive lookaheads were not
1603 being skipped when checking for a possible empty match (negative lookaheads
1604 and both kinds of lookbehind were skipped).
1605
1606 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1607 inclusion of <windows.h> to before rather than after the definition of
1608 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1609
1610 17. Specifying a possessive quantifier with a specific limit for a Unicode
1611 character property caused pcre_compile() to compile bad code, which led at
1612 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1613 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1614 caused the error; without that there was no problem.
1615
1616 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1617
1618 19. Added --enable-pcretest-libreadline.
1619
1620 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1621 RE::GlobalReplace(). As a result, the number of replacements returned was
1622 double what it should be. I removed one of the increments, but Craig sent a
1623 later patch that removed the other one (the right fix) and added unit tests
1624 that check the return values (which was not done before).
1625
1626 21. Several CMake things:
1627
1628 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1629 the names libpcre and libpcreposix, not just pcre and pcreposix.
1630
1631 (2) The above change means that pcretest and pcregrep are now correctly
1632 linked with the newly-built libraries, not previously installed ones.
1633
1634 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1635
1636 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1637 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1638 UTF-8 newline character). The key issue is that the pattern starts .*;
1639 this means that the match must be either at the beginning, or after a
1640 newline. The bug was in the code for advancing after a failed match and
1641 checking that the new position followed a newline. It was not taking
1642 account of UTF-8 characters correctly.
1643
1644 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1645 character classes. PCRE was not treating the sequence [:...:] as a
1646 character class unless the ... were all letters. Perl, however, seems to
1647 allow any characters between [: and :], though of course it rejects as
1648 unknown any "names" that contain non-letters, because all the known class
1649 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1650 for example, whereas PCRE did not - it did not recognize a POSIX character
1651 class. This seemed a bit dangerous, so the code has been changed to be
1652 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1653 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1654 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1655 Perl does, and where it didn't before.
1656
1657 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1658 Windows environments %n is disabled by default.
1659
1660
1661 Version 7.4 21-Sep-07
1662 ---------------------
1663
1664 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1665 means that a class such as [\s] counted as "explicit reference to CR or
1666 LF". That isn't really right - the whole point of the change was to try to
1667 help when there was an actual mention of one of the two characters. So now
1668 the change happens only if \r or \n (or a literal CR or LF) character is
1669 encountered.
1670
1671 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1672 of both had grown to the point where there were only 3 bits left.
1673 Fortunately, there was spare space in the data structure, and so I have
1674 moved the internal flags into a new 16-bit field to free up more option
1675 bits.
1676
1677 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1678 but did not set the internal JCHANGED flag - either of these is enough to
1679 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1680 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1681 start sets both bits.
1682
1683 4. Added options (at build time, compile time, exec time) to change \R from
1684 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1685
1686 5. doc/pcresyntax.html was missing from the distribution.
1687
1688 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1689 compatibility, even though it is no longer used.
1690
1691 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1692 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1693 windows.h file is present (where different names are used). [This was
1694 reversed later after testing - see 16 below.]
1695
1696 8. Changed all #include <config.h> to #include "config.h". There were also
1697 some further <pcre.h> cases that I changed to "pcre.h".
1698
1699 9. When pcregrep was used with the --colour option, it missed the line ending
1700 sequence off the lines that it output.
1701
1702 10. It was pointed out to me that arrays of string pointers cause lots of
1703 relocations when a shared library is dynamically loaded. A technique of
1704 using a single long string with a table of offsets can drastically reduce
1705 these. I have refactored PCRE in four places to do this. The result is
1706 dramatic:
1707
1708 Originally: 290
1709 After changing UCP table: 187
1710 After changing error message table: 43
1711 After changing table of "verbs" 36
1712 After changing table of Posix names 22
1713
1714 Thanks to the folks working on Gregex for glib for this insight.
1715
1716 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1717 unicode-properties was also set.
1718
1719 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1720
1721 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1722 checked only for CRLF.
1723
1724 14. Added casts to pcretest.c to avoid compiler warnings.
1725
1726 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1727
1728 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1729 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1730 entirely. This removes changes made in 7 above.
1731
1732 17. The CMake files have been updated, and there is now more information about
1733 building with CMake in the NON-UNIX-USE document.
1734
1735
1736 Version 7.3 28-Aug-07
1737 ---------------------
1738
1739 1. In the rejigging of the build system that eventually resulted in 7.1, the
1740 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1741 brackets there is not right, since it causes compilers to look for an
1742 installed pcre.h, not the version that is in the source that is being
1743 compiled (which of course may be different). I have changed it back to:
1744
1745 #include "pcre.h"
1746
1747 I have a vague recollection that the change was concerned with compiling in
1748 different directories, but in the new build system, that is taken care of
1749 by the VPATH setting the Makefile.
1750
1751 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1752 when the subject happened to end in the byte 0x85 (e.g. if the last
1753 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1754 characters but of course it shouldn't be taken as a newline when it is part
1755 of another character. The bug was that, for an unlimited repeat of . in
1756 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1757 characters when looking for a newline.
1758
1759 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1760
1761 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1762 in debug output.
1763
1764 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1765 long printing in the pcrecpp unittest when running under MinGW.
1766
1767 6. ESC_K was left out of the EBCDIC table.
1768
1769 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1770 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1771 limit also applies to "virtual nesting" when a pattern is recursive, and in
1772 this case 1000 isn't so big. I have been able to remove this limit at the
1773 expense of backing off one optimization in certain circumstances. Normally,
1774 when pcre_exec() would call its internal match() function recursively and
1775 immediately return the result unconditionally, it uses a "tail recursion"
1776 feature to save stack. However, when a subpattern that can match an empty
1777 string has an unlimited repetition quantifier, it no longer makes this
1778 optimization. That gives it a stack frame in which to save the data for
1779 checking that an empty string has been matched. Previously this was taken
1780 from the 1000-entry workspace that had been reserved. So now there is no
1781 explicit limit, but more stack is used.
1782
1783 8. Applied Daniel's patches to solve problems with the import/export magic
1784 syntax that is required for Windows, and which was going wrong for the
1785 pcreposix and pcrecpp parts of the library. These were overlooked when this
1786 problem was solved for the main library.
1787
1788 9. There were some crude static tests to avoid integer overflow when computing
1789 the size of patterns that contain repeated groups with explicit upper
1790 limits. As the maximum quantifier is 65535, the maximum group length was
1791 set at 30,000 so that the product of these two numbers did not overflow a
1792 32-bit integer. However, it turns out that people want to use groups that
1793 are longer than 30,000 bytes (though not repeat them that many times).
1794 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1795 made it possible to implement the integer overflow checks in a much more
1796 dynamic way, which I have now done. The artificial limitation on group
1797 length has been removed - we now have only the limit on the total length of
1798 the compiled pattern, which depends on the LINK_SIZE setting.
1799
1800 10. Fixed a bug in the documentation for get/copy named substring when
1801 duplicate names are permitted. If none of the named substrings are set, the
1802 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1803 empty string.
1804
1805 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1806 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1807 because the ] is interpreted as the first data character and the
1808 terminating ] is not found. PCRE has been made compatible with Perl in this
1809 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1810 cause memory overwriting.
1811
1812 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1813 string has been matched (to stop an infinite loop). It was not recognizing
1814 a conditional subpattern that could match an empty string if that
1815 subpattern was within another subpattern. For example, it looped when
1816 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1817 condition was not nested. This bug has been fixed.
1818
1819 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1820 past the start of the subject in the presence of bytes with the top bit
1821 set, for example "\x8aBCD".
1822
1823 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1824 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1825
1826 14. Optimized (?!) to (*FAIL).
1827
1828 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1829 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1830 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1831 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1832 does: it's just the validity check that is more restrictive.
1833
1834 16. Inserted checks for integer overflows during escape sequence (backslash)
1835 processing, and also fixed erroneous offset values for syntax errors during
1836 backslash processing.
1837
1838 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1839 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1840
1841 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1842 caused an overrun.
1843
1844 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1845 something other than just ASCII characters) inside a group that had an
1846 unlimited repeat caused a loop at compile time (while checking to see
1847 whether the group could match an empty string).
1848
1849 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1850 [\P{Any}] did so. (Error in the code for printing property names.)
1851
1852 21. An orphan \E inside a character class could cause a crash.
1853
1854 22. A repeated capturing bracket such as (A)? could cause a wild memory
1855 reference during compilation.
1856
1857 23. There are several functions in pcre_compile() that scan along a compiled
1858 expression for various reasons (e.g. to see if it's fixed length for look
1859 behind). There were bugs in these functions when a repeated \p or \P was
1860 present in the pattern. These operators have additional parameters compared
1861 with \d, etc, and these were not being taken into account when moving along
1862 the compiled data. Specifically:
1863
1864 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1865 length.
1866
1867 (b) An item such as \pL+ within a repeated group could cause crashes or
1868 loops.
1869
1870 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1871 "reference to non-existent subpattern" error.
1872
1873 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1874
1875 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1876 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1877
1878 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1879
1880 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1881 character were causing crashes (broken optimization).
1882
1883 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1884 \p or \P) caused a compile-time loop.
1885
1886 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1887 break. For example, the unstudied pattern [\r\n]A does not match the string
1888 "\r\nA" because change 7.0/46 below moves the current point on by two
1889 characters after failing to match at the start. However, the pattern \nA
1890 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1891 the same is true. There doesn't seem any very clean way out of this, but
1892 what I have chosen to do makes the common cases work: PCRE now takes note
1893 of whether there can be an explicit match for \r or \n anywhere in the
1894 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1895 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1896 pattern has explicit CR or LF references.
1897
1898 29. Added (*CR) etc for changing newline setting at start of pattern.
1899
1900
1901 Version 7.2 19-Jun-07
1902 ---------------------
1903
1904 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1905 which is apparently normally available under Windows.
1906
1907 2. Re-jig the pcregrep tests with different newline settings in an attempt
1908 to make them independent of the local environment's newline setting.
1909
1910 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1911
1912 4. Some of the "internals" tests were previously cut out when the link size
1913 was not 2, because the output contained actual offsets. The recent new
1914 "Z" feature of pcretest means that these can be cut out, making the tests
1915 usable with all link sizes.
1916
1917 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1918 stack recursion. This gives a massive performance boost under BSD, but just
1919 a small improvement under Linux. However, it saves one field in the frame
1920 in all cases.
1921
1922 6. Added more features from the forthcoming Perl 5.10:
1923
1924 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1925 recursion call. It refers to the nth most recently opened parentheses.
1926
1927 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1928 to be opened parentheses.
1929
1930 (c) Conditions that refer to capturing parentheses can be specified
1931 relatively, for example, (?(-2)... or (?(+3)...
1932
1933 (d) \K resets the start of the current match so that everything before
1934 is not part of it.
1935
1936 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1937
1938 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1939 reference syntax.
1940
1941 (g) (?| introduces a group in which the numbering of parentheses in each
1942 alternative starts with the same number.
1943
1944 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1945
1946 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1947 PCRE_INFO_JCHANGED.
1948
1949 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1950 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1951 for detecting groups that can match an empty string.
1952
1953 9. A pattern with a very large number of alternatives (more than several
1954 hundred) was running out of internal workspace during the pre-compile
1955 phase, where pcre_compile() figures out how much memory will be needed. A
1956 bit of new cunning has reduced the workspace needed for groups with
1957 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1958 workspace instead of running out of the 4096 that are available.
1959
1960 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1961
1962 11. Applied patch from Google to remove an optimization that didn't quite work.
1963 The report of the bug said:
1964
1965 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1966 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1967 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1968
1969 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1970 it matched the wrong number of bytes.
1971
1972
1973 Version 7.1 24-Apr-07
1974 ---------------------
1975
1976 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1977 that is more "standard", making use of automake and other Autotools. There
1978 is some re-arrangement of the files and adjustment of comments consequent
1979 on this.
1980
1981 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1982 for recursive directory scanning broke on some systems because the files
1983 are not scanned in any specific order and on different systems the order
1984 was different. A call to "sort" has been inserted into RunGrepTest for the
1985 approprate test as a short-term fix. In the longer term there may be an
1986 alternative.
1987
1988 3. I had an email from Eric Raymond about problems translating some of PCRE's
1989 man pages to HTML (despite the fact that I distribute HTML pages, some
1990 people do their own conversions for various reasons). The problems
1991 concerned the use of low-level troff macros .br and .in. I have therefore
1992 removed all such uses from the man pages (some were redundant, some could
1993 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1994 HTML has been updated to handle .nf/.fi and to complain if it encounters
1995 .br or .in.
1996
1997 4. Updated comments in configure.ac that get placed in config.h.in and also
1998 arranged for config.h to be included in the distribution, with the name
1999 config.h.generic, for the benefit of those who have to compile without
2000 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
2001
2002 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
2003 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
2004 makevp.bat for the current PCRE, using the additional files
2005 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
2006
2007 6. A Windows user reported a minor discrepancy with test 2, which turned out
2008 to be caused by a trailing space on an input line that had got lost in his
2009 copy. The trailing space was an accident, so I've just removed it.
2010
2011 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
2012 that is needed.
2013
2014 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
2015 as "const" (a) because they are and (b) because it helps the PHP
2016 maintainers who have recently made a script to detect big data structures
2017 in the php code that should be moved to the .rodata section. I remembered
2018 to update Builducptable as well, so it won't revert if ucptable.h is ever
2019 re-created.
2020
2021 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
2022 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
2023 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2024 support is not required. This saves 1.5-2K of code, which is important in
2025 some applications.
2026
2027 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2028 so as not to refer to the tables, even though these functions will never be
2029 called when UTF-8 support is disabled. Otherwise there are problems with a
2030 shared library.
2031
2032 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2033
2034 (a) It was defining its arguments as char * instead of void *.
2035
2036 (b) It was assuming that all moves were upwards in memory; this was true
2037 a long time ago when I wrote it, but is no longer the case.
2038
2039 The emulated memove() is provided for those environments that have neither
2040 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2041 is clearly not the case, as these two bugs were recently reported.
2042
2043 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2044 and Detrail to create the HTML documentation, the .txt form of the man
2045 pages, and it removes trailing spaces from listed files. It also creates
2046 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2047 case, it wraps all the #defines with #ifndefs. This script should be run
2048 before "make dist".
2049
2050 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2051 with Unicode property support.
2052
2053 (a) For a maximizing quantifier, if the two different cases of the
2054 character were of different lengths in their UTF-8 codings (there are
2055 some cases like this - I found 11), and the matching function had to
2056 back up over a mixture of the two cases, it incorrectly assumed they
2057 were both the same length.
2058
2059 (b) When PCRE was configured to use the heap rather than the stack for
2060 recursion during matching, it was not correctly preserving the data for
2061 the other case of a UTF-8 character when checking ahead for a match
2062 while processing a minimizing repeat. If the check also involved
2063 matching a wide character, but failed, corruption could cause an
2064 erroneous result when trying to check for a repeat of the original
2065 character.
2066
2067 13. Some tidying changes to the testing mechanism:
2068
2069 (a) The RunTest script now detects the internal link size and whether there
2070 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2071 values substituted by "configure". (The RunGrepTest script already did
2072 this for UTF-8.) The configure.ac script no longer substitutes the
2073 relevant variables.
2074
2075 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2076 with length and offset values. This means that the output is different
2077 for different internal link sizes. Test 2 is skipped for link sizes
2078 other than 2 because of this, bypassing the problem. Unfortunately,
2079 there was also a test in test 3 (the locale tests) that used /B and
2080 failed for link sizes other than 2. Rather than cut the whole test out,
2081 I have added a new /Z option to pcretest that replaces the length and
2082 offset values with spaces. This is now used to make test 3 independent
2083 of link size. (Test 2 will be tidied up later.)
2084
2085 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2086 segmentation fault instead of returning the appropriate error message.
2087
2088 15. In multiline mode when the newline sequence was set to "any", the pattern
2089 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2090 This doesn't seem right; it now treats the CRLF combination as the line
2091 ending, and so does not match in that case. It's only a pattern such as ^$
2092 that would hit this one: something like ^ABC$ would have failed after \r
2093 and then tried again after \r\n.
2094
2095 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2096 in an attempt to make files that differ only in their line terminators
2097 compare equal. This works on Linux.
2098
2099 17. Under certain error circumstances pcregrep might try to free random memory
2100 as it exited. This is now fixed, thanks to valgrind.
2101
2102 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2103 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2104 was because its rules for how to advance for /g after matching an empty
2105 string at the end of a line did not allow for this case. They now check for
2106 it specially.
2107
2108 20. pcretest is supposed to handle patterns and data of any length, by
2109 extending its buffers when necessary. It was getting this wrong when the
2110 buffer for a data line had to be extended.
2111
2112 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2113 CRLF as a newline sequence.
2114
2115 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2116 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2117 I have nevertheless tidied it up.
2118
2119 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2120
2121 24. Added a man page for pcre-config.
2122
2123
2124 Version 7.0 19-Dec-06
2125 ---------------------
2126
2127 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2128 moving to gcc 4.1.1.
2129
2130 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2131 sys/time.h, which is documented as needed for this function. It doesn't
2132 seem to matter on Linux, but it showed up on some releases of OS X.
2133
2134 3. It seems that there are systems where bytes whose values are greater than
2135 127 match isprint() in the "C" locale. The "C" locale should be the
2136 default when a C program starts up. In most systems, only ASCII printing
2137 characters match isprint(). This difference caused the output from pcretest
2138 to vary, making some of the tests fail. I have changed pcretest so that:
2139
2140 (a) When it is outputting text in the compiled version of a pattern, bytes
2141 other than 32-126 are always shown as hex escapes.
2142
2143 (b) When it is outputting text that is a matched part of a subject string,
2144 it does the same, unless a different locale has been set for the match
2145 (using the /L modifier). In this case, it uses isprint() to decide.
2146
2147 4. Fixed a major bug that caused incorrect computation of the amount of memory
2148 required for a compiled pattern when options that changed within the
2149 pattern affected the logic of the preliminary scan that determines the
2150 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2151 that the computed length was too small. The symptoms of this bug were
2152 either the PCRE error "internal error: code overflow" from pcre_compile(),
2153 or a glibc crash with a message such as "pcretest: free(): invalid next
2154 size (fast)". Examples of patterns that provoked this bug (shown in
2155 pcretest format) are:
2156
2157 /(?-x: )/x
2158 /(?x)(?-x: \s*#\s*)/
2159 /((?i)[\x{c0}])/8
2160 /(?i:[\x{c0}])/8
2161
2162 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2163 is now done differently.
2164
2165 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2166 wrapper classes; (b) implement a new function in the C++ scanner that is
2167 more efficient than the old way of doing things because it avoids levels of
2168 recursion in the regex matching; (c) add a paragraph to the documentation
2169 for the FullMatch() function.
2170
2171 6. The escape sequence \n was being treated as whatever was defined as
2172 "newline". Not only was this contrary to the documentation, which states
2173 that \n is character 10 (hex 0A), but it also went horribly wrong when
2174 "newline" was defined as CRLF. This has been fixed.
2175
2176 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2177 was being set to -1 for the "end of line" case (supposedly a value that no
2178 character can have). Though this value is never used (the check for end of
2179 line is "zero bytes in current character"), it caused compiler complaints.
2180 I've changed it to 0xffffffff.
2181
2182 8. In pcre_version.c, the version string was being built by a sequence of
2183 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2184 string (as it is for production releases) called a macro with an empty
2185 argument. The C standard says the result of this is undefined. The gcc
2186 compiler treats it as an empty string (which was what was wanted) but it is
2187 reported that Visual C gives an error. The source has been hacked around to
2188 avoid this problem.
2189
2190 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2191 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2192 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2193 of them did).
2194
2195 10. Originally, pcretest opened its input and output without "b"; then I was
2196 told that "b" was needed in some environments, so it was added for release
2197 5.0 to both the input and output. (It makes no difference on Unix-like
2198 systems.) Later I was told that it is wrong for the input on Windows. I've
2199 now abstracted the modes into two macros, to make it easier to fiddle with
2200 them, and removed "b" from the input mode under Windows.
2201
2202 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2203
2204 12. Added -help and --help to pcretest as an official way of being reminded
2205 of the options.
2206
2207 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2208 and pcrecpp.cc because they annoy compilers at high warning levels.
2209
2210 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2211
2212 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2213 scripts are not C programs :-) and which was not noticed because it works
2214 on Linux.
2215
2216 16. pcretest is supposed to handle any length of pattern and data line (as one
2217 line or as a continued sequence of lines) by extending its input buffer if
2218 necessary. This feature was broken for very long pattern lines, leading to
2219 a string of junk being passed to pcre_compile() if the pattern was longer
2220 than about 50K.
2221
2222 17. I have done a major re-factoring of the way pcre_compile() computes the
2223 amount of memory needed for a compiled pattern. Previously, there was code
2224 that made a preliminary scan of the pattern in order to do this. That was
2225 OK when PCRE was new, but as the facilities have expanded, it has become
2226 harder and harder to keep it in step with the real compile phase, and there
2227 have been a number of bugs (see for example, 4 above). I have now found a
2228 cunning way of running the real compile function in a "fake" mode that
2229 enables it to compute how much memory it would need, while actually only
2230 ever using a few hundred bytes of working memory and without too many
2231 tests of the mode. This should make future maintenance and development
2232 easier. A side effect of this work is that the limit of 200 on the nesting
2233 depth of parentheses has been removed (though this was never a serious
2234 limitation, I suspect). However, there is a downside: pcre_compile() now
2235 runs more slowly than before (30% or more, depending on the pattern). I
2236 hope this isn't a big issue. There is no effect on runtime performance.
2237
2238 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2239 newline (only possible for the last line of a file) and it was a
2240 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2241
2242 19. Added additional timing features to pcretest. (1) The -tm option now times
2243 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2244 separate command line item, by a number that specifies the number of
2245 repeats to use when timing. The default is 50000; this gives better
2246 precision, but takes uncomfortably long for very large patterns.
2247
2248 20. Extended pcre_study() to be more clever in cases where a branch of a
2249 subpattern has no definite first character. For example, (a*|b*)[cd] would
2250 previously give no result from pcre_study(). Now it recognizes that the
2251 first character must be a, b, c, or d.
2252
2253 21. There was an incorrect error "recursive call could loop indefinitely" if
2254 a subpattern (or the entire pattern) that was being tested for matching an
2255 empty string contained only one non-empty item after a nested subpattern.
2256 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2257 incorrectly, because the \d was being skipped in the check.
2258
2259 22. The pcretest program now has a new pattern option /B and a command line
2260 option -b, which is equivalent to adding /B to every pattern. This causes
2261 it to show the compiled bytecode, without the additional information that
2262 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2263 is the same as /B/I).
2264
2265 23. A new optimization is now able automatically to treat some sequences such
2266 as a*b as a*+b. More specifically, if something simple (such as a character
2267 or a simple class like \d) has an unlimited quantifier, and is followed by
2268 something that cannot possibly match the quantified thing, the quantifier
2269 is automatically "possessified".
2270
2271 24. A recursive reference to a subpattern whose number was greater than 39
2272 went wrong under certain circumstances in UTF-8 mode. This bug could also
2273 have affected the operation of pcre_study().
2274
2275 25. Realized that a little bit of performance could be had by replacing
2276 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2277
2278 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2279
2280 27. Possessive quantifiers such as a++ were previously implemented by turning
2281 them into atomic groups such as ($>a+). Now they have their own opcodes,
2282 which improves performance. This includes the automatically created ones
2283 from 23 above.
2284
2285 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2286 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2287 the first matched character to be a colon. This applied both to named and
2288 numbered groups.
2289
2290 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2291
2292 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2293 building PCRE on Windows easier, so I have included it in the distribution.
2294
2295 31. There is now a check in pcretest against a ridiculously large number being
2296 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2297 loop, the loop is abandoned.
2298
2299 32. Forward references to subpatterns in conditions such as (?(2)...) where
2300 subpattern 2 is defined later cause pcre_compile() to search forwards in
2301 the pattern for the relevant set of parentheses. This search went wrong
2302 when there were unescaped parentheses in a character class, parentheses
2303 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2304
2305 33. "Subroutine" calls and backreferences were previously restricted to
2306 referencing subpatterns earlier in the regex. This restriction has now
2307 been removed.
2308
2309 34. Added a number of extra features that are going to be in Perl 5.10. On the
2310 whole, these are just syntactic alternatives for features that PCRE had
2311 previously implemented using the Python syntax or my own invention. The
2312 other formats are all retained for compatibility.
2313
2314 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2315 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2316 also .NET compatible.
2317
2318 (b) A recursion or subroutine call to a named group can now be defined as
2319 (?&name) as well as (?P>name).
2320
2321 (c) A backreference to a named group can now be defined as \k<name> or
2322 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2323 5.10, are also .NET compatible.
2324
2325 (d) A conditional reference to a named group can now use the syntax
2326 (?(<name>) or (?('name') as well as (?(name).
2327
2328 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2329 groups (named and numbered) that are never evaluated inline, but can be
2330 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2331 is always false. There may be only one alternative in such a group.
2332
2333 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2334 as the simple (?(R). The condition is true only if the most recent
2335 recursion is that of the given number or name. It does not search out
2336 through the entire recursion stack.
2337
2338 (g) The escape \gN or \g{N} has been added, where N is a positive or
2339 negative number, specifying an absolute or relative reference.
2340
2341 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2342 some "unreachable code" warnings.
2343
2344 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2345 things, this adds five new scripts.
2346
2347 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2348 There were also incompatibilities regarding the handling of \Q..\E inside
2349 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2350 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2351
2352 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2353 matches an empty string, and forcibly breaks the loop. There were bugs in
2354 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2355 against aaaa the result was just "a" rather than "aaaa", for example. Two
2356 separate and independent bugs (that affected different cases) have been
2357 fixed.
2358
2359 39. Refactored the code to abolish the use of different opcodes for small
2360 capturing bracket numbers. This is a tidy that I avoided doing when I
2361 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2362 The new approach is not only tidier, it makes it possible to reduce the
2363 memory needed to fix the previous bug (38).
2364
2365 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2366 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2367 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2368 mode.
2369
2370 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2371 report.
2372
2373 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2374 copy construction and assignment in the C++ wrapper.
2375
2376 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2377 couple of bugs that could have given wrong results in the "--newline=crlf"
2378 case.
2379
2380 44. Added a number of casts and did some reorganization of signed/unsigned int
2381 variables following suggestions from Dair Grant. Also renamed the variable
2382 "this" as "item" because it is a C++ keyword.
2383
2384 45. Arranged for dftables to add
2385
2386 #include "pcre_internal.h"
2387
2388 to pcre_chartables.c because without it, gcc 4.x may remove the array
2389 definition from the final binary if PCRE is built into a static library and
2390 dead code stripping is activated.
2391
2392 46. For an unanchored pattern, if a match attempt fails at the start of a
2393 newline sequence, and the newline setting is CRLF or ANY, and the next two
2394 characters are CRLF, advance by two characters instead of one.
2395
2396
2397 Version 6.7 04-Jul-06
2398 ---------------------
2399
2400 1. In order to handle tests when input lines are enormously long, pcretest has
2401 been re-factored so that it automatically extends its buffers when
2402 necessary. The code is crude, but this _is_ just a test program. The
2403 default size has been increased from 32K to 50K.
2404
2405 2. The code in pcre_study() was using the value of the re argument before
2406 testing it for NULL. (Of course, in any sensible call of the function, it
2407 won't be NULL.)
2408
2409 3. The memmove() emulation function in pcre_internal.h, which is used on
2410 systems that lack both memmove() and bcopy() - that is, hardly ever -
2411 was missing a "static" storage class specifier.
2412
2413 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2414 containing an extended class (one that cannot be represented by a bitmap
2415 because it contains high-valued characters or Unicode property items, e.g.
2416 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2417 pattern, but PCRE should not loop if you do not (it no longer does).
2418 [Detail: two cases were found: (a) a repeated subpattern containing an
2419 extended class; (b) a recursive reference to a subpattern that followed a
2420 previous extended class. It wasn't skipping over the extended class
2421 correctly when UTF-8 mode was not set.]
2422
2423 5. A negated single-character class was not being recognized as fixed-length
2424 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2425 compile error "lookbehind assertion is not fixed length".
2426
2427 6. The RunPerlTest auxiliary script was showing an unexpected difference
2428 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2429 write a Perl script that can interpret lines of an input file either as
2430 byte characters or as UTF-8, which is what "perltest" was being required to
2431 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2432 can't do is switch easily at run time between having the "use utf8;" pragma
2433 or not. In the end, I fudged it by using the RunPerlTest script to insert
2434 "use utf8;" explicitly for the UTF-8 tests.
2435
2436 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2437 the end of the subject string, contrary to the documentation and to what
2438 Perl does. This was true of both matching functions. Now it matches only at
2439 the start of the subject and immediately after *internal* newlines.
2440
2441 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2442 a pointer to an int instead of a pointer to an unsigned long int. This
2443 caused problems on 64-bit systems.
2444
2445 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2446 instance of the 'standard' template library not being so standard".
2447
2448 10. There was no check on the number of named subpatterns nor the maximum
2449 length of a subpattern name. The product of these values is used to compute
2450 the size of the memory block for a compiled pattern. By supplying a very
2451 long subpattern name and a large number of named subpatterns, the size
2452 computation could be caused to overflow. This is now prevented by limiting
2453 the length of names to 32 characters, and the number of named subpatterns
2454 to 10,000.
2455
2456 11. Subpatterns that are repeated with specific counts have to be replicated in
2457 the compiled pattern. The size of memory for this was computed from the
2458 length of the subpattern and the repeat count. The latter is limited to
2459 65535, but there was no limit on the former, meaning that integer overflow
2460 could in principle occur. The compiled length of a repeated subpattern is
2461 now limited to 30,000 bytes in order to prevent this.
2462
2463 12. Added the optional facility to have named substrings with the same name.
2464
2465 13. Added the ability to use a named substring as a condition, using the
2466 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2467 are numbers (not recommended). Forward references are permitted.
2468
2469 14. Added forward references in named backreferences (if you see what I mean).
2470
2471 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2472 pattern could run off the end of the subject. For example, the pattern
2473 "(?s)(.{1,5})"8 did this with the subject "ab".
2474
2475 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2476 PCRE_CASELESS was set when matching characters that were quantified with ?
2477 or *.
2478
2479 17. A character class other than a single negated character that had a minimum
2480 but no maximum quantifier - for example [ab]{6,} - was not handled
2481 correctly by pce_dfa_exec(). It would match only one character.
2482
2483 18. A valid (though odd) pattern that looked like a POSIX character
2484 class but used an invalid character after [ (for example [[,abc,]]) caused
2485 pcre_compile() to give the error "Failed: internal error: code overflow" or
2486 in some cases to crash with a glibc free() error. This could even happen if
2487 the pattern terminated after [[ but there just happened to be a sequence of
2488 letters, a binary zero, and a closing ] in the memory that followed.
2489
2490 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2491 over the years. Originally (before any Unicode support), just the bottom 8
2492 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2493 output from "man perlunicode" includes this:
2494
2495 The regular expression compiler produces polymorphic opcodes. That
2496 is, the pattern adapts to the data and automatically switches to
2497 the Unicode character scheme when presented with Unicode data--or
2498 instead uses a traditional byte scheme when presented with byte
2499 data.
2500
2501 Sadly, a wide octal escape does not cause a switch, and in a string with
2502 no other multibyte characters, these octal escapes are treated as before.
2503 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2504 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2505 Unicode string.
2506
2507 I have not perpetrated such confusion in PCRE. Up till now, it took just
2508 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2509 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2510 translate to the appropriate multibyte character.
2511
2512 29. Applied some refactoring to reduce the number of warnings from Microsoft
2513 and Borland compilers. This has included removing the fudge introduced
2514 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2515 a warning about an unused variable.
2516
2517 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2518 characters since release 4.0, because Perl (from release 5.004) does not.
2519 [Or at least, is documented not to: some releases seem to be in conflict
2520 with the documentation.] However, when a pattern was studied with
2521 pcre_study() and all its branches started with \s, PCRE still included VT
2522 as a possible starting character. Of course, this did no harm; it just
2523 caused an unnecessary match attempt.
2524
2525 22. Removed a now-redundant internal flag bit that recorded the fact that case
2526 dependency changed within the pattern. This was once needed for "required
2527 byte" processing, but is no longer used. This recovers a now-scarce options
2528 bit. Also moved the least significant internal flag bit to the most-
2529 significant bit of the word, which was not previously used (hangover from
2530 the days when it was an int rather than a uint) to free up another bit for
2531 the future.
2532
2533 23. Added support for CRLF line endings as well as CR and LF. As well as the
2534 default being selectable at build time, it can now be changed at runtime
2535 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2536 specify that it is scanning data with non-default line endings.
2537
2538 24. Changed the definition of CXXLINK to make it agree with the definition of
2539 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2540
2541 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2542 recursions. This makes a big different to stack usage for some patterns.
2543
2544 26. If a subpattern containing a named recursion or subroutine reference such
2545 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2546 the space required for the compiled pattern went wrong and gave too small a
2547 value. Depending on the environment, this could lead to "Failed: internal
2548 error: code overflow at offset 49" or "glibc detected double free or
2549 corruption" errors.
2550
2551 27. Applied patches from Google (a) to support the new newline modes and (b) to
2552 advance over multibyte UTF-8 characters in GlobalReplace.
2553
2554 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2555 difference for some implementation of PCRE in some Windows version.
2556
2557 29. Added some extra testing facilities to pcretest:
2558
2559 \q<number> in a data line sets the "match limit" value
2560 \Q<number> in a data line sets the "match recursion limt" value
2561 -S <number> sets the stack size, where <number> is in megabytes
2562
2563 The -S option isn't available for Windows.
2564
2565
2566 Version 6.6 06-Feb-06
2567 ---------------------
2568
2569 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2570 in pcreposix.h. I have copied the definition from pcre.h.
2571
2572 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2573 because pcre.h is no longer a built file.
2574
2575 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2576 not normally included in the compiled code.
2577
2578
2579 Version 6.5 01-Feb-06
2580 ---------------------
2581
2582 1. When using the partial match feature with pcre_dfa_exec(), it was not
2583 anchoring the second and subsequent partial matches at the new starting
2584 point. This could lead to incorrect results. For example, with the pattern
2585 /1234/, partially matching against "123" and then "a4" gave a match.
2586
2587 2. Changes to pcregrep:
2588
2589 (a) All non-match returns from pcre_exec() were being treated as failures
2590 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2591 error message is output. Some extra information is given for the
2592 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2593 probably the only errors that are likely to be caused by users (by
2594 specifying a regex that has nested indefinite repeats, for instance).
2595 If there are more than 20 of these errors, pcregrep is abandoned.
2596
2597 (b) A binary zero was treated as data while matching, but terminated the
2598 output line if it was written out. This has been fixed: binary zeroes
2599 are now no different to any other data bytes.
2600
2601 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2602 used to set a locale for matching. The --locale=xxxx long option has
2603 been added (no short equivalent) to specify a locale explicitly on the
2604 pcregrep command, overriding the environment variables.
2605
2606 (d) When -B was used with -n, some line numbers in the output were one less
2607 than they should have been.
2608
2609 (e) Added the -o (--only-matching) option.
2610
2611 (f) If -A or -C was used with -c (count only), some lines of context were
2612 accidentally printed for the final match.
2613
2614 (g) Added the -H (--with-filename) option.
2615
2616 (h) The combination of options -rh failed to suppress file names for files
2617 that were found from directory arguments.
2618
2619 (i) Added the -D (--devices) and -d (--directories) options.
2620
2621 (j) Added the -F (--fixed-strings) option.
2622
2623 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2624
2625 (l) Added the --colo(u)r option.
2626
2627 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2628 is not present by default.
2629
2630 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2631 items such as (?R) or (?1), when the recursion could match a number of
2632 alternatives. If it matched one of the alternatives, but subsequently,
2633 outside the recursion, there was a failure, the code tried to back up into
2634 the recursion. However, because of the way PCRE is implemented, this is not
2635 possible, and the result was an incorrect result from the match.
2636
2637 In order to prevent this happening, the specification of recursion has
2638 been changed so that all such subpatterns are automatically treated as
2639 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2640
2641 4. I had overlooked the fact that, in some locales, there are characters for
2642 which isalpha() is true but neither isupper() nor islower() are true. In
2643 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2644 and ordfeminine) are like this. This affected the treatment of \w and \W
2645 when they appeared in character classes, but not when they appeared outside
2646 a character class. The bit map for "word" characters is now created
2647 separately from the results of isalnum() instead of just taking it from the
2648 upper, lower, and digit maps. (Plus the underscore character, of course.)
2649
2650 5. The above bug also affected the handling of POSIX character classes such as
2651 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2652 permanent tables. Instead, the bit maps for such a class were previously
2653 created as the appropriate unions of the upper, lower, and digit bitmaps.
2654 Now they are created by subtraction from the [[:word:]] class, which has
2655 its own bitmap.
2656
2657 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2658 It is created by subtracting the vertical space characters (\x09, \x0a,
2659 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2660 subtraction was done in the overall bitmap for a character class, meaning
2661 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2662 be recognized. This bug has been fixed.
2663
2664 7. Patches from the folks at Google:
2665
2666 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2667 real life, but is still worth protecting against".
2668
2669 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2670 regular expressions".
2671
2672 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2673 have it.
2674
2675 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2676 "configure" and the latter not, in order to fix a problem somebody had
2677 with compiling the Arg class on HP-UX.
2678
2679 (e) Improve the error-handling of the C++ wrapper a little bit.
2680
2681 (f) New tests for checking recursion limiting.
2682
2683 8. The pcre_memmove() function, which is used only if the environment does not
2684 have a standard memmove() function (and is therefore rarely compiled),
2685 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2686 returning a result (though PCRE never actually uses the result).
2687
2688 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2689 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2690 returned instead of calling malloc() with an overflowing number that would
2691 most likely cause subsequent chaos.
2692
2693 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2694
2695 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2696 with this option is matched, the nmatch and pmatch options of regexec() are
2697 ignored.
2698
2699 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2700 provided in case anyone wants to the the POSIX interface with UTF-8
2701 strings.
2702
2703 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2704 C++ linking (needed for some HP-UX environments).
2705
2706 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2707 (unused parameter) and in the pcre_printint() function (omitted "default"
2708 switch label when the default is to do nothing).
2709
2710 15. Added some code to make it possible, when PCRE is compiled as a C++
2711 library, to replace subject pointers for pcre_exec() with a smart pointer
2712 class, thus making it possible to process discontinuous strings.
2713
2714 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2715 much the same function. They were added by different people who were trying
2716 to make PCRE easy to compile on non-Unix systems. It has been suggested
2717 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2718 for compiling on Windows systems. I have therefore replaced it with
2719 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2720 defaults to "extern" for C or "extern C" for C++, which works fine on
2721 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2722 SCOPE with something explicit in config.h. In addition:
2723
2724 (a) pcreposix.h still had just "extern" instead of either of these macros;
2725 I have replaced it with PCRE_DATA_SCOPE.
2726
2727 (b) Functions such as _pcre_xclass(), which are internal to the library,
2728 but external in the C sense, all had PCRE_EXPORT in their definitions.
2729 This is apparently wrong for the Windows case, so I have removed it.
2730 (It makes no difference on Unix-like systems.)
2731
2732 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2733 of recursive calls to match(). This is different to MATCH_LIMIT because
2734 that limits the total number of calls to match(), not all of which increase
2735 the depth of recursion. Limiting the recursion depth limits the amount of
2736 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2737 when PCRE is compiled, and changed at run time. A patch from Google adds
2738 this functionality to the C++ interface.
2739
2740 18. Changes to the handling of Unicode character properties:
2741
2742 (a) Updated the table to Unicode 4.1.0.
2743
2744 (b) Recognize characters that are not in the table as "Cn" (undefined).
2745
2746 (c) I revised the way the table is implemented to a much improved format
2747 which includes recognition of ranges. It now supports the ranges that
2748 are defined in UnicodeData.txt, and it also amalgamates other
2749 characters into ranges. This has reduced the number of entries in the
2750 table from around 16,000 to around 3,000, thus reducing its size
2751 considerably. I realized I did not need to use a tree structure after
2752 all - a binary chop search is just as efficient. Having reduced the
2753 number of entries, I extended their size from 6 bytes to 8 bytes to
2754 allow for more data.
2755
2756 (d) Added support for Unicode script names via properties such as \p{Han}.
2757
2758 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2759 matching that character.
2760
2761 20. When matching a repeated Unicode property with a minimum greater than zero,
2762 (for example \pL{2,}), PCRE could look past the end of the subject if it
2763 reached it while seeking the minimum number of characters. This could
2764 happen only if some of the characters were more than one byte long, because
2765 there is a check for at least the minimum number of bytes.
2766
2767 21. Refactored the implementation of \p and \P so as to be more general, to
2768 allow for more different types of property in future. This has changed the
2769 compiled form incompatibly. Anybody with saved compiled patterns that use
2770 \p or \P will have to recompile them.
2771
2772 22. Added "Any" and "L&" to the supported property types.
2773
2774 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2775 but give a compile time error if the value is greater than 0xff.
2776
2777 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2778 accidentally not being installed or uninstalled.
2779
2780 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2781 made were to insert the current release number. This seemed silly, because
2782 it made things harder for people building PCRE on systems that don't run
2783 "configure". I have turned pcre.h into a distributed file, no longer built
2784 by "configure", with the version identification directly included. There is
2785 no longer a pcre.h.in file.
2786
2787 However, this change necessitated a change to the pcre-config script as
2788 well. It is built from pcre-config.in, and one of the substitutions was the
2789 release number. I have updated configure.ac so that ./configure now finds
2790 the release number by grepping pcre.h.
2791
2792 26. Added the ability to run the tests under valgrind.
2793
2794
2795 Version 6.4 05-Sep-05
2796 ---------------------
2797
2798 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2799 "--" to be printed when multiple files were scanned, even when none of the
2800 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2801 consider it to be a bug, and have restored the previous behaviour.
2802
2803 2. A couple of code tidies to get rid of compiler warnings.
2804
2805 3. The pcretest program used to cheat by referring to symbols in the library
2806 whose names begin with _pcre_. These are internal symbols that are not
2807 really supposed to be visible externally, and in some environments it is
2808 possible to suppress them. The cheating is now confined to including
2809 certain files from the library's source, which is a bit cleaner.
2810
2811 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2812 file's purpose clearer.
2813
2814 5. Reorganized pcre_ucp_findchar().
2815
2816
2817 Version 6.3 15-Aug-05
2818 ---------------------
2819
2820 1. The file libpcre.pc.in did not have general read permission in the tarball.
2821
2822 2. There were some problems when building without C++ support:
2823
2824 (a) If C++ support was not built, "make install" and "make test" still
2825 tried to test it.
2826
2827 (b) There were problems when the value of CXX was explicitly set. Some
2828 changes have been made to try to fix these, and ...
2829
2830 (c) --disable-cpp can now be used to explicitly disable C++ support.
2831
2832 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2833 backslash in a target when C++ was disabled. This confuses some
2834 versions of "make", apparently. Using an intermediate variable solves
2835 this. (Same for CPP_LOBJ.)
2836
2837 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2838 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2839 necessary on certain architectures.
2840
2841 4. Added a setting of -export-symbols-regex to the link command to remove
2842 those symbols that are exported in the C sense, but actually are local
2843 within the library, and not documented. Their names all begin with
2844 "_pcre_". This is not a perfect job, because (a) we have to except some
2845 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2846 available (and never for static libraries). I have made a note to try to
2847 find a way round (a) in the future.
2848
2849
2850 Version 6.2 01-Aug-05
2851 ---------------------
2852
2853 1. There was no test for integer overflow of quantifier values. A construction
2854 such as {1111111111111111} would give undefined results. What is worse, if
2855 a minimum quantifier for a parenthesized subpattern overflowed and became
2856 negative, the calculation of the memory size went wrong. This could have
2857 led to memory overwriting.
2858
2859 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2860
2861 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2862 operating environments where this matters.
2863
2864 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2865 PCRE options from within the C++ wrapper.
2866
2867 5. Named capturing subpatterns were not being correctly counted when a pattern
2868 was compiled. This caused two problems: (a) If there were more than 100
2869 such subpatterns, the calculation of the memory needed for the whole
2870 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2871 back references of the form \12, where the number was greater than 9, were
2872 not recognized as back references, even though there were sufficient
2873 previous subpatterns.
2874
2875 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2876 versions of gcc, e.g. 2.95.4.
2877
2878
2879 Version 6.1 21-Jun-05
2880 ---------------------
2881
2882 1. There was one reference to the variable "posix" in pcretest.c that was not
2883 surrounded by "#if !defined NOPOSIX".
2884
2885 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2886 the cross-check on the old pcre_info() function, for the benefit of the
2887 cut-down version of PCRE that is currently imported into Exim.
2888
2889 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2890 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2891 patterns that start (?i) but I don't think that matters. The use of (?i) is
2892 just an example; this all applies to the other options as well.
2893
2894 4. Since libtool seems to echo the compile commands it is issuing, the output
2895 from "make" can be reduced a bit by putting "@" in front of each libtool
2896 compile command.
2897
2898 5. Patch from the folks at Google for configure.in to be a bit more thorough
2899 in checking for a suitable C++ installation before trying to compile the
2900 C++ stuff. This should fix a reported problem when a compiler was present,
2901 but no suitable headers.
2902
2903 6. The man pages all had just "PCRE" as their title. I have changed them to
2904 be the relevant file name. I have also arranged that these names are
2905 retained in the file doc/pcre.txt, which is a concatenation in text format
2906 of all the man pages except the little individual ones for each function.
2907
2908 7. The NON-UNIX-USE file had not been updated for the different set of source
2909 files that come with release 6. I also added a few comments about the C++
2910 wrapper.
2911
2912
2913 Version 6.0 07-Jun-05
2914 ---------------------
2915
2916 1. Some minor internal re-organization to help with my DFA experiments.
2917
2918 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2919 didn't matter for the library itself when fully configured, but did matter
2920 when compiling without UCP support, or within Exim, where the ucp files are
2921 not imported.
2922
2923 3. Refactoring of the library code to split up the various functions into
2924 different source modules. The addition of the new DFA matching code (see
2925 below) to a single monolithic source would have made it really too
2926 unwieldy, quite apart from causing all the code to be include in a
2927 statically linked application, when only some functions are used. This is
2928 relevant even without the DFA addition now that patterns can be compiled in
2929 one application and matched in another.
2930
2931 The downside of splitting up is that there have to be some external
2932 functions and data tables that are used internally in different modules of
2933 the library but which are not part of the API. These have all had their
2934 names changed to start with "_pcre_" so that they are unlikely to clash
2935 with other external names.
2936
2937 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2938 a different (DFA) algorithm. Although it is slower than the original
2939 function, it does have some advantages for certain types of matching
2940 problem.
2941
2942 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2943 including restarting after a partial match.
2944
2945 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2946 defined when compiling for Windows was sent to me. I have put it into the
2947 code, though I have no means of testing or verifying it.
2948
2949 7. Added the pcre_refcount() auxiliary function.
2950
2951 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2952 match before or at the first newline in the subject string. In pcretest,
2953 the /f option on a pattern can be used to set this.
2954
2955 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2956 would behave wrongly. This has been present in PCRE since release 4.0.
2957
2958 10. A number of changes to the pcregrep command:
2959
2960 (a) Refactored how -x works; insert ^(...)$ instead of setting
2961 PCRE_ANCHORED and checking the length, in preparation for adding
2962 something similar for -w.
2963
2964 (b) Added the -w (match as a word) option.
2965
2966 (c) Refactored the way lines are read and buffered so as to have more
2967 than one at a time available.
2968
2969 (d) Implemented a pcregrep test script.
2970
2971 (e) Added the -M (multiline match) option. This allows patterns to match
2972 over several lines of the subject. The buffering ensures that at least
2973 8K, or the rest of the document (whichever is the shorter) is available
2974 for matching (and similarly the previous 8K for lookbehind assertions).
2975
2976 (f) Changed the --help output so that it now says
2977
2978 -w, --word-regex(p)
2979
2980 instead of two lines, one with "regex" and the other with "regexp"
2981 because that confused at least one person since the short forms are the
2982 same. (This required a bit of code, as the output is generated
2983 automatically from a table. It wasn't just a text change.)
2984
2985 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2986 option but starts with a hyphen. Could be a pattern or a path name
2987 starting with a hyphen, for instance.
2988
2989 (h) "-" can be given as a file name to represent stdin.
2990
2991 (i) When file names are being printed, "(standard input)" is used for
2992 the standard input, for compatibility with GNU grep. Previously
2993 "<stdin>" was used.
2994
2995 (j) The option --label=xxx can be used to supply a name to be used for
2996 stdin when file names are being printed. There is no short form.
2997
2998 (k) Re-factored the options decoding logic because we are going to add
2999 two more options that take data. Such options can now be given in four
3000 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
3001
3002 (l) Added the -A, -B, and -C options for requesting that lines of context
3003 around matches be printed.
3004
3005 (m) Added the -L option to print the names of files that do not contain
3006 any matching lines, that is, the complement of -l.
3007
3008 (n) The return code is 2 if any file cannot be opened, but pcregrep does
3009 continue to scan other files.
3010
3011 (o) The -s option was incorrectly implemented. For compatibility with other
3012 greps, it now suppresses the error message for a non-existent or non-
3013 accessible file (but not the return code). There is a new option called
3014 -q that suppresses the output of matching lines, which was what -s was
3015 previously doing.
3016
3017 (p) Added --include and --exclude options to specify files for inclusion
3018 and exclusion when recursing.
3019
3020 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
3021 Hopefully, it now does.
3022
3023 12. Missing cast in pcre_study().
3024
3025 13. Added an "uninstall" target to the makefile.
3026
3027 14. Replaced "extern" in the function prototypes in Makefile.in with
3028 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3029 world, but is set differently for Windows.
3030
3031 15. Added a second compiling function called pcre_compile2(). The only
3032 difference is that it has an extra argument, which is a pointer to an
3033 integer error code. When there is a compile-time failure, this is set
3034 non-zero, in addition to the error test pointer being set to point to an
3035 error message. The new argument may be NULL if no error number is required
3036 (but then you may as well call pcre_compile(), which is now just a
3037 wrapper). This facility is provided because some applications need a
3038 numeric error indication, but it has also enabled me to tidy up the way
3039 compile-time errors are handled in the POSIX wrapper.
3040
3041 16. Added VPATH=.libs to the makefile; this should help when building with one
3042 prefix path and installing with another. (Or so I'm told by someone who
3043 knows more about this stuff than I do.)
3044
3045 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3046 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3047 match everything, including newlines. This is not POSIX-compatible, but
3048 somebody wanted the feature. From pcretest it can be activated by using
3049 both the P and the s flags.
3050
3051 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3052
3053 19. libpcre.pc was being incorrectly installed as executable.
3054
3055 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3056 it now also looks for '\r' so that it will work unmodified on Windows.
3057
3058 21. Added Google's contributed C++ wrapper to the distribution.
3059
3060 22. Added some untidy missing memory free() calls in pcretest, to keep
3061 Electric Fence happy when testing.
3062
3063
3064
3065 Version 5.0 13-Sep-04
3066 ---------------------
3067
3068 1. Internal change: literal characters are no longer packed up into items
3069 containing multiple characters in a single byte-string. Each character
3070 is now matched using a separate opcode. However, there may be more than one
3071 byte in the character in UTF-8 mode.
3072
3073 2. The pcre_callout_block structure has two new fields: pattern_position and
3074 next_item_length. These contain the offset in the pattern to the next match
3075 item, and its length, respectively.
3076
3077 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3078 insertion of callouts before each pattern item. Added the /C option to
3079 pcretest to make use of this.
3080
3081 4. On the advice of a Windows user, the lines
3082
3083 #if defined(_WIN32) || defined(WIN32)
3084 _setmode( _fileno( stdout ), 0x8000 );
3085 #endif /* defined(_WIN32) || defined(WIN32) */
3086
3087 have been added to the source of pcretest. This apparently does useful
3088 magic in relation to line terminators.
3089
3090 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3091 for the benefit of those environments where the "b" makes a difference.
3092
3093 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3094 to know about it. I have put a hack into configure.in that adds in code
3095 to set GCC=yes if CC=icc. This seems to end up at a point in the
3096 generated configure script that is early enough to affect the setting of
3097 compiler options, which is what is needed, but I have no means of testing
3098 whether it really works. (The user who reported this had patched the
3099 generated configure script, which of course I cannot do.)
3100
3101 LATER: After change 22 below (new libtool files), the configure script
3102 seems to know about icc (and also ecc). Therefore, I have commented out
3103 this hack in configure.in.
3104
3105 7. Added support for pkg-config (2 patches were sent in).
3106
3107 8. Negated POSIX character classes that used a combination of internal tables
3108 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3109 [[:^ascii]]. Typically, they would match almost any characters. The other
3110 POSIX classes were not broken in this way.
3111
3112 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3113 to find the match, as PCRE was deluded into thinking that the match had to
3114 start at the start point or following a newline. The same bug applied to
3115 patterns with negative forward assertions or any backward assertions
3116 preceding ".*" at the start, unless the pattern required a fixed first
3117 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3118
3119 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3120 starting at the last subject character, bytes beyond the end of the subject
3121 string were read.
3122
3123 11. Renamed the variable "class" as "classbits" to make life easier for C++
3124 users. (Previously there was a macro definition, but it apparently wasn't
3125 enough.)
3126
3127 12. Added the new field "tables" to the extra data so that tables can be passed
3128 in at exec time, or the internal tables can be re-selected. This allows
3129 a compiled regex to be saved and re-used at a later time by a different
3130 program that might have everything at different addresses.
3131
3132 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3133 -R library as well as a -L library.
3134
3135 14. The debugging options of pcretest (-d on the command line or D on a
3136 pattern) showed incorrect output for anything following an extended class
3137 that contained multibyte characters and which was followed by a quantifier.
3138
3139 15. Added optional support for general category Unicode character properties
3140 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3141 support. It adds about 90K to the size of the library. The meanings of the
3142 inbuilt class escapes such as \d and \s have NOT been changed.
3143
3144 16. Updated pcredemo.c to include calls to free() to release the memory for the
3145 compiled pattern.
3146
3147 17. The generated file chartables.c was being created in the source directory
3148 instead of in the building directory. This caused the build to fail if the
3149 source directory was different from the building directory, and was
3150 read-only.
3151
3152 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3153 file. No doubt somebody will tell me if they don't make sense... Also added
3154 Dan Mooney's comments about building on OpenVMS.
3155
3156 19. Added support for partial matching via the PCRE_PARTIAL option for
3157 pcre_exec() and the \P data escape in pcretest.
3158
3159 20. Extended pcretest with 3 new pattern features:
3160
3161 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3162 write the compiled pattern to the file whose name is "rest-of-line".
3163 This is a straight binary dump of the data, with the saved pointer to
3164 the character tables forced to be NULL. The study data, if any, is
3165 written too. After writing, pcretest reads a new pattern.
3166
3167 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3168 compiled pattern from the given file. There must not be any
3169 occurrences of "<" in the file name (pretty unlikely); if there are,
3170 pcretest will instead treat the initial "<" as a pattern delimiter.
3171 After reading in the pattern, pcretest goes on to read data lines as
3172 usual.
3173
3174 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3175 and 16-bit fields in a compiled pattern, to simulate a pattern that
3176 was compiled on a host of opposite endianness.
3177
3178 21. The pcre-exec() function can now cope with patterns that were compiled on
3179 hosts of opposite endianness, with this restriction:
3180
3181 As for any compiled expression that is saved and used later, the tables
3182 pointer field cannot be preserved; the extra_data field in the arguments
3183 to pcre_exec() should be used to pass in a tables address if a value
3184 other than the default internal tables were used at compile time.
3185
3186 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3187 now diagnosed as an error. Previously, most of the time, a negative number
3188 would have been treated as zero, but if in addition "ovector" was passed as
3189 NULL, a crash could occur.
3190
3191 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3192 new versions from the libtool 1.5 distribution (the last one is a copy of
3193 a file called libtool.m4). This seems to have fixed the need to patch
3194 "configure" to support Darwin 1.3 (which I used to do). However, I still
3195 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3196 workstation).
3197
3198 24. Changed the PCRE licence to be the more standard "BSD" licence.
3199
3200
3201 Version 4.5 01-Dec-03
3202 ---------------------
3203
3204 1. There has been some re-arrangement of the code for the match() function so
3205 that it can be compiled in a version that does not call itself recursively.
3206 Instead, it keeps those local variables that need separate instances for
3207 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3208 needs to "recurse". Keeping track of where control must go is done by means
3209 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3210 hide most of the details from the main code, and operates only if
3211 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3212 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3213 operating.
3214
3215 To make it easier for callers to provide specially tailored get/free
3216 functions for this usage, two new functions, pcre_stack_malloc, and
3217 pcre_stack_free, are used. They are always called in strict stacking order,
3218 and the size of block requested is always the same.
3219
3220 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3221 PCRE has been compiled to use the stack or the heap for recursion. The
3222 -C option of pcretest uses this to show which version is compiled.
3223
3224 A new data escape \S, is added to pcretest; it causes the amounts of store
3225 obtained and freed by both kinds of malloc/free at match time to be added
3226 to the output.
3227
3228 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3229 what's available on my current Linux desktop machine.
3230
3231 3. When matching a UTF-8 string, the test for a valid string at the start has
3232 been extended. If start_offset is not zero, PCRE now checks that it points
3233 to a byte that is the start of a UTF-8 character. If not, it returns
3234 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3235 this is necessary because there may be backward assertions in the pattern.
3236 When matching the same subject several times, it may save resources to use
3237 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3238
3239 4. The code for checking the validity of UTF-8 strings has been tightened so
3240 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3241 containing "overlong sequences".
3242
3243 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3244 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3245 should have been "&", but it just so happened that all the cases this let
3246 through by mistake were picked up later in the function.
3247
3248 6. I had used a variable called "isblank" - this is a C99 function, causing
3249 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3250
3251 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3252 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3253 so the version is shown; (c) stop "make test" echoing "./RunTest".
3254
3255 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3256
3257 9. The prototype for memmove() for systems that don't have it was using
3258 size_t, but the inclusion of the header that defines size_t was later. I've
3259 moved the #includes for the C headers earlier to avoid this.
3260
3261 10. Added some adjustments to the code to make it easier to compiler on certain
3262 special systems:
3263
3264 (a) Some "const" qualifiers were missing.
3265 (b) Added the macro EXPORT before all exported functions; by default this
3266 is defined to be empty.
3267 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3268 that it reads its output file name as an argument instead of writing
3269 to the standard output and assuming this can be redirected.
3270
3271 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3272 class containing characters with values greater than 255, PCRE compilation
3273 went into a loop.
3274
3275 12. A recursive reference to a subpattern that was within another subpattern
3276 that had a minimum quantifier of zero caused PCRE to crash. For example,
3277 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3278 recursion. If the recursively-called subpattern itself had a zero repeat,
3279 that was OK.
3280
3281 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3282 buffer into which it was copied (for escape processing) was still set at
3283 1024, so long lines caused crashes.
3284
3285 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3286 "internal error: code overflow...". This applied to any character class
3287 that was followed by a possessive quantifier.
3288
3289 15. Modified the Makefile to add libpcre.la as a prerequisite for
3290 libpcreposix.la because I was told this is needed for a parallel build to
3291 work.
3292
3293 16. If a pattern that contained .* following optional items at the start was
3294 studied, the wrong optimizing data was generated, leading to matching
3295 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3296 matching string must start with a or b or c. The correct conclusion for
3297 this pattern is that a match can start with any character.
3298
3299
3300 Version 4.4 13-Aug-03
3301 ---------------------
3302
3303 1. In UTF-8 mode, a character class containing characters with values between
3304 127 and 255 was not handled correctly if the compiled pattern was studied.
3305 In fixing this, I have also improved the studying algorithm for such
3306 classes (slightly).
3307
3308 2. Three internal functions had redundant arguments passed to them. Removal
3309 might give a very teeny performance improvement.
3310
3311 3. Documentation bug: the value of the capture_top field in a callout is *one
3312 more than* the number of the hightest numbered captured substring.
3313
3314 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3315 in incorrectly linking with a previously installed version. They now link
3316 explicitly with libpcre.la.
3317
3318 5. configure.in no longer needs to recognize Cygwin specially.
3319
3320 6. A problem in pcre.in for Windows platforms is fixed.
3321
3322 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3323 pcretest, it used to include the size of the study block as part of its
3324 output. Unfortunately, the structure contains a field that has a different
3325 size on different hardware architectures. This meant that the tests that
3326 showed this size failed. As the block is currently always of a fixed size,
3327 this information isn't actually particularly useful in pcretest output, so
3328 I have just removed it.
3329
3330 8. Three pre-processor statements accidentally did not start in column 1.
3331 Sadly, there are *still* compilers around that complain, even though
3332 standard C has not required this for well over a decade. Sigh.
3333
3334 9. In pcretest, the code for checking callouts passed small integers in the
3335 callout_data field, which is a void * field. However, some picky compilers
3336 complained about the casts involved for this on 64-bit systems. Now
3337 pcretest passes the address of the small integer instead, which should get
3338 rid of the warnings.
3339
3340 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3341 both compile and run time, and gives an error if an invalid UTF-8 sequence
3342 is found. There is a option for disabling this check in cases where the
3343 string is known to be correct and/or the maximum performance is wanted.
3344
3345 11. In response to a bug report, I changed one line in Makefile.in from
3346
3347 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3348 to
3349 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3350
3351 to look similar to other lines, but I have no way of telling whether this
3352 is the right thing to do, as I do not use Windows. No doubt I'll get told
3353 if it's wrong...
3354
3355
3356 Version 4.3 21-May-03
3357 ---------------------
3358
3359 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3360 Makefile.
3361
3362 2. Some refactoring to improve the quality of the code:
3363
3364 (i) The utf8_table... variables are now declared "const".
3365
3366 (ii) The code for \cx, which used the "case flipping" table to upper case
3367 lower case letters, now just substracts 32. This is ASCII-specific,
3368 but the whole concept of \cx is ASCII-specific, so it seems
3369 reasonable.
3370
3371 (iii) PCRE was using its character types table to recognize decimal and
3372 hexadecimal digits in the pattern. This is silly, because it handles
3373 only 0-9, a-f, and A-F, but the character types table is locale-
3374 specific, which means strange things might happen. A private
3375 table is now used for this - though it costs 256 bytes, a table is
3376 much faster than multiple explicit tests. Of course, the standard
3377 character types table is still used for matching digits in subject
3378 strings against \d.
3379
3380 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3381 ending in _t are). So I've renamed it as ESC_tee.
3382
3383 3. The first argument for regexec() in the POSIX wrapper should have been
3384 defined as "const".
3385
3386 4. Changed pcretest to use malloc() for its buffers so that they can be
3387 Electric Fenced for debugging.
3388
3389 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3390 to read one or more bytes before the start of the subject string. Often this
3391 had no effect on PCRE's behaviour, but in some circumstances it could
3392 provoke a segmentation fault.
3393
3394 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3395 to try to read one or more bytes before the start of the subject string.
3396
3397 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3398 UTF-8 support could misbehave in various ways if the subject string
3399 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3400 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3401 back over UTF-8 characters.)
3402
3403
3404 Version 4.2 14-Apr-03
3405 ---------------------
3406
3407 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3408
3409 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3410 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3411 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3412 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3413 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3414 and BUILD_EXEEXT
3415 Note: automatic setting of the BUILD variables is not yet working
3416 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3417 compile-time but not at link-time
3418 [LINK]: use for linking executables only
3419 make different versions for Windows and non-Windows
3420 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3421 libraries
3422 [LINK_FOR_BUILD]: new variable
3423 [OBJEXT]: use throughout
3424 [EXEEXT]: use throughout
3425 <winshared>: new target
3426 <wininstall>: new target
3427 <dftables.o>: use native compiler
3428 <dftables>: use native linker
3429 <install>: handle Windows platform correctly
3430 <clean>: ditto
3431 <check>: ditto
3432 copy DLL to top builddir before testing
3433
3434 As part of these changes, -no-undefined was removed again. This was reported
3435 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3436 in any case.
3437
3438 3. Some tidies to get rid of compiler warnings:
3439
3440 . In the match_data structure, match_limit was an unsigned long int, whereas
3441 match_call_count was an int. I've made them both unsigned long ints.
3442
3443 . In pcretest the fact that a const uschar * doesn't automatically cast to
3444 a void * provoked a warning.
3445
3446 . Turning on some more compiler warnings threw up some "shadow" variables
3447 and a few more missing casts.
3448
3449 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3450 option, a class that contained a single character with a value between 128
3451 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3452
3453 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3454 option, a class that contained several characters, but with at least one
3455 whose value was between 128 and 255 caused PCRE to crash.
3456
3457
3458 Version 4.1 12-Mar-03
3459 ---------------------
3460
3461 1. Compiling with gcc -pedantic found a couple of places where casts were
3462 needed, and a string in dftables.c that was longer than standard compilers are
3463 required to support.
3464
3465 2. Compiling with Sun's compiler found a few more places where the code could
3466 be tidied up in order to avoid warnings.
3467
3468 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3469 first of these names is deprecated in the latest Autoconf in favour of the name
3470 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3471 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3472 analogy I have changed it to CFLAGS_FOR_BUILD.
3473
3474 4. Added -no-undefined to the linking command in the Makefile, because this is
3475 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3476 linking step for the pcreposix library.
3477
3478 5. PCRE was failing to diagnose the case of two named groups with the same
3479 name.
3480
3481 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3482 literal character that is needed in the subject for a match, and scans along to
3483 ensure that it is present before embarking on the full matching process. This
3484 saves time in cases of nested unlimited repeats that are never going to match.
3485 Problem: the scan can take a lot of time if the subject is very long (e.g.
3486 megabytes), thus penalizing straightforward matches. It is now done only if the
3487 amount of subject to be scanned is less than 1000 bytes.
3488
3489 7. A lesser problem with the same optimization is that it was recording the
3490 first character of an anchored pattern as "needed", thus provoking a search
3491 right along the subject, even when the first match of the pattern was going to
3492 fail. The "needed" character is now not set for anchored patterns, unless it
3493 follows something in the pattern that is of non-fixed length. Thus, it still
3494 fulfils its original purpose of finding quick non-matches in cases of nested
3495 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3496
3497
3498 Version 4.0 17-Feb-03
3499 ---------------------
3500
3501 1. If a comment in an extended regex that started immediately after a meta-item
3502 extended to the end of string, PCRE compiled incorrect data. This could lead to
3503 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3504
3505 2. Moved to autoconf 2.53 and libtool 1.4.2.
3506
3507 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3508 the special perltest8 script is no longer needed - all the tests can be run
3509 from a single perltest script.
3510
3511 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3512 by \s. It has now been removed in PCRE. This means it isn't recognized as
3513 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3514 class [:space:] *does* include VT, thereby creating a mess.
3515
3516 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3517 space and tab.
3518
3519 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3520 its new features into the main test script, reducing the number of scripts.
3521
3522 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3523 were backward compatible, and made the (?i) apply to the whole pattern, as if
3524 /i were given. Now it behaves more logically, and applies the option setting
3525 only to what follows. PCRE has been changed to follow suit. However, if it
3526 finds options settings right at the start of the pattern, it extracts them into
3527 the global options, as before. Thus, they show up in the info data.
3528
3529 8. Added support for the \Q...\E escape sequence. Characters in between are
3530 treated as literals. This is slightly different from Perl in that $ and @ are
3531 also handled as literals inside the quotes. In Perl, they will cause variable
3532 interpolation. Note the following examples:
3533
3534 Pattern PCRE matches Perl matches
3535
3536 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3537 \Qabc\$xyz\E abc\$xyz abc\$xyz
3538 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3539
3540 For compatibility with Perl, \Q...\E sequences are recognized inside character
3541 classes as well as outside them.
3542
3543 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3544 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3545 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3546 signed/unsigned warnings.
3547
3548 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3549 option for pcretest, so I've replaced it by a simple function that does just
3550 that job.
3551
3552 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3553 "pcregrep -".
3554
3555 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3556 Java package. This provides some syntactic sugar for simple cases of what my
3557 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3558 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3559 item, you can use this simplified notation. Note that only makes sense with
3560 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3561 greediness, whatever the setting of the PCRE_UNGREEDY option.
3562
3563 13. A change of greediness default within a pattern was not taking effect at
3564 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3565 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3566 was abstracted outside.
3567
3568 14. PCRE now supports the \G assertion. It is true when the current matching
3569 position is at the start point of the match. This differs from \A when the
3570 starting offset is non-zero. Used with the /g option of pcretest (or similar
3571 code), it works in the same way as it does for Perl's /g option. If all
3572 alternatives of a regex begin with \G, the expression is anchored to the start
3573 match position, and the "anchored" flag is set in the compiled expression.
3574
3575 15. Some bugs concerning the handling of certain option changes within patterns
3576 have been fixed. These applied to options other than (?ims). For example,
3577 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3578 been the other way round. Some of this was related to change 7 above.
3579
3580 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3581 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3582 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3583 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3584
3585 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3586 mode. Unlike ".", it always matches newline, whatever the setting of
3587 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3588 assertions. Perl allows it, but it doesn't (in general) work because it can't
3589 calculate the length of the lookbehind. At least, that's the case for Perl
3590 5.8.0 - I've been told they are going to document that it doesn't work in
3591 future.
3592
3593 18. Added an error diagnosis for escapes that PCRE does not support: these are
3594 \L, \l, \N, \P, \p, \U, \u, and \X.
3595
3596 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3597 reading past the end of the pattern in cases such as /[abcd/.
3598
3599 20. PCRE was getting more memory than necessary for patterns with classes that
3600 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3601
3602 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3603 compiling PCRE for use with Virtual Pascal.
3604
3605 22. Small fix to the Makefile to make it work properly if the build is done
3606 outside the source tree.
3607
3608 23. Added a new extension: a condition to go with recursion. If a conditional
3609 subpattern starts with (?(R) the "true" branch is used if recursion has
3610 happened, whereas the "false" branch is used only at the top level.
3611
3612 24. When there was a very long string of literal characters (over 255 bytes
3613 without UTF support, over 250 bytes with UTF support), the computation of how
3614 much memory was required could be incorrect, leading to segfaults or other
3615 strange effects.
3616
3617 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3618 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3619 there was a subsequent back reference to those brackets. This meant that, for
3620 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3621 possible to check for precisely this case. All we can do is abandon the
3622 optimization if .* occurs inside capturing brackets when there are any back
3623 references whatsoever. (See below for a better fix that came later.)
3624
3625 26. The handling of the optimization for finding the first character of a
3626 non-anchored pattern, and for finding a character that is required later in the
3627 match were failing in some cases. This didn't break the matching; it just
3628 failed to optimize when it could. The way this is done has been re-implemented.
3629
3630 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3631
3632 28. Added a new feature that provides some of the functionality that Perl
3633 provides with (?{...}). The facility is termed a "callout". The way it is done
3634 in PCRE is for the caller to provide an optional function, by setting
3635 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3636 global variable. By default it is unset, which disables all calling out. To get
3637 the function called, the regex must include (?C) at appropriate points. This
3638 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3639 This provides a means of identifying different callout points. When PCRE
3640 reaches such a point in the regex, if pcre_callout has been set, the external
3641 function is called. It is provided with data in a structure called
3642 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3643 matching continues; if it returns a non-zero value, the match at the current
3644 point fails. However, backtracking will occur if possible. [This was changed
3645 later and other features added - see item 49 below.]
3646
3647 29. pcretest is upgraded to test the callout functionality. It provides a
3648 callout function that displays information. By default, it shows the start of
3649 the match and the current position in the text. There are some new data escapes
3650 to vary what happens:
3651
3652 \C+ in addition, show current contents of captured substrings
3653 \C- do not supply a callout function
3654 \C!n return 1 when callout number n is reached
3655 \C!n!m return 1 when callout number n is reached for the mth time
3656
3657 30. If pcregrep was called with the -l option and just a single file name, it
3658 output "<stdin>" if a match was found, instead of the file name.
3659
3660 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3661 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3662 pcre_exec(). This saves a malloc/free per call. The default value of
3663 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3664 when configuring.
3665
3666 32. The default maximum size of a compiled pattern is 64K. There have been a
3667 few cases of people hitting this limit. The code now uses macros to handle the
3668 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3669 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3670 configuring. Tests 2 and 5 work only with 2-byte links because they output
3671 debugging information about compiled patterns.
3672
3673 33. Internal code re-arrangements:
3674
3675 (a) Moved the debugging function for printing out a compiled regex into
3676 its own source file (printint.c) and used #include to pull it into
3677 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3678 separate copies.
3679
3680 (b) Defined the list of op-code names for debugging as a macro in
3681 internal.h so that it is next to the definition of the opcodes.
3682
3683 (c) Defined a table of op-code lengths for simpler skipping along compiled
3684 code. This is again a macro in internal.h so that it is next to the
3685 definition of the opcodes.
3686
3687 34. Added support for recursive calls to individual subpatterns, along the
3688 lines of Robin Houston's patch (but implemented somewhat differently).
3689
3690 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3691 allow it to read and process whole directories in Win32. This code was
3692 contributed by Lionel Fourquaux; it has not been tested by me.
3693
3694 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3695 used to name a group. Names consist of alphanumerics and underscores, and must
3696 be unique. Back references use the syntax (?P=name) and recursive calls use
3697 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3698 numbers. The function pcre_fullinfo() can be used after compilation to extract
3699 a name/number map. There are three relevant calls:
3700
3701 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3702 PCRE_INFO_NAMECOUNT yields the number of entries
3703 PCRE_INFO_NAMETABLE yields a pointer to the map.
3704
3705 The map is a vector of fixed-size entries. The size of each entry depends on
3706 the length of the longest name used. The first two bytes of each entry are the
3707 group number, most significant byte first. There follows the corresponding
3708 name, zero terminated. The names are in alphabetical order.
3709
3710 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3711 case instead of 255. Making it the same both with and without UTF-8 support
3712 means that the same test output works with both.
3713
3714 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3715 calling malloc() with a zero argument.
3716
3717 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3718 optimization. I've improved things by keeping a bitmap of backreferences with
3719 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3720 fact referenced, the optimization can be applied. It is unlikely that a
3721 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3722 the match to follow \n) will appear inside brackets with a number greater than
3723 31, but if it does, any back reference > 31 suppresses the optimization.
3724
3725 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3726 of disabling numbered capturing parentheses. Any opening parenthesis that is
3727 not followed by ? behaves as if it were followed by ?: but named parentheses
3728 can still be used for capturing (and they will acquire numbers in the usual
3729 way).
3730
3731 41. Redesigned the return codes from the match() function into yes/no/error so
3732 that errors can be passed back from deep inside the nested calls. A malloc
3733 failure while inside a recursive subpattern call now causes the
3734 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3735
3736 42. It is now possible to set a limit on the number of times the match()
3737 function is called in a call to pcre_exec(). This facility makes it possible to
3738 limit the amount of recursion and backtracking, though not in a directly
3739 obvious way, because the match() function is used in a number of different
3740 circumstances. The count starts from zero for each position in the subject
3741 string (for non-anchored patterns). The default limit is, for compatibility, a
3742 large number, namely 10 000 000. You can change this in two ways:
3743
3744 (a) When configuring PCRE before making, you can use --with-match-limit=n
3745 to set a default value for the compiled library.
3746
3747 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3748 a different value is set. See 45 below.
3749
3750 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3751
3752 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3753 of things that can be changed at compile time. The first argument specifies
3754 what is wanted and the second points to where the information is to be placed.
3755 The current list of available information is:
3756
3757 PCRE_CONFIG_UTF8
3758
3759 The output is an integer that is set to one if UTF-8 support is available;
3760 otherwise it is set to zero.
3761
3762 PCRE_CONFIG_NEWLINE
3763
3764 The output is an integer that it set to the value of the code that is used for
3765 newline. It is either LF (10) or CR (13).
3766
3767 PCRE_CONFIG_LINK_SIZE
3768
3769 The output is an integer that contains the number of bytes used for internal
3770 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3771
3772 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3773
3774 The output is an integer that contains the threshold above which the POSIX
3775 interface uses malloc() for output vectors. See item 31 above.
3776
3777 PCRE_CONFIG_MATCH_LIMIT
3778
3779 The output is an unsigned integer that contains the default limit of the number
3780 of match() calls in a pcre_exec() execution. See 42 above.
3781
3782 44. pcretest has been upgraded by the addition of the -C option. This causes it
3783 to extract all the available output from the new pcre_config() function, and to
3784 output it. The program then exits immediately.
3785
3786 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3787 order to support additional features. One way would have been to define
3788 pcre_exec2() (for example) with extra arguments, but this would not have been
3789 extensible, and would also have required all calls to the original function to
3790 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3791 is used for passing in "extra" data from pcre_study().
3792
3793 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3794 contains the following fields:
3795
3796 flags a bitmap indicating which of the following fields are set
3797 study_data opaque data from pcre_study()
3798 match_limit a way of specifying a limit on match() calls for a specific
3799 call to pcre_exec()
3800 callout_data data for callouts (see 49 below)
3801
3802 The flag bits are also defined in pcre.h, and are
3803
3804 PCRE_EXTRA_STUDY_DATA
3805 PCRE_EXTRA_MATCH_LIMIT
3806 PCRE_EXTRA_CALLOUT_DATA
3807
3808 The pcre_study() function now returns one of these new pcre_extra blocks, with
3809 the actual study data pointed to by the study_data field, and the
3810 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3811 before. That is, this change is entirely upwards-compatible and requires no
3812 change to existing code.
3813
3814 If you want to pass in additional data to pcre_exec(), you can either place it
3815 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3816 block.
3817
3818 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3819 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3820 times with different match limits, until it finds the minimum value needed for
3821 pcre_exec() to complete. The value is then output. This can be instructive; for
3822 most simple matches the number is quite small, but for pathological cases it
3823 gets very large very quickly.
3824
3825 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3826 returns the size of the data block pointed to by the study_data field in a
3827 pcre_extra block, that is, the value that was passed as the argument to
3828 pcre_malloc() when PCRE was getting memory in which to place the information
3829 created by pcre_study(). The fourth argument should point to a size_t variable.
3830 pcretest has been extended so that this information is shown after a successful
3831 pcre_study() call when information about the compiled regex is being displayed.
3832
3833 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3834 because what follows is always an absolute path. (Later: it turns out that this
3835 is more than cosmetic for MinGW, because it doesn't like empty path
3836 components.)
3837
3838 49. Some changes have been made to the callout feature (see 28 above):
3839
3840 (i) A callout function now has three choices for what it returns:
3841
3842 0 => success, carry on matching
3843 > 0 => failure at this point, but backtrack if possible
3844 < 0 => serious error, return this value from pcre_exec()
3845
3846 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3847 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3848 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3849 use by callout functions. It will never be used by PCRE itself.
3850
3851 (ii) The pcre_extra structure (see 45 above) has a void * field called
3852 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3853 pcre_callout_block structure has a field of the same name. The contents of
3854 the field passed in the pcre_extra structure are passed to the callout
3855 function in the corresponding field in the callout block. This makes it
3856 easier to use the same callout-containing regex from multiple threads. For
3857 testing, the pcretest program has a new data escape
3858
3859 \C*n pass the number n (may be negative) as callout_data
3860
3861 If the callout function in pcretest receives a non-zero value as
3862 callout_data, it returns that value.
3863
3864 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3865 there were some redundant $(CFLAGS) in commands that are now specified as
3866 $(LINK), which already includes $(CFLAGS).
3867
3868 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3869 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3870 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3871 one-byte characters throughout. Note that case-insensitive matching applies
3872 only to characters whose values are less than 256. PCRE doesn't support the
3873 notion of cases for higher-valued characters.
3874
3875 (i) A character class whose characters are all within 0-255 is handled as
3876 a bit map, and the map is inverted for negative classes. Previously, a
3877 character > 255 always failed to match such a class; however it should
3878 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3879
3880 (ii) A negated character class with a single character < 255 is coded as
3881 "not this character" (OP_NOT). This wasn't working properly when the test
3882 character was multibyte, either singly or repeated.
3883
3884 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3885 mode, for example: \x{100}{2,3}.
3886
3887 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3888 singly or repeated) now correctly test multibyte characters. However,
3889 PCRE doesn't recognize any characters with values greater than 255 as
3890 digits, spaces, or word characters. Such characters always match \D, \S,
3891 and \W, and never match \d, \s, or \w.
3892
3893 (v) Classes may now contain characters and character ranges with values
3894 greater than 255. For example: [ab\x{100}-\x{400}].
3895
3896 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3897 PCRE in UTF-8 mode.
3898
3899 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3900 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3901 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3902 value.)
3903
3904 53. The single man page has become too large. I have therefore split it up into
3905 a number of separate man pages. These also give rise to individual HTML pages;
3906 these are now put in a separate directory, and there is an index.html page that
3907 lists them all. Some hyperlinking between the pages has been installed.
3908
3909 54. Added convenience functions for handling named capturing parentheses.
3910
3911 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3912 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3913 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3914 are faulted.
3915
3916 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3917 calling configure. These values are used when compiling the dftables.c program
3918 which is run to generate the source of the default character tables. They
3919 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3920 you will need to set these values.
3921
3922 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3923
3924
3925 Version 3.9 02-Jan-02
3926 ---------------------
3927
3928 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3929
3930 2. If --disable-static was given, the building process failed when trying to
3931 build pcretest and pcregrep. (For some reason it was using libtool to compile
3932 them, which is not right, as they aren't part of the library.)
3933
3934
3935 Version 3.8 18-Dec-01
3936 ---------------------
3937
3938 1. The experimental UTF-8 code was completely screwed up. It was packing the
3939 bytes in the wrong order. How dumb can you get?
3940
3941
3942 Version 3.7 29-Oct-01
3943 ---------------------
3944
3945 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3946 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3947 this didn't happen under Solaris 8, where I normally test things.
3948
3949 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3950 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3951 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3952 this fix an hour or so after the initial 3.7 release.)
3953
3954
3955 Version 3.6 23-Oct-01
3956 ---------------------
3957
3958 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3959 offsets passed as NULL with zero offset count.
3960
3961 2. The config.guess and config.sub files had not been updated when I moved to
3962 the latest autoconf.
3963
3964
3965 Version 3.5 15-Aug-01
3966 ---------------------
3967
3968 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3969 had been forgotten.
3970
3971 2. By using declared but undefined structures, we can avoid using "void"
3972 definitions in pcre.h while keeping the internal definitions of the structures
3973 private.
3974
3975 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3976 user point of view, this means that both static and shared libraries are built
3977 by default, but this can be individually controlled. More of the work of
3978 handling this static/shared cases is now inside libtool instead of PCRE's make
3979 file.
3980
3981 4. The pcretest utility is now installed along with pcregrep because it is
3982 useful for users (to test regexs) and by doing this, it automatically gets
3983 relinked by libtool. The documentation has been turned into a man page, so
3984 there are now .1, .txt, and .html versions in /doc.
3985
3986 5. Upgrades to pcregrep:
3987 (i) Added long-form option names like gnu grep.
3988 (ii) Added --help to list all options with an explanatory phrase.
3989 (iii) Added -r, --recursive to recurse into sub-directories.
3990 (iv) Added -f, --file to read patterns from a file.
3991
3992 6. pcre_exec() was referring to its "code" argument before testing that
3993 argument for NULL (and giving an error if it was NULL).
3994
3995 7. Upgraded Makefile.in to allow for compiling in a different directory from
3996 the source directory.
3997
3998 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3999 options bits, the pointer it was passed was to an int instead of to an unsigned
4000 long int. This mattered only on 64-bit systems.
4001
4002 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
4003 generated) instead of pcre.in, which it its source. Also made the same change
4004 in several of the .c files.
4005
4006 10. A new release of gcc defines printf() as a macro, which broke pcretest
4007 because it had an ifdef in the middle of a string argument for printf(). Fixed
4008 by using separate calls to printf().
4009
4010 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
4011 script, to force use of CR or LF instead of \n in the source. On non-Unix
4012 systems, the value can be set in config.h.
4013
4014 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
4015 absolute limit. Changed the text of the error message to make this clear, and
4016 likewise updated the man page.
4017
4018 13. The limit of 99 on the number of capturing subpatterns has been removed.
4019 The new limit is 65535, which I hope will not be a "real" limit.
4020
4021
4022 Version 3.4 22-Aug-00
4023 ---------------------
4024
4025 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4026
4027 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4028
4029
4030 Version 3.3 01-Aug-00
4031 ---------------------
4032
4033 1. If an octal character was given, but the value was greater than \377, it
4034 was not getting masked to the least significant bits, as documented. This could
4035 lead to crashes in some systems.
4036
4037 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4038 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4039
4040 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4041 These just pass their arguments on to (pcre_free)(), but they are provided
4042 because some uses of PCRE bind it to non-C systems that can call its functions,
4043 but cannot call free() or pcre_free() directly.
4044
4045 4. Add "make test" as a synonym for "make check". Corrected some comments in
4046 the Makefile.
4047
4048 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4049 Makefile.
4050
4051 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4052 command called pgrep for grepping around the active processes.
4053
4054 7. Added the beginnings of support for UTF-8 character strings.
4055
4056 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4057 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4058 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4059 out for the ar command.)
4060
4061
4062 Version 3.2 12-May-00
4063 ---------------------
4064
4065 This is purely a bug fixing release.
4066
4067 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4068 of ZA. This was just one example of several cases that could provoke this bug,
4069 which was introduced by change 9 of version 2.00. The code for breaking
4070 infinite loops after an iteration that matches an empty string was't working
4071 correctly.
4072
4073 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4074 when matched against abbab (for example). After matching an empty string, it
4075 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4076 caused it to match further down the string than it should.
4077
4078 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4079 was there because it doesn't seem to be needed, and it causes trouble on some
4080 systems, as it is not a Standard C header. It has been removed.
4081
4082 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4083 were reported on the Macintosh. The changes were from
4084
4085 while ((c = *(++ptr)) != 0 && c != '\n');
4086 to
4087 while ((c = *(++ptr)) != 0 && c != '\n') ;
4088
4089 Totally extraordinary, but if that's what it takes...
4090
4091 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4092 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4093 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4094 assumes the way PCRE uses memmove() (always moving upwards).
4095
4096 6. PCRE is being used in one environment where strchr() is not available. There
4097 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4098 faster code anyway.
4099
4100
4101 Version 3.1 09-Feb-00
4102 ---------------------
4103
4104 The only change in this release is the fixing of some bugs in Makefile.in for
4105 the "install" target:
4106
4107 (1) It was failing to install pcreposix.h.
4108
4109 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4110
4111
4112 Version 3.0 01-Feb-00
4113 ---------------------
4114
4115 1. Add support for the /+ modifier to perltest (to output $` like it does in
4116 pcretest).
4117
4118 2. Add support for the /g modifier to perltest.
4119
4120 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4121 matches null strings.
4122
4123 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4124 pattern. Perl treats empty patterns specially - it reuses the most recent
4125 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4126 effect.
4127
4128 5. The POSIX interface was broken in that it was just handing over the POSIX
4129 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4130 required a bigger vector, with some working space on the end. This means that
4131 the POSIX wrapper now has to get and free some memory, and copy the results.
4132
4133 6. Added some simple autoconf support, placing the test data and the
4134 documentation in separate directories, re-organizing some of the
4135 information files, and making it build pcre-config (a GNU standard). Also added
4136 libtool support for building PCRE as a shared library, which is now the
4137 default.
4138
4139 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4140 09 are not valid octal constants. Single digits will be used for minor values
4141 less than 10.
4142
4143 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4144 existing programs that set these in the POSIX interface can use PCRE without
4145 modification.
4146
4147 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4148 return all that pcre_info() returns, plus additional data. The pcre_info()
4149 function is retained for compatibility, but is considered to be obsolete.
4150
4151 10. Added experimental recursion feature (?R) to handle one common case that
4152 Perl 5.6 will be able to do with (?p{...}).
4153
4154 11. Added support for POSIX character classes like [:alpha:], which Perl is
4155 adopting.
4156
4157
4158 Version 2.08 31-Aug-99
4159 ----------------------
4160
4161 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4162 trying to match at the startoffset position, but instead was moving forward to
4163 the next newline as if a previous match had failed.
4164
4165 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4166 and could get into a loop if a null string was matched other than at the start
4167 of the subject.
4168
4169 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4170 be distinguished at compile time, and for completeness also added PCRE_DATE.
4171
4172 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4173 in GnuWin32 environments.
4174
4175
4176 Version 2.07 29-Jul-99
4177 ----------------------
4178
4179 1. The documentation is now supplied in plain text form and HTML as well as in
4180 the form of man page sources.
4181
4182 2. C++ compilers don't like assigning (void *) values to other pointer types.
4183 In particular this affects malloc(). Although there is no problem in Standard
4184 C, I've put in casts to keep C++ compilers happy.
4185
4186 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4187 should be (const char *).
4188
4189 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4190 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4191 However, I haven't made this a standard facility. The documentation doesn't
4192 mention it, and the Makefile doesn't support it.
4193
4194 5. The Makefile now contains an "install" target, with editable destinations at
4195 the top of the file. The pcretest program is not installed.
4196
4197 6. pgrep -V now gives the PCRE version number and date.
4198
4199 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4200 causing the entire string to be ignored, instead of just the last character.
4201
4202 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4203 non-matching string, it can take a very, very long time, even for strings of
4204 quite modest length, because of the nested recursion. PCRE now does better in
4205 some of these cases. It does this by remembering the last required literal
4206 character in the pattern, and pre-searching the subject to ensure it is present
4207 before running the real match. In other words, it applies a heuristic to detect
4208 some types of certain failure quickly, and in the above example, if presented
4209 with a string that has no trailing " it gives "no match" very quickly.
4210
4211 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4212 other alternatives are tried instead.
4213
4214
4215 Version 2.06 09-Jun-99
4216 ----------------------
4217
4218 1. Change pcretest's output for amount of store used to show just the code
4219 space, because the remainder (the data block) varies in size between 32-bit and
4220 64-bit systems.
4221
4222 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4223 start matching at. This allows lookbehinds to work when searching for multiple
4224 occurrences in a string.
4225
4226 3. Added additional options to pcretest for testing multiple occurrences:
4227
4228 /+ outputs the rest of the string that follows a match
4229 /g loops for multiple occurrences, using the new startoffset argument
4230 /G loops for multiple occurrences by passing an incremented pointer
4231
4232 4. PCRE wasn't doing the "first character" optimization for patterns starting
4233 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4234 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4235 the letter 'x'. On long subject strings, this gives a significant speed-up.
4236
4237
4238 Version 2.05 21-Apr-99
4239 ----------------------
4240
4241 1. Changed the type of magic_number from int to long int so that it works
4242 properly on 16-bit systems.
4243
4244 2. Fixed a bug which caused patterns starting with .* not to work correctly
4245 when the subject string contained newline characters. PCRE was assuming
4246 anchoring for such patterns in all cases, which is not correct because .* will
4247 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4248 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4249 must be retried after every newline in the subject.
4250
4251
4252 Version 2.04 18-Feb-99
4253 ----------------------
4254
4255 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4256 computation of the store needed to hold the pattern was incorrect (too large).
4257 If such patterns were nested a few deep, this could multiply and become a real
4258 problem.
4259
4260 2. Added /M option to pcretest to show the memory requirement of a specific
4261 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4262
4263 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4264 compiled in such a way that the backtracking after subsequent failure was
4265 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4266 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4267
4268
4269 Version 2.03 02-Feb-99
4270 ----------------------
4271
4272 1. Fixed typo and small mistake in man page.
4273
4274 2. Added 4th condition (GPL supersedes if conflict) and created separate
4275 LICENCE file containing the conditions.
4276
4277 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4278 Perl, that is the internal \ allows the delimiter to be included in the
4279 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4280 the final delimiter, add \ to the end of the pattern (to test the error).
4281
4282 4. Added the convenience functions for extracting substrings after a successful
4283 match. Updated pcretest to make it able to test these functions.
4284
4285
4286 Version 2.02 14-Jan-99
4287 ----------------------
4288
4289 1. Initialized the working variables associated with each extraction so that
4290 their saving and restoring doesn't refer to uninitialized store.
4291
4292 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4293 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4294 fix the problem.
4295
4296 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4297 calls, and wasn't printing the correct value for compiling calls. Increased the
4298 default value of LOOPREPEAT, and the number of significant figures in the
4299 times.
4300
4301 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4302
4303 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4304 a building problem on Windows NT with a FAT file system.
4305
4306
4307 Version 2.01 21-Oct-98
4308 ----------------------
4309
4310 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4311 to character tables built by pcre_maketables() in the current locale. If NULL
4312 is passed, the default tables are used.
4313
4314
4315 Version 2.00 24-Sep-98
4316 ----------------------
4317
4318 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4319 it any more.
4320
4321 2. Allow quantification of (?>) groups, and make it work correctly.
4322
4323 3. The first character computation wasn't working for (?>) groups.
4324
4325 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4326 end of the subject) and add 5.005's \z, which really does match only at the
4327 very end of the subject.
4328
4329 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4330
4331 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4332 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4333 localized options. All options to pcre_study() were also removed.
4334
4335 7. Add other new features from 5.005:
4336
4337 $(?<= positive lookbehind
4338 $(?<! negative lookbehind
4339 (?imsx-imsx) added the unsetting capability
4340 such a setting is global if at outer level; local otherwise
4341 (?imsx-imsx:) non-capturing groups with option setting
4342 (?(cond)re|re) conditional pattern matching
4343
4344 A backreference to itself in a repeated group matches the previous
4345 captured string.
4346
4347 8. General tidying up of studying (both automatic and via "study")
4348 consequential on the addition of new assertions.
4349
4350 9. As in 5.005, unlimited repeated groups that could match an empty substring
4351 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4352 runtime if any iteration does actually match an empty substring.
4353
4354 10. Include the RunTest script in the distribution.
4355
4356 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4357 discrepancies, some of which were old and were also with respect to 5.004. They
4358 have now been fixed.
4359
4360
4361 Version 1.09 28-Apr-98
4362 ----------------------
4363
4364 1. A negated single character class followed by a quantifier with a minimum
4365 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4366 program crashes, or just wrong answers. This did not apply to negated classes
4367 containing more than one character, or to minima other than one.
4368
4369
4370 Version 1.08 27-Mar-98
4371 ----------------------
4372
4373 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4374
4375 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4376 latter must appear before anything that relies on it in the pattern.
4377
4378
4379 Version 1.07 16-Feb-98
4380 ----------------------
4381
4382 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4383 repeat of a potentially empty string).
4384
4385
4386 Version 1.06 23-Jan-98
4387 ----------------------
4388
4389 1. Added Markus Oberhumer's little patches for C++.
4390
4391 2. Literal strings longer than 255 characters were broken.
4392
4393
4394 Version 1.05 23-Dec-97
4395 ----------------------
4396
4397 1. Negated character classes containing more than one character were failing if
4398 PCRE_CASELESS was set at run time.
4399
4400
4401 Version 1.04 19-Dec-97
4402 ----------------------
4403
4404 1. Corrected the man page, where some "const" qualifiers had been omitted.
4405
4406 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4407 input syntax.
4408
4409 3. Fixed memory leak which occurred when a regex with back references was
4410 matched with an offsets vector that wasn't big enough. The temporary memory
4411 that is used in this case wasn't being freed if the match failed.
4412
4413 4. Tidied pcretest to ensure it frees memory that it gets.
4414
4415 5. Temporary memory was being obtained in the case where the passed offsets
4416 vector was exactly big enough.
4417
4418 6. Corrected definition of offsetof() from change 5 below.
4419
4420 7. I had screwed up change 6 below and broken the rules for the use of
4421 setjmp(). Now fixed.
4422
4423
4424 Version 1.03 18-Dec-97
4425 ----------------------
4426
4427 1. A erroneous regex with a missing opening parenthesis was correctly
4428 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4429 on some systems.
4430
4431 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4432 it was reported that one broken compiler failed on the former because "code" is
4433 also an independent variable.
4434
4435 3. The erroneous regex a[]b caused an array overrun reference.
4436
4437 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4438 fail on data ending with that character. (It was going on too far, and checking
4439 the next character, typically a binary zero.) This was specific to the
4440 optimized code for single-character negative classes.
4441
4442 5. Added a contributed patch from the TIN world which does the following:
4443
4444 + Add an undef for memmove, in case the the system defines a macro for it.
4445
4446 + Add a definition of offsetof(), in case there isn't one. (I don't know
4447 the reason behind this - offsetof() is part of the ANSI standard - but
4448 it does no harm).
4449
4450 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4451 most of the places where whitespace preceded '#'. I have given up and
4452 allowed the remaining 2 cases to be at the margin.
4453
4454 + Rename some variables in pcre to eliminate shadowing. This seems very
4455 pedantic, but does no harm, of course.
4456
4457 6. Moved the call to setjmp() into its own function, to get rid of warnings
4458 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4459
4460 7. Constructs such as \d{8,} were compiling into the equivalent of
4461 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4462 outcome, but in this particular case used more store than had been allocated,
4463 which caused the bug to be discovered because it threw up an internal error.
4464
4465 8. The debugging code in both pcre and pcretest for outputting the compiled
4466 form of a regex was going wrong in the case of back references followed by
4467 curly-bracketed repeats.
4468
4469
4470 Version 1.02 12-Dec-97
4471 ----------------------
4472
4473 1. Typos in pcre.3 and comments in the source fixed.
4474
4475 2. Applied a contributed patch to get rid of places where it used to remove
4476 'const' from variables, and fixed some signed/unsigned and uninitialized
4477 variable warnings.
4478
4479 3. Added the "runtest" target to Makefile.
4480
4481 4. Set default compiler flag to -O2 rather than just -O.
4482
4483
4484 Version 1.01 19-Nov-97
4485 ----------------------
4486
4487 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4488 like /([ab]*)*/, that is, for classes with more than one character in them.
4489
4490 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4491 as /((?>a*))*/ (a PCRE_EXTRA facility).
4492
4493
4494 Version 1.00 18-Nov-97
4495 ----------------------
4496
4497 1. Added compile-time macros to support systems such as SunOS4 which don't have
4498 memmove() or strerror() but have other things that can be used instead.
4499
4500 2. Arranged that "make clean" removes the executables.
4501
4502
4503 Version 0.99 27-Oct-97
4504 ----------------------
4505
4506 1. Fixed bug in code for optimizing classes with only one character. It was
4507 initializing a 32-byte map regardless, which could cause it to run off the end
4508 of the memory it had got.
4509
4510 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4511
4512
4513 Version 0.98 22-Oct-97
4514 ----------------------
4515
4516 1. Fixed bug in code for handling temporary memory usage when there are more
4517 back references than supplied space in the ovector. This could cause segfaults.
4518
4519
4520 Version 0.97 21-Oct-97
4521 ----------------------
4522
4523 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4524
4525 2. Optimized negated single characters not to use a bit map.
4526
4527 3. Brought error texts together as macro definitions; clarified some of them;
4528 fixed one that was wrong - it said "range out of order" when it meant "invalid
4529 escape sequence".
4530
4531 4. Changed some char * arguments to const char *.
4532
4533 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4534
4535 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4536 pcretest.
4537
4538
4539 Version 0.96 16-Oct-97
4540 ----------------------
4541
4542 1. Added a simple "pgrep" utility to the distribution.
4543
4544 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4545 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4546 where "ddd" means "one or more decimal digits".
4547
4548 3. Fixed serious bug. If a pattern had a back reference, but the call to
4549 pcre_exec() didn't supply a large enough ovector to record the related
4550 identifying subpattern, the match always failed. PCRE now remembers the number
4551 of the largest back reference, and gets some temporary memory in which to save
4552 the offsets during matching if necessary, in order to ensure that
4553 backreferences always work.
4554
4555 4. Increased the compatibility with Perl in a number of ways:
4556
4557 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4558 to request this handling. The option can be set at compile or exec time.
4559
4560 (b) $ matches before a terminating newline by default; an option
4561 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4562 mode). The option can be set at compile or exec time.
4563
4564 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4565 the same as Perl's. If the decimal number it represents is less than 10
4566 or there aren't that many previous left capturing parentheses, an octal
4567 escape is read. Inside a character class, it's always an octal escape,
4568 even if it is a single digit.
4569
4570 (d) An escaped but undefined alphabetic character is taken as a literal,
4571 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4572 escapes.
4573
4574 (e) {0} is now permitted. (The previous item is removed from the compiled
4575 pattern).
4576
4577 5. Changed all the names of code files so that the basic parts are no longer
4578 than 10 characters, and abolished the teeny "globals.c" file.
4579
4580 6. Changed the handling of character classes; they are now done with a 32-byte
4581 bit map always.
4582
4583 7. Added the -d and /D options to pcretest to make it possible to look at the
4584 internals of compilation without having to recompile pcre.
4585
4586
4587 Version 0.95 23-Sep-97
4588 ----------------------
4589
4590 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4591 \x20 at the start of a run of normal characters. These were being treated as
4592 real characters, instead of the source characters being re-checked.
4593
4594
4595 Version 0.94 18-Sep-97
4596 ----------------------
4597
4598 1. The functions are now thread-safe, with the caveat that the global variables
4599 containing pointers to malloc() and free() or alternative functions are the
4600 same for all threads.
4601
4602 2. Get pcre_study() to generate a bitmap of initial characters for non-
4603 anchored patterns when this is possible, and use it if passed to pcre_exec().
4604
4605
4606 Version 0.93 15-Sep-97
4607 ----------------------
4608
4609 1. /(b)|(:+)/ was computing an incorrect first character.
4610
4611 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4612 but not actually doing anything yet.
4613
4614 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4615 as Perl does (e.g. [-az] or [az-]).
4616
4617 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4618 all possible positions.
4619
4620 5. Split up into different modules to avoid including unneeded functions in a
4621 compiled binary. However, compile and exec are still in one module. The "study"
4622 function is split off.
4623
4624 6. The character tables are now in a separate module whose source is generated
4625 by an auxiliary program - but can then be edited by hand if required. There are
4626 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4627 toupper() in the code.
4628
4629 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4630 make them global. Abolish the function for setting them, as the caller can now
4631 set them directly.
4632
4633
4634 Version 0.92 11-Sep-97
4635 ----------------------
4636
4637 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4638 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4639
4640 2. Caseless matching was not working in character classes if the characters in
4641 the pattern were in upper case.
4642
4643 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4644
4645 4. Make PCRE_ANCHORED public and accept as a compile option.
4646
4647 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4648 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4649 pass them.
4650
4651 6. Give an error if bad option bits passed at compile or run time.
4652
4653 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4654 pcretest to cause it to pass that flag.
4655
4656 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4657 options, and the first character, if set.
4658
4659 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4660
4661
4662 Version 0.91 10-Sep-97
4663 ----------------------
4664
4665 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4666 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4667
4668 2. PCRE was looping on encountering an indefinitely repeated back reference to
4669 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4670 Perl does - treats the match as successful.
4671
4672 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5