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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 881 - (show annotations)
Sun Jan 15 18:07:05 2012 UTC (3 years, 5 months ago) by ph10
File size: 224833 byte(s)
Error occurred while calculating annotation data.
Get rid of a number of -Wunused-but-set-variable compiler warnings.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.30
5 ------------
6
7 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
8 name is defined in ctype.h.
9
10 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
11 only in quite long subpatterns.
12
13 3. Removed the function pcre_info(), which has been obsolete and deprecated
14 since it was replaced by pcre_fullinfo() in February 2000.
15
16 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
17 match a (*MARK), and the match failed at the start of the subject, a
18 reference to memory before the start of the subject could occur. This bug
19 was introduced by fix 17 of release 8.21.
20
21 5. A reference to an unset group with zero minimum repetition was giving
22 totally wrong answers (in non-JavaScript-compatibility mode). For example,
23 /(another)?(\1?)test/ matched against "hello world test". This bug was
24 introduced in release 8.13.
25
26 6. Add support for 16-bit character strings (a large amount of work involving
27 many changes and refactorings).
28
29 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31 from a file.
32
33 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34 rounding is not applied in this particular case).
35
36 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
37 if they appear, or are escaped, in patterns.
38
39 10. Get rid of a number of -Wunused-but-set-variable warnings.
40
41
42 Version 8.21 12-Dec-2011
43 ------------------------
44
45 1. Updating the JIT compiler.
46
47 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
48 are added as well.
49
50 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
51 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
52 calling _pcre_jit_exec. Some extra comments are added.
53
54 4. (*MARK) settings inside atomic groups that do not contain any capturing
55 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
56 was introduced by change 18 for 8.20.
57
58 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
59 ECMA-262 standard.
60
61 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
62 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
63 This bug was probably introduced by change 9 of 8.13.
64
65 7. While fixing 6 above, I noticed that a number of other items were being
66 incorrectly rejected as "not fixed length". This arose partly because newer
67 opcodes had not been added to the fixed-length checking code. I have (a)
68 corrected the bug and added tests for these items, and (b) arranged for an
69 error to occur if an unknown opcode is encountered while checking for fixed
70 length instead of just assuming "not fixed length". The items that were
71 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
72 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
73 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
74
75 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
76 being incorrectly compiled and would have given unpredicatble results.
77
78 9. A possessively repeated subpattern with minimum repeat count greater than
79 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
80 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
81 the first (A) could occur when it should not.
82
83 10. Add a cast and remove a redundant test from the code.
84
85 11. JIT should use pcre_malloc/pcre_free for allocation.
86
87 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
88 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
89 is anything other than /usr, -L is still shown).
90
91 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
92
93 14. Perl does not support \N without a following name in a [] class; PCRE now
94 also gives an error.
95
96 15. If a forward reference was repeated with an upper limit of around 2000,
97 it caused the error "internal error: overran compiling workspace". The
98 maximum number of forward references (including repeats) was limited by the
99 internal workspace, and dependent on the LINK_SIZE. The code has been
100 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
101 the default depends on LINK_SIZE. There is a new upper limit (for safety)
102 of around 200,000 forward references. While doing this, I also speeded up
103 the filling in of repeated forward references.
104
105 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
106 incorrectly expecting the subject to contain another "a" after the start.
107
108 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
109 in the match, the SKIP should be ignored. This was not happening; instead
110 the SKIP was being treated as NOMATCH. For patterns such as
111 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
112 tested.
113
114 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
115 now much more compatible with Perl, in particular in cases where the result
116 is a non-match for a non-anchored pattern. For example, if
117 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
118 "m", where previously it did not return a name. A side effect of this
119 change is that for partial matches, the last encountered mark name is
120 returned, as for non matches. A number of tests that were previously not
121 Perl-compatible have been moved into the Perl-compatible test files. The
122 refactoring has had the pleasing side effect of removing one argument from
123 the match() function, thus reducing its stack requirements.
124
125 19. If the /S+ option was used in pcretest to study a pattern using JIT,
126 subsequent uses of /S (without +) incorrectly behaved like /S+.
127
128 21. Retrieve executable code size support for the JIT compiler and fixing
129 some warnings.
130
131 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
132 not work when the shorter character appeared right at the end of the
133 subject string.
134
135 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
136 systems.
137
138 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
139 output it when the /M option is used in pcretest.
140
141 25. The CheckMan script was not being included in the distribution. Also, added
142 an explicit "perl" to run Perl scripts from the PrepareRelease script
143 because this is reportedly needed in Windows.
144
145 26. If study data was being save in a file and studying had not found a set of
146 "starts with" bytes for the pattern, the data written to the file (though
147 never used) was taken from uninitialized memory and so caused valgrind to
148 complain.
149
150 27. Updated RunTest.bat as provided by Sheri Pierce.
151
152 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
153
154 29. Computation of memory usage for the table of capturing group names was
155 giving an unnecessarily large value.
156
157
158 Version 8.20 21-Oct-2011
159 ------------------------
160
161 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
162 a POSIX class. After further experiments with Perl, which convinced me that
163 Perl has bugs and confusions, a closing square bracket is no longer allowed
164 in a POSIX name. This bug also affected patterns with classes that started
165 with full stops.
166
167 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
168 captured substring, but while checking the failing first alternative,
169 substring 1 is temporarily captured. If the output vector supplied to
170 pcre_exec() was not big enough for this capture, the yield of the function
171 was still zero ("insufficient space for captured substrings"). This cannot
172 be totally fixed without adding another stack variable, which seems a lot
173 of expense for a edge case. However, I have improved the situation in cases
174 such as /(a)(b)x|abc/ matched against "abc", where the return code
175 indicates that fewer than the maximum number of slots in the ovector have
176 been set.
177
178 3. Related to (2) above: when there are more back references in a pattern than
179 slots in the output vector, pcre_exec() uses temporary memory during
180 matching, and copies in the captures as far as possible afterwards. It was
181 using the entire output vector, but this conflicts with the specification
182 that only 2/3 is used for passing back captured substrings. Now it uses
183 only the first 2/3, for compatibility. This is, of course, another edge
184 case.
185
186 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
187 main code base, and can be used by building with --enable-jit. When this is
188 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
189 runtime --no-jit option is given.
190
191 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
192 ovector, the return from the function was zero, implying that there were
193 other matches that did not fit. The correct "exactly full" value is now
194 returned.
195
196 6. If a subpattern that was called recursively or as a subroutine contained
197 (*PRUNE) or any other control that caused it to give a non-standard return,
198 invalid errors such as "Error -26 (nested recursion at the same subject
199 position)" or even infinite loops could occur.
200
201 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
202 computing the minimum length on reaching *ACCEPT, and so ended up with the
203 wrong value of 1 rather than 0. Further investigation indicates that
204 computing a minimum subject length in the presence of *ACCEPT is difficult
205 (think back references, subroutine calls), and so I have changed the code
206 so that no minimum is registered for a pattern that contains *ACCEPT.
207
208 8. If (*THEN) was present in the first (true) branch of a conditional group,
209 it was not handled as intended. [But see 16 below.]
210
211 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
212 Sheri Pierce.
213
214 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
215 the first byte in a match must be "a".
216
217 11. Change 17 for 8.13 increased the recursion depth for patterns like
218 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
219 pattern contains any instances of (*THEN). If it does not, the old
220 optimizations are restored. It would be nice to do this on a per-group
221 basis, but at the moment that is not feasible.
222
223 12. In some environments, the output of pcretest -C is CRLF terminated. This
224 broke RunTest's code that checks for the link size. A single white space
225 character after the value is now allowed for.
226
227 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
228 For "fr", it uses the Windows-specific input and output files.
229
230 14. If (*THEN) appeared in a group that was called recursively or as a
231 subroutine, it did not work as intended. [But see next item.]
232
233 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
234 pattern fragments (but not containing any | characters). If A and B are
235 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
236 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
237 D. In other words, Perl considers parentheses that do not contain any |
238 characters to be part of a surrounding alternative, whereas PCRE was
239 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
240 differently. PCRE now behaves in the same way as Perl, except in the case
241 of subroutine/recursion calls such as (?1) which have in any case always
242 been different (but PCRE had them first :-).
243
244 16. Related to 15 above: Perl does not treat the | in a conditional group as
245 creating alternatives. Such a group is treated in the same way as an
246 ordinary group without any | characters when processing (*THEN). PCRE has
247 been changed to match Perl's behaviour.
248
249 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
250 RunGrepTest script failed.
251
252 18. Change 22 for version 13 caused atomic groups to use more stack. This is
253 inevitable for groups that contain captures, but it can lead to a lot of
254 stack use in large patterns. The old behaviour has been restored for atomic
255 groups that do not contain any capturing parentheses.
256
257 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
258 suppress the check for a minimum subject length at run time. (If it was
259 given to pcre_exec() or pcre_dfa_exec() it did work.)
260
261 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
262 fail to work when decoding hex characters in data strings in EBCDIC
263 environments.
264
265 21. It appears that in at least one Mac OS environment, the isxdigit() function
266 is implemented as a macro that evaluates to its argument more than once,
267 contravening the C 90 Standard (I haven't checked a later standard). There
268 was an instance in pcretest which caused it to go wrong when processing
269 \x{...} escapes in subject strings. The has been rewritten to avoid using
270 things like p++ in the argument of isxdigit().
271
272
273 Version 8.13 16-Aug-2011
274 ------------------------
275
276 1. The Unicode data tables have been updated to Unicode 6.0.0.
277
278 2. Two minor typos in pcre_internal.h have been fixed.
279
280 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
281 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
282 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
283
284 4. There were a number of related bugs in the code for matching backrefences
285 caselessly in UTF-8 mode when codes for the characters concerned were
286 different numbers of bytes. For example, U+023A and U+2C65 are an upper
287 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
288 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
289 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
290 2-byte code at the end of the subject (it thought there wasn't enough data
291 left).
292
293 5. Comprehensive information about what went wrong is now returned by
294 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
295 as the output vector has at least 2 elements. The offset of the start of
296 the failing character and a reason code are placed in the vector.
297
298 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
299 now returned is for the first byte of the failing character, instead of the
300 last byte inspected. This is an incompatible change, but I hope it is small
301 enough not to be a problem. It makes the returned offset consistent with
302 pcre_exec() and pcre_dfa_exec().
303
304 7. pcretest now gives a text phrase as well as the error number when
305 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
306 failure, the offset and reason code are output.
307
308 8. When \R was used with a maximizing quantifier it failed to skip backwards
309 over a \r\n pair if the subsequent match failed. Instead, it just skipped
310 back over a single character (\n). This seems wrong (because it treated the
311 two characters as a single entity when going forwards), conflicts with the
312 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
313 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
314 has been changed.
315
316 9. Some internal refactoring has changed the processing so that the handling
317 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
318 time (the PCRE_DOTALL option was changed this way some time ago: version
319 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
320 which was always a bit of a fudge. It also means that there is one less
321 argument for the match() function, which reduces its stack requirements
322 slightly. This change also fixes an incompatibility with Perl: the pattern
323 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
324
325 10. More internal refactoring has drastically reduced the number of recursive
326 calls to match() for possessively repeated groups such as (abc)++ when
327 using pcre_exec().
328
329 11. While implementing 10, a number of bugs in the handling of groups were
330 discovered and fixed:
331
332 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
333 (a|)*(?1) gave a compile-time internal error.
334 ((a|)+)+ did not notice that the outer group could match an empty string.
335 (^a|^)+ was not marked as anchored.
336 (.*a|.*)+ was not marked as matching at start or after a newline.
337
338 12. Yet more internal refactoring has removed another argument from the match()
339 function. Special calls to this function are now indicated by setting a
340 value in a variable in the "match data" data block.
341
342 13. Be more explicit in pcre_study() instead of relying on "default" for
343 opcodes that mean there is no starting character; this means that when new
344 ones are added and accidentally left out of pcre_study(), testing should
345 pick them up.
346
347 14. The -s option of pcretest has been documented for ages as being an old
348 synonym of -m (show memory usage). I have changed it to mean "force study
349 for every regex", that is, assume /S for every regex. This is similar to -i
350 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
351 using it. It makes it easier to run collections of tests with and without
352 study enabled, and thereby test pcre_study() more easily. All the standard
353 tests are now run with and without -s (but some patterns can be marked as
354 "never study" - see 20 below).
355
356 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
357 restoration of the capturing data to the outer values was not happening
358 correctly.
359
360 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
361 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
362 pattern had matched an empty string, and so incorrectly returned a no
363 match.
364
365 17. There was optimizing code for the last branch of non-capturing parentheses,
366 and also for the obeyed branch of a conditional subexpression, which used
367 tail recursion to cut down on stack usage. Unfortunately, now that there is
368 the possibility of (*THEN) occurring in these branches, tail recursion is
369 no longer possible because the return has to be checked for (*THEN). These
370 two optimizations have therefore been removed. [But see 8.20/11 above.]
371
372 18. If a pattern containing \R was studied, it was assumed that \R always
373 matched two bytes, thus causing the minimum subject length to be
374 incorrectly computed because \R can also match just one byte.
375
376 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
377 was incorrectly computed.
378
379 20. If /S is present twice on a test pattern in pcretest input, it now
380 *disables* studying, thereby overriding the use of -s on the command line
381 (see 14 above). This is necessary for one or two tests to keep the output
382 identical in both cases.
383
384 21. When (*ACCEPT) was used in an assertion that matched an empty string and
385 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
386
387 22. When an atomic group that contained a capturing parenthesis was
388 successfully matched, but the branch in which it appeared failed, the
389 capturing was not being forgotten if a higher numbered group was later
390 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
391 group 1 to "a", when in fact it should be unset. This applied to multi-
392 branched capturing and non-capturing groups, repeated or not, and also to
393 positive assertions (capturing in negative assertions does not happen
394 in PCRE) and also to nested atomic groups.
395
396 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
397 subject after a captured substring, to make it easier to tell which of a
398 number of identical substrings has been captured.
399
400 24. The way atomic groups are processed by pcre_exec() has been changed so that
401 if they are repeated, backtracking one repetition now resets captured
402 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
403 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
404 "aaa". Previously, it would have been "a". As part of this code
405 refactoring, the way recursive calls are handled has also been changed.
406
407 25. If an assertion condition captured any substrings, they were not passed
408 back unless some other capturing happened later. For example, if
409 (?(?=(a))a) was matched against "a", no capturing was returned.
410
411 26. When studying a pattern that contained subroutine calls or assertions,
412 the code for finding the minimum length of a possible match was handling
413 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
414 group 1 called group 2 while simultaneously a separate group 2 called group
415 1). A stack overflow occurred in this case. I have fixed this by limiting
416 the recursion depth to 10.
417
418 27. Updated RunTest.bat in the distribution to the version supplied by Tom
419 Fortmann. This supports explicit test numbers on the command line, and has
420 argument validation and error reporting.
421
422 28. An instance of \X with an unlimited repeat could fail if at any point the
423 first character it looked at was a mark character.
424
425 29. Some minor code refactoring concerning Unicode properties and scripts
426 should reduce the stack requirement of match() slightly.
427
428 30. Added the '=' option to pcretest to check the setting of unused capturing
429 slots at the end of the pattern, which are documented as being -1, but are
430 not included in the return count.
431
432 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
433 compiled something random. Now it gives a compile-time error (as does
434 Perl).
435
436 32. A *MARK encountered during the processing of a positive assertion is now
437 recorded and passed back (compatible with Perl).
438
439 33. If --only-matching or --colour was set on a pcregrep call whose pattern
440 had alternative anchored branches, the search for a second match in a line
441 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
442 matched the line "0102" twice. The same bug affected patterns that started
443 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
444 twice.
445
446 34. Previously, PCRE did not allow quantification of assertions. However, Perl
447 does, and because of capturing effects, quantifying parenthesized
448 assertions may at times be useful. Quantifiers are now allowed for
449 parenthesized assertions.
450
451 35. A minor code tidy in pcre_compile() when checking options for \R usage.
452
453 36. \g was being checked for fancy things in a character class, when it should
454 just be a literal "g".
455
456 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
457 appearance of a nested POSIX class supersedes an apparent external class.
458 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
459 unescaped square brackets may also appear as part of class names. For
460 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
461 more like Perl. (But see 8.20/1 above.)
462
463 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
464 was because it thought it was \N{name}, which is not supported).
465
466 39. Add minix to OS list not supporting the -S option in pcretest.
467
468 40. PCRE tries to detect cases of infinite recursion at compile time, but it
469 cannot analyze patterns in sufficient detail to catch mutual recursions
470 such as ((?1))((?2)). There is now a runtime test that gives an error if a
471 subgroup is called recursively as a subpattern for a second time at the
472 same position in the subject string. In previous releases this might have
473 been caught by the recursion limit, or it might have run out of stack.
474
475 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
476 happen only once. PCRE was, however incorrectly giving a compile time error
477 "recursive call could loop indefinitely" because it cannot analyze the
478 pattern in sufficient detail. The compile time test no longer happens when
479 PCRE is compiling a conditional subpattern, but actual runaway loops are
480 now caught at runtime (see 40 above).
481
482 42. It seems that Perl allows any characters other than a closing parenthesis
483 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
484 has been changed to be the same.
485
486 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
487 as not to get warnings when autogen.sh is called. Also changed
488 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
489
490 44. To help people who use pcregrep to scan files containing exceedingly long
491 lines, the following changes have been made:
492
493 (a) The default value of the buffer size parameter has been increased from
494 8K to 20K. (The actual buffer used is three times this size.)
495
496 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
497 PCRE is built.
498
499 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
500 to be set at run time.
501
502 (d) Numerical values in pcregrep options can be followed by K or M, for
503 example --buffer-size=50K.
504
505 (e) If a line being scanned overflows pcregrep's buffer, an error is now
506 given and the return code is set to 2.
507
508 45. Add a pointer to the latest mark to the callout data block.
509
510 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
511 partial match of an empty string instead of no match. This was specific to
512 the use of ".".
513
514 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
515 complete match instead of a partial match. This bug was dependent on both
516 the PCRE_UTF8 and PCRE_DOTALL options being set.
517
518 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
519 starting byte set, because \b was not being ignored.
520
521
522 Version 8.12 15-Jan-2011
523 ------------------------
524
525 1. Fixed some typos in the markup of the man pages, and wrote a script that
526 checks for such things as part of the documentation building process.
527
528 2. On a big-endian 64-bit system, pcregrep did not correctly process the
529 --match-limit and --recursion-limit options (added for 8.11). In
530 particular, this made one of the standard tests fail. (The integer value
531 went into the wrong half of a long int.)
532
533 3. If the --colour option was given to pcregrep with -v (invert match), it
534 did strange things, either producing crazy output, or crashing. It should,
535 of course, ignore a request for colour when reporting lines that do not
536 match.
537
538 4. Another pcregrep bug caused similar problems if --colour was specified with
539 -M (multiline) and the pattern match finished with a line ending.
540
541 5. In pcregrep, when a pattern that ended with a literal newline sequence was
542 matched in multiline mode, the following line was shown as part of the
543 match. This seems wrong, so I have changed it.
544
545 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
546 the check for further matches in the same line (so they could be coloured)
547 to overrun the end of the current line. If another match was found, it was
548 incorrectly shown (and then shown again when found in the next line).
549
550 7. If pcregrep was compiled under Windows, there was a reference to the
551 function pcregrep_exit() before it was defined. I am assuming this was
552 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
553 reported by a user. I've moved the definition above the reference.
554
555
556 Version 8.11 10-Dec-2010
557 ------------------------
558
559 1. (*THEN) was not working properly if there were untried alternatives prior
560 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
561 backtracked to try for "b" instead of moving to the next alternative branch
562 at the same level (in this case, to look for "c"). The Perl documentation
563 is clear that when (*THEN) is backtracked onto, it goes to the "next
564 alternative in the innermost enclosing group".
565
566 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
567 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
568 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
569 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
570 (*THEN).
571
572 3. If \s appeared in a character class, it removed the VT character from
573 the class, even if it had been included by some previous item, for example
574 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
575 of \s, but is part of the POSIX "space" class.)
576
577 4. A partial match never returns an empty string (because you can always
578 match an empty string at the end of the subject); however the checking for
579 an empty string was starting at the "start of match" point. This has been
580 changed to the "earliest inspected character" point, because the returned
581 data for a partial match starts at this character. This means that, for
582 example, /(?<=abc)def/ gives a partial match for the subject "abc"
583 (previously it gave "no match").
584
585 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
586 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
587 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
588 has an implication that the given string is incomplete (because a partial
589 match is preferred over a full match). For this reason, these items now
590 give a partial match in this situation. [Aside: previously, the one case
591 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
592 match rather than a full match, which was wrong by the old rules, but is
593 now correct.]
594
595 6. There was a bug in the handling of #-introduced comments, recognized when
596 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
597 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
598 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
599 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
600 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
601 places in pcre_compile().
602
603 7. Related to (6) above, when pcre_compile() was skipping #-introduced
604 comments when looking ahead for named forward references to subpatterns,
605 the only newline sequence it recognized was NL. It now handles newlines
606 according to the set newline convention.
607
608 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
609 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
610 cater for a lack of strerror(). These oversights have been fixed.
611
612 9. Added --match-limit and --recursion-limit to pcregrep.
613
614 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
615
616 11. When the -o option was used, pcregrep was setting a return code of 1, even
617 when matches were found, and --line-buffered was not being honoured.
618
619 12. Added an optional parentheses number to the -o and --only-matching options
620 of pcregrep.
621
622 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
623 can match an empty string. The code to do it in pcretest and pcredemo
624 needed fixing:
625
626 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
627 only one byte after an empty string match just before CRLF (this case
628 just got forgotten; "any" and "anycrlf" were OK).
629
630 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
631 mode when an empty string match preceded an ASCII character followed by
632 a non-ASCII character. (The code for advancing by one character rather
633 than one byte was nonsense.)
634
635 (c) The pcredemo.c sample program did not have any code at all to handle
636 the cases when CRLF is a valid newline sequence.
637
638 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
639 as a starting offset was within the subject string. There is now a new
640 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
641 negative or greater than the length of the string. In order to test this,
642 pcretest is extended to allow the setting of negative starting offsets.
643
644 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
645 starting offset points to the beginning of a UTF-8 character was
646 unnecessarily clumsy. I tidied it up.
647
648 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
649 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
650
651 17. Nobody had reported that the --include_dir option, which was added in
652 release 7.7 should have been called --include-dir (hyphen, not underscore)
653 for compatibility with GNU grep. I have changed it to --include-dir, but
654 left --include_dir as an undocumented synonym, and the same for
655 --exclude-dir, though that is not available in GNU grep, at least as of
656 release 2.5.4.
657
658 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
659 characters from a string of bytes) have been redefined so as not to use
660 loops, in order to improve performance in some environments. At the same
661 time, I abstracted some of the common code into auxiliary macros to save
662 repetition (this should not affect the compiled code).
663
664 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
665 compile-time error is now given if \c is not followed by an ASCII
666 character, that is, a byte less than 128. (In EBCDIC mode, the code is
667 different, and any byte value is allowed.)
668
669 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
670 START_OPTIMIZE option, which is now allowed at compile time - but just
671 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
672 to pcregrep and other applications that have no direct access to PCRE
673 options. The new /Y option in pcretest sets this option when calling
674 pcre_compile().
675
676 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
677 back references. Groups containing recursive back references were forced to
678 be atomic by that change, but in the case of named groups, the amount of
679 memory required was incorrectly computed, leading to "Failed: internal
680 error: code overflow". This has been fixed.
681
682 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
683 pcretest.c, to avoid build problems in some Borland environments.
684
685
686 Version 8.10 25-Jun-2010
687 ------------------------
688
689 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
690 THEN.
691
692 2. (*ACCEPT) was not working when inside an atomic group.
693
694 3. Inside a character class, \B is treated as a literal by default, but
695 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
696 causes the error). The code is unchanged, but I tidied the documentation.
697
698 4. Inside a character class, PCRE always treated \R and \X as literals,
699 whereas Perl faults them if its -w option is set. I have changed PCRE so
700 that it faults them when PCRE_EXTRA is set.
701
702 5. Added support for \N, which always matches any character other than
703 newline. (It is the same as "." when PCRE_DOTALL is not set.)
704
705 6. When compiling pcregrep with newer versions of gcc which may have
706 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
707 declared with attribute warn_unused_result" were given. Just casting the
708 result to (void) does not stop the warnings; a more elaborate fudge is
709 needed. I've used a macro to implement this.
710
711 7. Minor change to pcretest.c to avoid a compiler warning.
712
713 8. Added four artifical Unicode properties to help with an option to make
714 \s etc use properties (see next item). The new properties are: Xan
715 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
716
717 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
718 use Unicode properties. (*UCP) at the start of a pattern can be used to set
719 this option. Modified pcretest to add /W to test this facility. Added
720 REG_UCP to make it available via the POSIX interface.
721
722 10. Added --line-buffered to pcregrep.
723
724 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
725 studied, and the match started with a letter with a code point greater than
726 127 whose first byte was different to the first byte of the other case of
727 the letter, the other case of this starting letter was not recognized
728 (#976).
729
730 12. If a pattern that was studied started with a repeated Unicode property
731 test, for example, \p{Nd}+, there was the theoretical possibility of
732 setting up an incorrect bitmap of starting bytes, but fortunately it could
733 not have actually happened in practice until change 8 above was made (it
734 added property types that matched character-matching opcodes).
735
736 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
737 possible starting bytes for non-anchored patterns.
738
739 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
740 \R, and also a number of cases that involve Unicode properties, both
741 explicit and implicit when PCRE_UCP is set.
742
743 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
744 input, it could crash or give wrong results if characters with values
745 greater than 0xc0 were present in the subject string. (Detail: it assumed
746 UTF-8 input when processing these items.)
747
748 16. Added a lot of (int) casts to avoid compiler warnings in systems where
749 size_t is 64-bit (#991).
750
751 17. Added a check for running out of memory when PCRE is compiled with
752 --disable-stack-for-recursion (#990).
753
754 18. If the last data line in a file for pcretest does not have a newline on
755 the end, a newline was missing in the output.
756
757 19. The default pcre_chartables.c file recognizes only ASCII characters (values
758 less than 128) in its various bitmaps. However, there is a facility for
759 generating tables according to the current locale when PCRE is compiled. It
760 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
761 characters, are recognized by isspace() and therefore were getting set in
762 these tables, and indeed these tables seem to approximate to ISO 8859. This
763 caused a problem in UTF-8 mode when pcre_study() was used to create a list
764 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
765 which of course cannot start UTF-8 characters. I have changed the code so
766 that only real ASCII characters (less than 128) and the correct starting
767 bytes for UTF-8 encodings are set for characters greater than 127 when in
768 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
769 altogether.)
770
771 20. Added the /T option to pcretest so as to be able to run tests with non-
772 standard character tables, thus making it possible to include the tests
773 used for 19 above in the standard set of tests.
774
775 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
776 reference to a subpattern the other side of a comment that contains an
777 opening parenthesis caused either an internal compiling error, or a
778 reference to the wrong subpattern.
779
780
781 Version 8.02 19-Mar-2010
782 ------------------------
783
784 1. The Unicode data tables have been updated to Unicode 5.2.0.
785
786 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
787 configured.
788
789 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
790 original author of that file, following a query about its status.
791
792 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
793 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
794
795 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
796 quantifier applied to a forward-referencing subroutine call, could compile
797 incorrect code or give the error "internal error: previously-checked
798 referenced subpattern not found".
799
800 6. Both MS Visual Studio and Symbian OS have problems with initializing
801 variables to point to external functions. For these systems, therefore,
802 pcre_malloc etc. are now initialized to local functions that call the
803 relevant global functions.
804
805 7. There were two entries missing in the vectors called coptable and poptable
806 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
807 I've fixed the data, and added a kludgy way of testing at compile time that
808 the lengths are correct (equal to the number of opcodes).
809
810 8. Following on from 7, I added a similar kludge to check the length of the
811 eint vector in pcreposix.c.
812
813 9. Error texts for pcre_compile() are held as one long string to avoid too
814 much relocation at load time. To find a text, the string is searched,
815 counting zeros. There was no check for running off the end of the string,
816 which could happen if a new error number was added without updating the
817 string.
818
819 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
820
821 11. \K was not working if it appeared in an atomic group or in a group that
822 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
823 \K is "not well defined" if used in an assertion. PCRE now accepts it if
824 the assertion is positive, but not if it is negative.
825
826 12. Change 11 fortuitously reduced the size of the stack frame used in the
827 "match()" function of pcre_exec.c by one pointer. Forthcoming
828 implementation of support for (*MARK) will need an extra pointer on the
829 stack; I have reserved it now, so that the stack frame size does not
830 decrease.
831
832 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
833 item in branch that calls a recursion is a subroutine call - as in the
834 second branch in the above example - was incorrectly given the compile-
835 time error "recursive call could loop indefinitely" because pcre_compile()
836 was not correctly checking the subroutine for matching a non-empty string.
837
838 14. The checks for overrunning compiling workspace could trigger after an
839 overrun had occurred. This is a "should never occur" error, but it can be
840 triggered by pathological patterns such as hundreds of nested parentheses.
841 The checks now trigger 100 bytes before the end of the workspace.
842
843 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
844
845
846 Version 8.01 19-Jan-2010
847 ------------------------
848
849 1. If a pattern contained a conditional subpattern with only one branch (in
850 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
851 computed the wrong minimum data length (which is of course zero for such
852 subpatterns). This could cause incorrect "no match" results.
853
854 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
855 the pattern is reset in the first branch, pcre_compile() failed with
856 "internal error: code overflow at offset...". This happened only when
857 the reset was to the original external option setting. (An optimization
858 abstracts leading options settings into an external setting, which was the
859 cause of this.)
860
861 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
862 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
863 assertion pattern did not match (meaning that the assertion was true), it
864 was incorrectly treated as false if the SKIP had been reached during the
865 matching. This also applied to assertions used as conditions.
866
867 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
868 assertion subpattern, including such a pattern used as a condition,
869 unpredictable results occurred, instead of the error return
870 PCRE_ERROR_DFA_UITEM.
871
872 5. The C++ GlobalReplace function was not working like Perl for the special
873 situation when an empty string is matched. It now does the fancy magic
874 stuff that is necessary.
875
876 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
877 removed. (These were left over from very, very early versions of PCRE.)
878
879 7. Some cosmetic changes to the code to make life easier when compiling it
880 as part of something else:
881
882 (a) Change DEBUG to PCRE_DEBUG.
883
884 (b) In pcre_compile(), rename the member of the "branch_chain" structure
885 called "current" as "current_branch", to prevent a collision with the
886 Linux macro when compiled as a kernel module.
887
888 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
889 prevent a collision with the Linux macro when compiled as a kernel
890 module.
891
892 8. In pcre_compile() there are some checks for integer overflows that used to
893 cast potentially large values to (double). This has been changed to that
894 when building, a check for int64_t is made, and if it is found, it is used
895 instead, thus avoiding the use of floating point arithmetic. (There is no
896 other use of FP in PCRE.) If int64_t is not found, the fallback is to
897 double.
898
899 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
900 2005 (difference between two addresses compared to an unsigned value).
901
902 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
903 custom one, because of the following reported problem in Windows:
904
905 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
906 under Win32.
907 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
908 therefore missing the function definition.
909 - The compiler thus generates a "C" signature for the test function.
910 - The linker fails to find the "C" function.
911 - PCRE fails to configure if asked to do so against libbz2.
912
913 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
914 messages were output:
915
916 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
917 rerunning libtoolize, to keep the correct libtool macros in-tree.
918 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
919
920 I have done both of these things.
921
922 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
923 most of the time, it *can* run out if it is given a pattern that contains a
924 runaway infinite recursion. I updated the discussion in the pcrestack man
925 page.
926
927 13. Now that we have gone to the x.xx style of version numbers, the minor
928 version may start with zero. Using 08 or 09 is a bad idea because users
929 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
930 interpreted as invalid octal numbers. I've updated the previous comment in
931 configure.ac, and also added a check that gives an error if 08 or 09 are
932 used.
933
934 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
935 causing partial matching to fail when the end of the subject matched \W
936 in a UTF-8 pattern where \W was quantified with a minimum of 3.
937
938 15. There were some discrepancies between the declarations in pcre_internal.h
939 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
940 their definitions. The declarations used "const uschar *" and the
941 definitions used USPTR. Even though USPTR is normally defined as "const
942 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
943 reported that: "This difference in casting confuses some C++ compilers, for
944 example, SunCC recognizes above declarations as different functions and
945 generates broken code for hbpcre." I have changed the declarations to use
946 USPTR.
947
948 16. GNU libtool is named differently on some systems. The autogen.sh script now
949 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
950 (FreeBSD).
951
952 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
953 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
954 comment: "Figure out how to create a longlong from a string: strtoll and
955 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
956 instance, but it only takes 2 args instead of 3!"
957
958 18. A subtle bug concerned with back references has been fixed by a change of
959 specification, with a corresponding code fix. A pattern such as
960 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
961 refers, was giving matches when it shouldn't. For example, xa=xaaa would
962 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
963 same bug. Such groups have to be quantified to be useful, or contained
964 inside another quantified group. (If there's no repetition, the reference
965 can never match.) The problem arises because, having left the group and
966 moved on to the rest of the pattern, a later failure that backtracks into
967 the group uses the captured value from the final iteration of the group
968 rather than the correct earlier one. I have fixed this in PCRE by forcing
969 any group that contains a reference to itself to be an atomic group; that
970 is, there cannot be any backtracking into it once it has completed. This is
971 similar to recursive and subroutine calls.
972
973
974 Version 8.00 19-Oct-09
975 ----------------------
976
977 1. The table for translating pcre_compile() error codes into POSIX error codes
978 was out-of-date, and there was no check on the pcre_compile() error code
979 being within the table. This could lead to an OK return being given in
980 error.
981
982 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
983 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
984 in a Windows environment.
985
986 3. The pcregrep --count option prints the count for each file even when it is
987 zero, as does GNU grep. However, pcregrep was also printing all files when
988 --files-with-matches was added. Now, when both options are given, it prints
989 counts only for those files that have at least one match. (GNU grep just
990 prints the file name in this circumstance, but including the count seems
991 more useful - otherwise, why use --count?) Also ensured that the
992 combination -clh just lists non-zero counts, with no names.
993
994 4. The long form of the pcregrep -F option was incorrectly implemented as
995 --fixed_strings instead of --fixed-strings. This is an incompatible change,
996 but it seems right to fix it, and I didn't think it was worth preserving
997 the old behaviour.
998
999 5. The command line items --regex=pattern and --regexp=pattern were not
1000 recognized by pcregrep, which required --regex pattern or --regexp pattern
1001 (with a space rather than an '='). The man page documented the '=' forms,
1002 which are compatible with GNU grep; these now work.
1003
1004 6. No libpcreposix.pc file was created for pkg-config; there was just
1005 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1006
1007 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1008 when UCP support is not needed, by modifying the Python script that
1009 generates it from Unicode data files. This should not matter if the module
1010 is correctly used as a library, but I received one complaint about 50K of
1011 unwanted data. My guess is that the person linked everything into his
1012 program rather than using a library. Anyway, it does no harm.
1013
1014 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1015 was a minimum greater than 1 for a wide character in a possessive
1016 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1017 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1018 character. Chaos in the form of incorrect output or a compiling loop could
1019 result.
1020
1021 9. The restrictions on what a pattern can contain when partial matching is
1022 requested for pcre_exec() have been removed. All patterns can now be
1023 partially matched by this function. In addition, if there are at least two
1024 slots in the offset vector, the offset of the earliest inspected character
1025 for the match and the offset of the end of the subject are set in them when
1026 PCRE_ERROR_PARTIAL is returned.
1027
1028 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1029 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1030 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1031 and may be more useful for multi-segment matching.
1032
1033 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1034 used to be given if ever the end of the subject was reached; now it is
1035 given only if matching could not proceed because another character was
1036 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1037 string "Z", which now yields "no match" instead of "partial match". In the
1038 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1039 final character ended with (*FAIL).
1040
1041 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1042 if the pattern had a "must contain" character that was already found in the
1043 earlier partial match, unless partial matching was again requested. For
1044 example, with the pattern /dog.(body)?/, the "must contain" character is
1045 "g". If the first part-match was for the string "dog", restarting with
1046 "sbody" failed. This bug has been fixed.
1047
1048 13. The string returned by pcre_dfa_exec() after a partial match has been
1049 changed so that it starts at the first inspected character rather than the
1050 first character of the match. This makes a difference only if the pattern
1051 starts with a lookbehind assertion or \b or \B (\K is not supported by
1052 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1053 matching functions compatible, and I think it's the right thing to do.
1054
1055 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1056 so that the demonstration program is easily available in environments where
1057 PCRE has not been installed from source.
1058
1059 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1060 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1061 library.
1062
1063 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1064 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1065 is not the first non-POSIX option to be added. Clearly some people find
1066 these options useful.
1067
1068 17. If a caller to the POSIX matching function regexec() passes a non-zero
1069 value for nmatch with a NULL value for pmatch, the value of
1070 nmatch is forced to zero.
1071
1072 18. RunGrepTest did not have a test for the availability of the -u option of
1073 the diff command, as RunTest does. It now checks in the same way as
1074 RunTest, and also checks for the -b option.
1075
1076 19. If an odd number of negated classes containing just a single character
1077 interposed, within parentheses, between a forward reference to a named
1078 subpattern and the definition of the subpattern, compilation crashed with
1079 an internal error, complaining that it could not find the referenced
1080 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1081 [The bug was that it was starting one character too far in when skipping
1082 over the character class, thus treating the ] as data rather than
1083 terminating the class. This meant it could skip too much.]
1084
1085 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1086 /g option in pcretest when the pattern contains \K, which makes it possible
1087 to have an empty string match not at the start, even when the pattern is
1088 anchored. Updated pcretest and pcredemo to use this option.
1089
1090 21. If the maximum number of capturing subpatterns in a recursion was greater
1091 than the maximum at the outer level, the higher number was returned, but
1092 with unset values at the outer level. The correct (outer level) value is
1093 now given.
1094
1095 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1096 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1097 make it do so. The string so far is captured, making this feature
1098 compatible with Perl.
1099
1100 23. The tests have been re-organized, adding tests 11 and 12, to make it
1101 possible to check the Perl 5.10 features against Perl 5.10.
1102
1103 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1104 pattern matches a fixed length string. PCRE did not allow this; now it
1105 does. Neither allows recursion.
1106
1107 25. I finally figured out how to implement a request to provide the minimum
1108 length of subject string that was needed in order to match a given pattern.
1109 (It was back references and recursion that I had previously got hung up
1110 on.) This code has now been added to pcre_study(); it finds a lower bound
1111 to the length of subject needed. It is not necessarily the greatest lower
1112 bound, but using it to avoid searching strings that are too short does give
1113 some useful speed-ups. The value is available to calling programs via
1114 pcre_fullinfo().
1115
1116 26. While implementing 25, I discovered to my embarrassment that pcretest had
1117 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1118 study optimizations had never been tested with that matching function.
1119 Oops. What is worse, even when it was passed study data, there was a bug in
1120 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1121 were also very few tests of studied patterns with pcre_dfa_exec().
1122
1123 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1124 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1125 on the other side of the coin, they are no longer allowed to have different
1126 names, because these cannot be distinguished in PCRE, and this has caused
1127 confusion. (This is a difference from Perl.)
1128
1129 28. When duplicate subpattern names are present (necessarily with different
1130 numbers, as required by 27 above), and a test is made by name in a
1131 conditional pattern, either for a subpattern having been matched, or for
1132 recursion in such a pattern, all the associated numbered subpatterns are
1133 tested, and the overall condition is true if the condition is true for any
1134 one of them. This is the way Perl works, and is also more like the way
1135 testing by number works.
1136
1137
1138 Version 7.9 11-Apr-09
1139 ---------------------
1140
1141 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1142 (pcretest), all targets were linked against these libraries. This included
1143 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1144 libraries. This caused unwanted dependencies to be created. This problem
1145 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1146 pcretest is linked with readline.
1147
1148 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1149 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1150 moved outside it again, because FALSE and TRUE are already defined in AIX,
1151 but BOOL is not.
1152
1153 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1154 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1155
1156 4. The pcregrep documentation said spaces were inserted as well as colons (or
1157 hyphens) following file names and line numbers when outputting matching
1158 lines. This is not true; no spaces are inserted. I have also clarified the
1159 wording for the --colour (or --color) option.
1160
1161 5. In pcregrep, when --colour was used with -o, the list of matching strings
1162 was not coloured; this is different to GNU grep, so I have changed it to be
1163 the same.
1164
1165 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1166 each matching line was coloured. Now it goes on to look for further matches
1167 of any of the test patterns, which is the same behaviour as GNU grep.
1168
1169 7. A pattern that could match an empty string could cause pcregrep to loop; it
1170 doesn't make sense to accept an empty string match in pcregrep, so I have
1171 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1172 seems to be how GNU grep behaves.
1173
1174 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1175 start or after a newline", because the conditional assertion was not being
1176 correctly handled. The rule now is that both the assertion and what follows
1177 in the first alternative must satisfy the test.
1178
1179 9. If auto-callout was enabled in a pattern with a conditional group whose
1180 condition was an assertion, PCRE could crash during matching, both with
1181 pcre_exec() and pcre_dfa_exec().
1182
1183 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1184 used for matching.
1185
1186 11. Unicode property support in character classes was not working for
1187 characters (bytes) greater than 127 when not in UTF-8 mode.
1188
1189 12. Added the -M command line option to pcretest.
1190
1191 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1192
1193 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1194
1195 16. Added comments and documentation about mis-use of no_arg in the C++
1196 wrapper.
1197
1198 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1199 from Martin Jerabek that uses macro names for all relevant character and
1200 string constants.
1201
1202 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1203 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1204 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1205 these, but not everybody uses configure.
1206
1207 19. A conditional group that had only one branch was not being correctly
1208 recognized as an item that could match an empty string. This meant that an
1209 enclosing group might also not be so recognized, causing infinite looping
1210 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1211 with the subject "ab", where knowledge that the repeated group can match
1212 nothing is needed in order to break the loop.
1213
1214 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1215 exec(), but without supplying a callout function, matching went wrong.
1216
1217 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1218 leak if the size of the offset vector was greater than 30. When the vector
1219 is smaller, the saved offsets during recursion go onto a local stack
1220 vector, but for larger vectors malloc() is used. It was failing to free
1221 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1222 error, in fact).
1223
1224 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1225 heapframe that is used only when UTF-8 support is enabled. This caused no
1226 problem, but was untidy.
1227
1228 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1229 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1230 included within another project.
1231
1232 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1233 slightly modified by me:
1234
1235 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1236 not building pcregrep.
1237
1238 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1239 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1240
1241 25. Forward references, both numeric and by name, in patterns that made use of
1242 duplicate group numbers, could behave incorrectly or give incorrect errors,
1243 because when scanning forward to find the reference group, PCRE was not
1244 taking into account the duplicate group numbers. A pattern such as
1245 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1246
1247 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1248 the feature of a custom pointer more persuasive (as requested by a user).
1249
1250 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1251 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1252 pre-defined.
1253
1254 28. Added support for (*UTF8) at the start of a pattern.
1255
1256 29. Arrange for flags added by the "release type" setting in CMake to be shown
1257 in the configuration summary.
1258
1259
1260 Version 7.8 05-Sep-08
1261 ---------------------
1262
1263 1. Replaced UCP searching code with optimized version as implemented for Ad
1264 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1265 stage table and inline lookup instead of a function, giving speed ups of 2
1266 to 5 times on some simple patterns that I tested. Permission was given to
1267 distribute the MultiStage2.py script that generates the tables (it's not in
1268 the tarball, but is in the Subversion repository).
1269
1270 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1271 scripts.
1272
1273 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1274 a group with a zero qualifier. The result of the study could be incorrect,
1275 or the function might crash, depending on the pattern.
1276
1277 4. Caseless matching was not working for non-ASCII characters in back
1278 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1279 It now works when Unicode Property Support is available.
1280
1281 5. In pcretest, an escape such as \x{de} in the data was always generating
1282 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1283 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1284 truncation.
1285
1286 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1287
1288 7. Added two (int) casts to pcregrep when printing the difference of two
1289 pointers, in case they are 64-bit values.
1290
1291 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1292 test 2 if it fails.
1293
1294 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1295 and a #define of that name to empty if it is not externally set. This is to
1296 allow users of MSVC to set it if necessary.
1297
1298 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1299 the convenience functions in the pcre_get.c source file.
1300
1301 11. An option change at the start of a pattern that had top-level alternatives
1302 could cause overwriting and/or a crash. This command provoked a crash in
1303 some environments:
1304
1305 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1306
1307 This potential security problem was recorded as CVE-2008-2371.
1308
1309 12. For a pattern where the match had to start at the beginning or immediately
1310 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1311 pcre_dfa_exec() could read past the end of the passed subject if there was
1312 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1313 pcretest so that it places the subject at the end of its malloc-ed buffer.
1314
1315 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1316 exec() might read past the end of the data buffer in UTF-8 mode.
1317
1318 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1319 the data contained the byte 0x85 as part of a UTF-8 character within its
1320 first line. This applied both to normal and DFA matching.
1321
1322 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1323 /^[^d]*?$/8 failed to match "abc".
1324
1325 16. Added a missing copyright notice to pcrecpp_internal.h.
1326
1327 17. Make it more clear in the documentation that values returned from
1328 pcre_exec() in ovector are byte offsets, not character counts.
1329
1330 18. Tidied a few places to stop certain compilers from issuing warnings.
1331
1332 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1333 supplied by Stefan Weber. I made a further small update for 7.8 because
1334 there is a change of source arrangements: the pcre_searchfuncs.c module is
1335 replaced by pcre_ucd.c.
1336
1337
1338 Version 7.7 07-May-08
1339 ---------------------
1340
1341 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1342 a string to a long long, pretend we don't even have a long long." This is
1343 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1344
1345 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1346 pre-7.6 versions, which defined a global no_arg variable instead of putting
1347 it in the RE class. (See also #8 below.)
1348
1349 3. Remove a line of dead code, identified by coverity and reported by Nuno
1350 Lopes.
1351
1352 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1353
1354 (1) The include/exclude patterns were being applied to the whole pathnames
1355 of files, instead of just to the final components.
1356
1357 (2) If there was more than one level of directory, the subdirectories were
1358 skipped unless they satisfied the include/exclude conditions. This is
1359 inconsistent with GNU grep (and could even be seen as contrary to the
1360 pcregrep specification - which I improved to make it absolutely clear).
1361 The action now is always to scan all levels of directory, and just
1362 apply the include/exclude patterns to regular files.
1363
1364 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1365 --exclude_dir in the tests to avoid scanning .svn directories.
1366
1367 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1368 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1369 doesn't support NULs in patterns.
1370
1371 7. Added some missing "const"s to declarations of static tables in
1372 pcre_compile.c and pcre_dfa_exec.c.
1373
1374 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1375 caused by fix #2 above. (Subsequently also a second patch to fix the
1376 first patch. And a third patch - this was a messy problem.)
1377
1378 9. Applied Craig's patch to remove the use of push_back().
1379
1380 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1381 matching function regexec().
1382
1383 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1384 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1385 references. PCRE supports relative numbers with this syntax (I don't think
1386 Oniguruma does).
1387
1388 12. Previously, a group with a zero repeat such as (...){0} was completely
1389 omitted from the compiled regex. However, this means that if the group
1390 was called as a subroutine from elsewhere in the pattern, things went wrong
1391 (an internal error was given). Such groups are now left in the compiled
1392 pattern, with a new opcode that causes them to be skipped at execution
1393 time.
1394
1395 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1396 to the way PCRE behaves:
1397
1398 (a) A lone ] character is dis-allowed (Perl treats it as data).
1399
1400 (b) A back reference to an unmatched subpattern matches an empty string
1401 (Perl fails the current match path).
1402
1403 (c) A data ] in a character class must be notated as \] because if the
1404 first data character in a class is ], it defines an empty class. (In
1405 Perl it is not possible to have an empty class.) The empty class []
1406 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1407 The negative empty class [^] matches any one character, independently
1408 of the DOTALL setting.
1409
1410 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1411 non-existent subpattern following a character class starting with ']' and
1412 containing () gave an internal compiling error instead of "reference to
1413 non-existent subpattern". Fortunately, when the pattern did exist, the
1414 compiled code was correct. (When scanning forwards to check for the
1415 existencd of the subpattern, it was treating the data ']' as terminating
1416 the class, so got the count wrong. When actually compiling, the reference
1417 was subsequently set up correctly.)
1418
1419 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1420 it was being rejected as not supported by pcre_dfa_exec(), even though
1421 other assertions are supported. I have made pcre_dfa_exec() support
1422 (*FAIL).
1423
1424 16. The implementation of 13c above involved the invention of a new opcode,
1425 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1426 cannot be changed at match time, I realized I could make a small
1427 improvement to matching performance by compiling OP_ALLANY instead of
1428 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1429 on the OP_ANY path.
1430
1431 17. Compiling pcretest on Windows with readline support failed without the
1432 following two fixes: (1) Make the unistd.h include conditional on
1433 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1434
1435 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1436 ncurses library to be included for pcretest when ReadLine support is
1437 requested, but also to allow for it to be overridden. This patch came from
1438 Daniel Bergström.
1439
1440 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1441 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1442 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1443 spotting this.
1444
1445
1446 Version 7.6 28-Jan-08
1447 ---------------------
1448
1449 1. A character class containing a very large number of characters with
1450 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1451 overflow.
1452
1453 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1454 HAVE_LONG_LONG is not defined.
1455
1456 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1457 bring them up to date and include new features. This patch includes:
1458
1459 - Fixed PH's badly added libz and libbz2 support.
1460 - Fixed a problem with static linking.
1461 - Added pcredemo. [But later removed - see 7 below.]
1462 - Fixed dftables problem and added an option.
1463 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1464 HAVE_LONG_LONG.
1465 - Added readline support for pcretest.
1466 - Added an listing of the option settings after cmake has run.
1467
1468 4. A user submitted a patch to Makefile that makes it easy to create
1469 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1470 Makefile.am that cause it to include this special target, without
1471 affecting anything else. Note that the same mingw target plus all
1472 the other distribution libraries and programs are now supported
1473 when configuring with CMake (see 6 below) instead of with
1474 Configure/Make.
1475
1476 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1477 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1478 exported in the Windows port". It has not yet been confirmed that the patch
1479 solves the problem, but it does no harm.
1480
1481 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1482 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1483 with CMake, and also correct the comment about stack recursion.
1484
1485 7. Remove the automatic building of pcredemo from the ./configure system and
1486 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1487 of a program that users should build themselves after PCRE is installed, so
1488 building it automatically is not really right. What is more, it gave
1489 trouble in some build environments.
1490
1491 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1492
1493
1494 Version 7.5 10-Jan-08
1495 ---------------------
1496
1497 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1498 values in parens when parsing an RE using the C++ wrapper."
1499
1500 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1501 Characters greater than 255 were excluded from the class instead of being
1502 included.
1503
1504 3. The same bug as (2) above applied to negated POSIX classes such as
1505 [:^space:].
1506
1507 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1508 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1509 I have changed it.
1510
1511 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1512 first named subpattern) and a construct such as (?&a) would reference the
1513 first named subpattern whose name started with "a" (in other words, the
1514 length check was missing). Both these problems are fixed. "Subpattern name
1515 expected" is now given for (?&) (a zero-length name), and this patch also
1516 makes it give the same error for \k'' (previously it complained that that
1517 was a reference to a non-existent subpattern).
1518
1519 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1520 this is right because (?- can be followed by option settings as well as by
1521 digits. I have, however, made the messages clearer.
1522
1523 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1524 than the number used in the conditional) now cause a compile-time error.
1525 This is actually not compatible with Perl, which accepts such patterns, but
1526 treats the conditional as always being FALSE (as PCRE used to), but it
1527 seems to me that giving a diagnostic is better.
1528
1529 8. Change "alphameric" to the more common word "alphanumeric" in comments
1530 and messages.
1531
1532 9. Fix two occurrences of "backslash" in comments that should have been
1533 "backspace".
1534
1535 10. Remove two redundant lines of code that can never be obeyed (their function
1536 was moved elsewhere).
1537
1538 11. The program that makes PCRE's Unicode character property table had a bug
1539 which caused it to generate incorrect table entries for sequences of
1540 characters that have the same character type, but are in different scripts.
1541 It amalgamated them into a single range, with the script of the first of
1542 them. In other words, some characters were in the wrong script. There were
1543 thirteen such cases, affecting characters in the following ranges:
1544
1545 U+002b0 - U+002c1
1546 U+0060c - U+0060d
1547 U+0061e - U+00612
1548 U+0064b - U+0065e
1549 U+0074d - U+0076d
1550 U+01800 - U+01805
1551 U+01d00 - U+01d77
1552 U+01d9b - U+01dbf
1553 U+0200b - U+0200f
1554 U+030fc - U+030fe
1555 U+03260 - U+0327f
1556 U+0fb46 - U+0fbb1
1557 U+10450 - U+1049d
1558
1559 12. The -o option (show only the matching part of a line) for pcregrep was not
1560 compatible with GNU grep in that, if there was more than one match in a
1561 line, it showed only the first of them. It now behaves in the same way as
1562 GNU grep.
1563
1564 13. If the -o and -v options were combined for pcregrep, it printed a blank
1565 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1566 does the same. The return code can be used to tell if there were any
1567 non-matching lines.
1568
1569 14. Added --file-offsets and --line-offsets to pcregrep.
1570
1571 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1572 infinitely looping recursion. The bug was that positive lookaheads were not
1573 being skipped when checking for a possible empty match (negative lookaheads
1574 and both kinds of lookbehind were skipped).
1575
1576 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1577 inclusion of <windows.h> to before rather than after the definition of
1578 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1579
1580 17. Specifying a possessive quantifier with a specific limit for a Unicode
1581 character property caused pcre_compile() to compile bad code, which led at
1582 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1583 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1584 caused the error; without that there was no problem.
1585
1586 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1587
1588 19. Added --enable-pcretest-libreadline.
1589
1590 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1591 RE::GlobalReplace(). As a result, the number of replacements returned was
1592 double what it should be. I removed one of the increments, but Craig sent a
1593 later patch that removed the other one (the right fix) and added unit tests
1594 that check the return values (which was not done before).
1595
1596 21. Several CMake things:
1597
1598 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1599 the names libpcre and libpcreposix, not just pcre and pcreposix.
1600
1601 (2) The above change means that pcretest and pcregrep are now correctly
1602 linked with the newly-built libraries, not previously installed ones.
1603
1604 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1605
1606 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1607 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1608 UTF-8 newline character). The key issue is that the pattern starts .*;
1609 this means that the match must be either at the beginning, or after a
1610 newline. The bug was in the code for advancing after a failed match and
1611 checking that the new position followed a newline. It was not taking
1612 account of UTF-8 characters correctly.
1613
1614 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1615 character classes. PCRE was not treating the sequence [:...:] as a
1616 character class unless the ... were all letters. Perl, however, seems to
1617 allow any characters between [: and :], though of course it rejects as
1618 unknown any "names" that contain non-letters, because all the known class
1619 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1620 for example, whereas PCRE did not - it did not recognize a POSIX character
1621 class. This seemed a bit dangerous, so the code has been changed to be
1622 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1623 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1624 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1625 Perl does, and where it didn't before.
1626
1627 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1628 Windows environments %n is disabled by default.
1629
1630
1631 Version 7.4 21-Sep-07
1632 ---------------------
1633
1634 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1635 means that a class such as [\s] counted as "explicit reference to CR or
1636 LF". That isn't really right - the whole point of the change was to try to
1637 help when there was an actual mention of one of the two characters. So now
1638 the change happens only if \r or \n (or a literal CR or LF) character is
1639 encountered.
1640
1641 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1642 of both had grown to the point where there were only 3 bits left.
1643 Fortunately, there was spare space in the data structure, and so I have
1644 moved the internal flags into a new 16-bit field to free up more option
1645 bits.
1646
1647 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1648 but did not set the internal JCHANGED flag - either of these is enough to
1649 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1650 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1651 start sets both bits.
1652
1653 4. Added options (at build time, compile time, exec time) to change \R from
1654 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1655
1656 5. doc/pcresyntax.html was missing from the distribution.
1657
1658 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1659 compatibility, even though it is no longer used.
1660
1661 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1662 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1663 windows.h file is present (where different names are used). [This was
1664 reversed later after testing - see 16 below.]
1665
1666 8. Changed all #include <config.h> to #include "config.h". There were also
1667 some further <pcre.h> cases that I changed to "pcre.h".
1668
1669 9. When pcregrep was used with the --colour option, it missed the line ending
1670 sequence off the lines that it output.
1671
1672 10. It was pointed out to me that arrays of string pointers cause lots of
1673 relocations when a shared library is dynamically loaded. A technique of
1674 using a single long string with a table of offsets can drastically reduce
1675 these. I have refactored PCRE in four places to do this. The result is
1676 dramatic:
1677
1678 Originally: 290
1679 After changing UCP table: 187
1680 After changing error message table: 43
1681 After changing table of "verbs" 36
1682 After changing table of Posix names 22
1683
1684 Thanks to the folks working on Gregex for glib for this insight.
1685
1686 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1687 unicode-properties was also set.
1688
1689 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1690
1691 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1692 checked only for CRLF.
1693
1694 14. Added casts to pcretest.c to avoid compiler warnings.
1695
1696 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1697
1698 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1699 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1700 entirely. This removes changes made in 7 above.
1701
1702 17. The CMake files have been updated, and there is now more information about
1703 building with CMake in the NON-UNIX-USE document.
1704
1705
1706 Version 7.3 28-Aug-07
1707 ---------------------
1708
1709 1. In the rejigging of the build system that eventually resulted in 7.1, the
1710 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1711 brackets there is not right, since it causes compilers to look for an
1712 installed pcre.h, not the version that is in the source that is being
1713 compiled (which of course may be different). I have changed it back to:
1714
1715 #include "pcre.h"
1716
1717 I have a vague recollection that the change was concerned with compiling in
1718 different directories, but in the new build system, that is taken care of
1719 by the VPATH setting the Makefile.
1720
1721 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1722 when the subject happened to end in the byte 0x85 (e.g. if the last
1723 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1724 characters but of course it shouldn't be taken as a newline when it is part
1725 of another character. The bug was that, for an unlimited repeat of . in
1726 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1727 characters when looking for a newline.
1728
1729 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1730
1731 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1732 in debug output.
1733
1734 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1735 long printing in the pcrecpp unittest when running under MinGW.
1736
1737 6. ESC_K was left out of the EBCDIC table.
1738
1739 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1740 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1741 limit also applies to "virtual nesting" when a pattern is recursive, and in
1742 this case 1000 isn't so big. I have been able to remove this limit at the
1743 expense of backing off one optimization in certain circumstances. Normally,
1744 when pcre_exec() would call its internal match() function recursively and
1745 immediately return the result unconditionally, it uses a "tail recursion"
1746 feature to save stack. However, when a subpattern that can match an empty
1747 string has an unlimited repetition quantifier, it no longer makes this
1748 optimization. That gives it a stack frame in which to save the data for
1749 checking that an empty string has been matched. Previously this was taken
1750 from the 1000-entry workspace that had been reserved. So now there is no
1751 explicit limit, but more stack is used.
1752
1753 8. Applied Daniel's patches to solve problems with the import/export magic
1754 syntax that is required for Windows, and which was going wrong for the
1755 pcreposix and pcrecpp parts of the library. These were overlooked when this
1756 problem was solved for the main library.
1757
1758 9. There were some crude static tests to avoid integer overflow when computing
1759 the size of patterns that contain repeated groups with explicit upper
1760 limits. As the maximum quantifier is 65535, the maximum group length was
1761 set at 30,000 so that the product of these two numbers did not overflow a
1762 32-bit integer. However, it turns out that people want to use groups that
1763 are longer than 30,000 bytes (though not repeat them that many times).
1764 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1765 made it possible to implement the integer overflow checks in a much more
1766 dynamic way, which I have now done. The artificial limitation on group
1767 length has been removed - we now have only the limit on the total length of
1768 the compiled pattern, which depends on the LINK_SIZE setting.
1769
1770 10. Fixed a bug in the documentation for get/copy named substring when
1771 duplicate names are permitted. If none of the named substrings are set, the
1772 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1773 empty string.
1774
1775 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1776 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1777 because the ] is interpreted as the first data character and the
1778 terminating ] is not found. PCRE has been made compatible with Perl in this
1779 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1780 cause memory overwriting.
1781
1782 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1783 string has been matched (to stop an infinite loop). It was not recognizing
1784 a conditional subpattern that could match an empty string if that
1785 subpattern was within another subpattern. For example, it looped when
1786 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1787 condition was not nested. This bug has been fixed.
1788
1789 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1790 past the start of the subject in the presence of bytes with the top bit
1791 set, for example "\x8aBCD".
1792
1793 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1794 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1795
1796 14. Optimized (?!) to (*FAIL).
1797
1798 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1799 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1800 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1801 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1802 does: it's just the validity check that is more restrictive.
1803
1804 16. Inserted checks for integer overflows during escape sequence (backslash)
1805 processing, and also fixed erroneous offset values for syntax errors during
1806 backslash processing.
1807
1808 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1809 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1810
1811 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1812 caused an overrun.
1813
1814 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1815 something other than just ASCII characters) inside a group that had an
1816 unlimited repeat caused a loop at compile time (while checking to see
1817 whether the group could match an empty string).
1818
1819 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1820 [\P{Any}] did so. (Error in the code for printing property names.)
1821
1822 21. An orphan \E inside a character class could cause a crash.
1823
1824 22. A repeated capturing bracket such as (A)? could cause a wild memory
1825 reference during compilation.
1826
1827 23. There are several functions in pcre_compile() that scan along a compiled
1828 expression for various reasons (e.g. to see if it's fixed length for look
1829 behind). There were bugs in these functions when a repeated \p or \P was
1830 present in the pattern. These operators have additional parameters compared
1831 with \d, etc, and these were not being taken into account when moving along
1832 the compiled data. Specifically:
1833
1834 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1835 length.
1836
1837 (b) An item such as \pL+ within a repeated group could cause crashes or
1838 loops.
1839
1840 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1841 "reference to non-existent subpattern" error.
1842
1843 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1844
1845 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1846 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1847
1848 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1849
1850 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1851 character were causing crashes (broken optimization).
1852
1853 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1854 \p or \P) caused a compile-time loop.
1855
1856 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1857 break. For example, the unstudied pattern [\r\n]A does not match the string
1858 "\r\nA" because change 7.0/46 below moves the current point on by two
1859 characters after failing to match at the start. However, the pattern \nA
1860 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1861 the same is true. There doesn't seem any very clean way out of this, but
1862 what I have chosen to do makes the common cases work: PCRE now takes note
1863 of whether there can be an explicit match for \r or \n anywhere in the
1864 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1865 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1866 pattern has explicit CR or LF references.
1867
1868 29. Added (*CR) etc for changing newline setting at start of pattern.
1869
1870
1871 Version 7.2 19-Jun-07
1872 ---------------------
1873
1874 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1875 which is apparently normally available under Windows.
1876
1877 2. Re-jig the pcregrep tests with different newline settings in an attempt
1878 to make them independent of the local environment's newline setting.
1879
1880 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1881
1882 4. Some of the "internals" tests were previously cut out when the link size
1883 was not 2, because the output contained actual offsets. The recent new
1884 "Z" feature of pcretest means that these can be cut out, making the tests
1885 usable with all link sizes.
1886
1887 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1888 stack recursion. This gives a massive performance boost under BSD, but just
1889 a small improvement under Linux. However, it saves one field in the frame
1890 in all cases.
1891
1892 6. Added more features from the forthcoming Perl 5.10:
1893
1894 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1895 recursion call. It refers to the nth most recently opened parentheses.
1896
1897 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1898 to be opened parentheses.
1899
1900 (c) Conditions that refer to capturing parentheses can be specified
1901 relatively, for example, (?(-2)... or (?(+3)...
1902
1903 (d) \K resets the start of the current match so that everything before
1904 is not part of it.
1905
1906 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1907
1908 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1909 reference syntax.
1910
1911 (g) (?| introduces a group in which the numbering of parentheses in each
1912 alternative starts with the same number.
1913
1914 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1915
1916 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1917 PCRE_INFO_JCHANGED.
1918
1919 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1920 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1921 for detecting groups that can match an empty string.
1922
1923 9. A pattern with a very large number of alternatives (more than several
1924 hundred) was running out of internal workspace during the pre-compile
1925 phase, where pcre_compile() figures out how much memory will be needed. A
1926 bit of new cunning has reduced the workspace needed for groups with
1927 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1928 workspace instead of running out of the 4096 that are available.
1929
1930 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1931
1932 11. Applied patch from Google to remove an optimization that didn't quite work.
1933 The report of the bug said:
1934
1935 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1936 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1937 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1938
1939 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1940 it matched the wrong number of bytes.
1941
1942
1943 Version 7.1 24-Apr-07
1944 ---------------------
1945
1946 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1947 that is more "standard", making use of automake and other Autotools. There
1948 is some re-arrangement of the files and adjustment of comments consequent
1949 on this.
1950
1951 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1952 for recursive directory scanning broke on some systems because the files
1953 are not scanned in any specific order and on different systems the order
1954 was different. A call to "sort" has been inserted into RunGrepTest for the
1955 approprate test as a short-term fix. In the longer term there may be an
1956 alternative.
1957
1958 3. I had an email from Eric Raymond about problems translating some of PCRE's
1959 man pages to HTML (despite the fact that I distribute HTML pages, some
1960 people do their own conversions for various reasons). The problems
1961 concerned the use of low-level troff macros .br and .in. I have therefore
1962 removed all such uses from the man pages (some were redundant, some could
1963 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1964 HTML has been updated to handle .nf/.fi and to complain if it encounters
1965 .br or .in.
1966
1967 4. Updated comments in configure.ac that get placed in config.h.in and also
1968 arranged for config.h to be included in the distribution, with the name
1969 config.h.generic, for the benefit of those who have to compile without
1970 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1971
1972 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1973 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1974 makevp.bat for the current PCRE, using the additional files
1975 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1976
1977 6. A Windows user reported a minor discrepancy with test 2, which turned out
1978 to be caused by a trailing space on an input line that had got lost in his
1979 copy. The trailing space was an accident, so I've just removed it.
1980
1981 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1982 that is needed.
1983
1984 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1985 as "const" (a) because they are and (b) because it helps the PHP
1986 maintainers who have recently made a script to detect big data structures
1987 in the php code that should be moved to the .rodata section. I remembered
1988 to update Builducptable as well, so it won't revert if ucptable.h is ever
1989 re-created.
1990
1991 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1992 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1993 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1994 support is not required. This saves 1.5-2K of code, which is important in
1995 some applications.
1996
1997 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1998 so as not to refer to the tables, even though these functions will never be
1999 called when UTF-8 support is disabled. Otherwise there are problems with a
2000 shared library.
2001
2002 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2003
2004 (a) It was defining its arguments as char * instead of void *.
2005
2006 (b) It was assuming that all moves were upwards in memory; this was true
2007 a long time ago when I wrote it, but is no longer the case.
2008
2009 The emulated memove() is provided for those environments that have neither
2010 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2011 is clearly not the case, as these two bugs were recently reported.
2012
2013 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2014 and Detrail to create the HTML documentation, the .txt form of the man
2015 pages, and it removes trailing spaces from listed files. It also creates
2016 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2017 case, it wraps all the #defines with #ifndefs. This script should be run
2018 before "make dist".
2019
2020 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2021 with Unicode property support.
2022
2023 (a) For a maximizing quantifier, if the two different cases of the
2024 character were of different lengths in their UTF-8 codings (there are
2025 some cases like this - I found 11), and the matching function had to
2026 back up over a mixture of the two cases, it incorrectly assumed they
2027 were both the same length.
2028
2029 (b) When PCRE was configured to use the heap rather than the stack for
2030 recursion during matching, it was not correctly preserving the data for
2031 the other case of a UTF-8 character when checking ahead for a match
2032 while processing a minimizing repeat. If the check also involved
2033 matching a wide character, but failed, corruption could cause an
2034 erroneous result when trying to check for a repeat of the original
2035 character.
2036
2037 13. Some tidying changes to the testing mechanism:
2038
2039 (a) The RunTest script now detects the internal link size and whether there
2040 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2041 values substituted by "configure". (The RunGrepTest script already did
2042 this for UTF-8.) The configure.ac script no longer substitutes the
2043 relevant variables.
2044
2045 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2046 with length and offset values. This means that the output is different
2047 for different internal link sizes. Test 2 is skipped for link sizes
2048 other than 2 because of this, bypassing the problem. Unfortunately,
2049 there was also a test in test 3 (the locale tests) that used /B and
2050 failed for link sizes other than 2. Rather than cut the whole test out,
2051 I have added a new /Z option to pcretest that replaces the length and
2052 offset values with spaces. This is now used to make test 3 independent
2053 of link size. (Test 2 will be tidied up later.)
2054
2055 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2056 segmentation fault instead of returning the appropriate error message.
2057
2058 15. In multiline mode when the newline sequence was set to "any", the pattern
2059 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2060 This doesn't seem right; it now treats the CRLF combination as the line
2061 ending, and so does not match in that case. It's only a pattern such as ^$
2062 that would hit this one: something like ^ABC$ would have failed after \r
2063 and then tried again after \r\n.
2064
2065 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2066 in an attempt to make files that differ only in their line terminators
2067 compare equal. This works on Linux.
2068
2069 17. Under certain error circumstances pcregrep might try to free random memory
2070 as it exited. This is now fixed, thanks to valgrind.
2071
2072 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2073 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2074 was because its rules for how to advance for /g after matching an empty
2075 string at the end of a line did not allow for this case. They now check for
2076 it specially.
2077
2078 20. pcretest is supposed to handle patterns and data of any length, by
2079 extending its buffers when necessary. It was getting this wrong when the
2080 buffer for a data line had to be extended.
2081
2082 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2083 CRLF as a newline sequence.
2084
2085 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2086 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2087 I have nevertheless tidied it up.
2088
2089 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2090
2091 24. Added a man page for pcre-config.
2092
2093
2094 Version 7.0 19-Dec-06
2095 ---------------------
2096
2097 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2098 moving to gcc 4.1.1.
2099
2100 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2101 sys/time.h, which is documented as needed for this function. It doesn't
2102 seem to matter on Linux, but it showed up on some releases of OS X.
2103
2104 3. It seems that there are systems where bytes whose values are greater than
2105 127 match isprint() in the "C" locale. The "C" locale should be the
2106 default when a C program starts up. In most systems, only ASCII printing
2107 characters match isprint(). This difference caused the output from pcretest
2108 to vary, making some of the tests fail. I have changed pcretest so that:
2109
2110 (a) When it is outputting text in the compiled version of a pattern, bytes
2111 other than 32-126 are always shown as hex escapes.
2112
2113 (b) When it is outputting text that is a matched part of a subject string,
2114 it does the same, unless a different locale has been set for the match
2115 (using the /L modifier). In this case, it uses isprint() to decide.
2116
2117 4. Fixed a major bug that caused incorrect computation of the amount of memory
2118 required for a compiled pattern when options that changed within the
2119 pattern affected the logic of the preliminary scan that determines the
2120 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2121 that the computed length was too small. The symptoms of this bug were
2122 either the PCRE error "internal error: code overflow" from pcre_compile(),
2123 or a glibc crash with a message such as "pcretest: free(): invalid next
2124 size (fast)". Examples of patterns that provoked this bug (shown in
2125 pcretest format) are:
2126
2127 /(?-x: )/x
2128 /(?x)(?-x: \s*#\s*)/
2129 /((?i)[\x{c0}])/8
2130 /(?i:[\x{c0}])/8
2131
2132 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2133 is now done differently.
2134
2135 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2136 wrapper classes; (b) implement a new function in the C++ scanner that is
2137 more efficient than the old way of doing things because it avoids levels of
2138 recursion in the regex matching; (c) add a paragraph to the documentation
2139 for the FullMatch() function.
2140
2141 6. The escape sequence \n was being treated as whatever was defined as
2142 "newline". Not only was this contrary to the documentation, which states
2143 that \n is character 10 (hex 0A), but it also went horribly wrong when
2144 "newline" was defined as CRLF. This has been fixed.
2145
2146 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2147 was being set to -1 for the "end of line" case (supposedly a value that no
2148 character can have). Though this value is never used (the check for end of
2149 line is "zero bytes in current character"), it caused compiler complaints.
2150 I've changed it to 0xffffffff.
2151
2152 8. In pcre_version.c, the version string was being built by a sequence of
2153 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2154 string (as it is for production releases) called a macro with an empty
2155 argument. The C standard says the result of this is undefined. The gcc
2156 compiler treats it as an empty string (which was what was wanted) but it is
2157 reported that Visual C gives an error. The source has been hacked around to
2158 avoid this problem.
2159
2160 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2161 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2162 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2163 of them did).
2164
2165 10. Originally, pcretest opened its input and output without "b"; then I was
2166 told that "b" was needed in some environments, so it was added for release
2167 5.0 to both the input and output. (It makes no difference on Unix-like
2168 systems.) Later I was told that it is wrong for the input on Windows. I've
2169 now abstracted the modes into two macros, to make it easier to fiddle with
2170 them, and removed "b" from the input mode under Windows.
2171
2172 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2173
2174 12. Added -help and --help to pcretest as an official way of being reminded
2175 of the options.
2176
2177 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2178 and pcrecpp.cc because they annoy compilers at high warning levels.
2179
2180 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2181
2182 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2183 scripts are not C programs :-) and which was not noticed because it works
2184 on Linux.
2185
2186 16. pcretest is supposed to handle any length of pattern and data line (as one
2187 line or as a continued sequence of lines) by extending its input buffer if
2188 necessary. This feature was broken for very long pattern lines, leading to
2189 a string of junk being passed to pcre_compile() if the pattern was longer
2190 than about 50K.
2191
2192 17. I have done a major re-factoring of the way pcre_compile() computes the
2193 amount of memory needed for a compiled pattern. Previously, there was code
2194 that made a preliminary scan of the pattern in order to do this. That was
2195 OK when PCRE was new, but as the facilities have expanded, it has become
2196 harder and harder to keep it in step with the real compile phase, and there
2197 have been a number of bugs (see for example, 4 above). I have now found a
2198 cunning way of running the real compile function in a "fake" mode that
2199 enables it to compute how much memory it would need, while actually only
2200 ever using a few hundred bytes of working memory and without too many
2201 tests of the mode. This should make future maintenance and development
2202 easier. A side effect of this work is that the limit of 200 on the nesting
2203 depth of parentheses has been removed (though this was never a serious
2204 limitation, I suspect). However, there is a downside: pcre_compile() now
2205 runs more slowly than before (30% or more, depending on the pattern). I
2206 hope this isn't a big issue. There is no effect on runtime performance.
2207
2208 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2209 newline (only possible for the last line of a file) and it was a
2210 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2211
2212 19. Added additional timing features to pcretest. (1) The -tm option now times
2213 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2214 separate command line item, by a number that specifies the number of
2215 repeats to use when timing. The default is 50000; this gives better
2216 precision, but takes uncomfortably long for very large patterns.
2217
2218 20. Extended pcre_study() to be more clever in cases where a branch of a
2219 subpattern has no definite first character. For example, (a*|b*)[cd] would
2220 previously give no result from pcre_study(). Now it recognizes that the
2221 first character must be a, b, c, or d.
2222
2223 21. There was an incorrect error "recursive call could loop indefinitely" if
2224 a subpattern (or the entire pattern) that was being tested for matching an
2225 empty string contained only one non-empty item after a nested subpattern.
2226 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2227 incorrectly, because the \d was being skipped in the check.
2228
2229 22. The pcretest program now has a new pattern option /B and a command line
2230 option -b, which is equivalent to adding /B to every pattern. This causes
2231 it to show the compiled bytecode, without the additional information that
2232 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2233 is the same as /B/I).
2234
2235 23. A new optimization is now able automatically to treat some sequences such
2236 as a*b as a*+b. More specifically, if something simple (such as a character
2237 or a simple class like \d) has an unlimited quantifier, and is followed by
2238 something that cannot possibly match the quantified thing, the quantifier
2239 is automatically "possessified".
2240
2241 24. A recursive reference to a subpattern whose number was greater than 39
2242 went wrong under certain circumstances in UTF-8 mode. This bug could also
2243 have affected the operation of pcre_study().
2244
2245 25. Realized that a little bit of performance could be had by replacing
2246 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2247
2248 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2249
2250 27. Possessive quantifiers such as a++ were previously implemented by turning
2251 them into atomic groups such as ($>a+). Now they have their own opcodes,
2252 which improves performance. This includes the automatically created ones
2253 from 23 above.
2254
2255 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2256 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2257 the first matched character to be a colon. This applied both to named and
2258 numbered groups.
2259
2260 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2261
2262 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2263 building PCRE on Windows easier, so I have included it in the distribution.
2264
2265 31. There is now a check in pcretest against a ridiculously large number being
2266 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2267 loop, the loop is abandoned.
2268
2269 32. Forward references to subpatterns in conditions such as (?(2)...) where
2270 subpattern 2 is defined later cause pcre_compile() to search forwards in
2271 the pattern for the relevant set of parentheses. This search went wrong
2272 when there were unescaped parentheses in a character class, parentheses
2273 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2274
2275 33. "Subroutine" calls and backreferences were previously restricted to
2276 referencing subpatterns earlier in the regex. This restriction has now
2277 been removed.
2278
2279 34. Added a number of extra features that are going to be in Perl 5.10. On the
2280 whole, these are just syntactic alternatives for features that PCRE had
2281 previously implemented using the Python syntax or my own invention. The
2282 other formats are all retained for compatibility.
2283
2284 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2285 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2286 also .NET compatible.
2287
2288 (b) A recursion or subroutine call to a named group can now be defined as
2289 (?&name) as well as (?P>name).
2290
2291 (c) A backreference to a named group can now be defined as \k<name> or
2292 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2293 5.10, are also .NET compatible.
2294
2295 (d) A conditional reference to a named group can now use the syntax
2296 (?(<name>) or (?('name') as well as (?(name).
2297
2298 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2299 groups (named and numbered) that are never evaluated inline, but can be
2300 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2301 is always false. There may be only one alternative in such a group.
2302
2303 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2304 as the simple (?(R). The condition is true only if the most recent
2305 recursion is that of the given number or name. It does not search out
2306 through the entire recursion stack.
2307
2308 (g) The escape \gN or \g{N} has been added, where N is a positive or
2309 negative number, specifying an absolute or relative reference.
2310
2311 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2312 some "unreachable code" warnings.
2313
2314 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2315 things, this adds five new scripts.
2316
2317 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2318 There were also incompatibilities regarding the handling of \Q..\E inside
2319 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2320 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2321
2322 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2323 matches an empty string, and forcibly breaks the loop. There were bugs in
2324 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2325 against aaaa the result was just "a" rather than "aaaa", for example. Two
2326 separate and independent bugs (that affected different cases) have been
2327 fixed.
2328
2329 39. Refactored the code to abolish the use of different opcodes for small
2330 capturing bracket numbers. This is a tidy that I avoided doing when I
2331 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2332 The new approach is not only tidier, it makes it possible to reduce the
2333 memory needed to fix the previous bug (38).
2334
2335 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2336 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2337 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2338 mode.
2339
2340 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2341 report.
2342
2343 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2344 copy construction and assignment in the C++ wrapper.
2345
2346 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2347 couple of bugs that could have given wrong results in the "--newline=crlf"
2348 case.
2349
2350 44. Added a number of casts and did some reorganization of signed/unsigned int
2351 variables following suggestions from Dair Grant. Also renamed the variable
2352 "this" as "item" because it is a C++ keyword.
2353
2354 45. Arranged for dftables to add
2355
2356 #include "pcre_internal.h"
2357
2358 to pcre_chartables.c because without it, gcc 4.x may remove the array
2359 definition from the final binary if PCRE is built into a static library and
2360 dead code stripping is activated.
2361
2362 46. For an unanchored pattern, if a match attempt fails at the start of a
2363 newline sequence, and the newline setting is CRLF or ANY, and the next two
2364 characters are CRLF, advance by two characters instead of one.
2365
2366
2367 Version 6.7 04-Jul-06
2368 ---------------------
2369
2370 1. In order to handle tests when input lines are enormously long, pcretest has
2371 been re-factored so that it automatically extends its buffers when
2372 necessary. The code is crude, but this _is_ just a test program. The
2373 default size has been increased from 32K to 50K.
2374
2375 2. The code in pcre_study() was using the value of the re argument before
2376 testing it for NULL. (Of course, in any sensible call of the function, it
2377 won't be NULL.)
2378
2379 3. The memmove() emulation function in pcre_internal.h, which is used on
2380 systems that lack both memmove() and bcopy() - that is, hardly ever -
2381 was missing a "static" storage class specifier.
2382
2383 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2384 containing an extended class (one that cannot be represented by a bitmap
2385 because it contains high-valued characters or Unicode property items, e.g.
2386 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2387 pattern, but PCRE should not loop if you do not (it no longer does).
2388 [Detail: two cases were found: (a) a repeated subpattern containing an
2389 extended class; (b) a recursive reference to a subpattern that followed a
2390 previous extended class. It wasn't skipping over the extended class
2391 correctly when UTF-8 mode was not set.]
2392
2393 5. A negated single-character class was not being recognized as fixed-length
2394 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2395 compile error "lookbehind assertion is not fixed length".
2396
2397 6. The RunPerlTest auxiliary script was showing an unexpected difference
2398 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2399 write a Perl script that can interpret lines of an input file either as
2400 byte characters or as UTF-8, which is what "perltest" was being required to
2401 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2402 can't do is switch easily at run time between having the "use utf8;" pragma
2403 or not. In the end, I fudged it by using the RunPerlTest script to insert
2404 "use utf8;" explicitly for the UTF-8 tests.
2405
2406 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2407 the end of the subject string, contrary to the documentation and to what
2408 Perl does. This was true of both matching functions. Now it matches only at
2409 the start of the subject and immediately after *internal* newlines.
2410
2411 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2412 a pointer to an int instead of a pointer to an unsigned long int. This
2413 caused problems on 64-bit systems.
2414
2415 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2416 instance of the 'standard' template library not being so standard".
2417
2418 10. There was no check on the number of named subpatterns nor the maximum
2419 length of a subpattern name. The product of these values is used to compute
2420 the size of the memory block for a compiled pattern. By supplying a very
2421 long subpattern name and a large number of named subpatterns, the size
2422 computation could be caused to overflow. This is now prevented by limiting
2423 the length of names to 32 characters, and the number of named subpatterns
2424 to 10,000.
2425
2426 11. Subpatterns that are repeated with specific counts have to be replicated in
2427 the compiled pattern. The size of memory for this was computed from the
2428 length of the subpattern and the repeat count. The latter is limited to
2429 65535, but there was no limit on the former, meaning that integer overflow
2430 could in principle occur. The compiled length of a repeated subpattern is
2431 now limited to 30,000 bytes in order to prevent this.
2432
2433 12. Added the optional facility to have named substrings with the same name.
2434
2435 13. Added the ability to use a named substring as a condition, using the
2436 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2437 are numbers (not recommended). Forward references are permitted.
2438
2439 14. Added forward references in named backreferences (if you see what I mean).
2440
2441 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2442 pattern could run off the end of the subject. For example, the pattern
2443 "(?s)(.{1,5})"8 did this with the subject "ab".
2444
2445 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2446 PCRE_CASELESS was set when matching characters that were quantified with ?
2447 or *.
2448
2449 17. A character class other than a single negated character that had a minimum
2450 but no maximum quantifier - for example [ab]{6,} - was not handled
2451 correctly by pce_dfa_exec(). It would match only one character.
2452
2453 18. A valid (though odd) pattern that looked like a POSIX character
2454 class but used an invalid character after [ (for example [[,abc,]]) caused
2455 pcre_compile() to give the error "Failed: internal error: code overflow" or
2456 in some cases to crash with a glibc free() error. This could even happen if
2457 the pattern terminated after [[ but there just happened to be a sequence of
2458 letters, a binary zero, and a closing ] in the memory that followed.
2459
2460 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2461 over the years. Originally (before any Unicode support), just the bottom 8
2462 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2463 output from "man perlunicode" includes this:
2464
2465 The regular expression compiler produces polymorphic opcodes. That
2466 is, the pattern adapts to the data and automatically switches to
2467 the Unicode character scheme when presented with Unicode data--or
2468 instead uses a traditional byte scheme when presented with byte
2469 data.
2470
2471 Sadly, a wide octal escape does not cause a switch, and in a string with
2472 no other multibyte characters, these octal escapes are treated as before.
2473 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2474 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2475 Unicode string.
2476
2477 I have not perpetrated such confusion in PCRE. Up till now, it took just
2478 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2479 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2480 translate to the appropriate multibyte character.
2481
2482 29. Applied some refactoring to reduce the number of warnings from Microsoft
2483 and Borland compilers. This has included removing the fudge introduced
2484 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2485 a warning about an unused variable.
2486
2487 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2488 characters since release 4.0, because Perl (from release 5.004) does not.
2489 [Or at least, is documented not to: some releases seem to be in conflict
2490 with the documentation.] However, when a pattern was studied with
2491 pcre_study() and all its branches started with \s, PCRE still included VT
2492 as a possible starting character. Of course, this did no harm; it just
2493 caused an unnecessary match attempt.
2494
2495 22. Removed a now-redundant internal flag bit that recorded the fact that case
2496 dependency changed within the pattern. This was once needed for "required
2497 byte" processing, but is no longer used. This recovers a now-scarce options
2498 bit. Also moved the least significant internal flag bit to the most-
2499 significant bit of the word, which was not previously used (hangover from
2500 the days when it was an int rather than a uint) to free up another bit for
2501 the future.
2502
2503 23. Added support for CRLF line endings as well as CR and LF. As well as the
2504 default being selectable at build time, it can now be changed at runtime
2505 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2506 specify that it is scanning data with non-default line endings.
2507
2508 24. Changed the definition of CXXLINK to make it agree with the definition of
2509 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2510
2511 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2512 recursions. This makes a big different to stack usage for some patterns.
2513
2514 26. If a subpattern containing a named recursion or subroutine reference such
2515 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2516 the space required for the compiled pattern went wrong and gave too small a
2517 value. Depending on the environment, this could lead to "Failed: internal
2518 error: code overflow at offset 49" or "glibc detected double free or
2519 corruption" errors.
2520
2521 27. Applied patches from Google (a) to support the new newline modes and (b) to
2522 advance over multibyte UTF-8 characters in GlobalReplace.
2523
2524 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2525 difference for some implementation of PCRE in some Windows version.
2526
2527 29. Added some extra testing facilities to pcretest:
2528
2529 \q<number> in a data line sets the "match limit" value
2530 \Q<number> in a data line sets the "match recursion limt" value
2531 -S <number> sets the stack size, where <number> is in megabytes
2532
2533 The -S option isn't available for Windows.
2534
2535
2536 Version 6.6 06-Feb-06
2537 ---------------------
2538
2539 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2540 in pcreposix.h. I have copied the definition from pcre.h.
2541
2542 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2543 because pcre.h is no longer a built file.
2544
2545 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2546 not normally included in the compiled code.
2547
2548
2549 Version 6.5 01-Feb-06
2550 ---------------------
2551
2552 1. When using the partial match feature with pcre_dfa_exec(), it was not
2553 anchoring the second and subsequent partial matches at the new starting
2554 point. This could lead to incorrect results. For example, with the pattern
2555 /1234/, partially matching against "123" and then "a4" gave a match.
2556
2557 2. Changes to pcregrep:
2558
2559 (a) All non-match returns from pcre_exec() were being treated as failures
2560 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2561 error message is output. Some extra information is given for the
2562 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2563 probably the only errors that are likely to be caused by users (by
2564 specifying a regex that has nested indefinite repeats, for instance).
2565 If there are more than 20 of these errors, pcregrep is abandoned.
2566
2567 (b) A binary zero was treated as data while matching, but terminated the
2568 output line if it was written out. This has been fixed: binary zeroes
2569 are now no different to any other data bytes.
2570
2571 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2572 used to set a locale for matching. The --locale=xxxx long option has
2573 been added (no short equivalent) to specify a locale explicitly on the
2574 pcregrep command, overriding the environment variables.
2575
2576 (d) When -B was used with -n, some line numbers in the output were one less
2577 than they should have been.
2578
2579 (e) Added the -o (--only-matching) option.
2580
2581 (f) If -A or -C was used with -c (count only), some lines of context were
2582 accidentally printed for the final match.
2583
2584 (g) Added the -H (--with-filename) option.
2585
2586 (h) The combination of options -rh failed to suppress file names for files
2587 that were found from directory arguments.
2588
2589 (i) Added the -D (--devices) and -d (--directories) options.
2590
2591 (j) Added the -F (--fixed-strings) option.
2592
2593 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2594
2595 (l) Added the --colo(u)r option.
2596
2597 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2598 is not present by default.
2599
2600 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2601 items such as (?R) or (?1), when the recursion could match a number of
2602 alternatives. If it matched one of the alternatives, but subsequently,
2603 outside the recursion, there was a failure, the code tried to back up into
2604 the recursion. However, because of the way PCRE is implemented, this is not
2605 possible, and the result was an incorrect result from the match.
2606
2607 In order to prevent this happening, the specification of recursion has
2608 been changed so that all such subpatterns are automatically treated as
2609 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2610
2611 4. I had overlooked the fact that, in some locales, there are characters for
2612 which isalpha() is true but neither isupper() nor islower() are true. In
2613 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2614 and ordfeminine) are like this. This affected the treatment of \w and \W
2615 when they appeared in character classes, but not when they appeared outside
2616 a character class. The bit map for "word" characters is now created
2617 separately from the results of isalnum() instead of just taking it from the
2618 upper, lower, and digit maps. (Plus the underscore character, of course.)
2619
2620 5. The above bug also affected the handling of POSIX character classes such as
2621 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2622 permanent tables. Instead, the bit maps for such a class were previously
2623 created as the appropriate unions of the upper, lower, and digit bitmaps.
2624 Now they are created by subtraction from the [[:word:]] class, which has
2625 its own bitmap.
2626
2627 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2628 It is created by subtracting the vertical space characters (\x09, \x0a,
2629 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2630 subtraction was done in the overall bitmap for a character class, meaning
2631 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2632 be recognized. This bug has been fixed.
2633
2634 7. Patches from the folks at Google:
2635
2636 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2637 real life, but is still worth protecting against".
2638
2639 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2640 regular expressions".
2641
2642 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2643 have it.
2644
2645 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2646 "configure" and the latter not, in order to fix a problem somebody had
2647 with compiling the Arg class on HP-UX.
2648
2649 (e) Improve the error-handling of the C++ wrapper a little bit.
2650
2651 (f) New tests for checking recursion limiting.
2652
2653 8. The pcre_memmove() function, which is used only if the environment does not
2654 have a standard memmove() function (and is therefore rarely compiled),
2655 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2656 returning a result (though PCRE never actually uses the result).
2657
2658 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2659 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2660 returned instead of calling malloc() with an overflowing number that would
2661 most likely cause subsequent chaos.
2662
2663 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2664
2665 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2666 with this option is matched, the nmatch and pmatch options of regexec() are
2667 ignored.
2668
2669 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2670 provided in case anyone wants to the the POSIX interface with UTF-8
2671 strings.
2672
2673 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2674 C++ linking (needed for some HP-UX environments).
2675
2676 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2677 (unused parameter) and in the pcre_printint() function (omitted "default"
2678 switch label when the default is to do nothing).
2679
2680 15. Added some code to make it possible, when PCRE is compiled as a C++
2681 library, to replace subject pointers for pcre_exec() with a smart pointer
2682 class, thus making it possible to process discontinuous strings.
2683
2684 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2685 much the same function. They were added by different people who were trying
2686 to make PCRE easy to compile on non-Unix systems. It has been suggested
2687 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2688 for compiling on Windows systems. I have therefore replaced it with
2689 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2690 defaults to "extern" for C or "extern C" for C++, which works fine on
2691 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2692 SCOPE with something explicit in config.h. In addition:
2693
2694 (a) pcreposix.h still had just "extern" instead of either of these macros;
2695 I have replaced it with PCRE_DATA_SCOPE.
2696
2697 (b) Functions such as _pcre_xclass(), which are internal to the library,
2698 but external in the C sense, all had PCRE_EXPORT in their definitions.
2699 This is apparently wrong for the Windows case, so I have removed it.
2700 (It makes no difference on Unix-like systems.)
2701
2702 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2703 of recursive calls to match(). This is different to MATCH_LIMIT because
2704 that limits the total number of calls to match(), not all of which increase
2705 the depth of recursion. Limiting the recursion depth limits the amount of
2706 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2707 when PCRE is compiled, and changed at run time. A patch from Google adds
2708 this functionality to the C++ interface.
2709
2710 18. Changes to the handling of Unicode character properties:
2711
2712 (a) Updated the table to Unicode 4.1.0.
2713
2714 (b) Recognize characters that are not in the table as "Cn" (undefined).
2715
2716 (c) I revised the way the table is implemented to a much improved format
2717 which includes recognition of ranges. It now supports the ranges that
2718 are defined in UnicodeData.txt, and it also amalgamates other
2719 characters into ranges. This has reduced the number of entries in the
2720 table from around 16,000 to around 3,000, thus reducing its size
2721 considerably. I realized I did not need to use a tree structure after
2722 all - a binary chop search is just as efficient. Having reduced the
2723 number of entries, I extended their size from 6 bytes to 8 bytes to
2724 allow for more data.
2725
2726 (d) Added support for Unicode script names via properties such as \p{Han}.
2727
2728 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2729 matching that character.
2730
2731 20. When matching a repeated Unicode property with a minimum greater than zero,
2732 (for example \pL{2,}), PCRE could look past the end of the subject if it
2733 reached it while seeking the minimum number of characters. This could
2734 happen only if some of the characters were more than one byte long, because
2735 there is a check for at least the minimum number of bytes.
2736
2737 21. Refactored the implementation of \p and \P so as to be more general, to
2738 allow for more different types of property in future. This has changed the
2739 compiled form incompatibly. Anybody with saved compiled patterns that use
2740 \p or \P will have to recompile them.
2741
2742 22. Added "Any" and "L&" to the supported property types.
2743
2744 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2745 but give a compile time error if the value is greater than 0xff.
2746
2747 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2748 accidentally not being installed or uninstalled.
2749
2750 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2751 made were to insert the current release number. This seemed silly, because
2752 it made things harder for people building PCRE on systems that don't run
2753 "configure". I have turned pcre.h into a distributed file, no longer built
2754 by "configure", with the version identification directly included. There is
2755 no longer a pcre.h.in file.
2756
2757 However, this change necessitated a change to the pcre-config script as
2758 well. It is built from pcre-config.in, and one of the substitutions was the
2759 release number. I have updated configure.ac so that ./configure now finds
2760 the release number by grepping pcre.h.
2761
2762 26. Added the ability to run the tests under valgrind.
2763
2764
2765 Version 6.4 05-Sep-05
2766 ---------------------
2767
2768 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2769 "--" to be printed when multiple files were scanned, even when none of the
2770 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2771 consider it to be a bug, and have restored the previous behaviour.
2772
2773 2. A couple of code tidies to get rid of compiler warnings.
2774
2775 3. The pcretest program used to cheat by referring to symbols in the library
2776 whose names begin with _pcre_. These are internal symbols that are not
2777 really supposed to be visible externally, and in some environments it is
2778 possible to suppress them. The cheating is now confined to including
2779 certain files from the library's source, which is a bit cleaner.
2780
2781 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2782 file's purpose clearer.
2783
2784 5. Reorganized pcre_ucp_findchar().
2785
2786
2787 Version 6.3 15-Aug-05
2788 ---------------------
2789
2790 1. The file libpcre.pc.in did not have general read permission in the tarball.
2791
2792 2. There were some problems when building without C++ support:
2793
2794 (a) If C++ support was not built, "make install" and "make test" still
2795 tried to test it.
2796
2797 (b) There were problems when the value of CXX was explicitly set. Some
2798 changes have been made to try to fix these, and ...
2799
2800 (c) --disable-cpp can now be used to explicitly disable C++ support.
2801
2802 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2803 backslash in a target when C++ was disabled. This confuses some
2804 versions of "make", apparently. Using an intermediate variable solves
2805 this. (Same for CPP_LOBJ.)
2806
2807 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2808 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2809 necessary on certain architectures.
2810
2811 4. Added a setting of -export-symbols-regex to the link command to remove
2812 those symbols that are exported in the C sense, but actually are local
2813 within the library, and not documented. Their names all begin with
2814 "_pcre_". This is not a perfect job, because (a) we have to except some
2815 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2816 available (and never for static libraries). I have made a note to try to
2817 find a way round (a) in the future.
2818
2819
2820 Version 6.2 01-Aug-05
2821 ---------------------
2822
2823 1. There was no test for integer overflow of quantifier values. A construction
2824 such as {1111111111111111} would give undefined results. What is worse, if
2825 a minimum quantifier for a parenthesized subpattern overflowed and became
2826 negative, the calculation of the memory size went wrong. This could have
2827 led to memory overwriting.
2828
2829 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2830
2831 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2832 operating environments where this matters.
2833
2834 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2835 PCRE options from within the C++ wrapper.
2836
2837 5. Named capturing subpatterns were not being correctly counted when a pattern
2838 was compiled. This caused two problems: (a) If there were more than 100
2839 such subpatterns, the calculation of the memory needed for the whole
2840 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2841 back references of the form \12, where the number was greater than 9, were
2842 not recognized as back references, even though there were sufficient
2843 previous subpatterns.
2844
2845 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2846 versions of gcc, e.g. 2.95.4.
2847
2848
2849 Version 6.1 21-Jun-05
2850 ---------------------
2851
2852 1. There was one reference to the variable "posix" in pcretest.c that was not
2853 surrounded by "#if !defined NOPOSIX".
2854
2855 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2856 the cross-check on the old pcre_info() function, for the benefit of the
2857 cut-down version of PCRE that is currently imported into Exim.
2858
2859 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2860 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2861 patterns that start (?i) but I don't think that matters. The use of (?i) is
2862 just an example; this all applies to the other options as well.
2863
2864 4. Since libtool seems to echo the compile commands it is issuing, the output
2865 from "make" can be reduced a bit by putting "@" in front of each libtool
2866 compile command.
2867
2868 5. Patch from the folks at Google for configure.in to be a bit more thorough
2869 in checking for a suitable C++ installation before trying to compile the
2870 C++ stuff. This should fix a reported problem when a compiler was present,
2871 but no suitable headers.
2872
2873 6. The man pages all had just "PCRE" as their title. I have changed them to
2874 be the relevant file name. I have also arranged that these names are
2875 retained in the file doc/pcre.txt, which is a concatenation in text format
2876 of all the man pages except the little individual ones for each function.
2877
2878 7. The NON-UNIX-USE file had not been updated for the different set of source
2879 files that come with release 6. I also added a few comments about the C++
2880 wrapper.
2881
2882
2883 Version 6.0 07-Jun-05
2884 ---------------------
2885
2886 1. Some minor internal re-organization to help with my DFA experiments.
2887
2888 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2889 didn't matter for the library itself when fully configured, but did matter
2890 when compiling without UCP support, or within Exim, where the ucp files are
2891 not imported.
2892
2893 3. Refactoring of the library code to split up the various functions into
2894 different source modules. The addition of the new DFA matching code (see
2895 below) to a single monolithic source would have made it really too
2896 unwieldy, quite apart from causing all the code to be include in a
2897 statically linked application, when only some functions are used. This is
2898 relevant even without the DFA addition now that patterns can be compiled in
2899 one application and matched in another.
2900
2901 The downside of splitting up is that there have to be some external
2902 functions and data tables that are used internally in different modules of
2903 the library but which are not part of the API. These have all had their
2904 names changed to start with "_pcre_" so that they are unlikely to clash
2905 with other external names.
2906
2907 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2908 a different (DFA) algorithm. Although it is slower than the original
2909 function, it does have some advantages for certain types of matching
2910 problem.
2911
2912 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2913 including restarting after a partial match.
2914
2915 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2916 defined when compiling for Windows was sent to me. I have put it into the
2917 code, though I have no means of testing or verifying it.
2918
2919 7. Added the pcre_refcount() auxiliary function.
2920
2921 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2922 match before or at the first newline in the subject string. In pcretest,
2923 the /f option on a pattern can be used to set this.
2924
2925 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2926 would behave wrongly. This has been present in PCRE since release 4.0.
2927
2928 10. A number of changes to the pcregrep command:
2929
2930 (a) Refactored how -x works; insert ^(...)$ instead of setting
2931 PCRE_ANCHORED and checking the length, in preparation for adding
2932 something similar for -w.
2933
2934 (b) Added the -w (match as a word) option.
2935
2936 (c) Refactored the way lines are read and buffered so as to have more
2937 than one at a time available.
2938
2939 (d) Implemented a pcregrep test script.
2940
2941 (e) Added the -M (multiline match) option. This allows patterns to match
2942 over several lines of the subject. The buffering ensures that at least
2943 8K, or the rest of the document (whichever is the shorter) is available
2944 for matching (and similarly the previous 8K for lookbehind assertions).
2945
2946 (f) Changed the --help output so that it now says
2947
2948 -w, --word-regex(p)
2949
2950 instead of two lines, one with "regex" and the other with "regexp"
2951 because that confused at least one person since the short forms are the
2952 same. (This required a bit of code, as the output is generated
2953 automatically from a table. It wasn't just a text change.)
2954
2955 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2956 option but starts with a hyphen. Could be a pattern or a path name
2957 starting with a hyphen, for instance.
2958
2959 (h) "-" can be given as a file name to represent stdin.
2960
2961 (i) When file names are being printed, "(standard input)" is used for
2962 the standard input, for compatibility with GNU grep. Previously
2963 "<stdin>" was used.
2964
2965 (j) The option --label=xxx can be used to supply a name to be used for
2966 stdin when file names are being printed. There is no short form.
2967
2968 (k) Re-factored the options decoding logic because we are going to add
2969 two more options that take data. Such options can now be given in four
2970 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2971
2972 (l) Added the -A, -B, and -C options for requesting that lines of context
2973 around matches be printed.
2974
2975 (m) Added the -L option to print the names of files that do not contain
2976 any matching lines, that is, the complement of -l.
2977
2978 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2979 continue to scan other files.
2980
2981 (o) The -s option was incorrectly implemented. For compatibility with other
2982 greps, it now suppresses the error message for a non-existent or non-
2983 accessible file (but not the return code). There is a new option called
2984 -q that suppresses the output of matching lines, which was what -s was
2985 previously doing.
2986
2987 (p) Added --include and --exclude options to specify files for inclusion
2988 and exclusion when recursing.
2989
2990 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2991 Hopefully, it now does.
2992
2993 12. Missing cast in pcre_study().
2994
2995 13. Added an "uninstall" target to the makefile.
2996
2997 14. Replaced "extern" in the function prototypes in Makefile.in with
2998 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2999 world, but is set differently for Windows.
3000
3001 15. Added a second compiling function called pcre_compile2(). The only
3002 difference is that it has an extra argument, which is a pointer to an
3003 integer error code. When there is a compile-time failure, this is set
3004 non-zero, in addition to the error test pointer being set to point to an
3005 error message. The new argument may be NULL if no error number is required
3006 (but then you may as well call pcre_compile(), which is now just a
3007 wrapper). This facility is provided because some applications need a
3008 numeric error indication, but it has also enabled me to tidy up the way
3009 compile-time errors are handled in the POSIX wrapper.
3010
3011 16. Added VPATH=.libs to the makefile; this should help when building with one
3012 prefix path and installing with another. (Or so I'm told by someone who
3013 knows more about this stuff than I do.)
3014
3015 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3016 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3017 match everything, including newlines. This is not POSIX-compatible, but
3018 somebody wanted the feature. From pcretest it can be activated by using
3019 both the P and the s flags.
3020
3021 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3022
3023 19. libpcre.pc was being incorrectly installed as executable.
3024
3025 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3026 it now also looks for '\r' so that it will work unmodified on Windows.
3027
3028 21. Added Google's contributed C++ wrapper to the distribution.
3029
3030 22. Added some untidy missing memory free() calls in pcretest, to keep
3031 Electric Fence happy when testing.
3032
3033
3034
3035 Version 5.0 13-Sep-04
3036 ---------------------
3037
3038 1. Internal change: literal characters are no longer packed up into items
3039 containing multiple characters in a single byte-string. Each character
3040 is now matched using a separate opcode. However, there may be more than one
3041 byte in the character in UTF-8 mode.
3042
3043 2. The pcre_callout_block structure has two new fields: pattern_position and
3044 next_item_length. These contain the offset in the pattern to the next match
3045 item, and its length, respectively.
3046
3047 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3048 insertion of callouts before each pattern item. Added the /C option to
3049 pcretest to make use of this.
3050
3051 4. On the advice of a Windows user, the lines
3052
3053 #if defined(_WIN32) || defined(WIN32)
3054 _setmode( _fileno( stdout ), 0x8000 );
3055 #endif /* defined(_WIN32) || defined(WIN32) */
3056
3057 have been added to the source of pcretest. This apparently does useful
3058 magic in relation to line terminators.
3059
3060 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3061 for the benefit of those environments where the "b" makes a difference.
3062
3063 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3064 to know about it. I have put a hack into configure.in that adds in code
3065 to set GCC=yes if CC=icc. This seems to end up at a point in the
3066 generated configure script that is early enough to affect the setting of
3067 compiler options, which is what is needed, but I have no means of testing
3068 whether it really works. (The user who reported this had patched the
3069 generated configure script, which of course I cannot do.)
3070
3071 LATER: After change 22 below (new libtool files), the configure script
3072 seems to know about icc (and also ecc). Therefore, I have commented out
3073 this hack in configure.in.
3074
3075 7. Added support for pkg-config (2 patches were sent in).
3076
3077 8. Negated POSIX character classes that used a combination of internal tables
3078 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3079 [[:^ascii]]. Typically, they would match almost any characters. The other
3080 POSIX classes were not broken in this way.
3081
3082 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3083 to find the match, as PCRE was deluded into thinking that the match had to
3084 start at the start point or following a newline. The same bug applied to
3085 patterns with negative forward assertions or any backward assertions
3086 preceding ".*" at the start, unless the pattern required a fixed first
3087 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3088
3089 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3090 starting at the last subject character, bytes beyond the end of the subject
3091 string were read.
3092
3093 11. Renamed the variable "class" as "classbits" to make life easier for C++
3094 users. (Previously there was a macro definition, but it apparently wasn't
3095 enough.)
3096
3097 12. Added the new field "tables" to the extra data so that tables can be passed
3098 in at exec time, or the internal tables can be re-selected. This allows
3099 a compiled regex to be saved and re-used at a later time by a different
3100 program that might have everything at different addresses.
3101
3102 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3103 -R library as well as a -L library.
3104
3105 14. The debugging options of pcretest (-d on the command line or D on a
3106 pattern) showed incorrect output for anything following an extended class
3107 that contained multibyte characters and which was followed by a quantifier.
3108
3109 15. Added optional support for general category Unicode character properties
3110 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3111 support. It adds about 90K to the size of the library. The meanings of the
3112 inbuilt class escapes such as \d and \s have NOT been changed.
3113
3114 16. Updated pcredemo.c to include calls to free() to release the memory for the
3115 compiled pattern.
3116
3117 17. The generated file chartables.c was being created in the source directory
3118 instead of in the building directory. This caused the build to fail if the
3119 source directory was different from the building directory, and was
3120 read-only.
3121
3122 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3123 file. No doubt somebody will tell me if they don't make sense... Also added
3124 Dan Mooney's comments about building on OpenVMS.
3125
3126 19. Added support for partial matching via the PCRE_PARTIAL option for
3127 pcre_exec() and the \P data escape in pcretest.
3128
3129 20. Extended pcretest with 3 new pattern features:
3130
3131 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3132 write the compiled pattern to the file whose name is "rest-of-line".
3133 This is a straight binary dump of the data, with the saved pointer to
3134 the character tables forced to be NULL. The study data, if any, is
3135 written too. After writing, pcretest reads a new pattern.
3136
3137 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3138 compiled pattern from the given file. There must not be any
3139 occurrences of "<" in the file name (pretty unlikely); if there are,
3140 pcretest will instead treat the initial "<" as a pattern delimiter.
3141 After reading in the pattern, pcretest goes on to read data lines as
3142 usual.
3143
3144 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3145 and 16-bit fields in a compiled pattern, to simulate a pattern that
3146 was compiled on a host of opposite endianness.
3147
3148 21. The pcre-exec() function can now cope with patterns that were compiled on
3149 hosts of opposite endianness, with this restriction:
3150
3151 As for any compiled expression that is saved and used later, the tables
3152 pointer field cannot be preserved; the extra_data field in the arguments
3153 to pcre_exec() should be used to pass in a tables address if a value
3154 other than the default internal tables were used at compile time.
3155
3156 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3157 now diagnosed as an error. Previously, most of the time, a negative number
3158 would have been treated as zero, but if in addition "ovector" was passed as
3159 NULL, a crash could occur.
3160
3161 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3162 new versions from the libtool 1.5 distribution (the last one is a copy of
3163 a file called libtool.m4). This seems to have fixed the need to patch
3164 "configure" to support Darwin 1.3 (which I used to do). However, I still
3165 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3166 workstation).
3167
3168 24. Changed the PCRE licence to be the more standard "BSD" licence.
3169
3170
3171 Version 4.5 01-Dec-03
3172 ---------------------
3173
3174 1. There has been some re-arrangement of the code for the match() function so
3175 that it can be compiled in a version that does not call itself recursively.
3176 Instead, it keeps those local variables that need separate instances for
3177 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3178 needs to "recurse". Keeping track of where control must go is done by means
3179 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3180 hide most of the details from the main code, and operates only if
3181 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3182 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3183 operating.
3184
3185 To make it easier for callers to provide specially tailored get/free
3186 functions for this usage, two new functions, pcre_stack_malloc, and
3187 pcre_stack_free, are used. They are always called in strict stacking order,
3188 and the size of block requested is always the same.
3189
3190 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3191 PCRE has been compiled to use the stack or the heap for recursion. The
3192 -C option of pcretest uses this to show which version is compiled.
3193
3194 A new data escape \S, is added to pcretest; it causes the amounts of store
3195 obtained and freed by both kinds of malloc/free at match time to be added
3196 to the output.
3197
3198 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3199 what's available on my current Linux desktop machine.
3200
3201 3. When matching a UTF-8 string, the test for a valid string at the start has
3202 been extended. If start_offset is not zero, PCRE now checks that it points
3203 to a byte that is the start of a UTF-8 character. If not, it returns
3204 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3205 this is necessary because there may be backward assertions in the pattern.
3206 When matching the same subject several times, it may save resources to use
3207 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3208
3209 4. The code for checking the validity of UTF-8 strings has been tightened so
3210 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3211 containing "overlong sequences".
3212
3213 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3214 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3215 should have been "&", but it just so happened that all the cases this let
3216 through by mistake were picked up later in the function.
3217
3218 6. I had used a variable called "isblank" - this is a C99 function, causing
3219 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3220
3221 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3222 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3223 so the version is shown; (c) stop "make test" echoing "./RunTest".
3224
3225 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3226
3227 9. The prototype for memmove() for systems that don't have it was using
3228 size_t, but the inclusion of the header that defines size_t was later. I've
3229 moved the #includes for the C headers earlier to avoid this.
3230
3231 10. Added some adjustments to the code to make it easier to compiler on certain
3232 special systems:
3233
3234 (a) Some "const" qualifiers were missing.
3235 (b) Added the macro EXPORT before all exported functions; by default this
3236 is defined to be empty.
3237 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3238 that it reads its output file name as an argument instead of writing
3239 to the standard output and assuming this can be redirected.
3240
3241 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3242 class containing characters with values greater than 255, PCRE compilation
3243 went into a loop.
3244
3245 12. A recursive reference to a subpattern that was within another subpattern
3246 that had a minimum quantifier of zero caused PCRE to crash. For example,
3247 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3248 recursion. If the recursively-called subpattern itself had a zero repeat,
3249 that was OK.
3250
3251 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3252 buffer into which it was copied (for escape processing) was still set at
3253 1024, so long lines caused crashes.
3254
3255 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3256 "internal error: code overflow...". This applied to any character class
3257 that was followed by a possessive quantifier.
3258
3259 15. Modified the Makefile to add libpcre.la as a prerequisite for
3260 libpcreposix.la because I was told this is needed for a parallel build to
3261 work.
3262
3263 16. If a pattern that contained .* following optional items at the start was
3264 studied, the wrong optimizing data was generated, leading to matching
3265 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3266 matching string must start with a or b or c. The correct conclusion for
3267 this pattern is that a match can start with any character.
3268
3269
3270 Version 4.4 13-Aug-03
3271 ---------------------
3272
3273 1. In UTF-8 mode, a character class containing characters with values between
3274 127 and 255 was not handled correctly if the compiled pattern was studied.
3275 In fixing this, I have also improved the studying algorithm for such
3276 classes (slightly).
3277
3278 2. Three internal functions had redundant arguments passed to them. Removal
3279 might give a very teeny performance improvement.
3280
3281 3. Documentation bug: the value of the capture_top field in a callout is *one
3282 more than* the number of the hightest numbered captured substring.
3283
3284 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3285 in incorrectly linking with a previously installed version. They now link
3286 explicitly with libpcre.la.
3287
3288 5. configure.in no longer needs to recognize Cygwin specially.
3289
3290 6. A problem in pcre.in for Windows platforms is fixed.
3291
3292 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3293 pcretest, it used to include the size of the study block as part of its
3294 output. Unfortunately, the structure contains a field that has a different
3295 size on different hardware architectures. This meant that the tests that
3296 showed this size failed. As the block is currently always of a fixed size,
3297 this information isn't actually particularly useful in pcretest output, so
3298 I have just removed it.
3299
3300 8. Three pre-processor statements accidentally did not start in column 1.
3301 Sadly, there are *still* compilers around that complain, even though
3302 standard C has not required this for well over a decade. Sigh.
3303
3304 9. In pcretest, the code for checking callouts passed small integers in the
3305 callout_data field, which is a void * field. However, some picky compilers
3306 complained about the casts involved for this on 64-bit systems. Now
3307 pcretest passes the address of the small integer instead, which should get
3308 rid of the warnings.
3309
3310 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3311 both compile and run time, and gives an error if an invalid UTF-8 sequence
3312 is found. There is a option for disabling this check in cases where the
3313 string is known to be correct and/or the maximum performance is wanted.
3314
3315 11. In response to a bug report, I changed one line in Makefile.in from
3316
3317 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3318 to
3319 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3320
3321 to look similar to other lines, but I have no way of telling whether this
3322 is the right thing to do, as I do not use Windows. No doubt I'll get told
3323 if it's wrong...
3324
3325
3326 Version 4.3 21-May-03
3327 ---------------------
3328
3329 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3330 Makefile.
3331
3332 2. Some refactoring to improve the quality of the code:
3333
3334 (i) The utf8_table... variables are now declared "const".
3335
3336 (ii) The code for \cx, which used the "case flipping" table to upper case
3337 lower case letters, now just substracts 32. This is ASCII-specific,
3338 but the whole concept of \cx is ASCII-specific, so it seems
3339 reasonable.
3340
3341 (iii) PCRE was using its character types table to recognize decimal and
3342 hexadecimal digits in the pattern. This is silly, because it handles
3343 only 0-9, a-f, and A-F, but the character types table is locale-
3344 specific, which means strange things might happen. A private
3345 table is now used for this - though it costs 256 bytes, a table is
3346 much faster than multiple explicit tests. Of course, the standard
3347 character types table is still used for matching digits in subject
3348 strings against \d.
3349
3350 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3351 ending in _t are). So I've renamed it as ESC_tee.
3352
3353 3. The first argument for regexec() in the POSIX wrapper should have been
3354 defined as "const".
3355
3356 4. Changed pcretest to use malloc() for its buffers so that they can be
3357 Electric Fenced for debugging.
3358
3359 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3360 to read one or more bytes before the start of the subject string. Often this
3361 had no effect on PCRE's behaviour, but in some circumstances it could
3362 provoke a segmentation fault.
3363
3364 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3365 to try to read one or more bytes before the start of the subject string.
3366
3367 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3368 UTF-8 support could misbehave in various ways if the subject string
3369 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3370 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3371 back over UTF-8 characters.)
3372
3373
3374 Version 4.2 14-Apr-03
3375 ---------------------
3376
3377 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3378
3379 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3380 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3381 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3382 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3383 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3384 and BUILD_EXEEXT
3385 Note: automatic setting of the BUILD variables is not yet working
3386 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3387 compile-time but not at link-time
3388 [LINK]: use for linking executables only
3389 make different versions for Windows and non-Windows
3390 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3391 libraries
3392 [LINK_FOR_BUILD]: new variable
3393 [OBJEXT]: use throughout
3394 [EXEEXT]: use throughout
3395 <winshared>: new target
3396 <wininstall>: new target
3397 <dftables.o>: use native compiler
3398 <dftables>: use native linker
3399 <install>: handle Windows platform correctly
3400 <clean>: ditto
3401 <check>: ditto
3402 copy DLL to top builddir before testing
3403
3404 As part of these changes, -no-undefined was removed again. This was reported
3405 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3406 in any case.
3407
3408 3. Some tidies to get rid of compiler warnings:
3409
3410 . In the match_data structure, match_limit was an unsigned long int, whereas
3411 match_call_count was an int. I've made them both unsigned long ints.
3412
3413 . In pcretest the fact that a const uschar * doesn't automatically cast to
3414 a void * provoked a warning.
3415
3416 . Turning on some more compiler warnings threw up some "shadow" variables
3417 and a few more missing casts.
3418
3419 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3420 option, a class that contained a single character with a value between 128
3421 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3422
3423 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3424 option, a class that contained several characters, but with at least one
3425 whose value was between 128 and 255 caused PCRE to crash.
3426
3427
3428 Version 4.1 12-Mar-03
3429 ---------------------
3430
3431 1. Compiling with gcc -pedantic found a couple of places where casts were
3432 needed, and a string in dftables.c that was longer than standard compilers are
3433 required to support.
3434
3435 2. Compiling with Sun's compiler found a few more places where the code could
3436 be tidied up in order to avoid warnings.
3437
3438 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3439 first of these names is deprecated in the latest Autoconf in favour of the name
3440 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3441 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3442 analogy I have changed it to CFLAGS_FOR_BUILD.
3443
3444 4. Added -no-undefined to the linking command in the Makefile, because this is
3445 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3446 linking step for the pcreposix library.
3447
3448 5. PCRE was failing to diagnose the case of two named groups with the same
3449 name.
3450
3451 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3452 literal character that is needed in the subject for a match, and scans along to
3453 ensure that it is present before embarking on the full matching process. This
3454 saves time in cases of nested unlimited repeats that are never going to match.
3455 Problem: the scan can take a lot of time if the subject is very long (e.g.
3456 megabytes), thus penalizing straightforward matches. It is now done only if the
3457 amount of subject to be scanned is less than 1000 bytes.
3458
3459 7. A lesser problem with the same optimization is that it was recording the
3460 first character of an anchored pattern as "needed", thus provoking a search
3461 right along the subject, even when the first match of the pattern was going to
3462 fail. The "needed" character is now not set for anchored patterns, unless it
3463 follows something in the pattern that is of non-fixed length. Thus, it still
3464 fulfils its original purpose of finding quick non-matches in cases of nested
3465 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3466
3467
3468 Version 4.0 17-Feb-03
3469 ---------------------
3470
3471 1. If a comment in an extended regex that started immediately after a meta-item
3472 extended to the end of string, PCRE compiled incorrect data. This could lead to
3473 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3474
3475 2. Moved to autoconf 2.53 and libtool 1.4.2.
3476
3477 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3478 the special perltest8 script is no longer needed - all the tests can be run
3479 from a single perltest script.
3480
3481 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3482 by \s. It has now been removed in PCRE. This means it isn't recognized as
3483 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3484 class [:space:] *does* include VT, thereby creating a mess.
3485
3486 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3487 space and tab.
3488
3489 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3490 its new features into the main test script, reducing the number of scripts.
3491
3492 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3493 were backward compatible, and made the (?i) apply to the whole pattern, as if
3494 /i were given. Now it behaves more logically, and applies the option setting
3495 only to what follows. PCRE has been changed to follow suit. However, if it
3496 finds options settings right at the start of the pattern, it extracts them into
3497 the global options, as before. Thus, they show up in the info data.
3498
3499 8. Added support for the \Q...\E escape sequence. Characters in between are
3500 treated as literals. This is slightly different from Perl in that $ and @ are
3501 also handled as literals inside the quotes. In Perl, they will cause variable
3502 interpolation. Note the following examples:
3503
3504 Pattern PCRE matches Perl matches
3505
3506 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3507 \Qabc\$xyz\E abc\$xyz abc\$xyz
3508 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3509
3510 For compatibility with Perl, \Q...\E sequences are recognized inside character
3511 classes as well as outside them.
3512
3513 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3514 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3515 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3516 signed/unsigned warnings.
3517
3518 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3519 option for pcretest, so I've replaced it by a simple function that does just
3520 that job.
3521
3522 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3523 "pcregrep -".
3524
3525 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3526 Java package. This provides some syntactic sugar for simple cases of what my
3527 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3528 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3529 item, you can use this simplified notation. Note that only makes sense with
3530 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3531 greediness, whatever the setting of the PCRE_UNGREEDY option.
3532
3533 13. A change of greediness default within a pattern was not taking effect at
3534 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3535 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3536 was abstracted outside.
3537
3538 14. PCRE now supports the \G assertion. It is true when the current matching
3539 position is at the start point of the match. This differs from \A when the
3540 starting offset is non-zero. Used with the /g option of pcretest (or similar
3541 code), it works in the same way as it does for Perl's /g option. If all
3542 alternatives of a regex begin with \G, the expression is anchored to the start
3543 match position, and the "anchored" flag is set in the compiled expression.
3544
3545 15. Some bugs concerning the handling of certain option changes within patterns
3546 have been fixed. These applied to options other than (?ims). For example,
3547 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3548 been the other way round. Some of this was related to change 7 above.
3549
3550 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3551 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3552 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3553 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3554
3555 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3556 mode. Unlike ".", it always matches newline, whatever the setting of
3557 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3558 assertions. Perl allows it, but it doesn't (in general) work because it can't
3559 calculate the length of the lookbehind. At least, that's the case for Perl
3560 5.8.0 - I've been told they are going to document that it doesn't work in
3561 future.
3562
3563 18. Added an error diagnosis for escapes that PCRE does not support: these are
3564 \L, \l, \N, \P, \p, \U, \u, and \X.
3565
3566 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3567 reading past the end of the pattern in cases such as /[abcd/.
3568
3569 20. PCRE was getting more memory than necessary for patterns with classes that
3570 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3571
3572 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3573 compiling PCRE for use with Virtual Pascal.
3574
3575 22. Small fix to the Makefile to make it work properly if the build is done
3576 outside the source tree.
3577
3578 23. Added a new extension: a condition to go with recursion. If a conditional
3579 subpattern starts with (?(R) the "true" branch is used if recursion has
3580 happened, whereas the "false" branch is used only at the top level.
3581
3582 24. When there was a very long string of literal characters (over 255 bytes
3583 without UTF support, over 250 bytes with UTF support), the computation of how
3584 much memory was required could be incorrect, leading to segfaults or other
3585 strange effects.
3586
3587 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3588 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3589 there was a subsequent back reference to those brackets. This meant that, for
3590 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3591 possible to check for precisely this case. All we can do is abandon the
3592 optimization if .* occurs inside capturing brackets when there are any back
3593 references whatsoever. (See below for a better fix that came later.)
3594
3595 26. The handling of the optimization for finding the first character of a
3596 non-anchored pattern, and for finding a character that is required later in the
3597 match were failing in some cases. This didn't break the matching; it just
3598 failed to optimize when it could. The way this is done has been re-implemented.
3599
3600 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3601
3602 28. Added a new feature that provides some of the functionality that Perl
3603 provides with (?{...}). The facility is termed a "callout". The way it is done
3604 in PCRE is for the caller to provide an optional function, by setting
3605 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3606 global variable. By default it is unset, which disables all calling out. To get
3607 the function called, the regex must include (?C) at appropriate points. This
3608 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3609 This provides a means of identifying different callout points. When PCRE
3610 reaches such a point in the regex, if pcre_callout has been set, the external
3611 function is called. It is provided with data in a structure called
3612 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3613 matching continues; if it returns a non-zero value, the match at the current
3614 point fails. However, backtracking will occur if possible. [This was changed
3615 later and other features added - see item 49 below.]
3616
3617 29. pcretest is upgraded to test the callout functionality. It provides a
3618 callout function that displays information. By default, it shows the start of
3619 the match and the current position in the text. There are some new data escapes
3620 to vary what happens:
3621
3622 \C+ in addition, show current contents of captured substrings
3623 \C- do not supply a callout function
3624 \C!n return 1 when callout number n is reached
3625 \C!n!m return 1 when callout number n is reached for the mth time
3626
3627 30. If pcregrep was called with the -l option and just a single file name, it
3628 output "<stdin>" if a match was found, instead of the file name.
3629
3630 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3631 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3632 pcre_exec(). This saves a malloc/free per call. The default value of
3633 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3634 when configuring.
3635
3636 32. The default maximum size of a compiled pattern is 64K. There have been a
3637 few cases of people hitting this limit. The code now uses macros to handle the
3638 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3639 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3640 configuring. Tests 2 and 5 work only with 2-byte links because they output
3641 debugging information about compiled patterns.
3642
3643 33. Internal code re-arrangements:
3644
3645 (a) Moved the debugging function for printing out a compiled regex into
3646 its own source file (printint.c) and used #include to pull it into
3647 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3648 separate copies.
3649
3650 (b) Defined the list of op-code names for debugging as a macro in
3651 internal.h so that it is next to the definition of the opcodes.
3652
3653 (c) Defined a table of op-code lengths for simpler skipping along compiled
3654 code. This is again a macro in internal.h so that it is next to the
3655 definition of the opcodes.
3656
3657 34. Added support for recursive calls to individual subpatterns, along the
3658 lines of Robin Houston's patch (but implemented somewhat differently).
3659
3660 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3661 allow it to read and process whole directories in Win32. This code was
3662 contributed by Lionel Fourquaux; it has not been tested by me.
3663
3664 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3665 used to name a group. Names consist of alphanumerics and underscores, and must
3666 be unique. Back references use the syntax (?P=name) and recursive calls use
3667 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3668 numbers. The function pcre_fullinfo() can be used after compilation to extract
3669 a name/number map. There are three relevant calls:
3670
3671 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3672 PCRE_INFO_NAMECOUNT yields the number of entries
3673 PCRE_INFO_NAMETABLE yields a pointer to the map.
3674
3675 The map is a vector of fixed-size entries. The size of each entry depends on
3676 the length of the longest name used. The first two bytes of each entry are the
3677 group number, most significant byte first. There follows the corresponding
3678 name, zero terminated. The names are in alphabetical order.
3679
3680 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3681 case instead of 255. Making it the same both with and without UTF-8 support
3682 means that the same test output works with both.
3683
3684 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3685 calling malloc() with a zero argument.
3686
3687 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3688 optimization. I've improved things by keeping a bitmap of backreferences with
3689 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3690 fact referenced, the optimization can be applied. It is unlikely that a
3691 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3692 the match to follow \n) will appear inside brackets with a number greater than
3693 31, but if it does, any back reference > 31 suppresses the optimization.
3694
3695 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3696 of disabling numbered capturing parentheses. Any opening parenthesis that is
3697 not followed by ? behaves as if it were followed by ?: but named parentheses
3698 can still be used for capturing (and they will acquire numbers in the usual
3699 way).
3700
3701 41. Redesigned the return codes from the match() function into yes/no/error so
3702 that errors can be passed back from deep inside the nested calls. A malloc
3703 failure while inside a recursive subpattern call now causes the
3704 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3705
3706 42. It is now possible to set a limit on the number of times the match()
3707 function is called in a call to pcre_exec(). This facility makes it possible to
3708 limit the amount of recursion and backtracking, though not in a directly
3709 obvious way, because the match() function is used in a number of different
3710 circumstances. The count starts from zero for each position in the subject
3711 string (for non-anchored patterns). The default limit is, for compatibility, a
3712 large number, namely 10 000 000. You can change this in two ways:
3713
3714 (a) When configuring PCRE before making, you can use --with-match-limit=n
3715 to set a default value for the compiled library.
3716
3717 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3718 a different value is set. See 45 below.
3719
3720 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3721
3722 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3723 of things that can be changed at compile time. The first argument specifies
3724 what is wanted and the second points to where the information is to be placed.
3725 The current list of available information is:
3726
3727 PCRE_CONFIG_UTF8
3728
3729 The output is an integer that is set to one if UTF-8 support is available;
3730 otherwise it is set to zero.
3731
3732 PCRE_CONFIG_NEWLINE
3733
3734 The output is an integer that it set to the value of the code that is used for
3735 newline. It is either LF (10) or CR (13).
3736
3737 PCRE_CONFIG_LINK_SIZE
3738
3739 The output is an integer that contains the number of bytes used for internal
3740 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3741
3742 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3743
3744 The output is an integer that contains the threshold above which the POSIX
3745 interface uses malloc() for output vectors. See item 31 above.
3746
3747 PCRE_CONFIG_MATCH_LIMIT
3748
3749 The output is an unsigned integer that contains the default limit of the number
3750 of match() calls in a pcre_exec() execution. See 42 above.
3751
3752 44. pcretest has been upgraded by the addition of the -C option. This causes it
3753 to extract all the available output from the new pcre_config() function, and to
3754 output it. The program then exits immediately.
3755
3756 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3757 order to support additional features. One way would have been to define
3758 pcre_exec2() (for example) with extra arguments, but this would not have been
3759 extensible, and would also have required all calls to the original function to
3760 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3761 is used for passing in "extra" data from pcre_study().
3762
3763 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3764 contains the following fields:
3765
3766 flags a bitmap indicating which of the following fields are set
3767 study_data opaque data from pcre_study()
3768 match_limit a way of specifying a limit on match() calls for a specific
3769 call to pcre_exec()
3770 callout_data data for callouts (see 49 below)
3771
3772 The flag bits are also defined in pcre.h, and are
3773
3774 PCRE_EXTRA_STUDY_DATA
3775 PCRE_EXTRA_MATCH_LIMIT
3776 PCRE_EXTRA_CALLOUT_DATA
3777
3778 The pcre_study() function now returns one of these new pcre_extra blocks, with
3779 the actual study data pointed to by the study_data field, and the
3780 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3781 before. That is, this change is entirely upwards-compatible and requires no
3782 change to existing code.
3783
3784 If you want to pass in additional data to pcre_exec(), you can either place it
3785 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3786 block.
3787
3788 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3789 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3790 times with different match limits, until it finds the minimum value needed for
3791 pcre_exec() to complete. The value is then output. This can be instructive; for
3792 most simple matches the number is quite small, but for pathological cases it
3793 gets very large very quickly.
3794
3795 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3796 returns the size of the data block pointed to by the study_data field in a
3797 pcre_extra block, that is, the value that was passed as the argument to
3798 pcre_malloc() when PCRE was getting memory in which to place the information
3799 created by pcre_study(). The fourth argument should point to a size_t variable.
3800 pcretest has been extended so that this information is shown after a successful
3801 pcre_study() call when information about the compiled regex is being displayed.
3802
3803 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3804 because what follows is always an absolute path. (Later: it turns out that this
3805 is more than cosmetic for MinGW, because it doesn't like empty path
3806 components.)
3807
3808 49. Some changes have been made to the callout feature (see 28 above):
3809
3810 (i) A callout function now has three choices for what it returns:
3811
3812 0 => success, carry on matching
3813 > 0 => failure at this point, but backtrack if possible
3814 < 0 => serious error, return this value from pcre_exec()
3815
3816 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3817 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3818 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3819 use by callout functions. It will never be used by PCRE itself.
3820
3821 (ii) The pcre_extra structure (see 45 above) has a void * field called
3822 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3823 pcre_callout_block structure has a field of the same name. The contents of
3824 the field passed in the pcre_extra structure are passed to the callout
3825 function in the corresponding field in the callout block. This makes it
3826 easier to use the same callout-containing regex from multiple threads. For
3827 testing, the pcretest program has a new data escape
3828
3829 \C*n pass the number n (may be negative) as callout_data
3830
3831 If the callout function in pcretest receives a non-zero value as
3832 callout_data, it returns that value.
3833
3834 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3835 there were some redundant $(CFLAGS) in commands that are now specified as
3836 $(LINK), which already includes $(CFLAGS).
3837
3838 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3839 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3840 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3841 one-byte characters throughout. Note that case-insensitive matching applies
3842 only to characters whose values are less than 256. PCRE doesn't support the
3843 notion of cases for higher-valued characters.
3844
3845 (i) A character class whose characters are all within 0-255 is handled as
3846 a bit map, and the map is inverted for negative classes. Previously, a
3847 character > 255 always failed to match such a class; however it should
3848 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3849
3850 (ii) A negated character class with a single character < 255 is coded as
3851 "not this character" (OP_NOT). This wasn't working properly when the test
3852 character was multibyte, either singly or repeated.
3853
3854 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3855 mode, for example: \x{100}{2,3}.
3856
3857 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3858 singly or repeated) now correctly test multibyte characters. However,
3859 PCRE doesn't recognize any characters with values greater than 255 as
3860 digits, spaces, or word characters. Such characters always match \D, \S,
3861 and \W, and never match \d, \s, or \w.
3862
3863 (v) Classes may now contain characters and character ranges with values
3864 greater than 255. For example: [ab\x{100}-\x{400}].
3865
3866 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3867 PCRE in UTF-8 mode.
3868
3869 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3870 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3871 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3872 value.)
3873
3874 53. The single man page has become too large. I have therefore split it up into
3875 a number of separate man pages. These also give rise to individual HTML pages;
3876 these are now put in a separate directory, and there is an index.html page that
3877 lists them all. Some hyperlinking between the pages has been installed.
3878
3879 54. Added convenience functions for handling named capturing parentheses.
3880
3881 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3882 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3883 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3884 are faulted.
3885
3886 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3887 calling configure. These values are used when compiling the dftables.c program
3888 which is run to generate the source of the default character tables. They
3889 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3890 you will need to set these values.
3891
3892 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3893
3894
3895 Version 3.9 02-Jan-02
3896 ---------------------
3897
3898 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3899
3900 2. If --disable-static was given, the building process failed when trying to
3901 build pcretest and pcregrep. (For some reason it was using libtool to compile
3902 them, which is not right, as they aren't part of the library.)
3903
3904
3905 Version 3.8 18-Dec-01
3906 ---------------------
3907
3908 1. The experimental UTF-8 code was completely screwed up. It was packing the
3909 bytes in the wrong order. How dumb can you get?
3910
3911
3912 Version 3.7 29-Oct-01
3913 ---------------------
3914
3915 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3916 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3917 this didn't happen under Solaris 8, where I normally test things.
3918
3919 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3920 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3921 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3922 this fix an hour or so after the initial 3.7 release.)
3923
3924
3925 Version 3.6 23-Oct-01
3926 ---------------------
3927
3928 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3929 offsets passed as NULL with zero offset count.
3930
3931 2. The config.guess and config.sub files had not been updated when I moved to
3932 the latest autoconf.
3933
3934
3935 Version 3.5 15-Aug-01
3936 ---------------------
3937
3938 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3939 had been forgotten.
3940
3941 2. By using declared but undefined structures, we can avoid using "void"
3942 definitions in pcre.h while keeping the internal definitions of the structures
3943 private.
3944
3945 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3946 user point of view, this means that both static and shared libraries are built
3947 by default, but this can be individually controlled. More of the work of
3948 handling this static/shared cases is now inside libtool instead of PCRE's make
3949 file.
3950
3951 4. The pcretest utility is now installed along with pcregrep because it is
3952 useful for users (to test regexs) and by doing this, it automatically gets
3953 relinked by libtool. The documentation has been turned into a man page, so
3954 there are now .1, .txt, and .html versions in /doc.
3955
3956 5. Upgrades to pcregrep:
3957 (i) Added long-form option names like gnu grep.
3958 (ii) Added --help to list all options with an explanatory phrase.
3959 (iii) Added -r, --recursive to recurse into sub-directories.
3960 (iv) Added -f, --file to read patterns from a file.
3961
3962 6. pcre_exec() was referring to its "code" argument before testing that
3963 argument for NULL (and giving an error if it was NULL).
3964
3965 7. Upgraded Makefile.in to allow for compiling in a different directory from
3966 the source directory.
3967
3968 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3969 options bits, the pointer it was passed was to an int instead of to an unsigned
3970 long int. This mattered only on 64-bit systems.
3971
3972 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3973 generated) instead of pcre.in, which it its source. Also made the same change
3974 in several of the .c files.
3975
3976 10. A new release of gcc defines printf() as a macro, which broke pcretest
3977 because it had an ifdef in the middle of a string argument for printf(). Fixed
3978 by using separate calls to printf().
3979
3980 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3981 script, to force use of CR or LF instead of \n in the source. On non-Unix
3982 systems, the value can be set in config.h.
3983
3984 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3985 absolute limit. Changed the text of the error message to make this clear, and
3986 likewise updated the man page.
3987
3988 13. The limit of 99 on the number of capturing subpatterns has been removed.
3989 The new limit is 65535, which I hope will not be a "real" limit.
3990
3991
3992 Version 3.4 22-Aug-00
3993 ---------------------
3994
3995 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3996
3997 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3998
3999
4000 Version 3.3 01-Aug-00
4001 ---------------------
4002
4003 1. If an octal character was given, but the value was greater than \377, it
4004 was not getting masked to the least significant bits, as documented. This could
4005 lead to crashes in some systems.
4006
4007 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4008 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4009
4010 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4011 These just pass their arguments on to (pcre_free)(), but they are provided
4012 because some uses of PCRE bind it to non-C systems that can call its functions,
4013 but cannot call free() or pcre_free() directly.
4014
4015 4. Add "make test" as a synonym for "make check". Corrected some comments in
4016 the Makefile.
4017
4018 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4019 Makefile.
4020
4021 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4022 command called pgrep for grepping around the active processes.
4023
4024 7. Added the beginnings of support for UTF-8 character strings.
4025
4026 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4027 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4028 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4029 out for the ar command.)
4030
4031
4032 Version 3.2 12-May-00
4033 ---------------------
4034
4035 This is purely a bug fixing release.
4036
4037 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4038 of ZA. This was just one example of several cases that could provoke this bug,
4039 which was introduced by change 9 of version 2.00. The code for breaking
4040 infinite loops after an iteration that matches an empty string was't working
4041 correctly.
4042
4043 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4044 when matched against abbab (for example). After matching an empty string, it
4045 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4046 caused it to match further down the string than it should.
4047
4048 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4049 was there because it doesn't seem to be needed, and it causes trouble on some
4050 systems, as it is not a Standard C header. It has been removed.
4051
4052 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4053 were reported on the Macintosh. The changes were from
4054
4055 while ((c = *(++ptr)) != 0 && c != '\n');
4056 to
4057 while ((c = *(++ptr)) != 0 && c != '\n') ;
4058
4059 Totally extraordinary, but if that's what it takes...
4060
4061 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4062 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4063 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4064 assumes the way PCRE uses memmove() (always moving upwards).
4065
4066 6. PCRE is being used in one environment where strchr() is not available. There
4067 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4068 faster code anyway.
4069
4070
4071 Version 3.1 09-Feb-00
4072 ---------------------
4073
4074 The only change in this release is the fixing of some bugs in Makefile.in for
4075 the "install" target:
4076
4077 (1) It was failing to install pcreposix.h.
4078
4079 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4080
4081
4082 Version 3.0 01-Feb-00
4083 ---------------------
4084
4085 1. Add support for the /+ modifier to perltest (to output $` like it does in
4086 pcretest).
4087
4088 2. Add support for the /g modifier to perltest.
4089
4090 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4091 matches null strings.
4092
4093 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4094 pattern. Perl treats empty patterns specially - it reuses the most recent
4095 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4096 effect.
4097
4098 5. The POSIX interface was broken in that it was just handing over the POSIX
4099 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4100 required a bigger vector, with some working space on the end. This means that
4101 the POSIX wrapper now has to get and free some memory, and copy the results.
4102
4103 6. Added some simple autoconf support, placing the test data and the
4104 documentation in separate directories, re-organizing some of the
4105 information files, and making it build pcre-config (a GNU standard). Also added
4106 libtool support for building PCRE as a shared library, which is now the
4107 default.
4108
4109 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4110 09 are not valid octal constants. Single digits will be used for minor values
4111 less than 10.
4112
4113 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4114 existing programs that set these in the POSIX interface can use PCRE without
4115 modification.
4116
4117 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4118 return all that pcre_info() returns, plus additional data. The pcre_info()
4119 function is retained for compatibility, but is considered to be obsolete.
4120
4121 10. Added experimental recursion feature (?R) to handle one common case that
4122 Perl 5.6 will be able to do with (?p{...}).
4123
4124 11. Added support for POSIX character classes like [:alpha:], which Perl is
4125 adopting.
4126
4127
4128 Version 2.08 31-Aug-99
4129 ----------------------
4130
4131 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4132 trying to match at the startoffset position, but instead was moving forward to
4133 the next newline as if a previous match had failed.
4134
4135 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4136 and could get into a loop if a null string was matched other than at the start
4137 of the subject.
4138
4139 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4140 be distinguished at compile time, and for completeness also added PCRE_DATE.
4141
4142 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4143 in GnuWin32 environments.
4144
4145
4146 Version 2.07 29-Jul-99
4147 ----------------------
4148
4149 1. The documentation is now supplied in plain text form and HTML as well as in
4150 the form of man page sources.
4151
4152 2. C++ compilers don't like assigning (void *) values to other pointer types.
4153 In particular this affects malloc(). Although there is no problem in Standard
4154 C, I've put in casts to keep C++ compilers happy.
4155
4156 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4157 should be (const char *).
4158
4159 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4160 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4161 However, I haven't made this a standard facility. The documentation doesn't
4162 mention it, and the Makefile doesn't support it.
4163
4164 5. The Makefile now contains an "install" target, with editable destinations at
4165 the top of the file. The pcretest program is not installed.
4166
4167 6. pgrep -V now gives the PCRE version number and date.
4168
4169 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4170 causing the entire string to be ignored, instead of just the last character.
4171
4172 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4173 non-matching string, it can take a very, very long time, even for strings of
4174 quite modest length, because of the nested recursion. PCRE now does better in
4175 some of these cases. It does this by remembering the last required literal
4176 character in the pattern, and pre-searching the subject to ensure it is present
4177 before running the real match. In other words, it applies a heuristic to detect
4178 some types of certain failure quickly, and in the above example, if presented
4179 with a string that has no trailing " it gives "no match" very quickly.
4180
4181 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4182 other alternatives are tried instead.
4183
4184
4185 Version 2.06 09-Jun-99
4186 ----------------------
4187
4188 1. Change pcretest's output for amount of store used to show just the code
4189 space, because the remainder (the data block) varies in size between 32-bit and
4190 64-bit systems.
4191
4192 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4193 start matching at. This allows lookbehinds to work when searching for multiple
4194 occurrences in a string.
4195
4196 3. Added additional options to pcretest for testing multiple occurrences:
4197
4198 /+ outputs the rest of the string that follows a match
4199 /g loops for multiple occurrences, using the new startoffset argument
4200 /G loops for multiple occurrences by passing an incremented pointer
4201
4202 4. PCRE wasn't doing the "first character" optimization for patterns starting
4203 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4204 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4205 the letter 'x'. On long subject strings, this gives a significant speed-up.
4206
4207
4208 Version 2.05 21-Apr-99
4209 ----------------------
4210
4211 1. Changed the type of magic_number from int to long int so that it works
4212 properly on 16-bit systems.
4213
4214 2. Fixed a bug which caused patterns starting with .* not to work correctly
4215 when the subject string contained newline characters. PCRE was assuming
4216 anchoring for such patterns in all cases, which is not correct because .* will
4217 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4218 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4219 must be retried after every newline in the subject.
4220
4221
4222 Version 2.04 18-Feb-99
4223 ----------------------
4224
4225 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4226 computation of the store needed to hold the pattern was incorrect (too large).
4227 If such patterns were nested a few deep, this could multiply and become a real
4228 problem.
4229
4230 2. Added /M option to pcretest to show the memory requirement of a specific
4231 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4232
4233 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4234 compiled in such a way that the backtracking after subsequent failure was
4235 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4236 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4237
4238
4239 Version 2.03 02-Feb-99
4240 ----------------------
4241
4242 1. Fixed typo and small mistake in man page.
4243
4244 2. Added 4th condition (GPL supersedes if conflict) and created separate
4245 LICENCE file containing the conditions.
4246
4247 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4248 Perl, that is the internal \ allows the delimiter to be included in the
4249 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4250 the final delimiter, add \ to the end of the pattern (to test the error).
4251
4252 4. Added the convenience functions for extracting substrings after a successful
4253 match. Updated pcretest to make it able to test these functions.
4254
4255
4256 Version 2.02 14-Jan-99
4257 ----------------------
4258
4259 1. Initialized the working variables associated with each extraction so that
4260 their saving and restoring doesn't refer to uninitialized store.
4261
4262 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4263 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4264 fix the problem.
4265
4266 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4267 calls, and wasn't printing the correct value for compiling calls. Increased the
4268 default value of LOOPREPEAT, and the number of significant figures in the
4269 times.
4270
4271 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4272
4273 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4274 a building problem on Windows NT with a FAT file system.
4275
4276
4277 Version 2.01 21-Oct-98
4278 ----------------------
4279
4280 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4281 to character tables built by pcre_maketables() in the current locale. If NULL
4282 is passed, the default tables are used.
4283
4284
4285 Version 2.00 24-Sep-98
4286 ----------------------
4287
4288 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4289 it any more.
4290
4291 2. Allow quantification of (?>) groups, and make it work correctly.
4292
4293 3. The first character computation wasn't working for (?>) groups.
4294
4295 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4296 end of the subject) and add 5.005's \z, which really does match only at the
4297 very end of the subject.
4298
4299 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4300
4301 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4302 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4303 localized options. All options to pcre_study() were also removed.
4304
4305 7. Add other new features from 5.005:
4306
4307 $(?<= positive lookbehind
4308 $(?<! negative lookbehind
4309 (?imsx-imsx) added the unsetting capability
4310 such a setting is global if at outer level; local otherwise
4311 (?imsx-imsx:) non-capturing groups with option setting
4312 (?(cond)re|re) conditional pattern matching
4313
4314 A backreference to itself in a repeated group matches the previous
4315 captured string.
4316
4317 8. General tidying up of studying (both automatic and via "study")
4318 consequential on the addition of new assertions.
4319
4320 9. As in 5.005, unlimited repeated groups that could match an empty substring
4321 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4322 runtime if any iteration does actually match an empty substring.
4323
4324 10. Include the RunTest script in the distribution.
4325
4326 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4327 discrepancies, some of which were old and were also with respect to 5.004. They
4328 have now been fixed.
4329
4330
4331 Version 1.09 28-Apr-98
4332 ----------------------
4333
4334 1. A negated single character class followed by a quantifier with a minimum
4335 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4336 program crashes, or just wrong answers. This did not apply to negated classes
4337 containing more than one character, or to minima other than one.
4338
4339
4340 Version 1.08 27-Mar-98
4341 ----------------------
4342
4343 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4344
4345 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4346 latter must appear before anything that relies on it in the pattern.
4347
4348
4349 Version 1.07 16-Feb-98
4350 ----------------------
4351
4352 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4353 repeat of a potentially empty string).
4354
4355
4356 Version 1.06 23-Jan-98
4357 ----------------------
4358
4359 1. Added Markus Oberhumer's little patches for C++.
4360
4361 2. Literal strings longer than 255 characters were broken.
4362
4363
4364 Version 1.05 23-Dec-97
4365 ----------------------
4366
4367 1. Negated character classes containing more than one character were failing if
4368 PCRE_CASELESS was set at run time.
4369
4370
4371 Version 1.04 19-Dec-97
4372 ----------------------
4373
4374 1. Corrected the man page, where some "const" qualifiers had been omitted.
4375
4376 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4377 input syntax.
4378
4379 3. Fixed memory leak which occurred when a regex with back references was
4380 matched with an offsets vector that wasn't big enough. The temporary memory
4381 that is used in this case wasn't being freed if the match failed.
4382
4383 4. Tidied pcretest to ensure it frees memory that it gets.
4384
4385 5. Temporary memory was being obtained in the case where the passed offsets
4386 vector was exactly big enough.
4387
4388 6. Corrected definition of offsetof() from change 5 below.
4389
4390 7. I had screwed up change 6 below and broken the rules for the use of
4391 setjmp(). Now fixed.
4392
4393
4394 Version 1.03 18-Dec-97
4395 ----------------------
4396
4397 1. A erroneous regex with a missing opening parenthesis was correctly
4398 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4399 on some systems.
4400
4401 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4402 it was reported that one broken compiler failed on the former because "code" is
4403 also an independent variable.
4404
4405 3. The erroneous regex a[]b caused an array overrun reference.
4406
4407 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4408 fail on data ending with that character. (It was going on too far, and checking
4409 the next character, typically a binary zero.) This was specific to the
4410 optimized code for single-character negative classes.
4411
4412 5. Added a contributed patch from the TIN world which does the following:
4413
4414 + Add an undef for memmove, in case the the system defines a macro for it.
4415
4416 + Add a definition of offsetof(), in case there isn't one. (I don't know
4417 the reason behind this - offsetof() is part of the ANSI standard - but
4418 it does no harm).
4419
4420 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4421 most of the places where whitespace preceded '#'. I have given up and
4422 allowed the remaining 2 cases to be at the margin.
4423
4424 + Rename some variables in pcre to eliminate shadowing. This seems very
4425 pedantic, but does no harm, of course.
4426
4427 6. Moved the call to setjmp() into its own function, to get rid of warnings
4428 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4429
4430 7. Constructs such as \d{8,} were compiling into the equivalent of
4431 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4432 outcome, but in this particular case used more store than had been allocated,
4433 which caused the bug to be discovered because it threw up an internal error.
4434
4435 8. The debugging code in both pcre and pcretest for outputting the compiled
4436 form of a regex was going wrong in the case of back references followed by
4437 curly-bracketed repeats.
4438
4439
4440 Version 1.02 12-Dec-97
4441 ----------------------
4442
4443 1. Typos in pcre.3 and comments in the source fixed.
4444
4445 2. Applied a contributed patch to get rid of places where it used to remove
4446 'const' from variables, and fixed some signed/unsigned and uninitialized
4447 variable warnings.
4448
4449 3. Added the "runtest" target to Makefile.
4450
4451 4. Set default compiler flag to -O2 rather than just -O.
4452
4453
4454 Version 1.01 19-Nov-97
4455 ----------------------
4456
4457 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4458 like /([ab]*)*/, that is, for classes with more than one character in them.
4459
4460 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4461 as /((?>a*))*/ (a PCRE_EXTRA facility).
4462
4463
4464 Version 1.00 18-Nov-97
4465 ----------------------
4466
4467 1. Added compile-time macros to support systems such as SunOS4 which don't have
4468 memmove() or strerror() but have other things that can be used instead.
4469
4470 2. Arranged that "make clean" removes the executables.
4471
4472
4473 Version 0.99 27-Oct-97
4474 ----------------------
4475
4476 1. Fixed bug in code for optimizing classes with only one character. It was
4477 initializing a 32-byte map regardless, which could cause it to run off the end
4478 of the memory it had got.
4479
4480 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4481
4482
4483 Version 0.98 22-Oct-97
4484 ----------------------
4485
4486 1. Fixed bug in code for handling temporary memory usage when there are more
4487 back references than supplied space in the ovector. This could cause segfaults.
4488
4489
4490 Version 0.97 21-Oct-97
4491 ----------------------
4492
4493 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4494
4495 2. Optimized negated single characters not to use a bit map.
4496
4497 3. Brought error texts together as macro definitions; clarified some of them;
4498 fixed one that was wrong - it said "range out of order" when it meant "invalid
4499 escape sequence".
4500
4501 4. Changed some char * arguments to const char *.
4502
4503 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4504
4505 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4506 pcretest.
4507
4508
4509 Version 0.96 16-Oct-97
4510 ----------------------
4511
4512 1. Added a simple "pgrep" utility to the distribution.
4513
4514 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4515 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4516 where "ddd" means "one or more decimal digits".
4517
4518 3. Fixed serious bug. If a pattern had a back reference, but the call to
4519 pcre_exec() didn't supply a large enough ovector to record the related
4520 identifying subpattern, the match always failed. PCRE now remembers the number
4521 of the largest back reference, and gets some temporary memory in which to save
4522 the offsets during matching if necessary, in order to ensure that
4523 backreferences always work.
4524
4525 4. Increased the compatibility with Perl in a number of ways:
4526
4527 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4528 to request this handling. The option can be set at compile or exec time.
4529
4530 (b) $ matches before a terminating newline by default; an option
4531 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4532 mode). The option can be set at compile or exec time.
4533
4534 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4535 the same as Perl's. If the decimal number it represents is less than 10
4536 or there aren't that many previous left capturing parentheses, an octal
4537 escape is read. Inside a character class, it's always an octal escape,
4538 even if it is a single digit.
4539
4540 (d) An escaped but undefined alphabetic character is taken as a literal,
4541 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4542 escapes.
4543
4544 (e) {0} is now permitted. (The previous item is removed from the compiled
4545 pattern).
4546
4547 5. Changed all the names of code files so that the basic parts are no longer
4548 than 10 characters, and abolished the teeny "globals.c" file.
4549
4550 6. Changed the handling of character classes; they are now done with a 32-byte
4551 bit map always.
4552
4553 7. Added the -d and /D options to pcretest to make it possible to look at the
4554 internals of compilation without having to recompile pcre.
4555
4556
4557 Version 0.95 23-Sep-97
4558 ----------------------
4559
4560 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4561 \x20 at the start of a run of normal characters. These were being treated as
4562 real characters, instead of the source characters being re-checked.
4563
4564
4565 Version 0.94 18-Sep-97
4566 ----------------------
4567
4568 1. The functions are now thread-safe, with the caveat that the global variables
4569 containing pointers to malloc() and free() or alternative functions are the
4570 same for all threads.
4571
4572 2. Get pcre_study() to generate a bitmap of initial characters for non-
4573 anchored patterns when this is possible, and use it if passed to pcre_exec().
4574
4575
4576 Version 0.93 15-Sep-97
4577 ----------------------
4578
4579 1. /(b)|(:+)/ was computing an incorrect first character.
4580
4581 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4582 but not actually doing anything yet.
4583
4584 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4585 as Perl does (e.g. [-az] or [az-]).
4586
4587 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4588 all possible positions.
4589
4590 5. Split up into different modules to avoid including unneeded functions in a
4591 compiled binary. However, compile and exec are still in one module. The "study"
4592 function is split off.
4593
4594 6. The character tables are now in a separate module whose source is generated
4595 by an auxiliary program - but can then be edited by hand if required. There are
4596 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4597 toupper() in the code.
4598
4599 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4600 make them global. Abolish the function for setting them, as the caller can now
4601 set them directly.
4602
4603
4604 Version 0.92 11-Sep-97
4605 ----------------------
4606
4607 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4608 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4609
4610 2. Caseless matching was not working in character classes if the characters in
4611 the pattern were in upper case.
4612
4613 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4614
4615 4. Make PCRE_ANCHORED public and accept as a compile option.
4616
4617 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4618 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4619 pass them.
4620
4621 6. Give an error if bad option bits passed at compile or run time.
4622
4623 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4624 pcretest to cause it to pass that flag.
4625
4626 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4627 options, and the first character, if set.
4628
4629 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4630
4631
4632 Version 0.91 10-Sep-97
4633 ----------------------
4634
4635 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4636 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4637
4638 2. PCRE was looping on encountering an indefinitely repeated back reference to
4639 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4640 Perl does - treats the match as successful.
4641
4642 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5