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

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log


Revision 787 - (show annotations)
Tue Dec 6 15:37:24 2011 UTC (3 years, 8 months ago) by ph10
File size: 223038 byte(s)
Error occurred while calculating annotation data.
Fix uninitialized memory use when writing study data to file if no starting 
byte set exists.
1 ChangeLog for PCRE
2 ------------------
3
4 Version 8.21 05-Dec-2011
5 ------------------------
6
7 1. Updating the JIT compiler.
8
9 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
10 are added as well.
11
12 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
13 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
14 calling _pcre_jit_exec. Some extra comments are added.
15
16 4. (*MARK) settings inside atomic groups that do not contain any capturing
17 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
18 was introduced by change 18 for 8.20.
19
20 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
21 ECMA-262 standard.
22
23 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
24 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
25 This bug was probably introduced by change 9 of 8.13.
26
27 7. While fixing 6 above, I noticed that a number of other items were being
28 incorrectly rejected as "not fixed length". This arose partly because newer
29 opcodes had not been added to the fixed-length checking code. I have (a)
30 corrected the bug and added tests for these items, and (b) arranged for an
31 error to occur if an unknown opcode is encountered while checking for fixed
32 length instead of just assuming "not fixed length". The items that were
33 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
34 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
35 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
36
37 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
38 being incorrectly compiled and would have given unpredicatble results.
39
40 9. A possessively repeated subpattern with minimum repeat count greater than
41 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
42 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
43 the first (A) could occur when it should not.
44
45 10. Add a cast and remove a redundant test from the code.
46
47 11. JIT should use pcre_malloc/pcre_free for allocation.
48
49 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
50 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
51 is anything other than /usr, -L is still shown).
52
53 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
54
55 14. Perl does not support \N without a following name in a [] class; PCRE now
56 also gives an error.
57
58 15. If a forward reference was repeated with an upper limit of around 2000,
59 it caused the error "internal error: overran compiling workspace". The
60 maximum number of forward references (including repeats) was limited by the
61 internal workspace, and dependent on the LINK_SIZE. The code has been
62 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
63 the default depends on LINK_SIZE. There is a new upper limit (for safety)
64 of around 200,000 forward references. While doing this, I also speeded up
65 the filling in of repeated forward references.
66
67 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
68 incorrectly expecting the subject to contain another "a" after the start.
69
70 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
71 in the match, the SKIP should be ignored. This was not happening; instead
72 the SKIP was being treated as NOMATCH. For patterns such as
73 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
74 tested.
75
76 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
77 now much more compatible with Perl, in particular in cases where the result
78 is a non-match for a non-anchored pattern. For example, if
79 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
80 "m", where previously it did not return a name. A side effect of this
81 change is that for partial matches, the last encountered mark name is
82 returned, as for non matches. A number of tests that were previously not
83 Perl-compatible have been moved into the Perl-compatible test files. The
84 refactoring has had the pleasing side effect of removing one argument from
85 the match() function, thus reducing its stack requirements.
86
87 19. If the /S+ option was used in pcretest to study a pattern using JIT,
88 subsequent uses of /S (without +) incorrectly behaved like /S+.
89
90 21. Retrieve executable code size support for the JIT compiler and fixing
91 some warnings.
92
93 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
94 not work when the shorter character appeared right at the end of the
95 subject string.
96
97 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
98 systems.
99
100 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
101 output it when the /M option is used in pcretest.
102
103 25. The CheckMan script was not being included in the distribution. Also, added
104 an explicit "perl" to run Perl scripts from the PrepareRelease script
105 because this is reportedly needed in Windows.
106
107 26. If study data was being save in a file and studying had not found a set of
108 "starts with" bytes for the pattern, the data written to the file (though
109 never used) was taken from uninitialized memory and so caused valgrind to
110 complain.
111
112
113 Version 8.20 21-Oct-2011
114 ------------------------
115
116 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
117 a POSIX class. After further experiments with Perl, which convinced me that
118 Perl has bugs and confusions, a closing square bracket is no longer allowed
119 in a POSIX name. This bug also affected patterns with classes that started
120 with full stops.
121
122 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
123 captured substring, but while checking the failing first alternative,
124 substring 1 is temporarily captured. If the output vector supplied to
125 pcre_exec() was not big enough for this capture, the yield of the function
126 was still zero ("insufficient space for captured substrings"). This cannot
127 be totally fixed without adding another stack variable, which seems a lot
128 of expense for a edge case. However, I have improved the situation in cases
129 such as /(a)(b)x|abc/ matched against "abc", where the return code
130 indicates that fewer than the maximum number of slots in the ovector have
131 been set.
132
133 3. Related to (2) above: when there are more back references in a pattern than
134 slots in the output vector, pcre_exec() uses temporary memory during
135 matching, and copies in the captures as far as possible afterwards. It was
136 using the entire output vector, but this conflicts with the specification
137 that only 2/3 is used for passing back captured substrings. Now it uses
138 only the first 2/3, for compatibility. This is, of course, another edge
139 case.
140
141 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
142 main code base, and can be used by building with --enable-jit. When this is
143 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
144 runtime --no-jit option is given.
145
146 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
147 ovector, the return from the function was zero, implying that there were
148 other matches that did not fit. The correct "exactly full" value is now
149 returned.
150
151 6. If a subpattern that was called recursively or as a subroutine contained
152 (*PRUNE) or any other control that caused it to give a non-standard return,
153 invalid errors such as "Error -26 (nested recursion at the same subject
154 position)" or even infinite loops could occur.
155
156 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
157 computing the minimum length on reaching *ACCEPT, and so ended up with the
158 wrong value of 1 rather than 0. Further investigation indicates that
159 computing a minimum subject length in the presence of *ACCEPT is difficult
160 (think back references, subroutine calls), and so I have changed the code
161 so that no minimum is registered for a pattern that contains *ACCEPT.
162
163 8. If (*THEN) was present in the first (true) branch of a conditional group,
164 it was not handled as intended. [But see 16 below.]
165
166 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
167 Sheri Pierce.
168
169 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
170 the first byte in a match must be "a".
171
172 11. Change 17 for 8.13 increased the recursion depth for patterns like
173 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
174 pattern contains any instances of (*THEN). If it does not, the old
175 optimizations are restored. It would be nice to do this on a per-group
176 basis, but at the moment that is not feasible.
177
178 12. In some environments, the output of pcretest -C is CRLF terminated. This
179 broke RunTest's code that checks for the link size. A single white space
180 character after the value is now allowed for.
181
182 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
183 For "fr", it uses the Windows-specific input and output files.
184
185 14. If (*THEN) appeared in a group that was called recursively or as a
186 subroutine, it did not work as intended. [But see next item.]
187
188 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
189 pattern fragments (but not containing any | characters). If A and B are
190 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
191 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
192 D. In other words, Perl considers parentheses that do not contain any |
193 characters to be part of a surrounding alternative, whereas PCRE was
194 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
195 differently. PCRE now behaves in the same way as Perl, except in the case
196 of subroutine/recursion calls such as (?1) which have in any case always
197 been different (but PCRE had them first :-).
198
199 16. Related to 15 above: Perl does not treat the | in a conditional group as
200 creating alternatives. Such a group is treated in the same way as an
201 ordinary group without any | characters when processing (*THEN). PCRE has
202 been changed to match Perl's behaviour.
203
204 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
205 RunGrepTest script failed.
206
207 18. Change 22 for version 13 caused atomic groups to use more stack. This is
208 inevitable for groups that contain captures, but it can lead to a lot of
209 stack use in large patterns. The old behaviour has been restored for atomic
210 groups that do not contain any capturing parentheses.
211
212 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
213 suppress the check for a minimum subject length at run time. (If it was
214 given to pcre_exec() or pcre_dfa_exec() it did work.)
215
216 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
217 fail to work when decoding hex characters in data strings in EBCDIC
218 environments.
219
220 21. It appears that in at least one Mac OS environment, the isxdigit() function
221 is implemented as a macro that evaluates to its argument more than once,
222 contravening the C 90 Standard (I haven't checked a later standard). There
223 was an instance in pcretest which caused it to go wrong when processing
224 \x{...} escapes in subject strings. The has been rewritten to avoid using
225 things like p++ in the argument of isxdigit().
226
227
228 Version 8.13 16-Aug-2011
229 ------------------------
230
231 1. The Unicode data tables have been updated to Unicode 6.0.0.
232
233 2. Two minor typos in pcre_internal.h have been fixed.
234
235 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
236 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
237 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
238
239 4. There were a number of related bugs in the code for matching backrefences
240 caselessly in UTF-8 mode when codes for the characters concerned were
241 different numbers of bytes. For example, U+023A and U+2C65 are an upper
242 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
243 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
244 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
245 2-byte code at the end of the subject (it thought there wasn't enough data
246 left).
247
248 5. Comprehensive information about what went wrong is now returned by
249 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
250 as the output vector has at least 2 elements. The offset of the start of
251 the failing character and a reason code are placed in the vector.
252
253 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
254 now returned is for the first byte of the failing character, instead of the
255 last byte inspected. This is an incompatible change, but I hope it is small
256 enough not to be a problem. It makes the returned offset consistent with
257 pcre_exec() and pcre_dfa_exec().
258
259 7. pcretest now gives a text phrase as well as the error number when
260 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
261 failure, the offset and reason code are output.
262
263 8. When \R was used with a maximizing quantifier it failed to skip backwards
264 over a \r\n pair if the subsequent match failed. Instead, it just skipped
265 back over a single character (\n). This seems wrong (because it treated the
266 two characters as a single entity when going forwards), conflicts with the
267 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
268 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
269 has been changed.
270
271 9. Some internal refactoring has changed the processing so that the handling
272 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
273 time (the PCRE_DOTALL option was changed this way some time ago: version
274 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
275 which was always a bit of a fudge. It also means that there is one less
276 argument for the match() function, which reduces its stack requirements
277 slightly. This change also fixes an incompatibility with Perl: the pattern
278 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
279
280 10. More internal refactoring has drastically reduced the number of recursive
281 calls to match() for possessively repeated groups such as (abc)++ when
282 using pcre_exec().
283
284 11. While implementing 10, a number of bugs in the handling of groups were
285 discovered and fixed:
286
287 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
288 (a|)*(?1) gave a compile-time internal error.
289 ((a|)+)+ did not notice that the outer group could match an empty string.
290 (^a|^)+ was not marked as anchored.
291 (.*a|.*)+ was not marked as matching at start or after a newline.
292
293 12. Yet more internal refactoring has removed another argument from the match()
294 function. Special calls to this function are now indicated by setting a
295 value in a variable in the "match data" data block.
296
297 13. Be more explicit in pcre_study() instead of relying on "default" for
298 opcodes that mean there is no starting character; this means that when new
299 ones are added and accidentally left out of pcre_study(), testing should
300 pick them up.
301
302 14. The -s option of pcretest has been documented for ages as being an old
303 synonym of -m (show memory usage). I have changed it to mean "force study
304 for every regex", that is, assume /S for every regex. This is similar to -i
305 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
306 using it. It makes it easier to run collections of tests with and without
307 study enabled, and thereby test pcre_study() more easily. All the standard
308 tests are now run with and without -s (but some patterns can be marked as
309 "never study" - see 20 below).
310
311 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
312 restoration of the capturing data to the outer values was not happening
313 correctly.
314
315 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
316 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
317 pattern had matched an empty string, and so incorrectly returned a no
318 match.
319
320 17. There was optimizing code for the last branch of non-capturing parentheses,
321 and also for the obeyed branch of a conditional subexpression, which used
322 tail recursion to cut down on stack usage. Unfortunately, now that there is
323 the possibility of (*THEN) occurring in these branches, tail recursion is
324 no longer possible because the return has to be checked for (*THEN). These
325 two optimizations have therefore been removed. [But see 8.20/11 above.]
326
327 18. If a pattern containing \R was studied, it was assumed that \R always
328 matched two bytes, thus causing the minimum subject length to be
329 incorrectly computed because \R can also match just one byte.
330
331 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
332 was incorrectly computed.
333
334 20. If /S is present twice on a test pattern in pcretest input, it now
335 *disables* studying, thereby overriding the use of -s on the command line
336 (see 14 above). This is necessary for one or two tests to keep the output
337 identical in both cases.
338
339 21. When (*ACCEPT) was used in an assertion that matched an empty string and
340 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
341
342 22. When an atomic group that contained a capturing parenthesis was
343 successfully matched, but the branch in which it appeared failed, the
344 capturing was not being forgotten if a higher numbered group was later
345 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
346 group 1 to "a", when in fact it should be unset. This applied to multi-
347 branched capturing and non-capturing groups, repeated or not, and also to
348 positive assertions (capturing in negative assertions does not happen
349 in PCRE) and also to nested atomic groups.
350
351 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
352 subject after a captured substring, to make it easier to tell which of a
353 number of identical substrings has been captured.
354
355 24. The way atomic groups are processed by pcre_exec() has been changed so that
356 if they are repeated, backtracking one repetition now resets captured
357 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
358 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
359 "aaa". Previously, it would have been "a". As part of this code
360 refactoring, the way recursive calls are handled has also been changed.
361
362 25. If an assertion condition captured any substrings, they were not passed
363 back unless some other capturing happened later. For example, if
364 (?(?=(a))a) was matched against "a", no capturing was returned.
365
366 26. When studying a pattern that contained subroutine calls or assertions,
367 the code for finding the minimum length of a possible match was handling
368 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
369 group 1 called group 2 while simultaneously a separate group 2 called group
370 1). A stack overflow occurred in this case. I have fixed this by limiting
371 the recursion depth to 10.
372
373 27. Updated RunTest.bat in the distribution to the version supplied by Tom
374 Fortmann. This supports explicit test numbers on the command line, and has
375 argument validation and error reporting.
376
377 28. An instance of \X with an unlimited repeat could fail if at any point the
378 first character it looked at was a mark character.
379
380 29. Some minor code refactoring concerning Unicode properties and scripts
381 should reduce the stack requirement of match() slightly.
382
383 30. Added the '=' option to pcretest to check the setting of unused capturing
384 slots at the end of the pattern, which are documented as being -1, but are
385 not included in the return count.
386
387 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
388 compiled something random. Now it gives a compile-time error (as does
389 Perl).
390
391 32. A *MARK encountered during the processing of a positive assertion is now
392 recorded and passed back (compatible with Perl).
393
394 33. If --only-matching or --colour was set on a pcregrep call whose pattern
395 had alternative anchored branches, the search for a second match in a line
396 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
397 matched the line "0102" twice. The same bug affected patterns that started
398 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
399 twice.
400
401 34. Previously, PCRE did not allow quantification of assertions. However, Perl
402 does, and because of capturing effects, quantifying parenthesized
403 assertions may at times be useful. Quantifiers are now allowed for
404 parenthesized assertions.
405
406 35. A minor code tidy in pcre_compile() when checking options for \R usage.
407
408 36. \g was being checked for fancy things in a character class, when it should
409 just be a literal "g".
410
411 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
412 appearance of a nested POSIX class supersedes an apparent external class.
413 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
414 unescaped square brackets may also appear as part of class names. For
415 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
416 more like Perl. (But see 8.20/1 above.)
417
418 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
419 was because it thought it was \N{name}, which is not supported).
420
421 39. Add minix to OS list not supporting the -S option in pcretest.
422
423 40. PCRE tries to detect cases of infinite recursion at compile time, but it
424 cannot analyze patterns in sufficient detail to catch mutual recursions
425 such as ((?1))((?2)). There is now a runtime test that gives an error if a
426 subgroup is called recursively as a subpattern for a second time at the
427 same position in the subject string. In previous releases this might have
428 been caught by the recursion limit, or it might have run out of stack.
429
430 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
431 happen only once. PCRE was, however incorrectly giving a compile time error
432 "recursive call could loop indefinitely" because it cannot analyze the
433 pattern in sufficient detail. The compile time test no longer happens when
434 PCRE is compiling a conditional subpattern, but actual runaway loops are
435 now caught at runtime (see 40 above).
436
437 42. It seems that Perl allows any characters other than a closing parenthesis
438 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
439 has been changed to be the same.
440
441 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
442 as not to get warnings when autogen.sh is called. Also changed
443 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
444
445 44. To help people who use pcregrep to scan files containing exceedingly long
446 lines, the following changes have been made:
447
448 (a) The default value of the buffer size parameter has been increased from
449 8K to 20K. (The actual buffer used is three times this size.)
450
451 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
452 PCRE is built.
453
454 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
455 to be set at run time.
456
457 (d) Numerical values in pcregrep options can be followed by K or M, for
458 example --buffer-size=50K.
459
460 (e) If a line being scanned overflows pcregrep's buffer, an error is now
461 given and the return code is set to 2.
462
463 45. Add a pointer to the latest mark to the callout data block.
464
465 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
466 partial match of an empty string instead of no match. This was specific to
467 the use of ".".
468
469 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
470 complete match instead of a partial match. This bug was dependent on both
471 the PCRE_UTF8 and PCRE_DOTALL options being set.
472
473 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
474 starting byte set, because \b was not being ignored.
475
476
477 Version 8.12 15-Jan-2011
478 ------------------------
479
480 1. Fixed some typos in the markup of the man pages, and wrote a script that
481 checks for such things as part of the documentation building process.
482
483 2. On a big-endian 64-bit system, pcregrep did not correctly process the
484 --match-limit and --recursion-limit options (added for 8.11). In
485 particular, this made one of the standard tests fail. (The integer value
486 went into the wrong half of a long int.)
487
488 3. If the --colour option was given to pcregrep with -v (invert match), it
489 did strange things, either producing crazy output, or crashing. It should,
490 of course, ignore a request for colour when reporting lines that do not
491 match.
492
493 4. Another pcregrep bug caused similar problems if --colour was specified with
494 -M (multiline) and the pattern match finished with a line ending.
495
496 5. In pcregrep, when a pattern that ended with a literal newline sequence was
497 matched in multiline mode, the following line was shown as part of the
498 match. This seems wrong, so I have changed it.
499
500 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
501 the check for further matches in the same line (so they could be coloured)
502 to overrun the end of the current line. If another match was found, it was
503 incorrectly shown (and then shown again when found in the next line).
504
505 7. If pcregrep was compiled under Windows, there was a reference to the
506 function pcregrep_exit() before it was defined. I am assuming this was
507 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
508 reported by a user. I've moved the definition above the reference.
509
510
511 Version 8.11 10-Dec-2010
512 ------------------------
513
514 1. (*THEN) was not working properly if there were untried alternatives prior
515 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
516 backtracked to try for "b" instead of moving to the next alternative branch
517 at the same level (in this case, to look for "c"). The Perl documentation
518 is clear that when (*THEN) is backtracked onto, it goes to the "next
519 alternative in the innermost enclosing group".
520
521 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
522 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
523 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
524 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
525 (*THEN).
526
527 3. If \s appeared in a character class, it removed the VT character from
528 the class, even if it had been included by some previous item, for example
529 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
530 of \s, but is part of the POSIX "space" class.)
531
532 4. A partial match never returns an empty string (because you can always
533 match an empty string at the end of the subject); however the checking for
534 an empty string was starting at the "start of match" point. This has been
535 changed to the "earliest inspected character" point, because the returned
536 data for a partial match starts at this character. This means that, for
537 example, /(?<=abc)def/ gives a partial match for the subject "abc"
538 (previously it gave "no match").
539
540 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
541 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
542 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
543 has an implication that the given string is incomplete (because a partial
544 match is preferred over a full match). For this reason, these items now
545 give a partial match in this situation. [Aside: previously, the one case
546 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
547 match rather than a full match, which was wrong by the old rules, but is
548 now correct.]
549
550 6. There was a bug in the handling of #-introduced comments, recognized when
551 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
552 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
553 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
554 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
555 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
556 places in pcre_compile().
557
558 7. Related to (6) above, when pcre_compile() was skipping #-introduced
559 comments when looking ahead for named forward references to subpatterns,
560 the only newline sequence it recognized was NL. It now handles newlines
561 according to the set newline convention.
562
563 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
564 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
565 cater for a lack of strerror(). These oversights have been fixed.
566
567 9. Added --match-limit and --recursion-limit to pcregrep.
568
569 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
570
571 11. When the -o option was used, pcregrep was setting a return code of 1, even
572 when matches were found, and --line-buffered was not being honoured.
573
574 12. Added an optional parentheses number to the -o and --only-matching options
575 of pcregrep.
576
577 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
578 can match an empty string. The code to do it in pcretest and pcredemo
579 needed fixing:
580
581 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
582 only one byte after an empty string match just before CRLF (this case
583 just got forgotten; "any" and "anycrlf" were OK).
584
585 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
586 mode when an empty string match preceded an ASCII character followed by
587 a non-ASCII character. (The code for advancing by one character rather
588 than one byte was nonsense.)
589
590 (c) The pcredemo.c sample program did not have any code at all to handle
591 the cases when CRLF is a valid newline sequence.
592
593 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
594 as a starting offset was within the subject string. There is now a new
595 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
596 negative or greater than the length of the string. In order to test this,
597 pcretest is extended to allow the setting of negative starting offsets.
598
599 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
600 starting offset points to the beginning of a UTF-8 character was
601 unnecessarily clumsy. I tidied it up.
602
603 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
604 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
605
606 17. Nobody had reported that the --include_dir option, which was added in
607 release 7.7 should have been called --include-dir (hyphen, not underscore)
608 for compatibility with GNU grep. I have changed it to --include-dir, but
609 left --include_dir as an undocumented synonym, and the same for
610 --exclude-dir, though that is not available in GNU grep, at least as of
611 release 2.5.4.
612
613 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
614 characters from a string of bytes) have been redefined so as not to use
615 loops, in order to improve performance in some environments. At the same
616 time, I abstracted some of the common code into auxiliary macros to save
617 repetition (this should not affect the compiled code).
618
619 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
620 compile-time error is now given if \c is not followed by an ASCII
621 character, that is, a byte less than 128. (In EBCDIC mode, the code is
622 different, and any byte value is allowed.)
623
624 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
625 START_OPTIMIZE option, which is now allowed at compile time - but just
626 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
627 to pcregrep and other applications that have no direct access to PCRE
628 options. The new /Y option in pcretest sets this option when calling
629 pcre_compile().
630
631 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
632 back references. Groups containing recursive back references were forced to
633 be atomic by that change, but in the case of named groups, the amount of
634 memory required was incorrectly computed, leading to "Failed: internal
635 error: code overflow". This has been fixed.
636
637 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
638 pcretest.c, to avoid build problems in some Borland environments.
639
640
641 Version 8.10 25-Jun-2010
642 ------------------------
643
644 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
645 THEN.
646
647 2. (*ACCEPT) was not working when inside an atomic group.
648
649 3. Inside a character class, \B is treated as a literal by default, but
650 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
651 causes the error). The code is unchanged, but I tidied the documentation.
652
653 4. Inside a character class, PCRE always treated \R and \X as literals,
654 whereas Perl faults them if its -w option is set. I have changed PCRE so
655 that it faults them when PCRE_EXTRA is set.
656
657 5. Added support for \N, which always matches any character other than
658 newline. (It is the same as "." when PCRE_DOTALL is not set.)
659
660 6. When compiling pcregrep with newer versions of gcc which may have
661 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
662 declared with attribute warn_unused_result" were given. Just casting the
663 result to (void) does not stop the warnings; a more elaborate fudge is
664 needed. I've used a macro to implement this.
665
666 7. Minor change to pcretest.c to avoid a compiler warning.
667
668 8. Added four artifical Unicode properties to help with an option to make
669 \s etc use properties (see next item). The new properties are: Xan
670 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
671
672 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
673 use Unicode properties. (*UCP) at the start of a pattern can be used to set
674 this option. Modified pcretest to add /W to test this facility. Added
675 REG_UCP to make it available via the POSIX interface.
676
677 10. Added --line-buffered to pcregrep.
678
679 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
680 studied, and the match started with a letter with a code point greater than
681 127 whose first byte was different to the first byte of the other case of
682 the letter, the other case of this starting letter was not recognized
683 (#976).
684
685 12. If a pattern that was studied started with a repeated Unicode property
686 test, for example, \p{Nd}+, there was the theoretical possibility of
687 setting up an incorrect bitmap of starting bytes, but fortunately it could
688 not have actually happened in practice until change 8 above was made (it
689 added property types that matched character-matching opcodes).
690
691 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
692 possible starting bytes for non-anchored patterns.
693
694 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
695 \R, and also a number of cases that involve Unicode properties, both
696 explicit and implicit when PCRE_UCP is set.
697
698 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
699 input, it could crash or give wrong results if characters with values
700 greater than 0xc0 were present in the subject string. (Detail: it assumed
701 UTF-8 input when processing these items.)
702
703 16. Added a lot of (int) casts to avoid compiler warnings in systems where
704 size_t is 64-bit (#991).
705
706 17. Added a check for running out of memory when PCRE is compiled with
707 --disable-stack-for-recursion (#990).
708
709 18. If the last data line in a file for pcretest does not have a newline on
710 the end, a newline was missing in the output.
711
712 19. The default pcre_chartables.c file recognizes only ASCII characters (values
713 less than 128) in its various bitmaps. However, there is a facility for
714 generating tables according to the current locale when PCRE is compiled. It
715 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
716 characters, are recognized by isspace() and therefore were getting set in
717 these tables, and indeed these tables seem to approximate to ISO 8859. This
718 caused a problem in UTF-8 mode when pcre_study() was used to create a list
719 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
720 which of course cannot start UTF-8 characters. I have changed the code so
721 that only real ASCII characters (less than 128) and the correct starting
722 bytes for UTF-8 encodings are set for characters greater than 127 when in
723 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
724 altogether.)
725
726 20. Added the /T option to pcretest so as to be able to run tests with non-
727 standard character tables, thus making it possible to include the tests
728 used for 19 above in the standard set of tests.
729
730 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
731 reference to a subpattern the other side of a comment that contains an
732 opening parenthesis caused either an internal compiling error, or a
733 reference to the wrong subpattern.
734
735
736 Version 8.02 19-Mar-2010
737 ------------------------
738
739 1. The Unicode data tables have been updated to Unicode 5.2.0.
740
741 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
742 configured.
743
744 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
745 original author of that file, following a query about its status.
746
747 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
748 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
749
750 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
751 quantifier applied to a forward-referencing subroutine call, could compile
752 incorrect code or give the error "internal error: previously-checked
753 referenced subpattern not found".
754
755 6. Both MS Visual Studio and Symbian OS have problems with initializing
756 variables to point to external functions. For these systems, therefore,
757 pcre_malloc etc. are now initialized to local functions that call the
758 relevant global functions.
759
760 7. There were two entries missing in the vectors called coptable and poptable
761 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
762 I've fixed the data, and added a kludgy way of testing at compile time that
763 the lengths are correct (equal to the number of opcodes).
764
765 8. Following on from 7, I added a similar kludge to check the length of the
766 eint vector in pcreposix.c.
767
768 9. Error texts for pcre_compile() are held as one long string to avoid too
769 much relocation at load time. To find a text, the string is searched,
770 counting zeros. There was no check for running off the end of the string,
771 which could happen if a new error number was added without updating the
772 string.
773
774 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
775
776 11. \K was not working if it appeared in an atomic group or in a group that
777 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
778 \K is "not well defined" if used in an assertion. PCRE now accepts it if
779 the assertion is positive, but not if it is negative.
780
781 12. Change 11 fortuitously reduced the size of the stack frame used in the
782 "match()" function of pcre_exec.c by one pointer. Forthcoming
783 implementation of support for (*MARK) will need an extra pointer on the
784 stack; I have reserved it now, so that the stack frame size does not
785 decrease.
786
787 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
788 item in branch that calls a recursion is a subroutine call - as in the
789 second branch in the above example - was incorrectly given the compile-
790 time error "recursive call could loop indefinitely" because pcre_compile()
791 was not correctly checking the subroutine for matching a non-empty string.
792
793 14. The checks for overrunning compiling workspace could trigger after an
794 overrun had occurred. This is a "should never occur" error, but it can be
795 triggered by pathological patterns such as hundreds of nested parentheses.
796 The checks now trigger 100 bytes before the end of the workspace.
797
798 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
799
800
801 Version 8.01 19-Jan-2010
802 ------------------------
803
804 1. If a pattern contained a conditional subpattern with only one branch (in
805 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
806 computed the wrong minimum data length (which is of course zero for such
807 subpatterns). This could cause incorrect "no match" results.
808
809 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
810 the pattern is reset in the first branch, pcre_compile() failed with
811 "internal error: code overflow at offset...". This happened only when
812 the reset was to the original external option setting. (An optimization
813 abstracts leading options settings into an external setting, which was the
814 cause of this.)
815
816 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
817 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
818 assertion pattern did not match (meaning that the assertion was true), it
819 was incorrectly treated as false if the SKIP had been reached during the
820 matching. This also applied to assertions used as conditions.
821
822 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
823 assertion subpattern, including such a pattern used as a condition,
824 unpredictable results occurred, instead of the error return
825 PCRE_ERROR_DFA_UITEM.
826
827 5. The C++ GlobalReplace function was not working like Perl for the special
828 situation when an empty string is matched. It now does the fancy magic
829 stuff that is necessary.
830
831 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
832 removed. (These were left over from very, very early versions of PCRE.)
833
834 7. Some cosmetic changes to the code to make life easier when compiling it
835 as part of something else:
836
837 (a) Change DEBUG to PCRE_DEBUG.
838
839 (b) In pcre_compile(), rename the member of the "branch_chain" structure
840 called "current" as "current_branch", to prevent a collision with the
841 Linux macro when compiled as a kernel module.
842
843 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
844 prevent a collision with the Linux macro when compiled as a kernel
845 module.
846
847 8. In pcre_compile() there are some checks for integer overflows that used to
848 cast potentially large values to (double). This has been changed to that
849 when building, a check for int64_t is made, and if it is found, it is used
850 instead, thus avoiding the use of floating point arithmetic. (There is no
851 other use of FP in PCRE.) If int64_t is not found, the fallback is to
852 double.
853
854 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
855 2005 (difference between two addresses compared to an unsigned value).
856
857 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
858 custom one, because of the following reported problem in Windows:
859
860 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
861 under Win32.
862 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
863 therefore missing the function definition.
864 - The compiler thus generates a "C" signature for the test function.
865 - The linker fails to find the "C" function.
866 - PCRE fails to configure if asked to do so against libbz2.
867
868 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
869 messages were output:
870
871 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
872 rerunning libtoolize, to keep the correct libtool macros in-tree.
873 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
874
875 I have done both of these things.
876
877 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
878 most of the time, it *can* run out if it is given a pattern that contains a
879 runaway infinite recursion. I updated the discussion in the pcrestack man
880 page.
881
882 13. Now that we have gone to the x.xx style of version numbers, the minor
883 version may start with zero. Using 08 or 09 is a bad idea because users
884 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
885 interpreted as invalid octal numbers. I've updated the previous comment in
886 configure.ac, and also added a check that gives an error if 08 or 09 are
887 used.
888
889 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
890 causing partial matching to fail when the end of the subject matched \W
891 in a UTF-8 pattern where \W was quantified with a minimum of 3.
892
893 15. There were some discrepancies between the declarations in pcre_internal.h
894 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
895 their definitions. The declarations used "const uschar *" and the
896 definitions used USPTR. Even though USPTR is normally defined as "const
897 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
898 reported that: "This difference in casting confuses some C++ compilers, for
899 example, SunCC recognizes above declarations as different functions and
900 generates broken code for hbpcre." I have changed the declarations to use
901 USPTR.
902
903 16. GNU libtool is named differently on some systems. The autogen.sh script now
904 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
905 (FreeBSD).
906
907 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
908 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
909 comment: "Figure out how to create a longlong from a string: strtoll and
910 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
911 instance, but it only takes 2 args instead of 3!"
912
913 18. A subtle bug concerned with back references has been fixed by a change of
914 specification, with a corresponding code fix. A pattern such as
915 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
916 refers, was giving matches when it shouldn't. For example, xa=xaaa would
917 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
918 same bug. Such groups have to be quantified to be useful, or contained
919 inside another quantified group. (If there's no repetition, the reference
920 can never match.) The problem arises because, having left the group and
921 moved on to the rest of the pattern, a later failure that backtracks into
922 the group uses the captured value from the final iteration of the group
923 rather than the correct earlier one. I have fixed this in PCRE by forcing
924 any group that contains a reference to itself to be an atomic group; that
925 is, there cannot be any backtracking into it once it has completed. This is
926 similar to recursive and subroutine calls.
927
928
929 Version 8.00 19-Oct-09
930 ----------------------
931
932 1. The table for translating pcre_compile() error codes into POSIX error codes
933 was out-of-date, and there was no check on the pcre_compile() error code
934 being within the table. This could lead to an OK return being given in
935 error.
936
937 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
938 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
939 in a Windows environment.
940
941 3. The pcregrep --count option prints the count for each file even when it is
942 zero, as does GNU grep. However, pcregrep was also printing all files when
943 --files-with-matches was added. Now, when both options are given, it prints
944 counts only for those files that have at least one match. (GNU grep just
945 prints the file name in this circumstance, but including the count seems
946 more useful - otherwise, why use --count?) Also ensured that the
947 combination -clh just lists non-zero counts, with no names.
948
949 4. The long form of the pcregrep -F option was incorrectly implemented as
950 --fixed_strings instead of --fixed-strings. This is an incompatible change,
951 but it seems right to fix it, and I didn't think it was worth preserving
952 the old behaviour.
953
954 5. The command line items --regex=pattern and --regexp=pattern were not
955 recognized by pcregrep, which required --regex pattern or --regexp pattern
956 (with a space rather than an '='). The man page documented the '=' forms,
957 which are compatible with GNU grep; these now work.
958
959 6. No libpcreposix.pc file was created for pkg-config; there was just
960 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
961
962 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
963 when UCP support is not needed, by modifying the Python script that
964 generates it from Unicode data files. This should not matter if the module
965 is correctly used as a library, but I received one complaint about 50K of
966 unwanted data. My guess is that the person linked everything into his
967 program rather than using a library. Anyway, it does no harm.
968
969 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
970 was a minimum greater than 1 for a wide character in a possessive
971 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
972 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
973 character. Chaos in the form of incorrect output or a compiling loop could
974 result.
975
976 9. The restrictions on what a pattern can contain when partial matching is
977 requested for pcre_exec() have been removed. All patterns can now be
978 partially matched by this function. In addition, if there are at least two
979 slots in the offset vector, the offset of the earliest inspected character
980 for the match and the offset of the end of the subject are set in them when
981 PCRE_ERROR_PARTIAL is returned.
982
983 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
984 synonymous with PCRE_PARTIAL, for backwards compatibility, and
985 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
986 and may be more useful for multi-segment matching.
987
988 11. Partial matching with pcre_exec() is now more intuitive. A partial match
989 used to be given if ever the end of the subject was reached; now it is
990 given only if matching could not proceed because another character was
991 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
992 string "Z", which now yields "no match" instead of "partial match". In the
993 case of pcre_dfa_exec(), "no match" is given if every matching path for the
994 final character ended with (*FAIL).
995
996 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
997 if the pattern had a "must contain" character that was already found in the
998 earlier partial match, unless partial matching was again requested. For
999 example, with the pattern /dog.(body)?/, the "must contain" character is
1000 "g". If the first part-match was for the string "dog", restarting with
1001 "sbody" failed. This bug has been fixed.
1002
1003 13. The string returned by pcre_dfa_exec() after a partial match has been
1004 changed so that it starts at the first inspected character rather than the
1005 first character of the match. This makes a difference only if the pattern
1006 starts with a lookbehind assertion or \b or \B (\K is not supported by
1007 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1008 matching functions compatible, and I think it's the right thing to do.
1009
1010 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1011 so that the demonstration program is easily available in environments where
1012 PCRE has not been installed from source.
1013
1014 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1015 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1016 library.
1017
1018 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1019 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1020 is not the first non-POSIX option to be added. Clearly some people find
1021 these options useful.
1022
1023 17. If a caller to the POSIX matching function regexec() passes a non-zero
1024 value for nmatch with a NULL value for pmatch, the value of
1025 nmatch is forced to zero.
1026
1027 18. RunGrepTest did not have a test for the availability of the -u option of
1028 the diff command, as RunTest does. It now checks in the same way as
1029 RunTest, and also checks for the -b option.
1030
1031 19. If an odd number of negated classes containing just a single character
1032 interposed, within parentheses, between a forward reference to a named
1033 subpattern and the definition of the subpattern, compilation crashed with
1034 an internal error, complaining that it could not find the referenced
1035 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1036 [The bug was that it was starting one character too far in when skipping
1037 over the character class, thus treating the ] as data rather than
1038 terminating the class. This meant it could skip too much.]
1039
1040 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1041 /g option in pcretest when the pattern contains \K, which makes it possible
1042 to have an empty string match not at the start, even when the pattern is
1043 anchored. Updated pcretest and pcredemo to use this option.
1044
1045 21. If the maximum number of capturing subpatterns in a recursion was greater
1046 than the maximum at the outer level, the higher number was returned, but
1047 with unset values at the outer level. The correct (outer level) value is
1048 now given.
1049
1050 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1051 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1052 make it do so. The string so far is captured, making this feature
1053 compatible with Perl.
1054
1055 23. The tests have been re-organized, adding tests 11 and 12, to make it
1056 possible to check the Perl 5.10 features against Perl 5.10.
1057
1058 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1059 pattern matches a fixed length string. PCRE did not allow this; now it
1060 does. Neither allows recursion.
1061
1062 25. I finally figured out how to implement a request to provide the minimum
1063 length of subject string that was needed in order to match a given pattern.
1064 (It was back references and recursion that I had previously got hung up
1065 on.) This code has now been added to pcre_study(); it finds a lower bound
1066 to the length of subject needed. It is not necessarily the greatest lower
1067 bound, but using it to avoid searching strings that are too short does give
1068 some useful speed-ups. The value is available to calling programs via
1069 pcre_fullinfo().
1070
1071 26. While implementing 25, I discovered to my embarrassment that pcretest had
1072 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1073 study optimizations had never been tested with that matching function.
1074 Oops. What is worse, even when it was passed study data, there was a bug in
1075 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1076 were also very few tests of studied patterns with pcre_dfa_exec().
1077
1078 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1079 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1080 on the other side of the coin, they are no longer allowed to have different
1081 names, because these cannot be distinguished in PCRE, and this has caused
1082 confusion. (This is a difference from Perl.)
1083
1084 28. When duplicate subpattern names are present (necessarily with different
1085 numbers, as required by 27 above), and a test is made by name in a
1086 conditional pattern, either for a subpattern having been matched, or for
1087 recursion in such a pattern, all the associated numbered subpatterns are
1088 tested, and the overall condition is true if the condition is true for any
1089 one of them. This is the way Perl works, and is also more like the way
1090 testing by number works.
1091
1092
1093 Version 7.9 11-Apr-09
1094 ---------------------
1095
1096 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1097 (pcretest), all targets were linked against these libraries. This included
1098 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1099 libraries. This caused unwanted dependencies to be created. This problem
1100 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1101 pcretest is linked with readline.
1102
1103 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1104 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1105 moved outside it again, because FALSE and TRUE are already defined in AIX,
1106 but BOOL is not.
1107
1108 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1109 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1110
1111 4. The pcregrep documentation said spaces were inserted as well as colons (or
1112 hyphens) following file names and line numbers when outputting matching
1113 lines. This is not true; no spaces are inserted. I have also clarified the
1114 wording for the --colour (or --color) option.
1115
1116 5. In pcregrep, when --colour was used with -o, the list of matching strings
1117 was not coloured; this is different to GNU grep, so I have changed it to be
1118 the same.
1119
1120 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1121 each matching line was coloured. Now it goes on to look for further matches
1122 of any of the test patterns, which is the same behaviour as GNU grep.
1123
1124 7. A pattern that could match an empty string could cause pcregrep to loop; it
1125 doesn't make sense to accept an empty string match in pcregrep, so I have
1126 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1127 seems to be how GNU grep behaves.
1128
1129 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1130 start or after a newline", because the conditional assertion was not being
1131 correctly handled. The rule now is that both the assertion and what follows
1132 in the first alternative must satisfy the test.
1133
1134 9. If auto-callout was enabled in a pattern with a conditional group whose
1135 condition was an assertion, PCRE could crash during matching, both with
1136 pcre_exec() and pcre_dfa_exec().
1137
1138 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1139 used for matching.
1140
1141 11. Unicode property support in character classes was not working for
1142 characters (bytes) greater than 127 when not in UTF-8 mode.
1143
1144 12. Added the -M command line option to pcretest.
1145
1146 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1147
1148 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1149
1150 16. Added comments and documentation about mis-use of no_arg in the C++
1151 wrapper.
1152
1153 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1154 from Martin Jerabek that uses macro names for all relevant character and
1155 string constants.
1156
1157 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1158 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1159 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1160 these, but not everybody uses configure.
1161
1162 19. A conditional group that had only one branch was not being correctly
1163 recognized as an item that could match an empty string. This meant that an
1164 enclosing group might also not be so recognized, causing infinite looping
1165 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1166 with the subject "ab", where knowledge that the repeated group can match
1167 nothing is needed in order to break the loop.
1168
1169 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1170 exec(), but without supplying a callout function, matching went wrong.
1171
1172 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1173 leak if the size of the offset vector was greater than 30. When the vector
1174 is smaller, the saved offsets during recursion go onto a local stack
1175 vector, but for larger vectors malloc() is used. It was failing to free
1176 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1177 error, in fact).
1178
1179 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1180 heapframe that is used only when UTF-8 support is enabled. This caused no
1181 problem, but was untidy.
1182
1183 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1184 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1185 included within another project.
1186
1187 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1188 slightly modified by me:
1189
1190 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1191 not building pcregrep.
1192
1193 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1194 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1195
1196 25. Forward references, both numeric and by name, in patterns that made use of
1197 duplicate group numbers, could behave incorrectly or give incorrect errors,
1198 because when scanning forward to find the reference group, PCRE was not
1199 taking into account the duplicate group numbers. A pattern such as
1200 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1201
1202 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1203 the feature of a custom pointer more persuasive (as requested by a user).
1204
1205 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1206 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1207 pre-defined.
1208
1209 28. Added support for (*UTF8) at the start of a pattern.
1210
1211 29. Arrange for flags added by the "release type" setting in CMake to be shown
1212 in the configuration summary.
1213
1214
1215 Version 7.8 05-Sep-08
1216 ---------------------
1217
1218 1. Replaced UCP searching code with optimized version as implemented for Ad
1219 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1220 stage table and inline lookup instead of a function, giving speed ups of 2
1221 to 5 times on some simple patterns that I tested. Permission was given to
1222 distribute the MultiStage2.py script that generates the tables (it's not in
1223 the tarball, but is in the Subversion repository).
1224
1225 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1226 scripts.
1227
1228 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1229 a group with a zero qualifier. The result of the study could be incorrect,
1230 or the function might crash, depending on the pattern.
1231
1232 4. Caseless matching was not working for non-ASCII characters in back
1233 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1234 It now works when Unicode Property Support is available.
1235
1236 5. In pcretest, an escape such as \x{de} in the data was always generating
1237 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1238 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1239 truncation.
1240
1241 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1242
1243 7. Added two (int) casts to pcregrep when printing the difference of two
1244 pointers, in case they are 64-bit values.
1245
1246 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1247 test 2 if it fails.
1248
1249 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1250 and a #define of that name to empty if it is not externally set. This is to
1251 allow users of MSVC to set it if necessary.
1252
1253 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1254 the convenience functions in the pcre_get.c source file.
1255
1256 11. An option change at the start of a pattern that had top-level alternatives
1257 could cause overwriting and/or a crash. This command provoked a crash in
1258 some environments:
1259
1260 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1261
1262 This potential security problem was recorded as CVE-2008-2371.
1263
1264 12. For a pattern where the match had to start at the beginning or immediately
1265 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1266 pcre_dfa_exec() could read past the end of the passed subject if there was
1267 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1268 pcretest so that it places the subject at the end of its malloc-ed buffer.
1269
1270 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1271 exec() might read past the end of the data buffer in UTF-8 mode.
1272
1273 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1274 the data contained the byte 0x85 as part of a UTF-8 character within its
1275 first line. This applied both to normal and DFA matching.
1276
1277 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1278 /^[^d]*?$/8 failed to match "abc".
1279
1280 16. Added a missing copyright notice to pcrecpp_internal.h.
1281
1282 17. Make it more clear in the documentation that values returned from
1283 pcre_exec() in ovector are byte offsets, not character counts.
1284
1285 18. Tidied a few places to stop certain compilers from issuing warnings.
1286
1287 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1288 supplied by Stefan Weber. I made a further small update for 7.8 because
1289 there is a change of source arrangements: the pcre_searchfuncs.c module is
1290 replaced by pcre_ucd.c.
1291
1292
1293 Version 7.7 07-May-08
1294 ---------------------
1295
1296 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1297 a string to a long long, pretend we don't even have a long long." This is
1298 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1299
1300 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1301 pre-7.6 versions, which defined a global no_arg variable instead of putting
1302 it in the RE class. (See also #8 below.)
1303
1304 3. Remove a line of dead code, identified by coverity and reported by Nuno
1305 Lopes.
1306
1307 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1308
1309 (1) The include/exclude patterns were being applied to the whole pathnames
1310 of files, instead of just to the final components.
1311
1312 (2) If there was more than one level of directory, the subdirectories were
1313 skipped unless they satisfied the include/exclude conditions. This is
1314 inconsistent with GNU grep (and could even be seen as contrary to the
1315 pcregrep specification - which I improved to make it absolutely clear).
1316 The action now is always to scan all levels of directory, and just
1317 apply the include/exclude patterns to regular files.
1318
1319 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1320 --exclude_dir in the tests to avoid scanning .svn directories.
1321
1322 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1323 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1324 doesn't support NULs in patterns.
1325
1326 7. Added some missing "const"s to declarations of static tables in
1327 pcre_compile.c and pcre_dfa_exec.c.
1328
1329 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1330 caused by fix #2 above. (Subsequently also a second patch to fix the
1331 first patch. And a third patch - this was a messy problem.)
1332
1333 9. Applied Craig's patch to remove the use of push_back().
1334
1335 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1336 matching function regexec().
1337
1338 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1339 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1340 references. PCRE supports relative numbers with this syntax (I don't think
1341 Oniguruma does).
1342
1343 12. Previously, a group with a zero repeat such as (...){0} was completely
1344 omitted from the compiled regex. However, this means that if the group
1345 was called as a subroutine from elsewhere in the pattern, things went wrong
1346 (an internal error was given). Such groups are now left in the compiled
1347 pattern, with a new opcode that causes them to be skipped at execution
1348 time.
1349
1350 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1351 to the way PCRE behaves:
1352
1353 (a) A lone ] character is dis-allowed (Perl treats it as data).
1354
1355 (b) A back reference to an unmatched subpattern matches an empty string
1356 (Perl fails the current match path).
1357
1358 (c) A data ] in a character class must be notated as \] because if the
1359 first data character in a class is ], it defines an empty class. (In
1360 Perl it is not possible to have an empty class.) The empty class []
1361 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1362 The negative empty class [^] matches any one character, independently
1363 of the DOTALL setting.
1364
1365 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1366 non-existent subpattern following a character class starting with ']' and
1367 containing () gave an internal compiling error instead of "reference to
1368 non-existent subpattern". Fortunately, when the pattern did exist, the
1369 compiled code was correct. (When scanning forwards to check for the
1370 existencd of the subpattern, it was treating the data ']' as terminating
1371 the class, so got the count wrong. When actually compiling, the reference
1372 was subsequently set up correctly.)
1373
1374 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1375 it was being rejected as not supported by pcre_dfa_exec(), even though
1376 other assertions are supported. I have made pcre_dfa_exec() support
1377 (*FAIL).
1378
1379 16. The implementation of 13c above involved the invention of a new opcode,
1380 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1381 cannot be changed at match time, I realized I could make a small
1382 improvement to matching performance by compiling OP_ALLANY instead of
1383 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1384 on the OP_ANY path.
1385
1386 17. Compiling pcretest on Windows with readline support failed without the
1387 following two fixes: (1) Make the unistd.h include conditional on
1388 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1389
1390 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1391 ncurses library to be included for pcretest when ReadLine support is
1392 requested, but also to allow for it to be overridden. This patch came from
1393 Daniel Bergström.
1394
1395 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1396 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1397 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1398 spotting this.
1399
1400
1401 Version 7.6 28-Jan-08
1402 ---------------------
1403
1404 1. A character class containing a very large number of characters with
1405 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1406 overflow.
1407
1408 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1409 HAVE_LONG_LONG is not defined.
1410
1411 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1412 bring them up to date and include new features. This patch includes:
1413
1414 - Fixed PH's badly added libz and libbz2 support.
1415 - Fixed a problem with static linking.
1416 - Added pcredemo. [But later removed - see 7 below.]
1417 - Fixed dftables problem and added an option.
1418 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1419 HAVE_LONG_LONG.
1420 - Added readline support for pcretest.
1421 - Added an listing of the option settings after cmake has run.
1422
1423 4. A user submitted a patch to Makefile that makes it easy to create
1424 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1425 Makefile.am that cause it to include this special target, without
1426 affecting anything else. Note that the same mingw target plus all
1427 the other distribution libraries and programs are now supported
1428 when configuring with CMake (see 6 below) instead of with
1429 Configure/Make.
1430
1431 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1432 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1433 exported in the Windows port". It has not yet been confirmed that the patch
1434 solves the problem, but it does no harm.
1435
1436 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1437 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1438 with CMake, and also correct the comment about stack recursion.
1439
1440 7. Remove the automatic building of pcredemo from the ./configure system and
1441 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1442 of a program that users should build themselves after PCRE is installed, so
1443 building it automatically is not really right. What is more, it gave
1444 trouble in some build environments.
1445
1446 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1447
1448
1449 Version 7.5 10-Jan-08
1450 ---------------------
1451
1452 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1453 values in parens when parsing an RE using the C++ wrapper."
1454
1455 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1456 Characters greater than 255 were excluded from the class instead of being
1457 included.
1458
1459 3. The same bug as (2) above applied to negated POSIX classes such as
1460 [:^space:].
1461
1462 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1463 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1464 I have changed it.
1465
1466 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1467 first named subpattern) and a construct such as (?&a) would reference the
1468 first named subpattern whose name started with "a" (in other words, the
1469 length check was missing). Both these problems are fixed. "Subpattern name
1470 expected" is now given for (?&) (a zero-length name), and this patch also
1471 makes it give the same error for \k'' (previously it complained that that
1472 was a reference to a non-existent subpattern).
1473
1474 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1475 this is right because (?- can be followed by option settings as well as by
1476 digits. I have, however, made the messages clearer.
1477
1478 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1479 than the number used in the conditional) now cause a compile-time error.
1480 This is actually not compatible with Perl, which accepts such patterns, but
1481 treats the conditional as always being FALSE (as PCRE used to), but it
1482 seems to me that giving a diagnostic is better.
1483
1484 8. Change "alphameric" to the more common word "alphanumeric" in comments
1485 and messages.
1486
1487 9. Fix two occurrences of "backslash" in comments that should have been
1488 "backspace".
1489
1490 10. Remove two redundant lines of code that can never be obeyed (their function
1491 was moved elsewhere).
1492
1493 11. The program that makes PCRE's Unicode character property table had a bug
1494 which caused it to generate incorrect table entries for sequences of
1495 characters that have the same character type, but are in different scripts.
1496 It amalgamated them into a single range, with the script of the first of
1497 them. In other words, some characters were in the wrong script. There were
1498 thirteen such cases, affecting characters in the following ranges:
1499
1500 U+002b0 - U+002c1
1501 U+0060c - U+0060d
1502 U+0061e - U+00612
1503 U+0064b - U+0065e
1504 U+0074d - U+0076d
1505 U+01800 - U+01805
1506 U+01d00 - U+01d77
1507 U+01d9b - U+01dbf
1508 U+0200b - U+0200f
1509 U+030fc - U+030fe
1510 U+03260 - U+0327f
1511 U+0fb46 - U+0fbb1
1512 U+10450 - U+1049d
1513
1514 12. The -o option (show only the matching part of a line) for pcregrep was not
1515 compatible with GNU grep in that, if there was more than one match in a
1516 line, it showed only the first of them. It now behaves in the same way as
1517 GNU grep.
1518
1519 13. If the -o and -v options were combined for pcregrep, it printed a blank
1520 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1521 does the same. The return code can be used to tell if there were any
1522 non-matching lines.
1523
1524 14. Added --file-offsets and --line-offsets to pcregrep.
1525
1526 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1527 infinitely looping recursion. The bug was that positive lookaheads were not
1528 being skipped when checking for a possible empty match (negative lookaheads
1529 and both kinds of lookbehind were skipped).
1530
1531 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1532 inclusion of <windows.h> to before rather than after the definition of
1533 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1534
1535 17. Specifying a possessive quantifier with a specific limit for a Unicode
1536 character property caused pcre_compile() to compile bad code, which led at
1537 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1538 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1539 caused the error; without that there was no problem.
1540
1541 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1542
1543 19. Added --enable-pcretest-libreadline.
1544
1545 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1546 RE::GlobalReplace(). As a result, the number of replacements returned was
1547 double what it should be. I removed one of the increments, but Craig sent a
1548 later patch that removed the other one (the right fix) and added unit tests
1549 that check the return values (which was not done before).
1550
1551 21. Several CMake things:
1552
1553 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1554 the names libpcre and libpcreposix, not just pcre and pcreposix.
1555
1556 (2) The above change means that pcretest and pcregrep are now correctly
1557 linked with the newly-built libraries, not previously installed ones.
1558
1559 (3) Added PCRE_SUPPORT_LIBREADLINE, PCRE_SUPPORT_LIBZ, PCRE_SUPPORT_LIBBZ2.
1560
1561 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1562 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1563 UTF-8 newline character). The key issue is that the pattern starts .*;
1564 this means that the match must be either at the beginning, or after a
1565 newline. The bug was in the code for advancing after a failed match and
1566 checking that the new position followed a newline. It was not taking
1567 account of UTF-8 characters correctly.
1568
1569 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1570 character classes. PCRE was not treating the sequence [:...:] as a
1571 character class unless the ... were all letters. Perl, however, seems to
1572 allow any characters between [: and :], though of course it rejects as
1573 unknown any "names" that contain non-letters, because all the known class
1574 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1575 for example, whereas PCRE did not - it did not recognize a POSIX character
1576 class. This seemed a bit dangerous, so the code has been changed to be
1577 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1578 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1579 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1580 Perl does, and where it didn't before.
1581
1582 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1583 Windows environments %n is disabled by default.
1584
1585
1586 Version 7.4 21-Sep-07
1587 ---------------------
1588
1589 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1590 means that a class such as [\s] counted as "explicit reference to CR or
1591 LF". That isn't really right - the whole point of the change was to try to
1592 help when there was an actual mention of one of the two characters. So now
1593 the change happens only if \r or \n (or a literal CR or LF) character is
1594 encountered.
1595
1596 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1597 of both had grown to the point where there were only 3 bits left.
1598 Fortunately, there was spare space in the data structure, and so I have
1599 moved the internal flags into a new 16-bit field to free up more option
1600 bits.
1601
1602 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1603 but did not set the internal JCHANGED flag - either of these is enough to
1604 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1605 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1606 start sets both bits.
1607
1608 4. Added options (at build time, compile time, exec time) to change \R from
1609 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1610
1611 5. doc/pcresyntax.html was missing from the distribution.
1612
1613 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1614 compatibility, even though it is no longer used.
1615
1616 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1617 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1618 windows.h file is present (where different names are used). [This was
1619 reversed later after testing - see 16 below.]
1620
1621 8. Changed all #include <config.h> to #include "config.h". There were also
1622 some further <pcre.h> cases that I changed to "pcre.h".
1623
1624 9. When pcregrep was used with the --colour option, it missed the line ending
1625 sequence off the lines that it output.
1626
1627 10. It was pointed out to me that arrays of string pointers cause lots of
1628 relocations when a shared library is dynamically loaded. A technique of
1629 using a single long string with a table of offsets can drastically reduce
1630 these. I have refactored PCRE in four places to do this. The result is
1631 dramatic:
1632
1633 Originally: 290
1634 After changing UCP table: 187
1635 After changing error message table: 43
1636 After changing table of "verbs" 36
1637 After changing table of Posix names 22
1638
1639 Thanks to the folks working on Gregex for glib for this insight.
1640
1641 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1642 unicode-properties was also set.
1643
1644 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1645
1646 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1647 checked only for CRLF.
1648
1649 14. Added casts to pcretest.c to avoid compiler warnings.
1650
1651 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1652
1653 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1654 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1655 entirely. This removes changes made in 7 above.
1656
1657 17. The CMake files have been updated, and there is now more information about
1658 building with CMake in the NON-UNIX-USE document.
1659
1660
1661 Version 7.3 28-Aug-07
1662 ---------------------
1663
1664 1. In the rejigging of the build system that eventually resulted in 7.1, the
1665 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1666 brackets there is not right, since it causes compilers to look for an
1667 installed pcre.h, not the version that is in the source that is being
1668 compiled (which of course may be different). I have changed it back to:
1669
1670 #include "pcre.h"
1671
1672 I have a vague recollection that the change was concerned with compiling in
1673 different directories, but in the new build system, that is taken care of
1674 by the VPATH setting the Makefile.
1675
1676 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1677 when the subject happened to end in the byte 0x85 (e.g. if the last
1678 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1679 characters but of course it shouldn't be taken as a newline when it is part
1680 of another character. The bug was that, for an unlimited repeat of . in
1681 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1682 characters when looking for a newline.
1683
1684 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1685
1686 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1687 in debug output.
1688
1689 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1690 long printing in the pcrecpp unittest when running under MinGW.
1691
1692 6. ESC_K was left out of the EBCDIC table.
1693
1694 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1695 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1696 limit also applies to "virtual nesting" when a pattern is recursive, and in
1697 this case 1000 isn't so big. I have been able to remove this limit at the
1698 expense of backing off one optimization in certain circumstances. Normally,
1699 when pcre_exec() would call its internal match() function recursively and
1700 immediately return the result unconditionally, it uses a "tail recursion"
1701 feature to save stack. However, when a subpattern that can match an empty
1702 string has an unlimited repetition quantifier, it no longer makes this
1703 optimization. That gives it a stack frame in which to save the data for
1704 checking that an empty string has been matched. Previously this was taken
1705 from the 1000-entry workspace that had been reserved. So now there is no
1706 explicit limit, but more stack is used.
1707
1708 8. Applied Daniel's patches to solve problems with the import/export magic
1709 syntax that is required for Windows, and which was going wrong for the
1710 pcreposix and pcrecpp parts of the library. These were overlooked when this
1711 problem was solved for the main library.
1712
1713 9. There were some crude static tests to avoid integer overflow when computing
1714 the size of patterns that contain repeated groups with explicit upper
1715 limits. As the maximum quantifier is 65535, the maximum group length was
1716 set at 30,000 so that the product of these two numbers did not overflow a
1717 32-bit integer. However, it turns out that people want to use groups that
1718 are longer than 30,000 bytes (though not repeat them that many times).
1719 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1720 made it possible to implement the integer overflow checks in a much more
1721 dynamic way, which I have now done. The artificial limitation on group
1722 length has been removed - we now have only the limit on the total length of
1723 the compiled pattern, which depends on the LINK_SIZE setting.
1724
1725 10. Fixed a bug in the documentation for get/copy named substring when
1726 duplicate names are permitted. If none of the named substrings are set, the
1727 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1728 empty string.
1729
1730 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1731 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1732 because the ] is interpreted as the first data character and the
1733 terminating ] is not found. PCRE has been made compatible with Perl in this
1734 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1735 cause memory overwriting.
1736
1737 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1738 string has been matched (to stop an infinite loop). It was not recognizing
1739 a conditional subpattern that could match an empty string if that
1740 subpattern was within another subpattern. For example, it looped when
1741 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1742 condition was not nested. This bug has been fixed.
1743
1744 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1745 past the start of the subject in the presence of bytes with the top bit
1746 set, for example "\x8aBCD".
1747
1748 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1749 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1750
1751 14. Optimized (?!) to (*FAIL).
1752
1753 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1754 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1755 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1756 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1757 does: it's just the validity check that is more restrictive.
1758
1759 16. Inserted checks for integer overflows during escape sequence (backslash)
1760 processing, and also fixed erroneous offset values for syntax errors during
1761 backslash processing.
1762
1763 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1764 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1765
1766 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1767 caused an overrun.
1768
1769 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1770 something other than just ASCII characters) inside a group that had an
1771 unlimited repeat caused a loop at compile time (while checking to see
1772 whether the group could match an empty string).
1773
1774 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1775 [\P{Any}] did so. (Error in the code for printing property names.)
1776
1777 21. An orphan \E inside a character class could cause a crash.
1778
1779 22. A repeated capturing bracket such as (A)? could cause a wild memory
1780 reference during compilation.
1781
1782 23. There are several functions in pcre_compile() that scan along a compiled
1783 expression for various reasons (e.g. to see if it's fixed length for look
1784 behind). There were bugs in these functions when a repeated \p or \P was
1785 present in the pattern. These operators have additional parameters compared
1786 with \d, etc, and these were not being taken into account when moving along
1787 the compiled data. Specifically:
1788
1789 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1790 length.
1791
1792 (b) An item such as \pL+ within a repeated group could cause crashes or
1793 loops.
1794
1795 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1796 "reference to non-existent subpattern" error.
1797
1798 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1799
1800 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1801 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1802
1803 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1804
1805 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1806 character were causing crashes (broken optimization).
1807
1808 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1809 \p or \P) caused a compile-time loop.
1810
1811 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1812 break. For example, the unstudied pattern [\r\n]A does not match the string
1813 "\r\nA" because change 7.0/46 below moves the current point on by two
1814 characters after failing to match at the start. However, the pattern \nA
1815 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1816 the same is true. There doesn't seem any very clean way out of this, but
1817 what I have chosen to do makes the common cases work: PCRE now takes note
1818 of whether there can be an explicit match for \r or \n anywhere in the
1819 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1820 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1821 pattern has explicit CR or LF references.
1822
1823 29. Added (*CR) etc for changing newline setting at start of pattern.
1824
1825
1826 Version 7.2 19-Jun-07
1827 ---------------------
1828
1829 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1830 which is apparently normally available under Windows.
1831
1832 2. Re-jig the pcregrep tests with different newline settings in an attempt
1833 to make them independent of the local environment's newline setting.
1834
1835 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1836
1837 4. Some of the "internals" tests were previously cut out when the link size
1838 was not 2, because the output contained actual offsets. The recent new
1839 "Z" feature of pcretest means that these can be cut out, making the tests
1840 usable with all link sizes.
1841
1842 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1843 stack recursion. This gives a massive performance boost under BSD, but just
1844 a small improvement under Linux. However, it saves one field in the frame
1845 in all cases.
1846
1847 6. Added more features from the forthcoming Perl 5.10:
1848
1849 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1850 recursion call. It refers to the nth most recently opened parentheses.
1851
1852 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1853 to be opened parentheses.
1854
1855 (c) Conditions that refer to capturing parentheses can be specified
1856 relatively, for example, (?(-2)... or (?(+3)...
1857
1858 (d) \K resets the start of the current match so that everything before
1859 is not part of it.
1860
1861 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1862
1863 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1864 reference syntax.
1865
1866 (g) (?| introduces a group in which the numbering of parentheses in each
1867 alternative starts with the same number.
1868
1869 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1870
1871 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1872 PCRE_INFO_JCHANGED.
1873
1874 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1875 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1876 for detecting groups that can match an empty string.
1877
1878 9. A pattern with a very large number of alternatives (more than several
1879 hundred) was running out of internal workspace during the pre-compile
1880 phase, where pcre_compile() figures out how much memory will be needed. A
1881 bit of new cunning has reduced the workspace needed for groups with
1882 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1883 workspace instead of running out of the 4096 that are available.
1884
1885 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1886
1887 11. Applied patch from Google to remove an optimization that didn't quite work.
1888 The report of the bug said:
1889
1890 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1891 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1892 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1893
1894 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1895 it matched the wrong number of bytes.
1896
1897
1898 Version 7.1 24-Apr-07
1899 ---------------------
1900
1901 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1902 that is more "standard", making use of automake and other Autotools. There
1903 is some re-arrangement of the files and adjustment of comments consequent
1904 on this.
1905
1906 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1907 for recursive directory scanning broke on some systems because the files
1908 are not scanned in any specific order and on different systems the order
1909 was different. A call to "sort" has been inserted into RunGrepTest for the
1910 approprate test as a short-term fix. In the longer term there may be an
1911 alternative.
1912
1913 3. I had an email from Eric Raymond about problems translating some of PCRE's
1914 man pages to HTML (despite the fact that I distribute HTML pages, some
1915 people do their own conversions for various reasons). The problems
1916 concerned the use of low-level troff macros .br and .in. I have therefore
1917 removed all such uses from the man pages (some were redundant, some could
1918 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1919 HTML has been updated to handle .nf/.fi and to complain if it encounters
1920 .br or .in.
1921
1922 4. Updated comments in configure.ac that get placed in config.h.in and also
1923 arranged for config.h to be included in the distribution, with the name
1924 config.h.generic, for the benefit of those who have to compile without
1925 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1926
1927 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1928 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1929 makevp.bat for the current PCRE, using the additional files
1930 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1931
1932 6. A Windows user reported a minor discrepancy with test 2, which turned out
1933 to be caused by a trailing space on an input line that had got lost in his
1934 copy. The trailing space was an accident, so I've just removed it.
1935
1936 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1937 that is needed.
1938
1939 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1940 as "const" (a) because they are and (b) because it helps the PHP
1941 maintainers who have recently made a script to detect big data structures
1942 in the php code that should be moved to the .rodata section. I remembered
1943 to update Builducptable as well, so it won't revert if ucptable.h is ever
1944 re-created.
1945
1946 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1947 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1948 order to be able to cut out the UTF-8 tables in the latter when UTF-8
1949 support is not required. This saves 1.5-2K of code, which is important in
1950 some applications.
1951
1952 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
1953 so as not to refer to the tables, even though these functions will never be
1954 called when UTF-8 support is disabled. Otherwise there are problems with a
1955 shared library.
1956
1957 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
1958
1959 (a) It was defining its arguments as char * instead of void *.
1960
1961 (b) It was assuming that all moves were upwards in memory; this was true
1962 a long time ago when I wrote it, but is no longer the case.
1963
1964 The emulated memove() is provided for those environments that have neither
1965 memmove() nor bcopy(). I didn't think anyone used it these days, but that
1966 is clearly not the case, as these two bugs were recently reported.
1967
1968 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
1969 and Detrail to create the HTML documentation, the .txt form of the man
1970 pages, and it removes trailing spaces from listed files. It also creates
1971 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
1972 case, it wraps all the #defines with #ifndefs. This script should be run
1973 before "make dist".
1974
1975 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
1976 with Unicode property support.
1977
1978 (a) For a maximizing quantifier, if the two different cases of the
1979 character were of different lengths in their UTF-8 codings (there are
1980 some cases like this - I found 11), and the matching function had to
1981 back up over a mixture of the two cases, it incorrectly assumed they
1982 were both the same length.
1983
1984 (b) When PCRE was configured to use the heap rather than the stack for
1985 recursion during matching, it was not correctly preserving the data for
1986 the other case of a UTF-8 character when checking ahead for a match
1987 while processing a minimizing repeat. If the check also involved
1988 matching a wide character, but failed, corruption could cause an
1989 erroneous result when trying to check for a repeat of the original
1990 character.
1991
1992 13. Some tidying changes to the testing mechanism:
1993
1994 (a) The RunTest script now detects the internal link size and whether there
1995 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
1996 values substituted by "configure". (The RunGrepTest script already did
1997 this for UTF-8.) The configure.ac script no longer substitutes the
1998 relevant variables.
1999
2000 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2001 with length and offset values. This means that the output is different
2002 for different internal link sizes. Test 2 is skipped for link sizes
2003 other than 2 because of this, bypassing the problem. Unfortunately,
2004 there was also a test in test 3 (the locale tests) that used /B and
2005 failed for link sizes other than 2. Rather than cut the whole test out,
2006 I have added a new /Z option to pcretest that replaces the length and
2007 offset values with spaces. This is now used to make test 3 independent
2008 of link size. (Test 2 will be tidied up later.)
2009
2010 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2011 segmentation fault instead of returning the appropriate error message.
2012
2013 15. In multiline mode when the newline sequence was set to "any", the pattern
2014 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2015 This doesn't seem right; it now treats the CRLF combination as the line
2016 ending, and so does not match in that case. It's only a pattern such as ^$
2017 that would hit this one: something like ^ABC$ would have failed after \r
2018 and then tried again after \r\n.
2019
2020 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2021 in an attempt to make files that differ only in their line terminators
2022 compare equal. This works on Linux.
2023
2024 17. Under certain error circumstances pcregrep might try to free random memory
2025 as it exited. This is now fixed, thanks to valgrind.
2026
2027 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2028 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2029 was because its rules for how to advance for /g after matching an empty
2030 string at the end of a line did not allow for this case. They now check for
2031 it specially.
2032
2033 20. pcretest is supposed to handle patterns and data of any length, by
2034 extending its buffers when necessary. It was getting this wrong when the
2035 buffer for a data line had to be extended.
2036
2037 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2038 CRLF as a newline sequence.
2039
2040 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2041 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2042 I have nevertheless tidied it up.
2043
2044 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2045
2046 24. Added a man page for pcre-config.
2047
2048
2049 Version 7.0 19-Dec-06
2050 ---------------------
2051
2052 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2053 moving to gcc 4.1.1.
2054
2055 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2056 sys/time.h, which is documented as needed for this function. It doesn't
2057 seem to matter on Linux, but it showed up on some releases of OS X.
2058
2059 3. It seems that there are systems where bytes whose values are greater than
2060 127 match isprint() in the "C" locale. The "C" locale should be the
2061 default when a C program starts up. In most systems, only ASCII printing
2062 characters match isprint(). This difference caused the output from pcretest
2063 to vary, making some of the tests fail. I have changed pcretest so that:
2064
2065 (a) When it is outputting text in the compiled version of a pattern, bytes
2066 other than 32-126 are always shown as hex escapes.
2067
2068 (b) When it is outputting text that is a matched part of a subject string,
2069 it does the same, unless a different locale has been set for the match
2070 (using the /L modifier). In this case, it uses isprint() to decide.
2071
2072 4. Fixed a major bug that caused incorrect computation of the amount of memory
2073 required for a compiled pattern when options that changed within the
2074 pattern affected the logic of the preliminary scan that determines the
2075 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2076 that the computed length was too small. The symptoms of this bug were
2077 either the PCRE error "internal error: code overflow" from pcre_compile(),
2078 or a glibc crash with a message such as "pcretest: free(): invalid next
2079 size (fast)". Examples of patterns that provoked this bug (shown in
2080 pcretest format) are:
2081
2082 /(?-x: )/x
2083 /(?x)(?-x: \s*#\s*)/
2084 /((?i)[\x{c0}])/8
2085 /(?i:[\x{c0}])/8
2086
2087 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2088 is now done differently.
2089
2090 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2091 wrapper classes; (b) implement a new function in the C++ scanner that is
2092 more efficient than the old way of doing things because it avoids levels of
2093 recursion in the regex matching; (c) add a paragraph to the documentation
2094 for the FullMatch() function.
2095
2096 6. The escape sequence \n was being treated as whatever was defined as
2097 "newline". Not only was this contrary to the documentation, which states
2098 that \n is character 10 (hex 0A), but it also went horribly wrong when
2099 "newline" was defined as CRLF. This has been fixed.
2100
2101 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2102 was being set to -1 for the "end of line" case (supposedly a value that no
2103 character can have). Though this value is never used (the check for end of
2104 line is "zero bytes in current character"), it caused compiler complaints.
2105 I've changed it to 0xffffffff.
2106
2107 8. In pcre_version.c, the version string was being built by a sequence of
2108 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2109 string (as it is for production releases) called a macro with an empty
2110 argument. The C standard says the result of this is undefined. The gcc
2111 compiler treats it as an empty string (which was what was wanted) but it is
2112 reported that Visual C gives an error. The source has been hacked around to
2113 avoid this problem.
2114
2115 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2116 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2117 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2118 of them did).
2119
2120 10. Originally, pcretest opened its input and output without "b"; then I was
2121 told that "b" was needed in some environments, so it was added for release
2122 5.0 to both the input and output. (It makes no difference on Unix-like
2123 systems.) Later I was told that it is wrong for the input on Windows. I've
2124 now abstracted the modes into two macros, to make it easier to fiddle with
2125 them, and removed "b" from the input mode under Windows.
2126
2127 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2128
2129 12. Added -help and --help to pcretest as an official way of being reminded
2130 of the options.
2131
2132 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2133 and pcrecpp.cc because they annoy compilers at high warning levels.
2134
2135 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2136
2137 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2138 scripts are not C programs :-) and which was not noticed because it works
2139 on Linux.
2140
2141 16. pcretest is supposed to handle any length of pattern and data line (as one
2142 line or as a continued sequence of lines) by extending its input buffer if
2143 necessary. This feature was broken for very long pattern lines, leading to
2144 a string of junk being passed to pcre_compile() if the pattern was longer
2145 than about 50K.
2146
2147 17. I have done a major re-factoring of the way pcre_compile() computes the
2148 amount of memory needed for a compiled pattern. Previously, there was code
2149 that made a preliminary scan of the pattern in order to do this. That was
2150 OK when PCRE was new, but as the facilities have expanded, it has become
2151 harder and harder to keep it in step with the real compile phase, and there
2152 have been a number of bugs (see for example, 4 above). I have now found a
2153 cunning way of running the real compile function in a "fake" mode that
2154 enables it to compute how much memory it would need, while actually only
2155 ever using a few hundred bytes of working memory and without too many
2156 tests of the mode. This should make future maintenance and development
2157 easier. A side effect of this work is that the limit of 200 on the nesting
2158 depth of parentheses has been removed (though this was never a serious
2159 limitation, I suspect). However, there is a downside: pcre_compile() now
2160 runs more slowly than before (30% or more, depending on the pattern). I
2161 hope this isn't a big issue. There is no effect on runtime performance.
2162
2163 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2164 newline (only possible for the last line of a file) and it was a
2165 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2166
2167 19. Added additional timing features to pcretest. (1) The -tm option now times
2168 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2169 separate command line item, by a number that specifies the number of
2170 repeats to use when timing. The default is 50000; this gives better
2171 precision, but takes uncomfortably long for very large patterns.
2172
2173 20. Extended pcre_study() to be more clever in cases where a branch of a
2174 subpattern has no definite first character. For example, (a*|b*)[cd] would
2175 previously give no result from pcre_study(). Now it recognizes that the
2176 first character must be a, b, c, or d.
2177
2178 21. There was an incorrect error "recursive call could loop indefinitely" if
2179 a subpattern (or the entire pattern) that was being tested for matching an
2180 empty string contained only one non-empty item after a nested subpattern.
2181 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2182 incorrectly, because the \d was being skipped in the check.
2183
2184 22. The pcretest program now has a new pattern option /B and a command line
2185 option -b, which is equivalent to adding /B to every pattern. This causes
2186 it to show the compiled bytecode, without the additional information that
2187 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2188 is the same as /B/I).
2189
2190 23. A new optimization is now able automatically to treat some sequences such
2191 as a*b as a*+b. More specifically, if something simple (such as a character
2192 or a simple class like \d) has an unlimited quantifier, and is followed by
2193 something that cannot possibly match the quantified thing, the quantifier
2194 is automatically "possessified".
2195
2196 24. A recursive reference to a subpattern whose number was greater than 39
2197 went wrong under certain circumstances in UTF-8 mode. This bug could also
2198 have affected the operation of pcre_study().
2199
2200 25. Realized that a little bit of performance could be had by replacing
2201 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2202
2203 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2204
2205 27. Possessive quantifiers such as a++ were previously implemented by turning
2206 them into atomic groups such as ($>a+). Now they have their own opcodes,
2207 which improves performance. This includes the automatically created ones
2208 from 23 above.
2209
2210 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2211 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2212 the first matched character to be a colon. This applied both to named and
2213 numbered groups.
2214
2215 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2216
2217 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2218 building PCRE on Windows easier, so I have included it in the distribution.
2219
2220 31. There is now a check in pcretest against a ridiculously large number being
2221 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2222 loop, the loop is abandoned.
2223
2224 32. Forward references to subpatterns in conditions such as (?(2)...) where
2225 subpattern 2 is defined later cause pcre_compile() to search forwards in
2226 the pattern for the relevant set of parentheses. This search went wrong
2227 when there were unescaped parentheses in a character class, parentheses
2228 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2229
2230 33. "Subroutine" calls and backreferences were previously restricted to
2231 referencing subpatterns earlier in the regex. This restriction has now
2232 been removed.
2233
2234 34. Added a number of extra features that are going to be in Perl 5.10. On the
2235 whole, these are just syntactic alternatives for features that PCRE had
2236 previously implemented using the Python syntax or my own invention. The
2237 other formats are all retained for compatibility.
2238
2239 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2240 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2241 also .NET compatible.
2242
2243 (b) A recursion or subroutine call to a named group can now be defined as
2244 (?&name) as well as (?P>name).
2245
2246 (c) A backreference to a named group can now be defined as \k<name> or
2247 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2248 5.10, are also .NET compatible.
2249
2250 (d) A conditional reference to a named group can now use the syntax
2251 (?(<name>) or (?('name') as well as (?(name).
2252
2253 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2254 groups (named and numbered) that are never evaluated inline, but can be
2255 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2256 is always false. There may be only one alternative in such a group.
2257
2258 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2259 as the simple (?(R). The condition is true only if the most recent
2260 recursion is that of the given number or name. It does not search out
2261 through the entire recursion stack.
2262
2263 (g) The escape \gN or \g{N} has been added, where N is a positive or
2264 negative number, specifying an absolute or relative reference.
2265
2266 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2267 some "unreachable code" warnings.
2268
2269 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2270 things, this adds five new scripts.
2271
2272 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2273 There were also incompatibilities regarding the handling of \Q..\E inside
2274 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2275 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2276
2277 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2278 matches an empty string, and forcibly breaks the loop. There were bugs in
2279 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2280 against aaaa the result was just "a" rather than "aaaa", for example. Two
2281 separate and independent bugs (that affected different cases) have been
2282 fixed.
2283
2284 39. Refactored the code to abolish the use of different opcodes for small
2285 capturing bracket numbers. This is a tidy that I avoided doing when I
2286 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2287 The new approach is not only tidier, it makes it possible to reduce the
2288 memory needed to fix the previous bug (38).
2289
2290 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2291 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2292 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2293 mode.
2294
2295 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2296 report.
2297
2298 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2299 copy construction and assignment in the C++ wrapper.
2300
2301 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2302 couple of bugs that could have given wrong results in the "--newline=crlf"
2303 case.
2304
2305 44. Added a number of casts and did some reorganization of signed/unsigned int
2306 variables following suggestions from Dair Grant. Also renamed the variable
2307 "this" as "item" because it is a C++ keyword.
2308
2309 45. Arranged for dftables to add
2310
2311 #include "pcre_internal.h"
2312
2313 to pcre_chartables.c because without it, gcc 4.x may remove the array
2314 definition from the final binary if PCRE is built into a static library and
2315 dead code stripping is activated.
2316
2317 46. For an unanchored pattern, if a match attempt fails at the start of a
2318 newline sequence, and the newline setting is CRLF or ANY, and the next two
2319 characters are CRLF, advance by two characters instead of one.
2320
2321
2322 Version 6.7 04-Jul-06
2323 ---------------------
2324
2325 1. In order to handle tests when input lines are enormously long, pcretest has
2326 been re-factored so that it automatically extends its buffers when
2327 necessary. The code is crude, but this _is_ just a test program. The
2328 default size has been increased from 32K to 50K.
2329
2330 2. The code in pcre_study() was using the value of the re argument before
2331 testing it for NULL. (Of course, in any sensible call of the function, it
2332 won't be NULL.)
2333
2334 3. The memmove() emulation function in pcre_internal.h, which is used on
2335 systems that lack both memmove() and bcopy() - that is, hardly ever -
2336 was missing a "static" storage class specifier.
2337
2338 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2339 containing an extended class (one that cannot be represented by a bitmap
2340 because it contains high-valued characters or Unicode property items, e.g.
2341 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2342 pattern, but PCRE should not loop if you do not (it no longer does).
2343 [Detail: two cases were found: (a) a repeated subpattern containing an
2344 extended class; (b) a recursive reference to a subpattern that followed a
2345 previous extended class. It wasn't skipping over the extended class
2346 correctly when UTF-8 mode was not set.]
2347
2348 5. A negated single-character class was not being recognized as fixed-length
2349 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2350 compile error "lookbehind assertion is not fixed length".
2351
2352 6. The RunPerlTest auxiliary script was showing an unexpected difference
2353 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2354 write a Perl script that can interpret lines of an input file either as
2355 byte characters or as UTF-8, which is what "perltest" was being required to
2356 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2357 can't do is switch easily at run time between having the "use utf8;" pragma
2358 or not. In the end, I fudged it by using the RunPerlTest script to insert
2359 "use utf8;" explicitly for the UTF-8 tests.
2360
2361 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2362 the end of the subject string, contrary to the documentation and to what
2363 Perl does. This was true of both matching functions. Now it matches only at
2364 the start of the subject and immediately after *internal* newlines.
2365
2366 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2367 a pointer to an int instead of a pointer to an unsigned long int. This
2368 caused problems on 64-bit systems.
2369
2370 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2371 instance of the 'standard' template library not being so standard".
2372
2373 10. There was no check on the number of named subpatterns nor the maximum
2374 length of a subpattern name. The product of these values is used to compute
2375 the size of the memory block for a compiled pattern. By supplying a very
2376 long subpattern name and a large number of named subpatterns, the size
2377 computation could be caused to overflow. This is now prevented by limiting
2378 the length of names to 32 characters, and the number of named subpatterns
2379 to 10,000.
2380
2381 11. Subpatterns that are repeated with specific counts have to be replicated in
2382 the compiled pattern. The size of memory for this was computed from the
2383 length of the subpattern and the repeat count. The latter is limited to
2384 65535, but there was no limit on the former, meaning that integer overflow
2385 could in principle occur. The compiled length of a repeated subpattern is
2386 now limited to 30,000 bytes in order to prevent this.
2387
2388 12. Added the optional facility to have named substrings with the same name.
2389
2390 13. Added the ability to use a named substring as a condition, using the
2391 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2392 are numbers (not recommended). Forward references are permitted.
2393
2394 14. Added forward references in named backreferences (if you see what I mean).
2395
2396 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2397 pattern could run off the end of the subject. For example, the pattern
2398 "(?s)(.{1,5})"8 did this with the subject "ab".
2399
2400 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2401 PCRE_CASELESS was set when matching characters that were quantified with ?
2402 or *.
2403
2404 17. A character class other than a single negated character that had a minimum
2405 but no maximum quantifier - for example [ab]{6,} - was not handled
2406 correctly by pce_dfa_exec(). It would match only one character.
2407
2408 18. A valid (though odd) pattern that looked like a POSIX character
2409 class but used an invalid character after [ (for example [[,abc,]]) caused
2410 pcre_compile() to give the error "Failed: internal error: code overflow" or
2411 in some cases to crash with a glibc free() error. This could even happen if
2412 the pattern terminated after [[ but there just happened to be a sequence of
2413 letters, a binary zero, and a closing ] in the memory that followed.
2414
2415 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2416 over the years. Originally (before any Unicode support), just the bottom 8
2417 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2418 output from "man perlunicode" includes this:
2419
2420 The regular expression compiler produces polymorphic opcodes. That
2421 is, the pattern adapts to the data and automatically switches to
2422 the Unicode character scheme when presented with Unicode data--or
2423 instead uses a traditional byte scheme when presented with byte
2424 data.
2425
2426 Sadly, a wide octal escape does not cause a switch, and in a string with
2427 no other multibyte characters, these octal escapes are treated as before.
2428 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2429 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2430 Unicode string.
2431
2432 I have not perpetrated such confusion in PCRE. Up till now, it took just
2433 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2434 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2435 translate to the appropriate multibyte character.
2436
2437 29. Applied some refactoring to reduce the number of warnings from Microsoft
2438 and Borland compilers. This has included removing the fudge introduced
2439 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2440 a warning about an unused variable.
2441
2442 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2443 characters since release 4.0, because Perl (from release 5.004) does not.
2444 [Or at least, is documented not to: some releases seem to be in conflict
2445 with the documentation.] However, when a pattern was studied with
2446 pcre_study() and all its branches started with \s, PCRE still included VT
2447 as a possible starting character. Of course, this did no harm; it just
2448 caused an unnecessary match attempt.
2449
2450 22. Removed a now-redundant internal flag bit that recorded the fact that case
2451 dependency changed within the pattern. This was once needed for "required
2452 byte" processing, but is no longer used. This recovers a now-scarce options
2453 bit. Also moved the least significant internal flag bit to the most-
2454 significant bit of the word, which was not previously used (hangover from
2455 the days when it was an int rather than a uint) to free up another bit for
2456 the future.
2457
2458 23. Added support for CRLF line endings as well as CR and LF. As well as the
2459 default being selectable at build time, it can now be changed at runtime
2460 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2461 specify that it is scanning data with non-default line endings.
2462
2463 24. Changed the definition of CXXLINK to make it agree with the definition of
2464 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2465
2466 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2467 recursions. This makes a big different to stack usage for some patterns.
2468
2469 26. If a subpattern containing a named recursion or subroutine reference such
2470 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2471 the space required for the compiled pattern went wrong and gave too small a
2472 value. Depending on the environment, this could lead to "Failed: internal
2473 error: code overflow at offset 49" or "glibc detected double free or
2474 corruption" errors.
2475
2476 27. Applied patches from Google (a) to support the new newline modes and (b) to
2477 advance over multibyte UTF-8 characters in GlobalReplace.
2478
2479 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2480 difference for some implementation of PCRE in some Windows version.
2481
2482 29. Added some extra testing facilities to pcretest:
2483
2484 \q<number> in a data line sets the "match limit" value
2485 \Q<number> in a data line sets the "match recursion limt" value
2486 -S <number> sets the stack size, where <number> is in megabytes
2487
2488 The -S option isn't available for Windows.
2489
2490
2491 Version 6.6 06-Feb-06
2492 ---------------------
2493
2494 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2495 in pcreposix.h. I have copied the definition from pcre.h.
2496
2497 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2498 because pcre.h is no longer a built file.
2499
2500 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2501 not normally included in the compiled code.
2502
2503
2504 Version 6.5 01-Feb-06
2505 ---------------------
2506
2507 1. When using the partial match feature with pcre_dfa_exec(), it was not
2508 anchoring the second and subsequent partial matches at the new starting
2509 point. This could lead to incorrect results. For example, with the pattern
2510 /1234/, partially matching against "123" and then "a4" gave a match.
2511
2512 2. Changes to pcregrep:
2513
2514 (a) All non-match returns from pcre_exec() were being treated as failures
2515 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2516 error message is output. Some extra information is given for the
2517 PCRE_ERROR_MATCHLIMIT and PCRE_ERROR_RECURSIONLIMIT errors, which are
2518 probably the only errors that are likely to be caused by users (by
2519 specifying a regex that has nested indefinite repeats, for instance).
2520 If there are more than 20 of these errors, pcregrep is abandoned.
2521
2522 (b) A binary zero was treated as data while matching, but terminated the
2523 output line if it was written out. This has been fixed: binary zeroes
2524 are now no different to any other data bytes.
2525
2526 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2527 used to set a locale for matching. The --locale=xxxx long option has
2528 been added (no short equivalent) to specify a locale explicitly on the
2529 pcregrep command, overriding the environment variables.
2530
2531 (d) When -B was used with -n, some line numbers in the output were one less
2532 than they should have been.
2533
2534 (e) Added the -o (--only-matching) option.
2535
2536 (f) If -A or -C was used with -c (count only), some lines of context were
2537 accidentally printed for the final match.
2538
2539 (g) Added the -H (--with-filename) option.
2540
2541 (h) The combination of options -rh failed to suppress file names for files
2542 that were found from directory arguments.
2543
2544 (i) Added the -D (--devices) and -d (--directories) options.
2545
2546 (j) Added the -F (--fixed-strings) option.
2547
2548 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2549
2550 (l) Added the --colo(u)r option.
2551
2552 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2553 is not present by default.
2554
2555 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2556 items such as (?R) or (?1), when the recursion could match a number of
2557 alternatives. If it matched one of the alternatives, but subsequently,
2558 outside the recursion, there was a failure, the code tried to back up into
2559 the recursion. However, because of the way PCRE is implemented, this is not
2560 possible, and the result was an incorrect result from the match.
2561
2562 In order to prevent this happening, the specification of recursion has
2563 been changed so that all such subpatterns are automatically treated as
2564 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2565
2566 4. I had overlooked the fact that, in some locales, there are characters for
2567 which isalpha() is true but neither isupper() nor islower() are true. In
2568 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2569 and ordfeminine) are like this. This affected the treatment of \w and \W
2570 when they appeared in character classes, but not when they appeared outside
2571 a character class. The bit map for "word" characters is now created
2572 separately from the results of isalnum() instead of just taking it from the
2573 upper, lower, and digit maps. (Plus the underscore character, of course.)
2574
2575 5. The above bug also affected the handling of POSIX character classes such as
2576 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2577 permanent tables. Instead, the bit maps for such a class were previously
2578 created as the appropriate unions of the upper, lower, and digit bitmaps.
2579 Now they are created by subtraction from the [[:word:]] class, which has
2580 its own bitmap.
2581
2582 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2583 It is created by subtracting the vertical space characters (\x09, \x0a,
2584 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2585 subtraction was done in the overall bitmap for a character class, meaning
2586 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2587 be recognized. This bug has been fixed.
2588
2589 7. Patches from the folks at Google:
2590
2591 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2592 real life, but is still worth protecting against".
2593
2594 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2595 regular expressions".
2596
2597 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2598 have it.
2599
2600 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2601 "configure" and the latter not, in order to fix a problem somebody had
2602 with compiling the Arg class on HP-UX.
2603
2604 (e) Improve the error-handling of the C++ wrapper a little bit.
2605
2606 (f) New tests for checking recursion limiting.
2607
2608 8. The pcre_memmove() function, which is used only if the environment does not
2609 have a standard memmove() function (and is therefore rarely compiled),
2610 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2611 returning a result (though PCRE never actually uses the result).
2612
2613 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2614 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2615 returned instead of calling malloc() with an overflowing number that would
2616 most likely cause subsequent chaos.
2617
2618 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2619
2620 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2621 with this option is matched, the nmatch and pmatch options of regexec() are
2622 ignored.
2623
2624 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2625 provided in case anyone wants to the the POSIX interface with UTF-8
2626 strings.
2627
2628 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2629 C++ linking (needed for some HP-UX environments).
2630
2631 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2632 (unused parameter) and in the pcre_printint() function (omitted "default"
2633 switch label when the default is to do nothing).
2634
2635 15. Added some code to make it possible, when PCRE is compiled as a C++
2636 library, to replace subject pointers for pcre_exec() with a smart pointer
2637 class, thus making it possible to process discontinuous strings.
2638
2639 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2640 much the same function. They were added by different people who were trying
2641 to make PCRE easy to compile on non-Unix systems. It has been suggested
2642 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2643 for compiling on Windows systems. I have therefore replaced it with
2644 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2645 defaults to "extern" for C or "extern C" for C++, which works fine on
2646 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2647 SCOPE with something explicit in config.h. In addition:
2648
2649 (a) pcreposix.h still had just "extern" instead of either of these macros;
2650 I have replaced it with PCRE_DATA_SCOPE.
2651
2652 (b) Functions such as _pcre_xclass(), which are internal to the library,
2653 but external in the C sense, all had PCRE_EXPORT in their definitions.
2654 This is apparently wrong for the Windows case, so I have removed it.
2655 (It makes no difference on Unix-like systems.)
2656
2657 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2658 of recursive calls to match(). This is different to MATCH_LIMIT because
2659 that limits the total number of calls to match(), not all of which increase
2660 the depth of recursion. Limiting the recursion depth limits the amount of
2661 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2662 when PCRE is compiled, and changed at run time. A patch from Google adds
2663 this functionality to the C++ interface.
2664
2665 18. Changes to the handling of Unicode character properties:
2666
2667 (a) Updated the table to Unicode 4.1.0.
2668
2669 (b) Recognize characters that are not in the table as "Cn" (undefined).
2670
2671 (c) I revised the way the table is implemented to a much improved format
2672 which includes recognition of ranges. It now supports the ranges that
2673 are defined in UnicodeData.txt, and it also amalgamates other
2674 characters into ranges. This has reduced the number of entries in the
2675 table from around 16,000 to around 3,000, thus reducing its size
2676 considerably. I realized I did not need to use a tree structure after
2677 all - a binary chop search is just as efficient. Having reduced the
2678 number of entries, I extended their size from 6 bytes to 8 bytes to
2679 allow for more data.
2680
2681 (d) Added support for Unicode script names via properties such as \p{Han}.
2682
2683 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2684 matching that character.
2685
2686 20. When matching a repeated Unicode property with a minimum greater than zero,
2687 (for example \pL{2,}), PCRE could look past the end of the subject if it
2688 reached it while seeking the minimum number of characters. This could
2689 happen only if some of the characters were more than one byte long, because
2690 there is a check for at least the minimum number of bytes.
2691
2692 21. Refactored the implementation of \p and \P so as to be more general, to
2693 allow for more different types of property in future. This has changed the
2694 compiled form incompatibly. Anybody with saved compiled patterns that use
2695 \p or \P will have to recompile them.
2696
2697 22. Added "Any" and "L&" to the supported property types.
2698
2699 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2700 but give a compile time error if the value is greater than 0xff.
2701
2702 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2703 accidentally not being installed or uninstalled.
2704
2705 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2706 made were to insert the current release number. This seemed silly, because
2707 it made things harder for people building PCRE on systems that don't run
2708 "configure". I have turned pcre.h into a distributed file, no longer built
2709 by "configure", with the version identification directly included. There is
2710 no longer a pcre.h.in file.
2711
2712 However, this change necessitated a change to the pcre-config script as
2713 well. It is built from pcre-config.in, and one of the substitutions was the
2714 release number. I have updated configure.ac so that ./configure now finds
2715 the release number by grepping pcre.h.
2716
2717 26. Added the ability to run the tests under valgrind.
2718
2719
2720 Version 6.4 05-Sep-05
2721 ---------------------
2722
2723 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2724 "--" to be printed when multiple files were scanned, even when none of the
2725 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2726 consider it to be a bug, and have restored the previous behaviour.
2727
2728 2. A couple of code tidies to get rid of compiler warnings.
2729
2730 3. The pcretest program used to cheat by referring to symbols in the library
2731 whose names begin with _pcre_. These are internal symbols that are not
2732 really supposed to be visible externally, and in some environments it is
2733 possible to suppress them. The cheating is now confined to including
2734 certain files from the library's source, which is a bit cleaner.
2735
2736 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2737 file's purpose clearer.
2738
2739 5. Reorganized pcre_ucp_findchar().
2740
2741
2742 Version 6.3 15-Aug-05
2743 ---------------------
2744
2745 1. The file libpcre.pc.in did not have general read permission in the tarball.
2746
2747 2. There were some problems when building without C++ support:
2748
2749 (a) If C++ support was not built, "make install" and "make test" still
2750 tried to test it.
2751
2752 (b) There were problems when the value of CXX was explicitly set. Some
2753 changes have been made to try to fix these, and ...
2754
2755 (c) --disable-cpp can now be used to explicitly disable C++ support.
2756
2757 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2758 backslash in a target when C++ was disabled. This confuses some
2759 versions of "make", apparently. Using an intermediate variable solves
2760 this. (Same for CPP_LOBJ.)
2761
2762 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2763 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2764 necessary on certain architectures.
2765
2766 4. Added a setting of -export-symbols-regex to the link command to remove
2767 those symbols that are exported in the C sense, but actually are local
2768 within the library, and not documented. Their names all begin with
2769 "_pcre_". This is not a perfect job, because (a) we have to except some
2770 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2771 available (and never for static libraries). I have made a note to try to
2772 find a way round (a) in the future.
2773
2774
2775 Version 6.2 01-Aug-05
2776 ---------------------
2777
2778 1. There was no test for integer overflow of quantifier values. A construction
2779 such as {1111111111111111} would give undefined results. What is worse, if
2780 a minimum quantifier for a parenthesized subpattern overflowed and became
2781 negative, the calculation of the memory size went wrong. This could have
2782 led to memory overwriting.
2783
2784 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2785
2786 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2787 operating environments where this matters.
2788
2789 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2790 PCRE options from within the C++ wrapper.
2791
2792 5. Named capturing subpatterns were not being correctly counted when a pattern
2793 was compiled. This caused two problems: (a) If there were more than 100
2794 such subpatterns, the calculation of the memory needed for the whole
2795 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2796 back references of the form \12, where the number was greater than 9, were
2797 not recognized as back references, even though there were sufficient
2798 previous subpatterns.
2799
2800 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2801 versions of gcc, e.g. 2.95.4.
2802
2803
2804 Version 6.1 21-Jun-05
2805 ---------------------
2806
2807 1. There was one reference to the variable "posix" in pcretest.c that was not
2808 surrounded by "#if !defined NOPOSIX".
2809
2810 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2811 the cross-check on the old pcre_info() function, for the benefit of the
2812 cut-down version of PCRE that is currently imported into Exim.
2813
2814 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2815 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2816 patterns that start (?i) but I don't think that matters. The use of (?i) is
2817 just an example; this all applies to the other options as well.
2818
2819 4. Since libtool seems to echo the compile commands it is issuing, the output
2820 from "make" can be reduced a bit by putting "@" in front of each libtool
2821 compile command.
2822
2823 5. Patch from the folks at Google for configure.in to be a bit more thorough
2824 in checking for a suitable C++ installation before trying to compile the
2825 C++ stuff. This should fix a reported problem when a compiler was present,
2826 but no suitable headers.
2827
2828 6. The man pages all had just "PCRE" as their title. I have changed them to
2829 be the relevant file name. I have also arranged that these names are
2830 retained in the file doc/pcre.txt, which is a concatenation in text format
2831 of all the man pages except the little individual ones for each function.
2832
2833 7. The NON-UNIX-USE file had not been updated for the different set of source
2834 files that come with release 6. I also added a few comments about the C++
2835 wrapper.
2836
2837
2838 Version 6.0 07-Jun-05
2839 ---------------------
2840
2841 1. Some minor internal re-organization to help with my DFA experiments.
2842
2843 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2844 didn't matter for the library itself when fully configured, but did matter
2845 when compiling without UCP support, or within Exim, where the ucp files are
2846 not imported.
2847
2848 3. Refactoring of the library code to split up the various functions into
2849 different source modules. The addition of the new DFA matching code (see
2850 below) to a single monolithic source would have made it really too
2851 unwieldy, quite apart from causing all the code to be include in a
2852 statically linked application, when only some functions are used. This is
2853 relevant even without the DFA addition now that patterns can be compiled in
2854 one application and matched in another.
2855
2856 The downside of splitting up is that there have to be some external
2857 functions and data tables that are used internally in different modules of
2858 the library but which are not part of the API. These have all had their
2859 names changed to start with "_pcre_" so that they are unlikely to clash
2860 with other external names.
2861
2862 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2863 a different (DFA) algorithm. Although it is slower than the original
2864 function, it does have some advantages for certain types of matching
2865 problem.
2866
2867 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2868 including restarting after a partial match.
2869
2870 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2871 defined when compiling for Windows was sent to me. I have put it into the
2872 code, though I have no means of testing or verifying it.
2873
2874 7. Added the pcre_refcount() auxiliary function.
2875
2876 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2877 match before or at the first newline in the subject string. In pcretest,
2878 the /f option on a pattern can be used to set this.
2879
2880 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2881 would behave wrongly. This has been present in PCRE since release 4.0.
2882
2883 10. A number of changes to the pcregrep command:
2884
2885 (a) Refactored how -x works; insert ^(...)$ instead of setting
2886 PCRE_ANCHORED and checking the length, in preparation for adding
2887 something similar for -w.
2888
2889 (b) Added the -w (match as a word) option.
2890
2891 (c) Refactored the way lines are read and buffered so as to have more
2892 than one at a time available.
2893
2894 (d) Implemented a pcregrep test script.
2895
2896 (e) Added the -M (multiline match) option. This allows patterns to match
2897 over several lines of the subject. The buffering ensures that at least
2898 8K, or the rest of the document (whichever is the shorter) is available
2899 for matching (and similarly the previous 8K for lookbehind assertions).
2900
2901 (f) Changed the --help output so that it now says
2902
2903 -w, --word-regex(p)
2904
2905 instead of two lines, one with "regex" and the other with "regexp"
2906 because that confused at least one person since the short forms are the
2907 same. (This required a bit of code, as the output is generated
2908 automatically from a table. It wasn't just a text change.)
2909
2910 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2911 option but starts with a hyphen. Could be a pattern or a path name
2912 starting with a hyphen, for instance.
2913
2914 (h) "-" can be given as a file name to represent stdin.
2915
2916 (i) When file names are being printed, "(standard input)" is used for
2917 the standard input, for compatibility with GNU grep. Previously
2918 "<stdin>" was used.
2919
2920 (j) The option --label=xxx can be used to supply a name to be used for
2921 stdin when file names are being printed. There is no short form.
2922
2923 (k) Re-factored the options decoding logic because we are going to add
2924 two more options that take data. Such options can now be given in four
2925 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2926
2927 (l) Added the -A, -B, and -C options for requesting that lines of context
2928 around matches be printed.
2929
2930 (m) Added the -L option to print the names of files that do not contain
2931 any matching lines, that is, the complement of -l.
2932
2933 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2934 continue to scan other files.
2935
2936 (o) The -s option was incorrectly implemented. For compatibility with other
2937 greps, it now suppresses the error message for a non-existent or non-
2938 accessible file (but not the return code). There is a new option called
2939 -q that suppresses the output of matching lines, which was what -s was
2940 previously doing.
2941
2942 (p) Added --include and --exclude options to specify files for inclusion
2943 and exclusion when recursing.
2944
2945 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2946 Hopefully, it now does.
2947
2948 12. Missing cast in pcre_study().
2949
2950 13. Added an "uninstall" target to the makefile.
2951
2952 14. Replaced "extern" in the function prototypes in Makefile.in with
2953 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
2954 world, but is set differently for Windows.
2955
2956 15. Added a second compiling function called pcre_compile2(). The only
2957 difference is that it has an extra argument, which is a pointer to an
2958 integer error code. When there is a compile-time failure, this is set
2959 non-zero, in addition to the error test pointer being set to point to an
2960 error message. The new argument may be NULL if no error number is required
2961 (but then you may as well call pcre_compile(), which is now just a
2962 wrapper). This facility is provided because some applications need a
2963 numeric error indication, but it has also enabled me to tidy up the way
2964 compile-time errors are handled in the POSIX wrapper.
2965
2966 16. Added VPATH=.libs to the makefile; this should help when building with one
2967 prefix path and installing with another. (Or so I'm told by someone who
2968 knows more about this stuff than I do.)
2969
2970 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
2971 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
2972 match everything, including newlines. This is not POSIX-compatible, but
2973 somebody wanted the feature. From pcretest it can be activated by using
2974 both the P and the s flags.
2975
2976 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
2977
2978 19. libpcre.pc was being incorrectly installed as executable.
2979
2980 20. A couple of places in pcretest check for end-of-line by looking for '\n';
2981 it now also looks for '\r' so that it will work unmodified on Windows.
2982
2983 21. Added Google's contributed C++ wrapper to the distribution.
2984
2985 22. Added some untidy missing memory free() calls in pcretest, to keep
2986 Electric Fence happy when testing.
2987
2988
2989
2990 Version 5.0 13-Sep-04
2991 ---------------------
2992
2993 1. Internal change: literal characters are no longer packed up into items
2994 containing multiple characters in a single byte-string. Each character
2995 is now matched using a separate opcode. However, there may be more than one
2996 byte in the character in UTF-8 mode.
2997
2998 2. The pcre_callout_block structure has two new fields: pattern_position and
2999 next_item_length. These contain the offset in the pattern to the next match
3000 item, and its length, respectively.
3001
3002 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3003 insertion of callouts before each pattern item. Added the /C option to
3004 pcretest to make use of this.
3005
3006 4. On the advice of a Windows user, the lines
3007
3008 #if defined(_WIN32) || defined(WIN32)
3009 _setmode( _fileno( stdout ), 0x8000 );
3010 #endif /* defined(_WIN32) || defined(WIN32) */
3011
3012 have been added to the source of pcretest. This apparently does useful
3013 magic in relation to line terminators.
3014
3015 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3016 for the benefit of those environments where the "b" makes a difference.
3017
3018 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3019 to know about it. I have put a hack into configure.in that adds in code
3020 to set GCC=yes if CC=icc. This seems to end up at a point in the
3021 generated configure script that is early enough to affect the setting of
3022 compiler options, which is what is needed, but I have no means of testing
3023 whether it really works. (The user who reported this had patched the
3024 generated configure script, which of course I cannot do.)
3025
3026 LATER: After change 22 below (new libtool files), the configure script
3027 seems to know about icc (and also ecc). Therefore, I have commented out
3028 this hack in configure.in.
3029
3030 7. Added support for pkg-config (2 patches were sent in).
3031
3032 8. Negated POSIX character classes that used a combination of internal tables
3033 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3034 [[:^ascii]]. Typically, they would match almost any characters. The other
3035 POSIX classes were not broken in this way.
3036
3037 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3038 to find the match, as PCRE was deluded into thinking that the match had to
3039 start at the start point or following a newline. The same bug applied to
3040 patterns with negative forward assertions or any backward assertions
3041 preceding ".*" at the start, unless the pattern required a fixed first
3042 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3043
3044 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3045 starting at the last subject character, bytes beyond the end of the subject
3046 string were read.
3047
3048 11. Renamed the variable "class" as "classbits" to make life easier for C++
3049 users. (Previously there was a macro definition, but it apparently wasn't
3050 enough.)
3051
3052 12. Added the new field "tables" to the extra data so that tables can be passed
3053 in at exec time, or the internal tables can be re-selected. This allows
3054 a compiled regex to be saved and re-used at a later time by a different
3055 program that might have everything at different addresses.
3056
3057 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3058 -R library as well as a -L library.
3059
3060 14. The debugging options of pcretest (-d on the command line or D on a
3061 pattern) showed incorrect output for anything following an extended class
3062 that contained multibyte characters and which was followed by a quantifier.
3063
3064 15. Added optional support for general category Unicode character properties
3065 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3066 support. It adds about 90K to the size of the library. The meanings of the
3067 inbuilt class escapes such as \d and \s have NOT been changed.
3068
3069 16. Updated pcredemo.c to include calls to free() to release the memory for the
3070 compiled pattern.
3071
3072 17. The generated file chartables.c was being created in the source directory
3073 instead of in the building directory. This caused the build to fail if the
3074 source directory was different from the building directory, and was
3075 read-only.
3076
3077 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3078 file. No doubt somebody will tell me if they don't make sense... Also added
3079 Dan Mooney's comments about building on OpenVMS.
3080
3081 19. Added support for partial matching via the PCRE_PARTIAL option for
3082 pcre_exec() and the \P data escape in pcretest.
3083
3084 20. Extended pcretest with 3 new pattern features:
3085
3086 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3087 write the compiled pattern to the file whose name is "rest-of-line".
3088 This is a straight binary dump of the data, with the saved pointer to
3089 the character tables forced to be NULL. The study data, if any, is
3090 written too. After writing, pcretest reads a new pattern.
3091
3092 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3093 compiled pattern from the given file. There must not be any
3094 occurrences of "<" in the file name (pretty unlikely); if there are,
3095 pcretest will instead treat the initial "<" as a pattern delimiter.
3096 After reading in the pattern, pcretest goes on to read data lines as
3097 usual.
3098
3099 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3100 and 16-bit fields in a compiled pattern, to simulate a pattern that
3101 was compiled on a host of opposite endianness.
3102
3103 21. The pcre-exec() function can now cope with patterns that were compiled on
3104 hosts of opposite endianness, with this restriction:
3105
3106 As for any compiled expression that is saved and used later, the tables
3107 pointer field cannot be preserved; the extra_data field in the arguments
3108 to pcre_exec() should be used to pass in a tables address if a value
3109 other than the default internal tables were used at compile time.
3110
3111 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3112 now diagnosed as an error. Previously, most of the time, a negative number
3113 would have been treated as zero, but if in addition "ovector" was passed as
3114 NULL, a crash could occur.
3115
3116 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3117 new versions from the libtool 1.5 distribution (the last one is a copy of
3118 a file called libtool.m4). This seems to have fixed the need to patch
3119 "configure" to support Darwin 1.3 (which I used to do). However, I still
3120 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3121 workstation).
3122
3123 24. Changed the PCRE licence to be the more standard "BSD" licence.
3124
3125
3126 Version 4.5 01-Dec-03
3127 ---------------------
3128
3129 1. There has been some re-arrangement of the code for the match() function so
3130 that it can be compiled in a version that does not call itself recursively.
3131 Instead, it keeps those local variables that need separate instances for
3132 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3133 needs to "recurse". Keeping track of where control must go is done by means
3134 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3135 hide most of the details from the main code, and operates only if
3136 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3137 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3138 operating.
3139
3140 To make it easier for callers to provide specially tailored get/free
3141 functions for this usage, two new functions, pcre_stack_malloc, and
3142 pcre_stack_free, are used. They are always called in strict stacking order,
3143 and the size of block requested is always the same.
3144
3145 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3146 PCRE has been compiled to use the stack or the heap for recursion. The
3147 -C option of pcretest uses this to show which version is compiled.
3148
3149 A new data escape \S, is added to pcretest; it causes the amounts of store
3150 obtained and freed by both kinds of malloc/free at match time to be added
3151 to the output.
3152
3153 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3154 what's available on my current Linux desktop machine.
3155
3156 3. When matching a UTF-8 string, the test for a valid string at the start has
3157 been extended. If start_offset is not zero, PCRE now checks that it points
3158 to a byte that is the start of a UTF-8 character. If not, it returns
3159 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3160 this is necessary because there may be backward assertions in the pattern.
3161 When matching the same subject several times, it may save resources to use
3162 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3163
3164 4. The code for checking the validity of UTF-8 strings has been tightened so
3165 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3166 containing "overlong sequences".
3167
3168 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3169 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3170 should have been "&", but it just so happened that all the cases this let
3171 through by mistake were picked up later in the function.
3172
3173 6. I had used a variable called "isblank" - this is a C99 function, causing
3174 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3175
3176 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3177 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3178 so the version is shown; (c) stop "make test" echoing "./RunTest".
3179
3180 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3181
3182 9. The prototype for memmove() for systems that don't have it was using
3183 size_t, but the inclusion of the header that defines size_t was later. I've
3184 moved the #includes for the C headers earlier to avoid this.
3185
3186 10. Added some adjustments to the code to make it easier to compiler on certain
3187 special systems:
3188
3189 (a) Some "const" qualifiers were missing.
3190 (b) Added the macro EXPORT before all exported functions; by default this
3191 is defined to be empty.
3192 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3193 that it reads its output file name as an argument instead of writing
3194 to the standard output and assuming this can be redirected.
3195
3196 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3197 class containing characters with values greater than 255, PCRE compilation
3198 went into a loop.
3199
3200 12. A recursive reference to a subpattern that was within another subpattern
3201 that had a minimum quantifier of zero caused PCRE to crash. For example,
3202 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3203 recursion. If the recursively-called subpattern itself had a zero repeat,
3204 that was OK.
3205
3206 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3207 buffer into which it was copied (for escape processing) was still set at
3208 1024, so long lines caused crashes.
3209
3210 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3211 "internal error: code overflow...". This applied to any character class
3212 that was followed by a possessive quantifier.
3213
3214 15. Modified the Makefile to add libpcre.la as a prerequisite for
3215 libpcreposix.la because I was told this is needed for a parallel build to
3216 work.
3217
3218 16. If a pattern that contained .* following optional items at the start was
3219 studied, the wrong optimizing data was generated, leading to matching
3220 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3221 matching string must start with a or b or c. The correct conclusion for
3222 this pattern is that a match can start with any character.
3223
3224
3225 Version 4.4 13-Aug-03
3226 ---------------------
3227
3228 1. In UTF-8 mode, a character class containing characters with values between
3229 127 and 255 was not handled correctly if the compiled pattern was studied.
3230 In fixing this, I have also improved the studying algorithm for such
3231 classes (slightly).
3232
3233 2. Three internal functions had redundant arguments passed to them. Removal
3234 might give a very teeny performance improvement.
3235
3236 3. Documentation bug: the value of the capture_top field in a callout is *one
3237 more than* the number of the hightest numbered captured substring.
3238
3239 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3240 in incorrectly linking with a previously installed version. They now link
3241 explicitly with libpcre.la.
3242
3243 5. configure.in no longer needs to recognize Cygwin specially.
3244
3245 6. A problem in pcre.in for Windows platforms is fixed.
3246
3247 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3248 pcretest, it used to include the size of the study block as part of its
3249 output. Unfortunately, the structure contains a field that has a different
3250 size on different hardware architectures. This meant that the tests that
3251 showed this size failed. As the block is currently always of a fixed size,
3252 this information isn't actually particularly useful in pcretest output, so
3253 I have just removed it.
3254
3255 8. Three pre-processor statements accidentally did not start in column 1.
3256 Sadly, there are *still* compilers around that complain, even though
3257 standard C has not required this for well over a decade. Sigh.
3258
3259 9. In pcretest, the code for checking callouts passed small integers in the
3260 callout_data field, which is a void * field. However, some picky compilers
3261 complained about the casts involved for this on 64-bit systems. Now
3262 pcretest passes the address of the small integer instead, which should get
3263 rid of the warnings.
3264
3265 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3266 both compile and run time, and gives an error if an invalid UTF-8 sequence
3267 is found. There is a option for disabling this check in cases where the
3268 string is known to be correct and/or the maximum performance is wanted.
3269
3270 11. In response to a bug report, I changed one line in Makefile.in from
3271
3272 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3273 to
3274 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3275
3276 to look similar to other lines, but I have no way of telling whether this
3277 is the right thing to do, as I do not use Windows. No doubt I'll get told
3278 if it's wrong...
3279
3280
3281 Version 4.3 21-May-03
3282 ---------------------
3283
3284 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3285 Makefile.
3286
3287 2. Some refactoring to improve the quality of the code:
3288
3289 (i) The utf8_table... variables are now declared "const".
3290
3291 (ii) The code for \cx, which used the "case flipping" table to upper case
3292 lower case letters, now just substracts 32. This is ASCII-specific,
3293 but the whole concept of \cx is ASCII-specific, so it seems
3294 reasonable.
3295
3296 (iii) PCRE was using its character types table to recognize decimal and
3297 hexadecimal digits in the pattern. This is silly, because it handles
3298 only 0-9, a-f, and A-F, but the character types table is locale-
3299 specific, which means strange things might happen. A private
3300 table is now used for this - though it costs 256 bytes, a table is
3301 much faster than multiple explicit tests. Of course, the standard
3302 character types table is still used for matching digits in subject
3303 strings against \d.
3304
3305 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3306 ending in _t are). So I've renamed it as ESC_tee.
3307
3308 3. The first argument for regexec() in the POSIX wrapper should have been
3309 defined as "const".
3310
3311 4. Changed pcretest to use malloc() for its buffers so that they can be
3312 Electric Fenced for debugging.
3313
3314 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3315 to read one or more bytes before the start of the subject string. Often this
3316 had no effect on PCRE's behaviour, but in some circumstances it could
3317 provoke a segmentation fault.
3318
3319 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3320 to try to read one or more bytes before the start of the subject string.
3321
3322 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3323 UTF-8 support could misbehave in various ways if the subject string
3324 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3325 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3326 back over UTF-8 characters.)
3327
3328
3329 Version 4.2 14-Apr-03
3330 ---------------------
3331
3332 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3333
3334 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3335 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3336 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3337 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3338 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3339 and BUILD_EXEEXT
3340 Note: automatic setting of the BUILD variables is not yet working
3341 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3342 compile-time but not at link-time
3343 [LINK]: use for linking executables only
3344 make different versions for Windows and non-Windows
3345 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3346 libraries
3347 [LINK_FOR_BUILD]: new variable
3348 [OBJEXT]: use throughout
3349 [EXEEXT]: use throughout
3350 <winshared>: new target
3351 <wininstall>: new target
3352 <dftables.o>: use native compiler
3353 <dftables>: use native linker
3354 <install>: handle Windows platform correctly
3355 <clean>: ditto
3356 <check>: ditto
3357 copy DLL to top builddir before testing
3358
3359 As part of these changes, -no-undefined was removed again. This was reported
3360 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3361 in any case.
3362
3363 3. Some tidies to get rid of compiler warnings:
3364
3365 . In the match_data structure, match_limit was an unsigned long int, whereas
3366 match_call_count was an int. I've made them both unsigned long ints.
3367
3368 . In pcretest the fact that a const uschar * doesn't automatically cast to
3369 a void * provoked a warning.
3370
3371 . Turning on some more compiler warnings threw up some "shadow" variables
3372 and a few more missing casts.
3373
3374 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3375 option, a class that contained a single character with a value between 128
3376 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3377
3378 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3379 option, a class that contained several characters, but with at least one
3380 whose value was between 128 and 255 caused PCRE to crash.
3381
3382
3383 Version 4.1 12-Mar-03
3384 ---------------------
3385
3386 1. Compiling with gcc -pedantic found a couple of places where casts were
3387 needed, and a string in dftables.c that was longer than standard compilers are
3388 required to support.
3389
3390 2. Compiling with Sun's compiler found a few more places where the code could
3391 be tidied up in order to avoid warnings.
3392
3393 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3394 first of these names is deprecated in the latest Autoconf in favour of the name
3395 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3396 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3397 analogy I have changed it to CFLAGS_FOR_BUILD.
3398
3399 4. Added -no-undefined to the linking command in the Makefile, because this is
3400 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3401 linking step for the pcreposix library.
3402
3403 5. PCRE was failing to diagnose the case of two named groups with the same
3404 name.
3405
3406 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3407 literal character that is needed in the subject for a match, and scans along to
3408 ensure that it is present before embarking on the full matching process. This
3409 saves time in cases of nested unlimited repeats that are never going to match.
3410 Problem: the scan can take a lot of time if the subject is very long (e.g.
3411 megabytes), thus penalizing straightforward matches. It is now done only if the
3412 amount of subject to be scanned is less than 1000 bytes.
3413
3414 7. A lesser problem with the same optimization is that it was recording the
3415 first character of an anchored pattern as "needed", thus provoking a search
3416 right along the subject, even when the first match of the pattern was going to
3417 fail. The "needed" character is now not set for anchored patterns, unless it
3418 follows something in the pattern that is of non-fixed length. Thus, it still
3419 fulfils its original purpose of finding quick non-matches in cases of nested
3420 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3421
3422
3423 Version 4.0 17-Feb-03
3424 ---------------------
3425
3426 1. If a comment in an extended regex that started immediately after a meta-item
3427 extended to the end of string, PCRE compiled incorrect data. This could lead to
3428 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3429
3430 2. Moved to autoconf 2.53 and libtool 1.4.2.
3431
3432 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3433 the special perltest8 script is no longer needed - all the tests can be run
3434 from a single perltest script.
3435
3436 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3437 by \s. It has now been removed in PCRE. This means it isn't recognized as
3438 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3439 class [:space:] *does* include VT, thereby creating a mess.
3440
3441 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3442 space and tab.
3443
3444 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3445 its new features into the main test script, reducing the number of scripts.
3446
3447 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3448 were backward compatible, and made the (?i) apply to the whole pattern, as if
3449 /i were given. Now it behaves more logically, and applies the option setting
3450 only to what follows. PCRE has been changed to follow suit. However, if it
3451 finds options settings right at the start of the pattern, it extracts them into
3452 the global options, as before. Thus, they show up in the info data.
3453
3454 8. Added support for the \Q...\E escape sequence. Characters in between are
3455 treated as literals. This is slightly different from Perl in that $ and @ are
3456 also handled as literals inside the quotes. In Perl, they will cause variable
3457 interpolation. Note the following examples:
3458
3459 Pattern PCRE matches Perl matches
3460
3461 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3462 \Qabc\$xyz\E abc\$xyz abc\$xyz
3463 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3464
3465 For compatibility with Perl, \Q...\E sequences are recognized inside character
3466 classes as well as outside them.
3467
3468 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3469 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3470 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3471 signed/unsigned warnings.
3472
3473 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3474 option for pcretest, so I've replaced it by a simple function that does just
3475 that job.
3476
3477 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3478 "pcregrep -".
3479
3480 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3481 Java package. This provides some syntactic sugar for simple cases of what my
3482 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3483 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3484 item, you can use this simplified notation. Note that only makes sense with
3485 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3486 greediness, whatever the setting of the PCRE_UNGREEDY option.
3487
3488 13. A change of greediness default within a pattern was not taking effect at
3489 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3490 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3491 was abstracted outside.
3492
3493 14. PCRE now supports the \G assertion. It is true when the current matching
3494 position is at the start point of the match. This differs from \A when the
3495 starting offset is non-zero. Used with the /g option of pcretest (or similar
3496 code), it works in the same way as it does for Perl's /g option. If all
3497 alternatives of a regex begin with \G, the expression is anchored to the start
3498 match position, and the "anchored" flag is set in the compiled expression.
3499
3500 15. Some bugs concerning the handling of certain option changes within patterns
3501 have been fixed. These applied to options other than (?ims). For example,
3502 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3503 been the other way round. Some of this was related to change 7 above.
3504
3505 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3506 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3507 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3508 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3509
3510 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3511 mode. Unlike ".", it always matches newline, whatever the setting of
3512 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3513 assertions. Perl allows it, but it doesn't (in general) work because it can't
3514 calculate the length of the lookbehind. At least, that's the case for Perl
3515 5.8.0 - I've been told they are going to document that it doesn't work in
3516 future.
3517
3518 18. Added an error diagnosis for escapes that PCRE does not support: these are
3519 \L, \l, \N, \P, \p, \U, \u, and \X.
3520
3521 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3522 reading past the end of the pattern in cases such as /[abcd/.
3523
3524 20. PCRE was getting more memory than necessary for patterns with classes that
3525 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3526
3527 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3528 compiling PCRE for use with Virtual Pascal.
3529
3530 22. Small fix to the Makefile to make it work properly if the build is done
3531 outside the source tree.
3532
3533 23. Added a new extension: a condition to go with recursion. If a conditional
3534 subpattern starts with (?(R) the "true" branch is used if recursion has
3535 happened, whereas the "false" branch is used only at the top level.
3536
3537 24. When there was a very long string of literal characters (over 255 bytes
3538 without UTF support, over 250 bytes with UTF support), the computation of how
3539 much memory was required could be incorrect, leading to segfaults or other
3540 strange effects.
3541
3542 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3543 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3544 there was a subsequent back reference to those brackets. This meant that, for
3545 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3546 possible to check for precisely this case. All we can do is abandon the
3547 optimization if .* occurs inside capturing brackets when there are any back
3548 references whatsoever. (See below for a better fix that came later.)
3549
3550 26. The handling of the optimization for finding the first character of a
3551 non-anchored pattern, and for finding a character that is required later in the
3552 match were failing in some cases. This didn't break the matching; it just
3553 failed to optimize when it could. The way this is done has been re-implemented.
3554
3555 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3556
3557 28. Added a new feature that provides some of the functionality that Perl
3558 provides with (?{...}). The facility is termed a "callout". The way it is done
3559 in PCRE is for the caller to provide an optional function, by setting
3560 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3561 global variable. By default it is unset, which disables all calling out. To get
3562 the function called, the regex must include (?C) at appropriate points. This
3563 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3564 This provides a means of identifying different callout points. When PCRE
3565 reaches such a point in the regex, if pcre_callout has been set, the external
3566 function is called. It is provided with data in a structure called
3567 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3568 matching continues; if it returns a non-zero value, the match at the current
3569 point fails. However, backtracking will occur if possible. [This was changed
3570 later and other features added - see item 49 below.]
3571
3572 29. pcretest is upgraded to test the callout functionality. It provides a
3573 callout function that displays information. By default, it shows the start of
3574 the match and the current position in the text. There are some new data escapes
3575 to vary what happens:
3576
3577 \C+ in addition, show current contents of captured substrings
3578 \C- do not supply a callout function
3579 \C!n return 1 when callout number n is reached
3580 \C!n!m return 1 when callout number n is reached for the mth time
3581
3582 30. If pcregrep was called with the -l option and just a single file name, it
3583 output "<stdin>" if a match was found, instead of the file name.
3584
3585 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3586 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3587 pcre_exec(). This saves a malloc/free per call. The default value of
3588 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3589 when configuring.
3590
3591 32. The default maximum size of a compiled pattern is 64K. There have been a
3592 few cases of people hitting this limit. The code now uses macros to handle the
3593 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3594 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3595 configuring. Tests 2 and 5 work only with 2-byte links because they output
3596 debugging information about compiled patterns.
3597
3598 33. Internal code re-arrangements:
3599
3600 (a) Moved the debugging function for printing out a compiled regex into
3601 its own source file (printint.c) and used #include to pull it into
3602 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3603 separate copies.
3604
3605 (b) Defined the list of op-code names for debugging as a macro in
3606 internal.h so that it is next to the definition of the opcodes.
3607
3608 (c) Defined a table of op-code lengths for simpler skipping along compiled
3609 code. This is again a macro in internal.h so that it is next to the
3610 definition of the opcodes.
3611
3612 34. Added support for recursive calls to individual subpatterns, along the
3613 lines of Robin Houston's patch (but implemented somewhat differently).
3614
3615 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3616 allow it to read and process whole directories in Win32. This code was
3617 contributed by Lionel Fourquaux; it has not been tested by me.
3618
3619 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3620 used to name a group. Names consist of alphanumerics and underscores, and must
3621 be unique. Back references use the syntax (?P=name) and recursive calls use
3622 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3623 numbers. The function pcre_fullinfo() can be used after compilation to extract
3624 a name/number map. There are three relevant calls:
3625
3626 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3627 PCRE_INFO_NAMECOUNT yields the number of entries
3628 PCRE_INFO_NAMETABLE yields a pointer to the map.
3629
3630 The map is a vector of fixed-size entries. The size of each entry depends on
3631 the length of the longest name used. The first two bytes of each entry are the
3632 group number, most significant byte first. There follows the corresponding
3633 name, zero terminated. The names are in alphabetical order.
3634
3635 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3636 case instead of 255. Making it the same both with and without UTF-8 support
3637 means that the same test output works with both.
3638
3639 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3640 calling malloc() with a zero argument.
3641
3642 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3643 optimization. I've improved things by keeping a bitmap of backreferences with
3644 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3645 fact referenced, the optimization can be applied. It is unlikely that a
3646 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3647 the match to follow \n) will appear inside brackets with a number greater than
3648 31, but if it does, any back reference > 31 suppresses the optimization.
3649
3650 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3651 of disabling numbered capturing parentheses. Any opening parenthesis that is
3652 not followed by ? behaves as if it were followed by ?: but named parentheses
3653 can still be used for capturing (and they will acquire numbers in the usual
3654 way).
3655
3656 41. Redesigned the return codes from the match() function into yes/no/error so
3657 that errors can be passed back from deep inside the nested calls. A malloc
3658 failure while inside a recursive subpattern call now causes the
3659 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3660
3661 42. It is now possible to set a limit on the number of times the match()
3662 function is called in a call to pcre_exec(). This facility makes it possible to
3663 limit the amount of recursion and backtracking, though not in a directly
3664 obvious way, because the match() function is used in a number of different
3665 circumstances. The count starts from zero for each position in the subject
3666 string (for non-anchored patterns). The default limit is, for compatibility, a
3667 large number, namely 10 000 000. You can change this in two ways:
3668
3669 (a) When configuring PCRE before making, you can use --with-match-limit=n
3670 to set a default value for the compiled library.
3671
3672 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3673 a different value is set. See 45 below.
3674
3675 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3676
3677 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3678 of things that can be changed at compile time. The first argument specifies
3679 what is wanted and the second points to where the information is to be placed.
3680 The current list of available information is:
3681
3682 PCRE_CONFIG_UTF8
3683
3684 The output is an integer that is set to one if UTF-8 support is available;
3685 otherwise it is set to zero.
3686
3687 PCRE_CONFIG_NEWLINE
3688
3689 The output is an integer that it set to the value of the code that is used for
3690 newline. It is either LF (10) or CR (13).
3691
3692 PCRE_CONFIG_LINK_SIZE
3693
3694 The output is an integer that contains the number of bytes used for internal
3695 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3696
3697 PCRE_CONFIG_POSIX_MALLOC_THRESHOLD
3698
3699 The output is an integer that contains the threshold above which the POSIX
3700 interface uses malloc() for output vectors. See item 31 above.
3701
3702 PCRE_CONFIG_MATCH_LIMIT
3703
3704 The output is an unsigned integer that contains the default limit of the number
3705 of match() calls in a pcre_exec() execution. See 42 above.
3706
3707 44. pcretest has been upgraded by the addition of the -C option. This causes it
3708 to extract all the available output from the new pcre_config() function, and to
3709 output it. The program then exits immediately.
3710
3711 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3712 order to support additional features. One way would have been to define
3713 pcre_exec2() (for example) with extra arguments, but this would not have been
3714 extensible, and would also have required all calls to the original function to
3715 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3716 is used for passing in "extra" data from pcre_study().
3717
3718 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3719 contains the following fields:
3720
3721 flags a bitmap indicating which of the following fields are set
3722 study_data opaque data from pcre_study()
3723 match_limit a way of specifying a limit on match() calls for a specific
3724 call to pcre_exec()
3725 callout_data data for callouts (see 49 below)
3726
3727 The flag bits are also defined in pcre.h, and are
3728
3729 PCRE_EXTRA_STUDY_DATA
3730 PCRE_EXTRA_MATCH_LIMIT
3731 PCRE_EXTRA_CALLOUT_DATA
3732
3733 The pcre_study() function now returns one of these new pcre_extra blocks, with
3734 the actual study data pointed to by the study_data field, and the
3735 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3736 before. That is, this change is entirely upwards-compatible and requires no
3737 change to existing code.
3738
3739 If you want to pass in additional data to pcre_exec(), you can either place it
3740 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3741 block.
3742
3743 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3744 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3745 times with different match limits, until it finds the minimum value needed for
3746 pcre_exec() to complete. The value is then output. This can be instructive; for
3747 most simple matches the number is quite small, but for pathological cases it
3748 gets very large very quickly.
3749
3750 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3751 returns the size of the data block pointed to by the study_data field in a
3752 pcre_extra block, that is, the value that was passed as the argument to
3753 pcre_malloc() when PCRE was getting memory in which to place the information
3754 created by pcre_study(). The fourth argument should point to a size_t variable.
3755 pcretest has been extended so that this information is shown after a successful
3756 pcre_study() call when information about the compiled regex is being displayed.
3757
3758 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3759 because what follows is always an absolute path. (Later: it turns out that this
3760 is more than cosmetic for MinGW, because it doesn't like empty path
3761 components.)
3762
3763 49. Some changes have been made to the callout feature (see 28 above):
3764
3765 (i) A callout function now has three choices for what it returns:
3766
3767 0 => success, carry on matching
3768 > 0 => failure at this point, but backtrack if possible
3769 < 0 => serious error, return this value from pcre_exec()
3770
3771 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3772 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3773 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3774 use by callout functions. It will never be used by PCRE itself.
3775
3776 (ii) The pcre_extra structure (see 45 above) has a void * field called
3777 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3778 pcre_callout_block structure has a field of the same name. The contents of
3779 the field passed in the pcre_extra structure are passed to the callout
3780 function in the corresponding field in the callout block. This makes it
3781 easier to use the same callout-containing regex from multiple threads. For
3782 testing, the pcretest program has a new data escape
3783
3784 \C*n pass the number n (may be negative) as callout_data
3785
3786 If the callout function in pcretest receives a non-zero value as
3787 callout_data, it returns that value.
3788
3789 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3790 there were some redundant $(CFLAGS) in commands that are now specified as
3791 $(LINK), which already includes $(CFLAGS).
3792
3793 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3794 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3795 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3796 one-byte characters throughout. Note that case-insensitive matching applies
3797 only to characters whose values are less than 256. PCRE doesn't support the
3798 notion of cases for higher-valued characters.
3799
3800 (i) A character class whose characters are all within 0-255 is handled as
3801 a bit map, and the map is inverted for negative classes. Previously, a
3802 character > 255 always failed to match such a class; however it should
3803 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3804
3805 (ii) A negated character class with a single character < 255 is coded as
3806 "not this character" (OP_NOT). This wasn't working properly when the test
3807 character was multibyte, either singly or repeated.
3808
3809 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3810 mode, for example: \x{100}{2,3}.
3811
3812 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3813 singly or repeated) now correctly test multibyte characters. However,
3814 PCRE doesn't recognize any characters with values greater than 255 as
3815 digits, spaces, or word characters. Such characters always match \D, \S,
3816 and \W, and never match \d, \s, or \w.
3817
3818 (v) Classes may now contain characters and character ranges with values
3819 greater than 255. For example: [ab\x{100}-\x{400}].
3820
3821 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3822 PCRE in UTF-8 mode.
3823
3824 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3825 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3826 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3827 value.)
3828
3829 53. The single man page has become too large. I have therefore split it up into
3830 a number of separate man pages. These also give rise to individual HTML pages;
3831 these are now put in a separate directory, and there is an index.html page that
3832 lists them all. Some hyperlinking between the pages has been installed.
3833
3834 54. Added convenience functions for handling named capturing parentheses.
3835
3836 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3837 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3838 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3839 are faulted.
3840
3841 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3842 calling configure. These values are used when compiling the dftables.c program
3843 which is run to generate the source of the default character tables. They
3844 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3845 you will need to set these values.
3846
3847 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3848
3849
3850 Version 3.9 02-Jan-02
3851 ---------------------
3852
3853 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3854
3855 2. If --disable-static was given, the building process failed when trying to
3856 build pcretest and pcregrep. (For some reason it was using libtool to compile
3857 them, which is not right, as they aren't part of the library.)
3858
3859
3860 Version 3.8 18-Dec-01
3861 ---------------------
3862
3863 1. The experimental UTF-8 code was completely screwed up. It was packing the
3864 bytes in the wrong order. How dumb can you get?
3865
3866
3867 Version 3.7 29-Oct-01
3868 ---------------------
3869
3870 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3871 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3872 this didn't happen under Solaris 8, where I normally test things.
3873
3874 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3875 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3876 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3877 this fix an hour or so after the initial 3.7 release.)
3878
3879
3880 Version 3.6 23-Oct-01
3881 ---------------------
3882
3883 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3884 offsets passed as NULL with zero offset count.
3885
3886 2. The config.guess and config.sub files had not been updated when I moved to
3887 the latest autoconf.
3888
3889
3890 Version 3.5 15-Aug-01
3891 ---------------------
3892
3893 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3894 had been forgotten.
3895
3896 2. By using declared but undefined structures, we can avoid using "void"
3897 definitions in pcre.h while keeping the internal definitions of the structures
3898 private.
3899
3900 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3901 user point of view, this means that both static and shared libraries are built
3902 by default, but this can be individually controlled. More of the work of
3903 handling this static/shared cases is now inside libtool instead of PCRE's make
3904 file.
3905
3906 4. The pcretest utility is now installed along with pcregrep because it is
3907 useful for users (to test regexs) and by doing this, it automatically gets
3908 relinked by libtool. The documentation has been turned into a man page, so
3909 there are now .1, .txt, and .html versions in /doc.
3910
3911 5. Upgrades to pcregrep:
3912 (i) Added long-form option names like gnu grep.
3913 (ii) Added --help to list all options with an explanatory phrase.
3914 (iii) Added -r, --recursive to recurse into sub-directories.
3915 (iv) Added -f, --file to read patterns from a file.
3916
3917 6. pcre_exec() was referring to its "code" argument before testing that
3918 argument for NULL (and giving an error if it was NULL).
3919
3920 7. Upgraded Makefile.in to allow for compiling in a different directory from
3921 the source directory.
3922
3923 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3924 options bits, the pointer it was passed was to an int instead of to an unsigned
3925 long int. This mattered only on 64-bit systems.
3926
3927 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3928 generated) instead of pcre.in, which it its source. Also made the same change
3929 in several of the .c files.
3930
3931 10. A new release of gcc defines printf() as a macro, which broke pcretest
3932 because it had an ifdef in the middle of a string argument for printf(). Fixed
3933 by using separate calls to printf().
3934
3935 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3936 script, to force use of CR or LF instead of \n in the source. On non-Unix
3937 systems, the value can be set in config.h.
3938
3939 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3940 absolute limit. Changed the text of the error message to make this clear, and
3941 likewise updated the man page.
3942
3943 13. The limit of 99 on the number of capturing subpatterns has been removed.
3944 The new limit is 65535, which I hope will not be a "real" limit.
3945
3946
3947 Version 3.4 22-Aug-00
3948 ---------------------
3949
3950 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
3951
3952 2. Diagnose condition (?(0) as an error instead of crashing on matching.
3953
3954
3955 Version 3.3 01-Aug-00
3956 ---------------------
3957
3958 1. If an octal character was given, but the value was greater than \377, it
3959 was not getting masked to the least significant bits, as documented. This could
3960 lead to crashes in some systems.
3961
3962 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
3963 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
3964
3965 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
3966 These just pass their arguments on to (pcre_free)(), but they are provided
3967 because some uses of PCRE bind it to non-C systems that can call its functions,
3968 but cannot call free() or pcre_free() directly.
3969
3970 4. Add "make test" as a synonym for "make check". Corrected some comments in
3971 the Makefile.
3972
3973 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
3974 Makefile.
3975
3976 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
3977 command called pgrep for grepping around the active processes.
3978
3979 7. Added the beginnings of support for UTF-8 character strings.
3980
3981 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
3982 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
3983 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
3984 out for the ar command.)
3985
3986
3987 Version 3.2 12-May-00
3988 ---------------------
3989
3990 This is purely a bug fixing release.
3991
3992 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
3993 of ZA. This was just one example of several cases that could provoke this bug,
3994 which was introduced by change 9 of version 2.00. The code for breaking
3995 infinite loops after an iteration that matches an empty string was't working
3996 correctly.
3997
3998 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
3999 when matched against abbab (for example). After matching an empty string, it
4000 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4001 caused it to match further down the string than it should.
4002
4003 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4004 was there because it doesn't seem to be needed, and it causes trouble on some
4005 systems, as it is not a Standard C header. It has been removed.
4006
4007 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4008 were reported on the Macintosh. The changes were from
4009
4010 while ((c = *(++ptr)) != 0 && c != '\n');
4011 to
4012 while ((c = *(++ptr)) != 0 && c != '\n') ;
4013
4014 Totally extraordinary, but if that's what it takes...
4015
4016 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4017 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4018 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4019 assumes the way PCRE uses memmove() (always moving upwards).
4020
4021 6. PCRE is being used in one environment where strchr() is not available. There
4022 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4023 faster code anyway.
4024
4025
4026 Version 3.1 09-Feb-00
4027 ---------------------
4028
4029 The only change in this release is the fixing of some bugs in Makefile.in for
4030 the "install" target:
4031
4032 (1) It was failing to install pcreposix.h.
4033
4034 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4035
4036
4037 Version 3.0 01-Feb-00
4038 ---------------------
4039
4040 1. Add support for the /+ modifier to perltest (to output $` like it does in
4041 pcretest).
4042
4043 2. Add support for the /g modifier to perltest.
4044
4045 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4046 matches null strings.
4047
4048 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4049 pattern. Perl treats empty patterns specially - it reuses the most recent
4050 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4051 effect.
4052
4053 5. The POSIX interface was broken in that it was just handing over the POSIX
4054 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4055 required a bigger vector, with some working space on the end. This means that
4056 the POSIX wrapper now has to get and free some memory, and copy the results.
4057
4058 6. Added some simple autoconf support, placing the test data and the
4059 documentation in separate directories, re-organizing some of the
4060 information files, and making it build pcre-config (a GNU standard). Also added
4061 libtool support for building PCRE as a shared library, which is now the
4062 default.
4063
4064 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4065 09 are not valid octal constants. Single digits will be used for minor values
4066 less than 10.
4067
4068 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4069 existing programs that set these in the POSIX interface can use PCRE without
4070 modification.
4071
4072 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4073 return all that pcre_info() returns, plus additional data. The pcre_info()
4074 function is retained for compatibility, but is considered to be obsolete.
4075
4076 10. Added experimental recursion feature (?R) to handle one common case that
4077 Perl 5.6 will be able to do with (?p{...}).
4078
4079 11. Added support for POSIX character classes like [:alpha:], which Perl is
4080 adopting.
4081
4082
4083 Version 2.08 31-Aug-99
4084 ----------------------
4085
4086 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4087 trying to match at the startoffset position, but instead was moving forward to
4088 the next newline as if a previous match had failed.
4089
4090 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4091 and could get into a loop if a null string was matched other than at the start
4092 of the subject.
4093
4094 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4095 be distinguished at compile time, and for completeness also added PCRE_DATE.
4096
4097 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4098 in GnuWin32 environments.
4099
4100
4101 Version 2.07 29-Jul-99
4102 ----------------------
4103
4104 1. The documentation is now supplied in plain text form and HTML as well as in
4105 the form of man page sources.
4106
4107 2. C++ compilers don't like assigning (void *) values to other pointer types.
4108 In particular this affects malloc(). Although there is no problem in Standard
4109 C, I've put in casts to keep C++ compilers happy.
4110
4111 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4112 should be (const char *).
4113
4114 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4115 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4116 However, I haven't made this a standard facility. The documentation doesn't
4117 mention it, and the Makefile doesn't support it.
4118
4119 5. The Makefile now contains an "install" target, with editable destinations at
4120 the top of the file. The pcretest program is not installed.
4121
4122 6. pgrep -V now gives the PCRE version number and date.
4123
4124 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4125 causing the entire string to be ignored, instead of just the last character.
4126
4127 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4128 non-matching string, it can take a very, very long time, even for strings of
4129 quite modest length, because of the nested recursion. PCRE now does better in
4130 some of these cases. It does this by remembering the last required literal
4131 character in the pattern, and pre-searching the subject to ensure it is present
4132 before running the real match. In other words, it applies a heuristic to detect
4133 some types of certain failure quickly, and in the above example, if presented
4134 with a string that has no trailing " it gives "no match" very quickly.
4135
4136 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4137 other alternatives are tried instead.
4138
4139
4140 Version 2.06 09-Jun-99
4141 ----------------------
4142
4143 1. Change pcretest's output for amount of store used to show just the code
4144 space, because the remainder (the data block) varies in size between 32-bit and
4145 64-bit systems.
4146
4147 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4148 start matching at. This allows lookbehinds to work when searching for multiple
4149 occurrences in a string.
4150
4151 3. Added additional options to pcretest for testing multiple occurrences:
4152
4153 /+ outputs the rest of the string that follows a match
4154 /g loops for multiple occurrences, using the new startoffset argument
4155 /G loops for multiple occurrences by passing an incremented pointer
4156
4157 4. PCRE wasn't doing the "first character" optimization for patterns starting
4158 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4159 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4160 the letter 'x'. On long subject strings, this gives a significant speed-up.
4161
4162
4163 Version 2.05 21-Apr-99
4164 ----------------------
4165
4166 1. Changed the type of magic_number from int to long int so that it works
4167 properly on 16-bit systems.
4168
4169 2. Fixed a bug which caused patterns starting with .* not to work correctly
4170 when the subject string contained newline characters. PCRE was assuming
4171 anchoring for such patterns in all cases, which is not correct because .* will
4172 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4173 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4174 must be retried after every newline in the subject.
4175
4176
4177 Version 2.04 18-Feb-99
4178 ----------------------
4179
4180 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4181 computation of the store needed to hold the pattern was incorrect (too large).
4182 If such patterns were nested a few deep, this could multiply and become a real
4183 problem.
4184
4185 2. Added /M option to pcretest to show the memory requirement of a specific
4186 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4187
4188 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4189 compiled in such a way that the backtracking after subsequent failure was
4190 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4191 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4192
4193
4194 Version 2.03 02-Feb-99
4195 ----------------------
4196
4197 1. Fixed typo and small mistake in man page.
4198
4199 2. Added 4th condition (GPL supersedes if conflict) and created separate
4200 LICENCE file containing the conditions.
4201
4202 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4203 Perl, that is the internal \ allows the delimiter to be included in the
4204 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4205 the final delimiter, add \ to the end of the pattern (to test the error).
4206
4207 4. Added the convenience functions for extracting substrings after a successful
4208 match. Updated pcretest to make it able to test these functions.
4209
4210
4211 Version 2.02 14-Jan-99
4212 ----------------------
4213
4214 1. Initialized the working variables associated with each extraction so that
4215 their saving and restoring doesn't refer to uninitialized store.
4216
4217 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4218 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4219 fix the problem.
4220
4221 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4222 calls, and wasn't printing the correct value for compiling calls. Increased the
4223 default value of LOOPREPEAT, and the number of significant figures in the
4224 times.
4225
4226 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4227
4228 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4229 a building problem on Windows NT with a FAT file system.
4230
4231
4232 Version 2.01 21-Oct-98
4233 ----------------------
4234
4235 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4236 to character tables built by pcre_maketables() in the current locale. If NULL
4237 is passed, the default tables are used.
4238
4239
4240 Version 2.00 24-Sep-98
4241 ----------------------
4242
4243 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4244 it any more.
4245
4246 2. Allow quantification of (?>) groups, and make it work correctly.
4247
4248 3. The first character computation wasn't working for (?>) groups.
4249
4250 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4251 end of the subject) and add 5.005's \z, which really does match only at the
4252 very end of the subject.
4253
4254 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4255
4256 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4257 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4258 localized options. All options to pcre_study() were also removed.
4259
4260 7. Add other new features from 5.005:
4261
4262 $(?<= positive lookbehind
4263 $(?<! negative lookbehind
4264 (?imsx-imsx) added the unsetting capability
4265 such a setting is global if at outer level; local otherwise
4266 (?imsx-imsx:) non-capturing groups with option setting
4267 (?(cond)re|re) conditional pattern matching
4268
4269 A backreference to itself in a repeated group matches the previous
4270 captured string.
4271
4272 8. General tidying up of studying (both automatic and via "study")
4273 consequential on the addition of new assertions.
4274
4275 9. As in 5.005, unlimited repeated groups that could match an empty substring
4276 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4277 runtime if any iteration does actually match an empty substring.
4278
4279 10. Include the RunTest script in the distribution.
4280
4281 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4282 discrepancies, some of which were old and were also with respect to 5.004. They
4283 have now been fixed.
4284
4285
4286 Version 1.09 28-Apr-98
4287 ----------------------
4288
4289 1. A negated single character class followed by a quantifier with a minimum
4290 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4291 program crashes, or just wrong answers. This did not apply to negated classes
4292 containing more than one character, or to minima other than one.
4293
4294
4295 Version 1.08 27-Mar-98
4296 ----------------------
4297
4298 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4299
4300 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4301 latter must appear before anything that relies on it in the pattern.
4302
4303
4304 Version 1.07 16-Feb-98
4305 ----------------------
4306
4307 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4308 repeat of a potentially empty string).
4309
4310
4311 Version 1.06 23-Jan-98
4312 ----------------------
4313
4314 1. Added Markus Oberhumer's little patches for C++.
4315
4316 2. Literal strings longer than 255 characters were broken.
4317
4318
4319 Version 1.05 23-Dec-97
4320 ----------------------
4321
4322 1. Negated character classes containing more than one character were failing if
4323 PCRE_CASELESS was set at run time.
4324
4325
4326 Version 1.04 19-Dec-97
4327 ----------------------
4328
4329 1. Corrected the man page, where some "const" qualifiers had been omitted.
4330
4331 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4332 input syntax.
4333
4334 3. Fixed memory leak which occurred when a regex with back references was
4335 matched with an offsets vector that wasn't big enough. The temporary memory
4336 that is used in this case wasn't being freed if the match failed.
4337
4338 4. Tidied pcretest to ensure it frees memory that it gets.
4339
4340 5. Temporary memory was being obtained in the case where the passed offsets
4341 vector was exactly big enough.
4342
4343 6. Corrected definition of offsetof() from change 5 below.
4344
4345 7. I had screwed up change 6 below and broken the rules for the use of
4346 setjmp(). Now fixed.
4347
4348
4349 Version 1.03 18-Dec-97
4350 ----------------------
4351
4352 1. A erroneous regex with a missing opening parenthesis was correctly
4353 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4354 on some systems.
4355
4356 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4357 it was reported that one broken compiler failed on the former because "code" is
4358 also an independent variable.
4359
4360 3. The erroneous regex a[]b caused an array overrun reference.
4361
4362 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4363 fail on data ending with that character. (It was going on too far, and checking
4364 the next character, typically a binary zero.) This was specific to the
4365 optimized code for single-character negative classes.
4366
4367 5. Added a contributed patch from the TIN world which does the following:
4368
4369 + Add an undef for memmove, in case the the system defines a macro for it.
4370
4371 + Add a definition of offsetof(), in case there isn't one. (I don't know
4372 the reason behind this - offsetof() is part of the ANSI standard - but
4373 it does no harm).
4374
4375 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4376 most of the places where whitespace preceded '#'. I have given up and
4377 allowed the remaining 2 cases to be at the margin.
4378
4379 + Rename some variables in pcre to eliminate shadowing. This seems very
4380 pedantic, but does no harm, of course.
4381
4382 6. Moved the call to setjmp() into its own function, to get rid of warnings
4383 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4384
4385 7. Constructs such as \d{8,} were compiling into the equivalent of
4386 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4387 outcome, but in this particular case used more store than had been allocated,
4388 which caused the bug to be discovered because it threw up an internal error.
4389
4390 8. The debugging code in both pcre and pcretest for outputting the compiled
4391 form of a regex was going wrong in the case of back references followed by
4392 curly-bracketed repeats.
4393
4394
4395 Version 1.02 12-Dec-97
4396 ----------------------
4397
4398 1. Typos in pcre.3 and comments in the source fixed.
4399
4400 2. Applied a contributed patch to get rid of places where it used to remove
4401 'const' from variables, and fixed some signed/unsigned and uninitialized
4402 variable warnings.
4403
4404 3. Added the "runtest" target to Makefile.
4405
4406 4. Set default compiler flag to -O2 rather than just -O.
4407
4408
4409 Version 1.01 19-Nov-97
4410 ----------------------
4411
4412 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4413 like /([ab]*)*/, that is, for classes with more than one character in them.
4414
4415 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4416 as /((?>a*))*/ (a PCRE_EXTRA facility).
4417
4418
4419 Version 1.00 18-Nov-97
4420 ----------------------
4421
4422 1. Added compile-time macros to support systems such as SunOS4 which don't have
4423 memmove() or strerror() but have other things that can be used instead.
4424
4425 2. Arranged that "make clean" removes the executables.
4426
4427
4428 Version 0.99 27-Oct-97
4429 ----------------------
4430
4431 1. Fixed bug in code for optimizing classes with only one character. It was
4432 initializing a 32-byte map regardless, which could cause it to run off the end
4433 of the memory it had got.
4434
4435 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4436
4437
4438 Version 0.98 22-Oct-97
4439 ----------------------
4440
4441 1. Fixed bug in code for handling temporary memory usage when there are more
4442 back references than supplied space in the ovector. This could cause segfaults.
4443
4444
4445 Version 0.97 21-Oct-97
4446 ----------------------
4447
4448 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4449
4450 2. Optimized negated single characters not to use a bit map.
4451
4452 3. Brought error texts together as macro definitions; clarified some of them;
4453 fixed one that was wrong - it said "range out of order" when it meant "invalid
4454 escape sequence".
4455
4456 4. Changed some char * arguments to const char *.
4457
4458 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4459
4460 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4461 pcretest.
4462
4463
4464 Version 0.96 16-Oct-97
4465 ----------------------
4466
4467 1. Added a simple "pgrep" utility to the distribution.
4468
4469 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4470 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4471 where "ddd" means "one or more decimal digits".
4472
4473 3. Fixed serious bug. If a pattern had a back reference, but the call to
4474 pcre_exec() didn't supply a large enough ovector to record the related
4475 identifying subpattern, the match always failed. PCRE now remembers the number
4476 of the largest back reference, and gets some temporary memory in which to save
4477 the offsets during matching if necessary, in order to ensure that
4478 backreferences always work.
4479
4480 4. Increased the compatibility with Perl in a number of ways:
4481
4482 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4483 to request this handling. The option can be set at compile or exec time.
4484
4485 (b) $ matches before a terminating newline by default; an option
4486 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4487 mode). The option can be set at compile or exec time.
4488
4489 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4490 the same as Perl's. If the decimal number it represents is less than 10
4491 or there aren't that many previous left capturing parentheses, an octal
4492 escape is read. Inside a character class, it's always an octal escape,
4493 even if it is a single digit.
4494
4495 (d) An escaped but undefined alphabetic character is taken as a literal,
4496 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4497 escapes.
4498
4499 (e) {0} is now permitted. (The previous item is removed from the compiled
4500 pattern).
4501
4502 5. Changed all the names of code files so that the basic parts are no longer
4503 than 10 characters, and abolished the teeny "globals.c" file.
4504
4505 6. Changed the handling of character classes; they are now done with a 32-byte
4506 bit map always.
4507
4508 7. Added the -d and /D options to pcretest to make it possible to look at the
4509 internals of compilation without having to recompile pcre.
4510
4511
4512 Version 0.95 23-Sep-97
4513 ----------------------
4514
4515 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4516 \x20 at the start of a run of normal characters. These were being treated as
4517 real characters, instead of the source characters being re-checked.
4518
4519
4520 Version 0.94 18-Sep-97
4521 ----------------------
4522
4523 1. The functions are now thread-safe, with the caveat that the global variables
4524 containing pointers to malloc() and free() or alternative functions are the
4525 same for all threads.
4526
4527 2. Get pcre_study() to generate a bitmap of initial characters for non-
4528 anchored patterns when this is possible, and use it if passed to pcre_exec().
4529
4530
4531 Version 0.93 15-Sep-97
4532 ----------------------
4533
4534 1. /(b)|(:+)/ was computing an incorrect first character.
4535
4536 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4537 but not actually doing anything yet.
4538
4539 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4540 as Perl does (e.g. [-az] or [az-]).
4541
4542 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4543 all possible positions.
4544
4545 5. Split up into different modules to avoid including unneeded functions in a
4546 compiled binary. However, compile and exec are still in one module. The "study"
4547 function is split off.
4548
4549 6. The character tables are now in a separate module whose source is generated
4550 by an auxiliary program - but can then be edited by hand if required. There are
4551 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4552 toupper() in the code.
4553
4554 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4555 make them global. Abolish the function for setting them, as the caller can now
4556 set them directly.
4557
4558
4559 Version 0.92 11-Sep-97
4560 ----------------------
4561
4562 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4563 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4564
4565 2. Caseless matching was not working in character classes if the characters in
4566 the pattern were in upper case.
4567
4568 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4569
4570 4. Make PCRE_ANCHORED public and accept as a compile option.
4571
4572 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4573 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4574 pass them.
4575
4576 6. Give an error if bad option bits passed at compile or run time.
4577
4578 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4579 pcretest to cause it to pass that flag.
4580
4581 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4582 options, and the first character, if set.
4583
4584 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4585
4586
4587 Version 0.91 10-Sep-97
4588 ----------------------
4589
4590 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4591 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4592
4593 2. PCRE was looping on encountering an indefinitely repeated back reference to
4594 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4595 Perl does - treats the match as successful.
4596
4597 ****

Properties

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

  ViewVC Help
Powered by ViewVC 1.1.5