ViewVC logotype

Contents of /code/trunk/ChangeLog

Parent Directory Parent Directory | Revision Log Revision Log

Revision 882 - (show annotations)
Sun Jan 15 18:45:27 2012 UTC (3 years, 8 months ago) by ph10
File size: 225205 byte(s)
Error occurred while calculating annotation data.
Fix nested *MARK bug (nothing shown for /(?=(*:x))((*:y)q|)/ etc.)
1 ChangeLog for PCRE
2 ------------------
4 Version 8.30
5 ------------
7 1. Renamed "isnumber" as "is_a_number" because in some Mac environments this
8 name is defined in ctype.h.
10 2. Fixed a bug in fixed-length calculation for lookbehinds that would show up
11 only in quite long subpatterns.
13 3. Removed the function pcre_info(), which has been obsolete and deprecated
14 since it was replaced by pcre_fullinfo() in February 2000.
16 4. For a non-anchored pattern, if (*SKIP) was given with a name that did not
17 match a (*MARK), and the match failed at the start of the subject, a
18 reference to memory before the start of the subject could occur. This bug
19 was introduced by fix 17 of release 8.21.
21 5. A reference to an unset group with zero minimum repetition was giving
22 totally wrong answers (in non-JavaScript-compatibility mode). For example,
23 /(another)?(\1?)test/ matched against "hello world test". This bug was
24 introduced in release 8.13.
26 6. Add support for 16-bit character strings (a large amount of work involving
27 many changes and refactorings).
29 7. RunGrepTest failed on msys because \r\n was replaced by whitespace when the
30 command "pattern=`printf 'xxx\r\njkl'`" was run. The pattern is now taken
31 from a file.
33 8. Ovector size of 2 is also supported by JIT based pcre_exec (the ovector size
34 rounding is not applied in this particular case).
36 9. The invalid Unicode surrogate codepoints U+D800 to U+DFFF are now rejected
37 if they appear, or are escaped, in patterns.
39 10. Get rid of a number of -Wunused-but-set-variable warnings.
41 11. The pattern /(?=(*:x))(q|)/ matches an empty string, and returns the mark
42 "x". The similar pattern /(?=(*:x))((*:y)q|)/ did not return a mark at all.
43 Oddly, Perl behaves the same way. PCRE has been fixed so that this pattern
44 also returns the mark "x". This bug applied to capturing parentheses,
45 non-capturing parentheses, and atomic parentheses.
48 Version 8.21 12-Dec-2011
49 ------------------------
51 1. Updating the JIT compiler.
53 2. JIT compiler now supports OP_NCREF, OP_RREF and OP_NRREF. New test cases
54 are added as well.
56 3. Fix cache-flush issue on PowerPC (It is still an experimental JIT port).
57 PCRE_EXTRA_TABLES is not suported by JIT, and should be checked before
58 calling _pcre_jit_exec. Some extra comments are added.
60 4. (*MARK) settings inside atomic groups that do not contain any capturing
61 parentheses, for example, (?>a(*:m)), were not being passed out. This bug
62 was introduced by change 18 for 8.20.
64 5. Supporting of \x, \U and \u in JavaScript compatibility mode based on the
65 ECMA-262 standard.
67 6. Lookbehinds such as (?<=a{2}b) that contained a fixed repetition were
68 erroneously being rejected as "not fixed length" if PCRE_CASELESS was set.
69 This bug was probably introduced by change 9 of 8.13.
71 7. While fixing 6 above, I noticed that a number of other items were being
72 incorrectly rejected as "not fixed length". This arose partly because newer
73 opcodes had not been added to the fixed-length checking code. I have (a)
74 corrected the bug and added tests for these items, and (b) arranged for an
75 error to occur if an unknown opcode is encountered while checking for fixed
76 length instead of just assuming "not fixed length". The items that were
77 rejected were: (*ACCEPT), (*COMMIT), (*FAIL), (*MARK), (*PRUNE), (*SKIP),
78 (*THEN), \h, \H, \v, \V, and single character negative classes with fixed
79 repetitions, e.g. [^a]{3}, with and without PCRE_CASELESS.
81 8. A possessively repeated conditional subpattern such as (?(?=c)c|d)++ was
82 being incorrectly compiled and would have given unpredicatble results.
84 9. A possessively repeated subpattern with minimum repeat count greater than
85 one behaved incorrectly. For example, (A){2,}+ behaved as if it was
86 (A)(A)++ which meant that, after a subsequent mismatch, backtracking into
87 the first (A) could occur when it should not.
89 10. Add a cast and remove a redundant test from the code.
91 11. JIT should use pcre_malloc/pcre_free for allocation.
93 12. Updated pcre-config so that it no longer shows -L/usr/lib, which seems
94 best practice nowadays, and helps with cross-compiling. (If the exec_prefix
95 is anything other than /usr, -L is still shown).
97 13. In non-UTF-8 mode, \C is now supported in lookbehinds and DFA matching.
99 14. Perl does not support \N without a following name in a [] class; PCRE now
100 also gives an error.
102 15. If a forward reference was repeated with an upper limit of around 2000,
103 it caused the error "internal error: overran compiling workspace". The
104 maximum number of forward references (including repeats) was limited by the
105 internal workspace, and dependent on the LINK_SIZE. The code has been
106 rewritten so that the workspace expands (via pcre_malloc) if necessary, and
107 the default depends on LINK_SIZE. There is a new upper limit (for safety)
108 of around 200,000 forward references. While doing this, I also speeded up
109 the filling in of repeated forward references.
111 16. A repeated forward reference in a pattern such as (a)(?2){2}(.) was
112 incorrectly expecting the subject to contain another "a" after the start.
114 17. When (*SKIP:name) is activated without a corresponding (*MARK:name) earlier
115 in the match, the SKIP should be ignored. This was not happening; instead
116 the SKIP was being treated as NOMATCH. For patterns such as
117 /A(*MARK:A)A+(*SKIP:B)Z|AAC/ this meant that the AAC branch was never
118 tested.
120 18. The behaviour of (*MARK), (*PRUNE), and (*THEN) has been reworked and is
121 now much more compatible with Perl, in particular in cases where the result
122 is a non-match for a non-anchored pattern. For example, if
123 /b(*:m)f|a(*:n)w/ is matched against "abc", the non-match returns the name
124 "m", where previously it did not return a name. A side effect of this
125 change is that for partial matches, the last encountered mark name is
126 returned, as for non matches. A number of tests that were previously not
127 Perl-compatible have been moved into the Perl-compatible test files. The
128 refactoring has had the pleasing side effect of removing one argument from
129 the match() function, thus reducing its stack requirements.
131 19. If the /S+ option was used in pcretest to study a pattern using JIT,
132 subsequent uses of /S (without +) incorrectly behaved like /S+.
134 21. Retrieve executable code size support for the JIT compiler and fixing
135 some warnings.
137 22. A caseless match of a UTF-8 character whose other case uses fewer bytes did
138 not work when the shorter character appeared right at the end of the
139 subject string.
141 23. Added some (int) casts to non-JIT modules to reduce warnings on 64-bit
142 systems.
144 24. Added PCRE_INFO_JITSIZE to pass on the value from (21) above, and also
145 output it when the /M option is used in pcretest.
147 25. The CheckMan script was not being included in the distribution. Also, added
148 an explicit "perl" to run Perl scripts from the PrepareRelease script
149 because this is reportedly needed in Windows.
151 26. If study data was being save in a file and studying had not found a set of
152 "starts with" bytes for the pattern, the data written to the file (though
153 never used) was taken from uninitialized memory and so caused valgrind to
154 complain.
156 27. Updated RunTest.bat as provided by Sheri Pierce.
158 28. Fixed a possible uninitialized memory bug in pcre_jit_compile.c.
160 29. Computation of memory usage for the table of capturing group names was
161 giving an unnecessarily large value.
164 Version 8.20 21-Oct-2011
165 ------------------------
167 1. Change 37 of 8.13 broke patterns like [:a]...[b:] because it thought it had
168 a POSIX class. After further experiments with Perl, which convinced me that
169 Perl has bugs and confusions, a closing square bracket is no longer allowed
170 in a POSIX name. This bug also affected patterns with classes that started
171 with full stops.
173 2. If a pattern such as /(a)b|ac/ is matched against "ac", there is no
174 captured substring, but while checking the failing first alternative,
175 substring 1 is temporarily captured. If the output vector supplied to
176 pcre_exec() was not big enough for this capture, the yield of the function
177 was still zero ("insufficient space for captured substrings"). This cannot
178 be totally fixed without adding another stack variable, which seems a lot
179 of expense for a edge case. However, I have improved the situation in cases
180 such as /(a)(b)x|abc/ matched against "abc", where the return code
181 indicates that fewer than the maximum number of slots in the ovector have
182 been set.
184 3. Related to (2) above: when there are more back references in a pattern than
185 slots in the output vector, pcre_exec() uses temporary memory during
186 matching, and copies in the captures as far as possible afterwards. It was
187 using the entire output vector, but this conflicts with the specification
188 that only 2/3 is used for passing back captured substrings. Now it uses
189 only the first 2/3, for compatibility. This is, of course, another edge
190 case.
192 4. Zoltan Herczeg's just-in-time compiler support has been integrated into the
193 main code base, and can be used by building with --enable-jit. When this is
194 done, pcregrep automatically uses it unless --disable-pcregrep-jit or the
195 runtime --no-jit option is given.
197 5. When the number of matches in a pcre_dfa_exec() run exactly filled the
198 ovector, the return from the function was zero, implying that there were
199 other matches that did not fit. The correct "exactly full" value is now
200 returned.
202 6. If a subpattern that was called recursively or as a subroutine contained
203 (*PRUNE) or any other control that caused it to give a non-standard return,
204 invalid errors such as "Error -26 (nested recursion at the same subject
205 position)" or even infinite loops could occur.
207 7. If a pattern such as /a(*SKIP)c|b(*ACCEPT)|/ was studied, it stopped
208 computing the minimum length on reaching *ACCEPT, and so ended up with the
209 wrong value of 1 rather than 0. Further investigation indicates that
210 computing a minimum subject length in the presence of *ACCEPT is difficult
211 (think back references, subroutine calls), and so I have changed the code
212 so that no minimum is registered for a pattern that contains *ACCEPT.
214 8. If (*THEN) was present in the first (true) branch of a conditional group,
215 it was not handled as intended. [But see 16 below.]
217 9. Replaced RunTest.bat and CMakeLists.txt with improved versions provided by
218 Sheri Pierce.
220 10. A pathological pattern such as /(*ACCEPT)a/ was miscompiled, thinking that
221 the first byte in a match must be "a".
223 11. Change 17 for 8.13 increased the recursion depth for patterns like
224 /a(?:.)*?a/ drastically. I've improved things by remembering whether a
225 pattern contains any instances of (*THEN). If it does not, the old
226 optimizations are restored. It would be nice to do this on a per-group
227 basis, but at the moment that is not feasible.
229 12. In some environments, the output of pcretest -C is CRLF terminated. This
230 broke RunTest's code that checks for the link size. A single white space
231 character after the value is now allowed for.
233 13. RunTest now checks for the "fr" locale as well as for "fr_FR" and "french".
234 For "fr", it uses the Windows-specific input and output files.
236 14. If (*THEN) appeared in a group that was called recursively or as a
237 subroutine, it did not work as intended. [But see next item.]
239 15. Consider the pattern /A (B(*THEN)C) | D/ where A, B, C, and D are complex
240 pattern fragments (but not containing any | characters). If A and B are
241 matched, but there is a failure in C so that it backtracks to (*THEN), PCRE
242 was behaving differently to Perl. PCRE backtracked into A, but Perl goes to
243 D. In other words, Perl considers parentheses that do not contain any |
244 characters to be part of a surrounding alternative, whereas PCRE was
245 treading (B(*THEN)C) the same as (B(*THEN)C|(*FAIL)) -- which Perl handles
246 differently. PCRE now behaves in the same way as Perl, except in the case
247 of subroutine/recursion calls such as (?1) which have in any case always
248 been different (but PCRE had them first :-).
250 16. Related to 15 above: Perl does not treat the | in a conditional group as
251 creating alternatives. Such a group is treated in the same way as an
252 ordinary group without any | characters when processing (*THEN). PCRE has
253 been changed to match Perl's behaviour.
255 17. If a user had set PCREGREP_COLO(U)R to something other than 1:31, the
256 RunGrepTest script failed.
258 18. Change 22 for version 13 caused atomic groups to use more stack. This is
259 inevitable for groups that contain captures, but it can lead to a lot of
260 stack use in large patterns. The old behaviour has been restored for atomic
261 groups that do not contain any capturing parentheses.
263 19. If the PCRE_NO_START_OPTIMIZE option was set for pcre_compile(), it did not
264 suppress the check for a minimum subject length at run time. (If it was
265 given to pcre_exec() or pcre_dfa_exec() it did work.)
267 20. Fixed an ASCII-dependent infelicity in pcretest that would have made it
268 fail to work when decoding hex characters in data strings in EBCDIC
269 environments.
271 21. It appears that in at least one Mac OS environment, the isxdigit() function
272 is implemented as a macro that evaluates to its argument more than once,
273 contravening the C 90 Standard (I haven't checked a later standard). There
274 was an instance in pcretest which caused it to go wrong when processing
275 \x{...} escapes in subject strings. The has been rewritten to avoid using
276 things like p++ in the argument of isxdigit().
279 Version 8.13 16-Aug-2011
280 ------------------------
282 1. The Unicode data tables have been updated to Unicode 6.0.0.
284 2. Two minor typos in pcre_internal.h have been fixed.
286 3. Added #include <string.h> to pcre_scanner_unittest.cc, pcrecpp.cc, and
287 pcrecpp_unittest.cc. They are needed for strcmp(), memset(), and strchr()
288 in some environments (e.g. Solaris 10/SPARC using Sun Studio 12U2).
290 4. There were a number of related bugs in the code for matching backrefences
291 caselessly in UTF-8 mode when codes for the characters concerned were
292 different numbers of bytes. For example, U+023A and U+2C65 are an upper
293 and lower case pair, using 2 and 3 bytes, respectively. The main bugs were:
294 (a) A reference to 3 copies of a 2-byte code matched only 2 of a 3-byte
295 code. (b) A reference to 2 copies of a 3-byte code would not match 2 of a
296 2-byte code at the end of the subject (it thought there wasn't enough data
297 left).
299 5. Comprehensive information about what went wrong is now returned by
300 pcre_exec() and pcre_dfa_exec() when the UTF-8 string check fails, as long
301 as the output vector has at least 2 elements. The offset of the start of
302 the failing character and a reason code are placed in the vector.
304 6. When the UTF-8 string check fails for pcre_compile(), the offset that is
305 now returned is for the first byte of the failing character, instead of the
306 last byte inspected. This is an incompatible change, but I hope it is small
307 enough not to be a problem. It makes the returned offset consistent with
308 pcre_exec() and pcre_dfa_exec().
310 7. pcretest now gives a text phrase as well as the error number when
311 pcre_exec() or pcre_dfa_exec() fails; if the error is a UTF-8 check
312 failure, the offset and reason code are output.
314 8. When \R was used with a maximizing quantifier it failed to skip backwards
315 over a \r\n pair if the subsequent match failed. Instead, it just skipped
316 back over a single character (\n). This seems wrong (because it treated the
317 two characters as a single entity when going forwards), conflicts with the
318 documentation that \R is equivalent to (?>\r\n|\n|...etc), and makes the
319 behaviour of \R* different to (\R)*, which also seems wrong. The behaviour
320 has been changed.
322 9. Some internal refactoring has changed the processing so that the handling
323 of the PCRE_CASELESS and PCRE_MULTILINE options is done entirely at compile
324 time (the PCRE_DOTALL option was changed this way some time ago: version
325 7.7 change 16). This has made it possible to abolish the OP_OPT op code,
326 which was always a bit of a fudge. It also means that there is one less
327 argument for the match() function, which reduces its stack requirements
328 slightly. This change also fixes an incompatibility with Perl: the pattern
329 (?i:([^b]))(?1) should not match "ab", but previously PCRE gave a match.
331 10. More internal refactoring has drastically reduced the number of recursive
332 calls to match() for possessively repeated groups such as (abc)++ when
333 using pcre_exec().
335 11. While implementing 10, a number of bugs in the handling of groups were
336 discovered and fixed:
338 (?<=(a)+) was not diagnosed as invalid (non-fixed-length lookbehind).
339 (a|)*(?1) gave a compile-time internal error.
340 ((a|)+)+ did not notice that the outer group could match an empty string.
341 (^a|^)+ was not marked as anchored.
342 (.*a|.*)+ was not marked as matching at start or after a newline.
344 12. Yet more internal refactoring has removed another argument from the match()
345 function. Special calls to this function are now indicated by setting a
346 value in a variable in the "match data" data block.
348 13. Be more explicit in pcre_study() instead of relying on "default" for
349 opcodes that mean there is no starting character; this means that when new
350 ones are added and accidentally left out of pcre_study(), testing should
351 pick them up.
353 14. The -s option of pcretest has been documented for ages as being an old
354 synonym of -m (show memory usage). I have changed it to mean "force study
355 for every regex", that is, assume /S for every regex. This is similar to -i
356 and -d etc. It's slightly incompatible, but I'm hoping nobody is still
357 using it. It makes it easier to run collections of tests with and without
358 study enabled, and thereby test pcre_study() more easily. All the standard
359 tests are now run with and without -s (but some patterns can be marked as
360 "never study" - see 20 below).
362 15. When (*ACCEPT) was used in a subpattern that was called recursively, the
363 restoration of the capturing data to the outer values was not happening
364 correctly.
366 16. If a recursively called subpattern ended with (*ACCEPT) and matched an
367 empty string, and PCRE_NOTEMPTY was set, pcre_exec() thought the whole
368 pattern had matched an empty string, and so incorrectly returned a no
369 match.
371 17. There was optimizing code for the last branch of non-capturing parentheses,
372 and also for the obeyed branch of a conditional subexpression, which used
373 tail recursion to cut down on stack usage. Unfortunately, now that there is
374 the possibility of (*THEN) occurring in these branches, tail recursion is
375 no longer possible because the return has to be checked for (*THEN). These
376 two optimizations have therefore been removed. [But see 8.20/11 above.]
378 18. If a pattern containing \R was studied, it was assumed that \R always
379 matched two bytes, thus causing the minimum subject length to be
380 incorrectly computed because \R can also match just one byte.
382 19. If a pattern containing (*ACCEPT) was studied, the minimum subject length
383 was incorrectly computed.
385 20. If /S is present twice on a test pattern in pcretest input, it now
386 *disables* studying, thereby overriding the use of -s on the command line
387 (see 14 above). This is necessary for one or two tests to keep the output
388 identical in both cases.
390 21. When (*ACCEPT) was used in an assertion that matched an empty string and
391 PCRE_NOTEMPTY was set, PCRE applied the non-empty test to the assertion.
393 22. When an atomic group that contained a capturing parenthesis was
394 successfully matched, but the branch in which it appeared failed, the
395 capturing was not being forgotten if a higher numbered group was later
396 captured. For example, /(?>(a))b|(a)c/ when matching "ac" set capturing
397 group 1 to "a", when in fact it should be unset. This applied to multi-
398 branched capturing and non-capturing groups, repeated or not, and also to
399 positive assertions (capturing in negative assertions does not happen
400 in PCRE) and also to nested atomic groups.
402 23. Add the ++ qualifier feature to pcretest, to show the remainder of the
403 subject after a captured substring, to make it easier to tell which of a
404 number of identical substrings has been captured.
406 24. The way atomic groups are processed by pcre_exec() has been changed so that
407 if they are repeated, backtracking one repetition now resets captured
408 values correctly. For example, if ((?>(a+)b)+aabab) is matched against
409 "aaaabaaabaabab" the value of captured group 2 is now correctly recorded as
410 "aaa". Previously, it would have been "a". As part of this code
411 refactoring, the way recursive calls are handled has also been changed.
413 25. If an assertion condition captured any substrings, they were not passed
414 back unless some other capturing happened later. For example, if
415 (?(?=(a))a) was matched against "a", no capturing was returned.
417 26. When studying a pattern that contained subroutine calls or assertions,
418 the code for finding the minimum length of a possible match was handling
419 direct recursions such as (xxx(?1)|yyy) but not mutual recursions (where
420 group 1 called group 2 while simultaneously a separate group 2 called group
421 1). A stack overflow occurred in this case. I have fixed this by limiting
422 the recursion depth to 10.
424 27. Updated RunTest.bat in the distribution to the version supplied by Tom
425 Fortmann. This supports explicit test numbers on the command line, and has
426 argument validation and error reporting.
428 28. An instance of \X with an unlimited repeat could fail if at any point the
429 first character it looked at was a mark character.
431 29. Some minor code refactoring concerning Unicode properties and scripts
432 should reduce the stack requirement of match() slightly.
434 30. Added the '=' option to pcretest to check the setting of unused capturing
435 slots at the end of the pattern, which are documented as being -1, but are
436 not included in the return count.
438 31. If \k was not followed by a braced, angle-bracketed, or quoted name, PCRE
439 compiled something random. Now it gives a compile-time error (as does
440 Perl).
442 32. A *MARK encountered during the processing of a positive assertion is now
443 recorded and passed back (compatible with Perl).
445 33. If --only-matching or --colour was set on a pcregrep call whose pattern
446 had alternative anchored branches, the search for a second match in a line
447 was done as if at the line start. Thus, for example, /^01|^02/ incorrectly
448 matched the line "0102" twice. The same bug affected patterns that started
449 with a backwards assertion. For example /\b01|\b02/ also matched "0102"
450 twice.
452 34. Previously, PCRE did not allow quantification of assertions. However, Perl
453 does, and because of capturing effects, quantifying parenthesized
454 assertions may at times be useful. Quantifiers are now allowed for
455 parenthesized assertions.
457 35. A minor code tidy in pcre_compile() when checking options for \R usage.
459 36. \g was being checked for fancy things in a character class, when it should
460 just be a literal "g".
462 37. PCRE was rejecting [:a[:digit:]] whereas Perl was not. It seems that the
463 appearance of a nested POSIX class supersedes an apparent external class.
464 For example, [:a[:digit:]b:] matches "a", "b", ":", or a digit. Also,
465 unescaped square brackets may also appear as part of class names. For
466 example, [:a[:abc]b:] gives unknown class "[:abc]b:]". PCRE now behaves
467 more like Perl. (But see 8.20/1 above.)
469 38. PCRE was giving an error for \N with a braced quantifier such as {1,} (this
470 was because it thought it was \N{name}, which is not supported).
472 39. Add minix to OS list not supporting the -S option in pcretest.
474 40. PCRE tries to detect cases of infinite recursion at compile time, but it
475 cannot analyze patterns in sufficient detail to catch mutual recursions
476 such as ((?1))((?2)). There is now a runtime test that gives an error if a
477 subgroup is called recursively as a subpattern for a second time at the
478 same position in the subject string. In previous releases this might have
479 been caught by the recursion limit, or it might have run out of stack.
481 41. A pattern such as /(?(R)a+|(?R)b)/ is quite safe, as the recursion can
482 happen only once. PCRE was, however incorrectly giving a compile time error
483 "recursive call could loop indefinitely" because it cannot analyze the
484 pattern in sufficient detail. The compile time test no longer happens when
485 PCRE is compiling a conditional subpattern, but actual runaway loops are
486 now caught at runtime (see 40 above).
488 42. It seems that Perl allows any characters other than a closing parenthesis
489 to be part of the NAME in (*MARK:NAME) and other backtracking verbs. PCRE
490 has been changed to be the same.
492 43. Updated configure.ac to put in more quoting round AC_LANG_PROGRAM etc. so
493 as not to get warnings when autogen.sh is called. Also changed
494 AC_PROG_LIBTOOL (deprecated) to LT_INIT (the current macro).
496 44. To help people who use pcregrep to scan files containing exceedingly long
497 lines, the following changes have been made:
499 (a) The default value of the buffer size parameter has been increased from
500 8K to 20K. (The actual buffer used is three times this size.)
502 (b) The default can be changed by ./configure --with-pcregrep-bufsize when
503 PCRE is built.
505 (c) A --buffer-size=n option has been added to pcregrep, to allow the size
506 to be set at run time.
508 (d) Numerical values in pcregrep options can be followed by K or M, for
509 example --buffer-size=50K.
511 (e) If a line being scanned overflows pcregrep's buffer, an error is now
512 given and the return code is set to 2.
514 45. Add a pointer to the latest mark to the callout data block.
516 46. The pattern /.(*F)/, when applied to "abc" with PCRE_PARTIAL_HARD, gave a
517 partial match of an empty string instead of no match. This was specific to
518 the use of ".".
520 47. The pattern /f.*/8s, when applied to "for" with PCRE_PARTIAL_HARD, gave a
521 complete match instead of a partial match. This bug was dependent on both
522 the PCRE_UTF8 and PCRE_DOTALL options being set.
524 48. For a pattern such as /\babc|\bdef/ pcre_study() was failing to set up the
525 starting byte set, because \b was not being ignored.
528 Version 8.12 15-Jan-2011
529 ------------------------
531 1. Fixed some typos in the markup of the man pages, and wrote a script that
532 checks for such things as part of the documentation building process.
534 2. On a big-endian 64-bit system, pcregrep did not correctly process the
535 --match-limit and --recursion-limit options (added for 8.11). In
536 particular, this made one of the standard tests fail. (The integer value
537 went into the wrong half of a long int.)
539 3. If the --colour option was given to pcregrep with -v (invert match), it
540 did strange things, either producing crazy output, or crashing. It should,
541 of course, ignore a request for colour when reporting lines that do not
542 match.
544 4. Another pcregrep bug caused similar problems if --colour was specified with
545 -M (multiline) and the pattern match finished with a line ending.
547 5. In pcregrep, when a pattern that ended with a literal newline sequence was
548 matched in multiline mode, the following line was shown as part of the
549 match. This seems wrong, so I have changed it.
551 6. Another pcregrep bug in multiline mode, when --colour was specified, caused
552 the check for further matches in the same line (so they could be coloured)
553 to overrun the end of the current line. If another match was found, it was
554 incorrectly shown (and then shown again when found in the next line).
556 7. If pcregrep was compiled under Windows, there was a reference to the
557 function pcregrep_exit() before it was defined. I am assuming this was
558 the cause of the "error C2371: 'pcregrep_exit' : redefinition;" that was
559 reported by a user. I've moved the definition above the reference.
562 Version 8.11 10-Dec-2010
563 ------------------------
565 1. (*THEN) was not working properly if there were untried alternatives prior
566 to it in the current branch. For example, in ((a|b)(*THEN)(*F)|c..) it
567 backtracked to try for "b" instead of moving to the next alternative branch
568 at the same level (in this case, to look for "c"). The Perl documentation
569 is clear that when (*THEN) is backtracked onto, it goes to the "next
570 alternative in the innermost enclosing group".
572 2. (*COMMIT) was not overriding (*THEN), as it does in Perl. In a pattern
573 such as (A(*COMMIT)B(*THEN)C|D) any failure after matching A should
574 result in overall failure. Similarly, (*COMMIT) now overrides (*PRUNE) and
575 (*SKIP), (*SKIP) overrides (*PRUNE) and (*THEN), and (*PRUNE) overrides
576 (*THEN).
578 3. If \s appeared in a character class, it removed the VT character from
579 the class, even if it had been included by some previous item, for example
580 in [\x00-\xff\s]. (This was a bug related to the fact that VT is not part
581 of \s, but is part of the POSIX "space" class.)
583 4. A partial match never returns an empty string (because you can always
584 match an empty string at the end of the subject); however the checking for
585 an empty string was starting at the "start of match" point. This has been
586 changed to the "earliest inspected character" point, because the returned
587 data for a partial match starts at this character. This means that, for
588 example, /(?<=abc)def/ gives a partial match for the subject "abc"
589 (previously it gave "no match").
591 5. Changes have been made to the way PCRE_PARTIAL_HARD affects the matching
592 of $, \z, \Z, \b, and \B. If the match point is at the end of the string,
593 previously a full match would be given. However, setting PCRE_PARTIAL_HARD
594 has an implication that the given string is incomplete (because a partial
595 match is preferred over a full match). For this reason, these items now
596 give a partial match in this situation. [Aside: previously, the one case
597 /t\b/ matched against "cat" with PCRE_PARTIAL_HARD set did return a partial
598 match rather than a full match, which was wrong by the old rules, but is
599 now correct.]
601 6. There was a bug in the handling of #-introduced comments, recognized when
602 PCRE_EXTENDED is set, when PCRE_NEWLINE_ANY and PCRE_UTF8 were also set.
603 If a UTF-8 multi-byte character included the byte 0x85 (e.g. +U0445, whose
604 UTF-8 encoding is 0xd1,0x85), this was misinterpreted as a newline when
605 scanning for the end of the comment. (*Character* 0x85 is an "any" newline,
606 but *byte* 0x85 is not, in UTF-8 mode). This bug was present in several
607 places in pcre_compile().
609 7. Related to (6) above, when pcre_compile() was skipping #-introduced
610 comments when looking ahead for named forward references to subpatterns,
611 the only newline sequence it recognized was NL. It now handles newlines
612 according to the set newline convention.
614 8. SunOS4 doesn't have strerror() or strtoul(); pcregrep dealt with the
615 former, but used strtoul(), whereas pcretest avoided strtoul() but did not
616 cater for a lack of strerror(). These oversights have been fixed.
618 9. Added --match-limit and --recursion-limit to pcregrep.
620 10. Added two casts needed to build with Visual Studio when NO_RECURSE is set.
622 11. When the -o option was used, pcregrep was setting a return code of 1, even
623 when matches were found, and --line-buffered was not being honoured.
625 12. Added an optional parentheses number to the -o and --only-matching options
626 of pcregrep.
628 13. Imitating Perl's /g action for multiple matches is tricky when the pattern
629 can match an empty string. The code to do it in pcretest and pcredemo
630 needed fixing:
632 (a) When the newline convention was "crlf", pcretest got it wrong, skipping
633 only one byte after an empty string match just before CRLF (this case
634 just got forgotten; "any" and "anycrlf" were OK).
636 (b) The pcretest code also had a bug, causing it to loop forever in UTF-8
637 mode when an empty string match preceded an ASCII character followed by
638 a non-ASCII character. (The code for advancing by one character rather
639 than one byte was nonsense.)
641 (c) The pcredemo.c sample program did not have any code at all to handle
642 the cases when CRLF is a valid newline sequence.
644 14. Neither pcre_exec() nor pcre_dfa_exec() was checking that the value given
645 as a starting offset was within the subject string. There is now a new
646 error, PCRE_ERROR_BADOFFSET, which is returned if the starting offset is
647 negative or greater than the length of the string. In order to test this,
648 pcretest is extended to allow the setting of negative starting offsets.
650 15. In both pcre_exec() and pcre_dfa_exec() the code for checking that the
651 starting offset points to the beginning of a UTF-8 character was
652 unnecessarily clumsy. I tidied it up.
654 16. Added PCRE_ERROR_SHORTUTF8 to make it possible to distinguish between a
655 bad UTF-8 sequence and one that is incomplete when using PCRE_PARTIAL_HARD.
657 17. Nobody had reported that the --include_dir option, which was added in
658 release 7.7 should have been called --include-dir (hyphen, not underscore)
659 for compatibility with GNU grep. I have changed it to --include-dir, but
660 left --include_dir as an undocumented synonym, and the same for
661 --exclude-dir, though that is not available in GNU grep, at least as of
662 release 2.5.4.
664 18. At a user's suggestion, the macros GETCHAR and friends (which pick up UTF-8
665 characters from a string of bytes) have been redefined so as not to use
666 loops, in order to improve performance in some environments. At the same
667 time, I abstracted some of the common code into auxiliary macros to save
668 repetition (this should not affect the compiled code).
670 19. If \c was followed by a multibyte UTF-8 character, bad things happened. A
671 compile-time error is now given if \c is not followed by an ASCII
672 character, that is, a byte less than 128. (In EBCDIC mode, the code is
673 different, and any byte value is allowed.)
675 20. Recognize (*NO_START_OPT) at the start of a pattern to set the PCRE_NO_
676 START_OPTIMIZE option, which is now allowed at compile time - but just
677 passed through to pcre_exec() or pcre_dfa_exec(). This makes it available
678 to pcregrep and other applications that have no direct access to PCRE
679 options. The new /Y option in pcretest sets this option when calling
680 pcre_compile().
682 21. Change 18 of release 8.01 broke the use of named subpatterns for recursive
683 back references. Groups containing recursive back references were forced to
684 be atomic by that change, but in the case of named groups, the amount of
685 memory required was incorrectly computed, leading to "Failed: internal
686 error: code overflow". This has been fixed.
688 22. Some patches to pcre_stringpiece.h, pcre_stringpiece_unittest.cc, and
689 pcretest.c, to avoid build problems in some Borland environments.
692 Version 8.10 25-Jun-2010
693 ------------------------
695 1. Added support for (*MARK:ARG) and for ARG additions to PRUNE, SKIP, and
696 THEN.
698 2. (*ACCEPT) was not working when inside an atomic group.
700 3. Inside a character class, \B is treated as a literal by default, but
701 faulted if PCRE_EXTRA is set. This mimics Perl's behaviour (the -w option
702 causes the error). The code is unchanged, but I tidied the documentation.
704 4. Inside a character class, PCRE always treated \R and \X as literals,
705 whereas Perl faults them if its -w option is set. I have changed PCRE so
706 that it faults them when PCRE_EXTRA is set.
708 5. Added support for \N, which always matches any character other than
709 newline. (It is the same as "." when PCRE_DOTALL is not set.)
711 6. When compiling pcregrep with newer versions of gcc which may have
712 FORTIFY_SOURCE set, several warnings "ignoring return value of 'fwrite',
713 declared with attribute warn_unused_result" were given. Just casting the
714 result to (void) does not stop the warnings; a more elaborate fudge is
715 needed. I've used a macro to implement this.
717 7. Minor change to pcretest.c to avoid a compiler warning.
719 8. Added four artifical Unicode properties to help with an option to make
720 \s etc use properties (see next item). The new properties are: Xan
721 (alphanumeric), Xsp (Perl space), Xps (POSIX space), and Xwd (word).
723 9. Added PCRE_UCP to make \b, \d, \s, \w, and certain POSIX character classes
724 use Unicode properties. (*UCP) at the start of a pattern can be used to set
725 this option. Modified pcretest to add /W to test this facility. Added
726 REG_UCP to make it available via the POSIX interface.
728 10. Added --line-buffered to pcregrep.
730 11. In UTF-8 mode, if a pattern that was compiled with PCRE_CASELESS was
731 studied, and the match started with a letter with a code point greater than
732 127 whose first byte was different to the first byte of the other case of
733 the letter, the other case of this starting letter was not recognized
734 (#976).
736 12. If a pattern that was studied started with a repeated Unicode property
737 test, for example, \p{Nd}+, there was the theoretical possibility of
738 setting up an incorrect bitmap of starting bytes, but fortunately it could
739 not have actually happened in practice until change 8 above was made (it
740 added property types that matched character-matching opcodes).
742 13. pcre_study() now recognizes \h, \v, and \R when constructing a bit map of
743 possible starting bytes for non-anchored patterns.
745 14. Extended the "auto-possessify" feature of pcre_compile(). It now recognizes
746 \R, and also a number of cases that involve Unicode properties, both
747 explicit and implicit when PCRE_UCP is set.
749 15. If a repeated Unicode property match (e.g. \p{Lu}*) was used with non-UTF-8
750 input, it could crash or give wrong results if characters with values
751 greater than 0xc0 were present in the subject string. (Detail: it assumed
752 UTF-8 input when processing these items.)
754 16. Added a lot of (int) casts to avoid compiler warnings in systems where
755 size_t is 64-bit (#991).
757 17. Added a check for running out of memory when PCRE is compiled with
758 --disable-stack-for-recursion (#990).
760 18. If the last data line in a file for pcretest does not have a newline on
761 the end, a newline was missing in the output.
763 19. The default pcre_chartables.c file recognizes only ASCII characters (values
764 less than 128) in its various bitmaps. However, there is a facility for
765 generating tables according to the current locale when PCRE is compiled. It
766 turns out that in some environments, 0x85 and 0xa0, which are Unicode space
767 characters, are recognized by isspace() and therefore were getting set in
768 these tables, and indeed these tables seem to approximate to ISO 8859. This
769 caused a problem in UTF-8 mode when pcre_study() was used to create a list
770 of bytes that can start a match. For \s, it was including 0x85 and 0xa0,
771 which of course cannot start UTF-8 characters. I have changed the code so
772 that only real ASCII characters (less than 128) and the correct starting
773 bytes for UTF-8 encodings are set for characters greater than 127 when in
774 UTF-8 mode. (When PCRE_UCP is set - see 9 above - the code is different
775 altogether.)
777 20. Added the /T option to pcretest so as to be able to run tests with non-
778 standard character tables, thus making it possible to include the tests
779 used for 19 above in the standard set of tests.
781 21. A pattern such as (?&t)(?#()(?(DEFINE)(?<t>a)) which has a forward
782 reference to a subpattern the other side of a comment that contains an
783 opening parenthesis caused either an internal compiling error, or a
784 reference to the wrong subpattern.
787 Version 8.02 19-Mar-2010
788 ------------------------
790 1. The Unicode data tables have been updated to Unicode 5.2.0.
792 2. Added the option --libs-cpp to pcre-config, but only when C++ support is
793 configured.
795 3. Updated the licensing terms in the pcregexp.pas file, as agreed with the
796 original author of that file, following a query about its status.
798 4. On systems that do not have stdint.h (e.g. Solaris), check for and include
799 inttypes.h instead. This fixes a bug that was introduced by change 8.01/8.
801 5. A pattern such as (?&t)*+(?(DEFINE)(?<t>.)) which has a possessive
802 quantifier applied to a forward-referencing subroutine call, could compile
803 incorrect code or give the error "internal error: previously-checked
804 referenced subpattern not found".
806 6. Both MS Visual Studio and Symbian OS have problems with initializing
807 variables to point to external functions. For these systems, therefore,
808 pcre_malloc etc. are now initialized to local functions that call the
809 relevant global functions.
811 7. There were two entries missing in the vectors called coptable and poptable
812 in pcre_dfa_exec.c. This could lead to memory accesses outsize the vectors.
813 I've fixed the data, and added a kludgy way of testing at compile time that
814 the lengths are correct (equal to the number of opcodes).
816 8. Following on from 7, I added a similar kludge to check the length of the
817 eint vector in pcreposix.c.
819 9. Error texts for pcre_compile() are held as one long string to avoid too
820 much relocation at load time. To find a text, the string is searched,
821 counting zeros. There was no check for running off the end of the string,
822 which could happen if a new error number was added without updating the
823 string.
825 10. \K gave a compile-time error if it appeared in a lookbehind assersion.
827 11. \K was not working if it appeared in an atomic group or in a group that
828 was called as a "subroutine", or in an assertion. Perl 5.11 documents that
829 \K is "not well defined" if used in an assertion. PCRE now accepts it if
830 the assertion is positive, but not if it is negative.
832 12. Change 11 fortuitously reduced the size of the stack frame used in the
833 "match()" function of pcre_exec.c by one pointer. Forthcoming
834 implementation of support for (*MARK) will need an extra pointer on the
835 stack; I have reserved it now, so that the stack frame size does not
836 decrease.
838 13. A pattern such as (?P<L1>(?P<L2>0)|(?P>L2)(?P>L1)) in which the only other
839 item in branch that calls a recursion is a subroutine call - as in the
840 second branch in the above example - was incorrectly given the compile-
841 time error "recursive call could loop indefinitely" because pcre_compile()
842 was not correctly checking the subroutine for matching a non-empty string.
844 14. The checks for overrunning compiling workspace could trigger after an
845 overrun had occurred. This is a "should never occur" error, but it can be
846 triggered by pathological patterns such as hundreds of nested parentheses.
847 The checks now trigger 100 bytes before the end of the workspace.
849 15. Fix typo in configure.ac: "srtoq" should be "strtoq".
852 Version 8.01 19-Jan-2010
853 ------------------------
855 1. If a pattern contained a conditional subpattern with only one branch (in
856 particular, this includes all (*DEFINE) patterns), a call to pcre_study()
857 computed the wrong minimum data length (which is of course zero for such
858 subpatterns). This could cause incorrect "no match" results.
860 2. For patterns such as (?i)a(?-i)b|c where an option setting at the start of
861 the pattern is reset in the first branch, pcre_compile() failed with
862 "internal error: code overflow at offset...". This happened only when
863 the reset was to the original external option setting. (An optimization
864 abstracts leading options settings into an external setting, which was the
865 cause of this.)
867 3. A pattern such as ^(?!a(*SKIP)b) where a negative assertion contained one
868 of the verbs SKIP, PRUNE, or COMMIT, did not work correctly. When the
869 assertion pattern did not match (meaning that the assertion was true), it
870 was incorrectly treated as false if the SKIP had been reached during the
871 matching. This also applied to assertions used as conditions.
873 4. If an item that is not supported by pcre_dfa_exec() was encountered in an
874 assertion subpattern, including such a pattern used as a condition,
875 unpredictable results occurred, instead of the error return
878 5. The C++ GlobalReplace function was not working like Perl for the special
879 situation when an empty string is matched. It now does the fancy magic
880 stuff that is necessary.
882 6. In pcre_internal.h, obsolete includes to setjmp.h and stdarg.h have been
883 removed. (These were left over from very, very early versions of PCRE.)
885 7. Some cosmetic changes to the code to make life easier when compiling it
886 as part of something else:
888 (a) Change DEBUG to PCRE_DEBUG.
890 (b) In pcre_compile(), rename the member of the "branch_chain" structure
891 called "current" as "current_branch", to prevent a collision with the
892 Linux macro when compiled as a kernel module.
894 (c) In pcre_study(), rename the function set_bit() as set_table_bit(), to
895 prevent a collision with the Linux macro when compiled as a kernel
896 module.
898 8. In pcre_compile() there are some checks for integer overflows that used to
899 cast potentially large values to (double). This has been changed to that
900 when building, a check for int64_t is made, and if it is found, it is used
901 instead, thus avoiding the use of floating point arithmetic. (There is no
902 other use of FP in PCRE.) If int64_t is not found, the fallback is to
903 double.
905 9. Added two casts to avoid signed/unsigned warnings from VS Studio Express
906 2005 (difference between two addresses compared to an unsigned value).
908 10. Change the standard AC_CHECK_LIB test for libbz2 in configure.ac to a
909 custom one, because of the following reported problem in Windows:
911 - libbz2 uses the Pascal calling convention (WINAPI) for the functions
912 under Win32.
913 - The standard autoconf AC_CHECK_LIB fails to include "bzlib.h",
914 therefore missing the function definition.
915 - The compiler thus generates a "C" signature for the test function.
916 - The linker fails to find the "C" function.
917 - PCRE fails to configure if asked to do so against libbz2.
919 11. When running libtoolize from libtool-2.2.6b as part of autogen.sh, these
920 messages were output:
922 Consider adding `AC_CONFIG_MACRO_DIR([m4])' to configure.ac and
923 rerunning libtoolize, to keep the correct libtool macros in-tree.
924 Consider adding `-I m4' to ACLOCAL_AMFLAGS in Makefile.am.
926 I have done both of these things.
928 12. Although pcre_dfa_exec() does not use nearly as much stack as pcre_exec()
929 most of the time, it *can* run out if it is given a pattern that contains a
930 runaway infinite recursion. I updated the discussion in the pcrestack man
931 page.
933 13. Now that we have gone to the x.xx style of version numbers, the minor
934 version may start with zero. Using 08 or 09 is a bad idea because users
935 might check the value of PCRE_MINOR in their code, and 08 or 09 may be
936 interpreted as invalid octal numbers. I've updated the previous comment in
937 configure.ac, and also added a check that gives an error if 08 or 09 are
938 used.
940 14. Change 8.00/11 was not quite complete: code had been accidentally omitted,
941 causing partial matching to fail when the end of the subject matched \W
942 in a UTF-8 pattern where \W was quantified with a minimum of 3.
944 15. There were some discrepancies between the declarations in pcre_internal.h
945 of _pcre_is_newline(), _pcre_was_newline(), and _pcre_valid_utf8() and
946 their definitions. The declarations used "const uschar *" and the
947 definitions used USPTR. Even though USPTR is normally defined as "const
948 unsigned char *" (and uschar is typedeffed as "unsigned char"), it was
949 reported that: "This difference in casting confuses some C++ compilers, for
950 example, SunCC recognizes above declarations as different functions and
951 generates broken code for hbpcre." I have changed the declarations to use
952 USPTR.
954 16. GNU libtool is named differently on some systems. The autogen.sh script now
955 tries several variants such as glibtoolize (MacOSX) and libtoolize1x
956 (FreeBSD).
958 17. Applied Craig's patch that fixes an HP aCC compile error in pcre 8.00
959 (strtoXX undefined when compiling pcrecpp.cc). The patch contains this
960 comment: "Figure out how to create a longlong from a string: strtoll and
961 equivalent. It's not enough to call AC_CHECK_FUNCS: hpux has a strtoll, for
962 instance, but it only takes 2 args instead of 3!"
964 18. A subtle bug concerned with back references has been fixed by a change of
965 specification, with a corresponding code fix. A pattern such as
966 ^(xa|=?\1a)+$ which contains a back reference inside the group to which it
967 refers, was giving matches when it shouldn't. For example, xa=xaaa would
968 match that pattern. Interestingly, Perl (at least up to 5.11.3) has the
969 same bug. Such groups have to be quantified to be useful, or contained
970 inside another quantified group. (If there's no repetition, the reference
971 can never match.) The problem arises because, having left the group and
972 moved on to the rest of the pattern, a later failure that backtracks into
973 the group uses the captured value from the final iteration of the group
974 rather than the correct earlier one. I have fixed this in PCRE by forcing
975 any group that contains a reference to itself to be an atomic group; that
976 is, there cannot be any backtracking into it once it has completed. This is
977 similar to recursive and subroutine calls.
980 Version 8.00 19-Oct-09
981 ----------------------
983 1. The table for translating pcre_compile() error codes into POSIX error codes
984 was out-of-date, and there was no check on the pcre_compile() error code
985 being within the table. This could lead to an OK return being given in
986 error.
988 2. Changed the call to open a subject file in pcregrep from fopen(pathname,
989 "r") to fopen(pathname, "rb"), which fixed a problem with some of the tests
990 in a Windows environment.
992 3. The pcregrep --count option prints the count for each file even when it is
993 zero, as does GNU grep. However, pcregrep was also printing all files when
994 --files-with-matches was added. Now, when both options are given, it prints
995 counts only for those files that have at least one match. (GNU grep just
996 prints the file name in this circumstance, but including the count seems
997 more useful - otherwise, why use --count?) Also ensured that the
998 combination -clh just lists non-zero counts, with no names.
1000 4. The long form of the pcregrep -F option was incorrectly implemented as
1001 --fixed_strings instead of --fixed-strings. This is an incompatible change,
1002 but it seems right to fix it, and I didn't think it was worth preserving
1003 the old behaviour.
1005 5. The command line items --regex=pattern and --regexp=pattern were not
1006 recognized by pcregrep, which required --regex pattern or --regexp pattern
1007 (with a space rather than an '='). The man page documented the '=' forms,
1008 which are compatible with GNU grep; these now work.
1010 6. No libpcreposix.pc file was created for pkg-config; there was just
1011 libpcre.pc and libpcrecpp.pc. The omission has been rectified.
1013 7. Added #ifndef SUPPORT_UCP into the pcre_ucd.c module, to reduce its size
1014 when UCP support is not needed, by modifying the Python script that
1015 generates it from Unicode data files. This should not matter if the module
1016 is correctly used as a library, but I received one complaint about 50K of
1017 unwanted data. My guess is that the person linked everything into his
1018 program rather than using a library. Anyway, it does no harm.
1020 8. A pattern such as /\x{123}{2,2}+/8 was incorrectly compiled; the trigger
1021 was a minimum greater than 1 for a wide character in a possessive
1022 repetition. The same bug could also affect patterns like /(\x{ff}{0,2})*/8
1023 which had an unlimited repeat of a nested, fixed maximum repeat of a wide
1024 character. Chaos in the form of incorrect output or a compiling loop could
1025 result.
1027 9. The restrictions on what a pattern can contain when partial matching is
1028 requested for pcre_exec() have been removed. All patterns can now be
1029 partially matched by this function. In addition, if there are at least two
1030 slots in the offset vector, the offset of the earliest inspected character
1031 for the match and the offset of the end of the subject are set in them when
1032 PCRE_ERROR_PARTIAL is returned.
1034 10. Partial matching has been split into two forms: PCRE_PARTIAL_SOFT, which is
1035 synonymous with PCRE_PARTIAL, for backwards compatibility, and
1036 PCRE_PARTIAL_HARD, which causes a partial match to supersede a full match,
1037 and may be more useful for multi-segment matching.
1039 11. Partial matching with pcre_exec() is now more intuitive. A partial match
1040 used to be given if ever the end of the subject was reached; now it is
1041 given only if matching could not proceed because another character was
1042 needed. This makes a difference in some odd cases such as Z(*FAIL) with the
1043 string "Z", which now yields "no match" instead of "partial match". In the
1044 case of pcre_dfa_exec(), "no match" is given if every matching path for the
1045 final character ended with (*FAIL).
1047 12. Restarting a match using pcre_dfa_exec() after a partial match did not work
1048 if the pattern had a "must contain" character that was already found in the
1049 earlier partial match, unless partial matching was again requested. For
1050 example, with the pattern /dog.(body)?/, the "must contain" character is
1051 "g". If the first part-match was for the string "dog", restarting with
1052 "sbody" failed. This bug has been fixed.
1054 13. The string returned by pcre_dfa_exec() after a partial match has been
1055 changed so that it starts at the first inspected character rather than the
1056 first character of the match. This makes a difference only if the pattern
1057 starts with a lookbehind assertion or \b or \B (\K is not supported by
1058 pcre_dfa_exec()). It's an incompatible change, but it makes the two
1059 matching functions compatible, and I think it's the right thing to do.
1061 14. Added a pcredemo man page, created automatically from the pcredemo.c file,
1062 so that the demonstration program is easily available in environments where
1063 PCRE has not been installed from source.
1065 15. Arranged to add -DPCRE_STATIC to cflags in libpcre.pc, libpcreposix.cp,
1066 libpcrecpp.pc and pcre-config when PCRE is not compiled as a shared
1067 library.
1069 16. Added REG_UNGREEDY to the pcreposix interface, at the request of a user.
1070 It maps to PCRE_UNGREEDY. It is not, of course, POSIX-compatible, but it
1071 is not the first non-POSIX option to be added. Clearly some people find
1072 these options useful.
1074 17. If a caller to the POSIX matching function regexec() passes a non-zero
1075 value for nmatch with a NULL value for pmatch, the value of
1076 nmatch is forced to zero.
1078 18. RunGrepTest did not have a test for the availability of the -u option of
1079 the diff command, as RunTest does. It now checks in the same way as
1080 RunTest, and also checks for the -b option.
1082 19. If an odd number of negated classes containing just a single character
1083 interposed, within parentheses, between a forward reference to a named
1084 subpattern and the definition of the subpattern, compilation crashed with
1085 an internal error, complaining that it could not find the referenced
1086 subpattern. An example of a crashing pattern is /(?&A)(([^m])(?<A>))/.
1087 [The bug was that it was starting one character too far in when skipping
1088 over the character class, thus treating the ] as data rather than
1089 terminating the class. This meant it could skip too much.]
1091 20. Added PCRE_NOTEMPTY_ATSTART in order to be able to correctly implement the
1092 /g option in pcretest when the pattern contains \K, which makes it possible
1093 to have an empty string match not at the start, even when the pattern is
1094 anchored. Updated pcretest and pcredemo to use this option.
1096 21. If the maximum number of capturing subpatterns in a recursion was greater
1097 than the maximum at the outer level, the higher number was returned, but
1098 with unset values at the outer level. The correct (outer level) value is
1099 now given.
1101 22. If (*ACCEPT) appeared inside capturing parentheses, previous releases of
1102 PCRE did not set those parentheses (unlike Perl). I have now found a way to
1103 make it do so. The string so far is captured, making this feature
1104 compatible with Perl.
1106 23. The tests have been re-organized, adding tests 11 and 12, to make it
1107 possible to check the Perl 5.10 features against Perl 5.10.
1109 24. Perl 5.10 allows subroutine calls in lookbehinds, as long as the subroutine
1110 pattern matches a fixed length string. PCRE did not allow this; now it
1111 does. Neither allows recursion.
1113 25. I finally figured out how to implement a request to provide the minimum
1114 length of subject string that was needed in order to match a given pattern.
1115 (It was back references and recursion that I had previously got hung up
1116 on.) This code has now been added to pcre_study(); it finds a lower bound
1117 to the length of subject needed. It is not necessarily the greatest lower
1118 bound, but using it to avoid searching strings that are too short does give
1119 some useful speed-ups. The value is available to calling programs via
1120 pcre_fullinfo().
1122 26. While implementing 25, I discovered to my embarrassment that pcretest had
1123 not been passing the result of pcre_study() to pcre_dfa_exec(), so the
1124 study optimizations had never been tested with that matching function.
1125 Oops. What is worse, even when it was passed study data, there was a bug in
1126 pcre_dfa_exec() that meant it never actually used it. Double oops. There
1127 were also very few tests of studied patterns with pcre_dfa_exec().
1129 27. If (?| is used to create subpatterns with duplicate numbers, they are now
1130 allowed to have the same name, even if PCRE_DUPNAMES is not set. However,
1131 on the other side of the coin, they are no longer allowed to have different
1132 names, because these cannot be distinguished in PCRE, and this has caused
1133 confusion. (This is a difference from Perl.)
1135 28. When duplicate subpattern names are present (necessarily with different
1136 numbers, as required by 27 above), and a test is made by name in a
1137 conditional pattern, either for a subpattern having been matched, or for
1138 recursion in such a pattern, all the associated numbered subpatterns are
1139 tested, and the overall condition is true if the condition is true for any
1140 one of them. This is the way Perl works, and is also more like the way
1141 testing by number works.
1144 Version 7.9 11-Apr-09
1145 ---------------------
1147 1. When building with support for bzlib/zlib (pcregrep) and/or readline
1148 (pcretest), all targets were linked against these libraries. This included
1149 libpcre, libpcreposix, and libpcrecpp, even though they do not use these
1150 libraries. This caused unwanted dependencies to be created. This problem
1151 has been fixed, and now only pcregrep is linked with bzlib/zlib and only
1152 pcretest is linked with readline.
1154 2. The "typedef int BOOL" in pcre_internal.h that was included inside the
1155 "#ifndef FALSE" condition by an earlier change (probably 7.8/18) has been
1156 moved outside it again, because FALSE and TRUE are already defined in AIX,
1157 but BOOL is not.
1159 3. The pcre_config() function was treating the PCRE_MATCH_LIMIT and
1160 PCRE_MATCH_LIMIT_RECURSION values as ints, when they should be long ints.
1162 4. The pcregrep documentation said spaces were inserted as well as colons (or
1163 hyphens) following file names and line numbers when outputting matching
1164 lines. This is not true; no spaces are inserted. I have also clarified the
1165 wording for the --colour (or --color) option.
1167 5. In pcregrep, when --colour was used with -o, the list of matching strings
1168 was not coloured; this is different to GNU grep, so I have changed it to be
1169 the same.
1171 6. When --colo(u)r was used in pcregrep, only the first matching substring in
1172 each matching line was coloured. Now it goes on to look for further matches
1173 of any of the test patterns, which is the same behaviour as GNU grep.
1175 7. A pattern that could match an empty string could cause pcregrep to loop; it
1176 doesn't make sense to accept an empty string match in pcregrep, so I have
1177 locked it out (using PCRE's PCRE_NOTEMPTY option). By experiment, this
1178 seems to be how GNU grep behaves.
1180 8. The pattern (?(?=.*b)b|^) was incorrectly compiled as "match must be at
1181 start or after a newline", because the conditional assertion was not being
1182 correctly handled. The rule now is that both the assertion and what follows
1183 in the first alternative must satisfy the test.
1185 9. If auto-callout was enabled in a pattern with a conditional group whose
1186 condition was an assertion, PCRE could crash during matching, both with
1187 pcre_exec() and pcre_dfa_exec().
1189 10. The PCRE_DOLLAR_ENDONLY option was not working when pcre_dfa_exec() was
1190 used for matching.
1192 11. Unicode property support in character classes was not working for
1193 characters (bytes) greater than 127 when not in UTF-8 mode.
1195 12. Added the -M command line option to pcretest.
1197 14. Added the non-standard REG_NOTEMPTY option to the POSIX interface.
1199 15. Added the PCRE_NO_START_OPTIMIZE match-time option.
1201 16. Added comments and documentation about mis-use of no_arg in the C++
1202 wrapper.
1204 17. Implemented support for UTF-8 encoding in EBCDIC environments, a patch
1205 from Martin Jerabek that uses macro names for all relevant character and
1206 string constants.
1208 18. Added to pcre_internal.h two configuration checks: (a) If both EBCDIC and
1209 SUPPORT_UTF8 are set, give an error; (b) If SUPPORT_UCP is set without
1210 SUPPORT_UTF8, define SUPPORT_UTF8. The "configure" script handles both of
1211 these, but not everybody uses configure.
1213 19. A conditional group that had only one branch was not being correctly
1214 recognized as an item that could match an empty string. This meant that an
1215 enclosing group might also not be so recognized, causing infinite looping
1216 (and probably a segfault) for patterns such as ^"((?(?=[a])[^"])|b)*"$
1217 with the subject "ab", where knowledge that the repeated group can match
1218 nothing is needed in order to break the loop.
1220 20. If a pattern that was compiled with callouts was matched using pcre_dfa_
1221 exec(), but without supplying a callout function, matching went wrong.
1223 21. If PCRE_ERROR_MATCHLIMIT occurred during a recursion, there was a memory
1224 leak if the size of the offset vector was greater than 30. When the vector
1225 is smaller, the saved offsets during recursion go onto a local stack
1226 vector, but for larger vectors malloc() is used. It was failing to free
1227 when the recursion yielded PCRE_ERROR_MATCH_LIMIT (or any other "abnormal"
1228 error, in fact).
1230 22. There was a missing #ifdef SUPPORT_UTF8 round one of the variables in the
1231 heapframe that is used only when UTF-8 support is enabled. This caused no
1232 problem, but was untidy.
1234 23. Steven Van Ingelgem's patch to CMakeLists.txt to change the name
1235 CMAKE_BINARY_DIR to PROJECT_BINARY_DIR so that it works when PCRE is
1236 included within another project.
1238 24. Steven Van Ingelgem's patches to add more options to the CMake support,
1239 slightly modified by me:
1241 (a) PCRE_BUILD_TESTS can be set OFF not to build the tests, including
1242 not building pcregrep.
1244 (b) PCRE_BUILD_PCREGREP can be see OFF not to build pcregrep, but only
1245 if PCRE_BUILD_TESTS is also set OFF, because the tests use pcregrep.
1247 25. Forward references, both numeric and by name, in patterns that made use of
1248 duplicate group numbers, could behave incorrectly or give incorrect errors,
1249 because when scanning forward to find the reference group, PCRE was not
1250 taking into account the duplicate group numbers. A pattern such as
1251 ^X(?3)(a)(?|(b)|(q))(Y) is an example.
1253 26. Changed a few more instances of "const unsigned char *" to USPTR, making
1254 the feature of a custom pointer more persuasive (as requested by a user).
1256 27. Wrapped the definitions of fileno and isatty for Windows, which appear in
1257 pcretest.c, inside #ifndefs, because it seems they are sometimes already
1258 pre-defined.
1260 28. Added support for (*UTF8) at the start of a pattern.
1262 29. Arrange for flags added by the "release type" setting in CMake to be shown
1263 in the configuration summary.
1266 Version 7.8 05-Sep-08
1267 ---------------------
1269 1. Replaced UCP searching code with optimized version as implemented for Ad
1270 Muncher (http://www.admuncher.com/) by Peter Kankowski. This uses a two-
1271 stage table and inline lookup instead of a function, giving speed ups of 2
1272 to 5 times on some simple patterns that I tested. Permission was given to
1273 distribute the MultiStage2.py script that generates the tables (it's not in
1274 the tarball, but is in the Subversion repository).
1276 2. Updated the Unicode datatables to Unicode 5.1.0. This adds yet more
1277 scripts.
1279 3. Change 12 for 7.7 introduced a bug in pcre_study() when a pattern contained
1280 a group with a zero qualifier. The result of the study could be incorrect,
1281 or the function might crash, depending on the pattern.
1283 4. Caseless matching was not working for non-ASCII characters in back
1284 references. For example, /(\x{de})\1/8i was not matching \x{de}\x{fe}.
1285 It now works when Unicode Property Support is available.
1287 5. In pcretest, an escape such as \x{de} in the data was always generating
1288 a UTF-8 string, even in non-UTF-8 mode. Now it generates a single byte in
1289 non-UTF-8 mode. If the value is greater than 255, it gives a warning about
1290 truncation.
1292 6. Minor bugfix in pcrecpp.cc (change "" == ... to NULL == ...).
1294 7. Added two (int) casts to pcregrep when printing the difference of two
1295 pointers, in case they are 64-bit values.
1297 8. Added comments about Mac OS X stack usage to the pcrestack man page and to
1298 test 2 if it fails.
1300 9. Added PCRE_CALL_CONVENTION just before the names of all exported functions,
1301 and a #define of that name to empty if it is not externally set. This is to
1302 allow users of MSVC to set it if necessary.
1304 10. The PCRE_EXP_DEFN macro which precedes exported functions was missing from
1305 the convenience functions in the pcre_get.c source file.
1307 11. An option change at the start of a pattern that had top-level alternatives
1308 could cause overwriting and/or a crash. This command provoked a crash in
1309 some environments:
1311 printf "/(?i)[\xc3\xa9\xc3\xbd]|[\xc3\xa9\xc3\xbdA]/8\n" | pcretest
1313 This potential security problem was recorded as CVE-2008-2371.
1315 12. For a pattern where the match had to start at the beginning or immediately
1316 after a newline (e.g /.*anything/ without the DOTALL flag), pcre_exec() and
1317 pcre_dfa_exec() could read past the end of the passed subject if there was
1318 no match. To help with detecting such bugs (e.g. with valgrind), I modified
1319 pcretest so that it places the subject at the end of its malloc-ed buffer.
1321 13. The change to pcretest in 12 above threw up a couple more cases when pcre_
1322 exec() might read past the end of the data buffer in UTF-8 mode.
1324 14. A similar bug to 7.3/2 existed when the PCRE_FIRSTLINE option was set and
1325 the data contained the byte 0x85 as part of a UTF-8 character within its
1326 first line. This applied both to normal and DFA matching.
1328 15. Lazy qualifiers were not working in some cases in UTF-8 mode. For example,
1329 /^[^d]*?$/8 failed to match "abc".
1331 16. Added a missing copyright notice to pcrecpp_internal.h.
1333 17. Make it more clear in the documentation that values returned from
1334 pcre_exec() in ovector are byte offsets, not character counts.
1336 18. Tidied a few places to stop certain compilers from issuing warnings.
1338 19. Updated the Virtual Pascal + BCC files to compile the latest v7.7, as
1339 supplied by Stefan Weber. I made a further small update for 7.8 because
1340 there is a change of source arrangements: the pcre_searchfuncs.c module is
1341 replaced by pcre_ucd.c.
1344 Version 7.7 07-May-08
1345 ---------------------
1347 1. Applied Craig's patch to sort out a long long problem: "If we can't convert
1348 a string to a long long, pretend we don't even have a long long." This is
1349 done by checking for the strtoq, strtoll, and _strtoi64 functions.
1351 2. Applied Craig's patch to pcrecpp.cc to restore ABI compatibility with
1352 pre-7.6 versions, which defined a global no_arg variable instead of putting
1353 it in the RE class. (See also #8 below.)
1355 3. Remove a line of dead code, identified by coverity and reported by Nuno
1356 Lopes.
1358 4. Fixed two related pcregrep bugs involving -r with --include or --exclude:
1360 (1) The include/exclude patterns were being applied to the whole pathnames
1361 of files, instead of just to the final components.
1363 (2) If there was more than one level of directory, the subdirectories were
1364 skipped unless they satisfied the include/exclude conditions. This is
1365 inconsistent with GNU grep (and could even be seen as contrary to the
1366 pcregrep specification - which I improved to make it absolutely clear).
1367 The action now is always to scan all levels of directory, and just
1368 apply the include/exclude patterns to regular files.
1370 5. Added the --include_dir and --exclude_dir patterns to pcregrep, and used
1371 --exclude_dir in the tests to avoid scanning .svn directories.
1373 6. Applied Craig's patch to the QuoteMeta function so that it escapes the
1374 NUL character as backslash + 0 rather than backslash + NUL, because PCRE
1375 doesn't support NULs in patterns.
1377 7. Added some missing "const"s to declarations of static tables in
1378 pcre_compile.c and pcre_dfa_exec.c.
1380 8. Applied Craig's patch to pcrecpp.cc to fix a problem in OS X that was
1381 caused by fix #2 above. (Subsequently also a second patch to fix the
1382 first patch. And a third patch - this was a messy problem.)
1384 9. Applied Craig's patch to remove the use of push_back().
1386 10. Applied Alan Lehotsky's patch to add REG_STARTEND support to the POSIX
1387 matching function regexec().
1389 11. Added support for the Oniguruma syntax \g<name>, \g<n>, \g'name', \g'n',
1390 which, however, unlike Perl's \g{...}, are subroutine calls, not back
1391 references. PCRE supports relative numbers with this syntax (I don't think
1392 Oniguruma does).
1394 12. Previously, a group with a zero repeat such as (...){0} was completely
1395 omitted from the compiled regex. However, this means that if the group
1396 was called as a subroutine from elsewhere in the pattern, things went wrong
1397 (an internal error was given). Such groups are now left in the compiled
1398 pattern, with a new opcode that causes them to be skipped at execution
1399 time.
1401 13. Added the PCRE_JAVASCRIPT_COMPAT option. This makes the following changes
1402 to the way PCRE behaves:
1404 (a) A lone ] character is dis-allowed (Perl treats it as data).
1406 (b) A back reference to an unmatched subpattern matches an empty string
1407 (Perl fails the current match path).
1409 (c) A data ] in a character class must be notated as \] because if the
1410 first data character in a class is ], it defines an empty class. (In
1411 Perl it is not possible to have an empty class.) The empty class []
1412 never matches; it forces failure and is equivalent to (*FAIL) or (?!).
1413 The negative empty class [^] matches any one character, independently
1414 of the DOTALL setting.
1416 14. A pattern such as /(?2)[]a()b](abc)/ which had a forward reference to a
1417 non-existent subpattern following a character class starting with ']' and
1418 containing () gave an internal compiling error instead of "reference to
1419 non-existent subpattern". Fortunately, when the pattern did exist, the
1420 compiled code was correct. (When scanning forwards to check for the
1421 existencd of the subpattern, it was treating the data ']' as terminating
1422 the class, so got the count wrong. When actually compiling, the reference
1423 was subsequently set up correctly.)
1425 15. The "always fail" assertion (?!) is optimzed to (*FAIL) by pcre_compile;
1426 it was being rejected as not supported by pcre_dfa_exec(), even though
1427 other assertions are supported. I have made pcre_dfa_exec() support
1428 (*FAIL).
1430 16. The implementation of 13c above involved the invention of a new opcode,
1431 OP_ALLANY, which is like OP_ANY but doesn't check the /s flag. Since /s
1432 cannot be changed at match time, I realized I could make a small
1433 improvement to matching performance by compiling OP_ALLANY instead of
1434 OP_ANY for "." when DOTALL was set, and then removing the runtime tests
1435 on the OP_ANY path.
1437 17. Compiling pcretest on Windows with readline support failed without the
1438 following two fixes: (1) Make the unistd.h include conditional on
1439 HAVE_UNISTD_H; (2) #define isatty and fileno as _isatty and _fileno.
1441 18. Changed CMakeLists.txt and cmake/FindReadline.cmake to arrange for the
1442 ncurses library to be included for pcretest when ReadLine support is
1443 requested, but also to allow for it to be overridden. This patch came from
1444 Daniel Bergström.
1446 19. There was a typo in the file ucpinternal.h where f0_rangeflag was defined
1447 as 0x00f00000 instead of 0x00800000. Luckily, this would not have caused
1448 any errors with the current Unicode tables. Thanks to Peter Kankowski for
1449 spotting this.
1452 Version 7.6 28-Jan-08
1453 ---------------------
1455 1. A character class containing a very large number of characters with
1456 codepoints greater than 255 (in UTF-8 mode, of course) caused a buffer
1457 overflow.
1459 2. Patch to cut out the "long long" test in pcrecpp_unittest when
1460 HAVE_LONG_LONG is not defined.
1462 3. Applied Christian Ehrlicher's patch to update the CMake build files to
1463 bring them up to date and include new features. This patch includes:
1465 - Fixed PH's badly added libz and libbz2 support.
1466 - Fixed a problem with static linking.
1467 - Added pcredemo. [But later removed - see 7 below.]
1468 - Fixed dftables problem and added an option.
1469 - Added a number of HAVE_XXX tests, including HAVE_WINDOWS_H and
1471 - Added readline support for pcretest.
1472 - Added an listing of the option settings after cmake has run.
1474 4. A user submitted a patch to Makefile that makes it easy to create
1475 "pcre.dll" under mingw when using Configure/Make. I added stuff to
1476 Makefile.am that cause it to include this special target, without
1477 affecting anything else. Note that the same mingw target plus all
1478 the other distribution libraries and programs are now supported
1479 when configuring with CMake (see 6 below) instead of with
1480 Configure/Make.
1482 5. Applied Craig's patch that moves no_arg into the RE class in the C++ code.
1483 This is an attempt to solve the reported problem "pcrecpp::no_arg is not
1484 exported in the Windows port". It has not yet been confirmed that the patch
1485 solves the problem, but it does no harm.
1487 6. Applied Sheri's patch to CMakeLists.txt to add NON_STANDARD_LIB_PREFIX and
1488 NON_STANDARD_LIB_SUFFIX for dll names built with mingw when configured
1489 with CMake, and also correct the comment about stack recursion.
1491 7. Remove the automatic building of pcredemo from the ./configure system and
1492 from CMakeLists.txt. The whole idea of pcredemo.c is that it is an example
1493 of a program that users should build themselves after PCRE is installed, so
1494 building it automatically is not really right. What is more, it gave
1495 trouble in some build environments.
1497 8. Further tidies to CMakeLists.txt from Sheri and Christian.
1500 Version 7.5 10-Jan-08
1501 ---------------------
1503 1. Applied a patch from Craig: "This patch makes it possible to 'ignore'
1504 values in parens when parsing an RE using the C++ wrapper."
1506 2. Negative specials like \S did not work in character classes in UTF-8 mode.
1507 Characters greater than 255 were excluded from the class instead of being
1508 included.
1510 3. The same bug as (2) above applied to negated POSIX classes such as
1511 [:^space:].
1513 4. PCRECPP_STATIC was referenced in pcrecpp_internal.h, but nowhere was it
1514 defined or documented. It seems to have been a typo for PCRE_STATIC, so
1515 I have changed it.
1517 5. The construct (?&) was not diagnosed as a syntax error (it referenced the
1518 first named subpattern) and a construct such as (?&a) would reference the
1519 first named subpattern whose name started with "a" (in other words, the
1520 length check was missing). Both these problems are fixed. "Subpattern name
1521 expected" is now given for (?&) (a zero-length name), and this patch also
1522 makes it give the same error for \k'' (previously it complained that that
1523 was a reference to a non-existent subpattern).
1525 6. The erroneous patterns (?+-a) and (?-+a) give different error messages;
1526 this is right because (?- can be followed by option settings as well as by
1527 digits. I have, however, made the messages clearer.
1529 7. Patterns such as (?(1)a|b) (a pattern that contains fewer subpatterns
1530 than the number used in the conditional) now cause a compile-time error.
1531 This is actually not compatible with Perl, which accepts such patterns, but
1532 treats the conditional as always being FALSE (as PCRE used to), but it
1533 seems to me that giving a diagnostic is better.
1535 8. Change "alphameric" to the more common word "alphanumeric" in comments
1536 and messages.
1538 9. Fix two occurrences of "backslash" in comments that should have been
1539 "backspace".
1541 10. Remove two redundant lines of code that can never be obeyed (their function
1542 was moved elsewhere).
1544 11. The program that makes PCRE's Unicode character property table had a bug
1545 which caused it to generate incorrect table entries for sequences of
1546 characters that have the same character type, but are in different scripts.
1547 It amalgamated them into a single range, with the script of the first of
1548 them. In other words, some characters were in the wrong script. There were
1549 thirteen such cases, affecting characters in the following ranges:
1551 U+002b0 - U+002c1
1552 U+0060c - U+0060d
1553 U+0061e - U+00612
1554 U+0064b - U+0065e
1555 U+0074d - U+0076d
1556 U+01800 - U+01805
1557 U+01d00 - U+01d77
1558 U+01d9b - U+01dbf
1559 U+0200b - U+0200f
1560 U+030fc - U+030fe
1561 U+03260 - U+0327f
1562 U+0fb46 - U+0fbb1
1563 U+10450 - U+1049d
1565 12. The -o option (show only the matching part of a line) for pcregrep was not
1566 compatible with GNU grep in that, if there was more than one match in a
1567 line, it showed only the first of them. It now behaves in the same way as
1568 GNU grep.
1570 13. If the -o and -v options were combined for pcregrep, it printed a blank
1571 line for every non-matching line. GNU grep prints nothing, and pcregrep now
1572 does the same. The return code can be used to tell if there were any
1573 non-matching lines.
1575 14. Added --file-offsets and --line-offsets to pcregrep.
1577 15. The pattern (?=something)(?R) was not being diagnosed as a potentially
1578 infinitely looping recursion. The bug was that positive lookaheads were not
1579 being skipped when checking for a possible empty match (negative lookaheads
1580 and both kinds of lookbehind were skipped).
1582 16. Fixed two typos in the Windows-only code in pcregrep.c, and moved the
1583 inclusion of <windows.h> to before rather than after the definition of
1584 INVALID_FILE_ATTRIBUTES (patch from David Byron).
1586 17. Specifying a possessive quantifier with a specific limit for a Unicode
1587 character property caused pcre_compile() to compile bad code, which led at
1588 runtime to PCRE_ERROR_INTERNAL (-14). Examples of patterns that caused this
1589 are: /\p{Zl}{2,3}+/8 and /\p{Cc}{2}+/8. It was the possessive "+" that
1590 caused the error; without that there was no problem.
1592 18. Added --enable-pcregrep-libz and --enable-pcregrep-libbz2.
1594 19. Added --enable-pcretest-libreadline.
1596 20. In pcrecpp.cc, the variable 'count' was incremented twice in
1597 RE::GlobalReplace(). As a result, the number of replacements returned was
1598 double what it should be. I removed one of the increments, but Craig sent a
1599 later patch that removed the other one (the right fix) and added unit tests
1600 that check the return values (which was not done before).
1602 21. Several CMake things:
1604 (1) Arranged that, when cmake is used on Unix, the libraries end up with
1605 the names libpcre and libpcreposix, not just pcre and pcreposix.
1607 (2) The above change means that pcretest and pcregrep are now correctly
1608 linked with the newly-built libraries, not previously installed ones.
1612 22. In UTF-8 mode, with newline set to "any", a pattern such as .*a.*=.b.*
1613 crashed when matching a string such as a\x{2029}b (note that \x{2029} is a
1614 UTF-8 newline character). The key issue is that the pattern starts .*;
1615 this means that the match must be either at the beginning, or after a
1616 newline. The bug was in the code for advancing after a failed match and
1617 checking that the new position followed a newline. It was not taking
1618 account of UTF-8 characters correctly.
1620 23. PCRE was behaving differently from Perl in the way it recognized POSIX
1621 character classes. PCRE was not treating the sequence [:...:] as a
1622 character class unless the ... were all letters. Perl, however, seems to
1623 allow any characters between [: and :], though of course it rejects as
1624 unknown any "names" that contain non-letters, because all the known class
1625 names consist only of letters. Thus, Perl gives an error for [[:1234:]],
1626 for example, whereas PCRE did not - it did not recognize a POSIX character
1627 class. This seemed a bit dangerous, so the code has been changed to be
1628 closer to Perl. The behaviour is not identical to Perl, because PCRE will
1629 diagnose an unknown class for, for example, [[:l\ower:]] where Perl will
1630 treat it as [[:lower:]]. However, PCRE does now give "unknown" errors where
1631 Perl does, and where it didn't before.
1633 24. Rewrite so as to remove the single use of %n from pcregrep because in some
1634 Windows environments %n is disabled by default.
1637 Version 7.4 21-Sep-07
1638 ---------------------
1640 1. Change 7.3/28 was implemented for classes by looking at the bitmap. This
1641 means that a class such as [\s] counted as "explicit reference to CR or
1642 LF". That isn't really right - the whole point of the change was to try to
1643 help when there was an actual mention of one of the two characters. So now
1644 the change happens only if \r or \n (or a literal CR or LF) character is
1645 encountered.
1647 2. The 32-bit options word was also used for 6 internal flags, but the numbers
1648 of both had grown to the point where there were only 3 bits left.
1649 Fortunately, there was spare space in the data structure, and so I have
1650 moved the internal flags into a new 16-bit field to free up more option
1651 bits.
1653 3. The appearance of (?J) at the start of a pattern set the DUPNAMES option,
1654 but did not set the internal JCHANGED flag - either of these is enough to
1655 control the way the "get" function works - but the PCRE_INFO_JCHANGED
1656 facility is supposed to tell if (?J) was ever used, so now (?J) at the
1657 start sets both bits.
1659 4. Added options (at build time, compile time, exec time) to change \R from
1660 matching any Unicode line ending sequence to just matching CR, LF, or CRLF.
1662 5. doc/pcresyntax.html was missing from the distribution.
1664 6. Put back the definition of PCRE_ERROR_NULLWSLIMIT, for backward
1665 compatibility, even though it is no longer used.
1667 7. Added macro for snprintf to pcrecpp_unittest.cc and also for strtoll and
1668 strtoull to pcrecpp.cc to select the available functions in WIN32 when the
1669 windows.h file is present (where different names are used). [This was
1670 reversed later after testing - see 16 below.]
1672 8. Changed all #include <config.h> to #include "config.h". There were also
1673 some further <pcre.h> cases that I changed to "pcre.h".
1675 9. When pcregrep was used with the --colour option, it missed the line ending
1676 sequence off the lines that it output.
1678 10. It was pointed out to me that arrays of string pointers cause lots of
1679 relocations when a shared library is dynamically loaded. A technique of
1680 using a single long string with a table of offsets can drastically reduce
1681 these. I have refactored PCRE in four places to do this. The result is
1682 dramatic:
1684 Originally: 290
1685 After changing UCP table: 187
1686 After changing error message table: 43
1687 After changing table of "verbs" 36
1688 After changing table of Posix names 22
1690 Thanks to the folks working on Gregex for glib for this insight.
1692 11. --disable-stack-for-recursion caused compiling to fail unless -enable-
1693 unicode-properties was also set.
1695 12. Updated the tests so that they work when \R is defaulted to ANYCRLF.
1697 13. Added checks for ANY and ANYCRLF to pcrecpp.cc where it previously
1698 checked only for CRLF.
1700 14. Added casts to pcretest.c to avoid compiler warnings.
1702 15. Added Craig's patch to various pcrecpp modules to avoid compiler warnings.
1704 16. Added Craig's patch to remove the WINDOWS_H tests, that were not working,
1705 and instead check for _strtoi64 explicitly, and avoid the use of snprintf()
1706 entirely. This removes changes made in 7 above.
1708 17. The CMake files have been updated, and there is now more information about
1709 building with CMake in the NON-UNIX-USE document.
1712 Version 7.3 28-Aug-07
1713 ---------------------
1715 1. In the rejigging of the build system that eventually resulted in 7.1, the
1716 line "#include <pcre.h>" was included in pcre_internal.h. The use of angle
1717 brackets there is not right, since it causes compilers to look for an
1718 installed pcre.h, not the version that is in the source that is being
1719 compiled (which of course may be different). I have changed it back to:
1721 #include "pcre.h"
1723 I have a vague recollection that the change was concerned with compiling in
1724 different directories, but in the new build system, that is taken care of
1725 by the VPATH setting the Makefile.
1727 2. The pattern .*$ when run in not-DOTALL UTF-8 mode with newline=any failed
1728 when the subject happened to end in the byte 0x85 (e.g. if the last
1729 character was \x{1ec5}). *Character* 0x85 is one of the "any" newline
1730 characters but of course it shouldn't be taken as a newline when it is part
1731 of another character. The bug was that, for an unlimited repeat of . in
1732 not-DOTALL UTF-8 mode, PCRE was advancing by bytes rather than by
1733 characters when looking for a newline.
1735 3. A small performance improvement in the DOTALL UTF-8 mode .* case.
1737 4. Debugging: adjusted the names of opcodes for different kinds of parentheses
1738 in debug output.
1740 5. Arrange to use "%I64d" instead of "%lld" and "%I64u" instead of "%llu" for
1741 long printing in the pcrecpp unittest when running under MinGW.
1743 6. ESC_K was left out of the EBCDIC table.
1745 7. Change 7.0/38 introduced a new limit on the number of nested non-capturing
1746 parentheses; I made it 1000, which seemed large enough. Unfortunately, the
1747 limit also applies to "virtual nesting" when a pattern is recursive, and in
1748 this case 1000 isn't so big. I have been able to remove this limit at the
1749 expense of backing off one optimization in certain circumstances. Normally,
1750 when pcre_exec() would call its internal match() function recursively and
1751 immediately return the result unconditionally, it uses a "tail recursion"
1752 feature to save stack. However, when a subpattern that can match an empty
1753 string has an unlimited repetition quantifier, it no longer makes this
1754 optimization. That gives it a stack frame in which to save the data for
1755 checking that an empty string has been matched. Previously this was taken
1756 from the 1000-entry workspace that had been reserved. So now there is no
1757 explicit limit, but more stack is used.
1759 8. Applied Daniel's patches to solve problems with the import/export magic
1760 syntax that is required for Windows, and which was going wrong for the
1761 pcreposix and pcrecpp parts of the library. These were overlooked when this
1762 problem was solved for the main library.
1764 9. There were some crude static tests to avoid integer overflow when computing
1765 the size of patterns that contain repeated groups with explicit upper
1766 limits. As the maximum quantifier is 65535, the maximum group length was
1767 set at 30,000 so that the product of these two numbers did not overflow a
1768 32-bit integer. However, it turns out that people want to use groups that
1769 are longer than 30,000 bytes (though not repeat them that many times).
1770 Change 7.0/17 (the refactoring of the way the pattern size is computed) has
1771 made it possible to implement the integer overflow checks in a much more
1772 dynamic way, which I have now done. The artificial limitation on group
1773 length has been removed - we now have only the limit on the total length of
1774 the compiled pattern, which depends on the LINK_SIZE setting.
1776 10. Fixed a bug in the documentation for get/copy named substring when
1777 duplicate names are permitted. If none of the named substrings are set, the
1778 functions return PCRE_ERROR_NOSUBSTRING (7); the doc said they returned an
1779 empty string.
1781 11. Because Perl interprets \Q...\E at a high level, and ignores orphan \E
1782 instances, patterns such as [\Q\E] or [\E] or even [^\E] cause an error,
1783 because the ] is interpreted as the first data character and the
1784 terminating ] is not found. PCRE has been made compatible with Perl in this
1785 regard. Previously, it interpreted [\Q\E] as an empty class, and [\E] could
1786 cause memory overwriting.
1788 10. Like Perl, PCRE automatically breaks an unlimited repeat after an empty
1789 string has been matched (to stop an infinite loop). It was not recognizing
1790 a conditional subpattern that could match an empty string if that
1791 subpattern was within another subpattern. For example, it looped when
1792 trying to match (((?(1)X|))*) but it was OK with ((?(1)X|)*) where the
1793 condition was not nested. This bug has been fixed.
1795 12. A pattern like \X?\d or \P{L}?\d in non-UTF-8 mode could cause a backtrack
1796 past the start of the subject in the presence of bytes with the top bit
1797 set, for example "\x8aBCD".
1799 13. Added Perl 5.10 experimental backtracking controls (*FAIL), (*F), (*PRUNE),
1800 (*SKIP), (*THEN), (*COMMIT), and (*ACCEPT).
1802 14. Optimized (?!) to (*FAIL).
1804 15. Updated the test for a valid UTF-8 string to conform to the later RFC 3629.
1805 This restricts code points to be within the range 0 to 0x10FFFF, excluding
1806 the "low surrogate" sequence 0xD800 to 0xDFFF. Previously, PCRE allowed the
1807 full range 0 to 0x7FFFFFFF, as defined by RFC 2279. Internally, it still
1808 does: it's just the validity check that is more restrictive.
1810 16. Inserted checks for integer overflows during escape sequence (backslash)
1811 processing, and also fixed erroneous offset values for syntax errors during
1812 backslash processing.
1814 17. Fixed another case of looking too far back in non-UTF-8 mode (cf 12 above)
1815 for patterns like [\PPP\x8a]{1,}\x80 with the subject "A\x80".
1817 18. An unterminated class in a pattern like (?1)\c[ with a "forward reference"
1818 caused an overrun.
1820 19. A pattern like (?:[\PPa*]*){8,} which had an "extended class" (one with
1821 something other than just ASCII characters) inside a group that had an
1822 unlimited repeat caused a loop at compile time (while checking to see
1823 whether the group could match an empty string).
1825 20. Debugging a pattern containing \p or \P could cause a crash. For example,
1826 [\P{Any}] did so. (Error in the code for printing property names.)
1828 21. An orphan \E inside a character class could cause a crash.
1830 22. A repeated capturing bracket such as (A)? could cause a wild memory
1831 reference during compilation.
1833 23. There are several functions in pcre_compile() that scan along a compiled
1834 expression for various reasons (e.g. to see if it's fixed length for look
1835 behind). There were bugs in these functions when a repeated \p or \P was
1836 present in the pattern. These operators have additional parameters compared
1837 with \d, etc, and these were not being taken into account when moving along
1838 the compiled data. Specifically:
1840 (a) A item such as \p{Yi}{3} in a lookbehind was not treated as fixed
1841 length.
1843 (b) An item such as \pL+ within a repeated group could cause crashes or
1844 loops.
1846 (c) A pattern such as \p{Yi}+(\P{Yi}+)(?1) could give an incorrect
1847 "reference to non-existent subpattern" error.
1849 (d) A pattern like (\P{Yi}{2}\277)? could loop at compile time.
1851 24. A repeated \S or \W in UTF-8 mode could give wrong answers when multibyte
1852 characters were involved (for example /\S{2}/8g with "A\x{a3}BC").
1854 25. Using pcregrep in multiline, inverted mode (-Mv) caused it to loop.
1856 26. Patterns such as [\P{Yi}A] which include \p or \P and just one other
1857 character were causing crashes (broken optimization).
1859 27. Patterns such as (\P{Yi}*\277)* (group with possible zero repeat containing
1860 \p or \P) caused a compile-time loop.
1862 28. More problems have arisen in unanchored patterns when CRLF is a valid line
1863 break. For example, the unstudied pattern [\r\n]A does not match the string
1864 "\r\nA" because change 7.0/46 below moves the current point on by two
1865 characters after failing to match at the start. However, the pattern \nA
1866 *does* match, because it doesn't start till \n, and if [\r\n]A is studied,
1867 the same is true. There doesn't seem any very clean way out of this, but
1868 what I have chosen to do makes the common cases work: PCRE now takes note
1869 of whether there can be an explicit match for \r or \n anywhere in the
1870 pattern, and if so, 7.0/46 no longer applies. As part of this change,
1871 there's a new PCRE_INFO_HASCRORLF option for finding out whether a compiled
1872 pattern has explicit CR or LF references.
1874 29. Added (*CR) etc for changing newline setting at start of pattern.
1877 Version 7.2 19-Jun-07
1878 ---------------------
1880 1. If the fr_FR locale cannot be found for test 3, try the "french" locale,
1881 which is apparently normally available under Windows.
1883 2. Re-jig the pcregrep tests with different newline settings in an attempt
1884 to make them independent of the local environment's newline setting.
1886 3. Add code to configure.ac to remove -g from the CFLAGS default settings.
1888 4. Some of the "internals" tests were previously cut out when the link size
1889 was not 2, because the output contained actual offsets. The recent new
1890 "Z" feature of pcretest means that these can be cut out, making the tests
1891 usable with all link sizes.
1893 5. Implemented Stan Switzer's goto replacement for longjmp() when not using
1894 stack recursion. This gives a massive performance boost under BSD, but just
1895 a small improvement under Linux. However, it saves one field in the frame
1896 in all cases.
1898 6. Added more features from the forthcoming Perl 5.10:
1900 (a) (?-n) (where n is a string of digits) is a relative subroutine or
1901 recursion call. It refers to the nth most recently opened parentheses.
1903 (b) (?+n) is also a relative subroutine call; it refers to the nth next
1904 to be opened parentheses.
1906 (c) Conditions that refer to capturing parentheses can be specified
1907 relatively, for example, (?(-2)... or (?(+3)...
1909 (d) \K resets the start of the current match so that everything before
1910 is not part of it.
1912 (e) \k{name} is synonymous with \k<name> and \k'name' (.NET compatible).
1914 (f) \g{name} is another synonym - part of Perl 5.10's unification of
1915 reference syntax.
1917 (g) (?| introduces a group in which the numbering of parentheses in each
1918 alternative starts with the same number.
1920 (h) \h, \H, \v, and \V match horizontal and vertical whitespace.
1922 7. Added two new calls to pcre_fullinfo(): PCRE_INFO_OKPARTIAL and
1925 8. A pattern such as (.*(.)?)* caused pcre_exec() to fail by either not
1926 terminating or by crashing. Diagnosed by Viktor Griph; it was in the code
1927 for detecting groups that can match an empty string.
1929 9. A pattern with a very large number of alternatives (more than several
1930 hundred) was running out of internal workspace during the pre-compile
1931 phase, where pcre_compile() figures out how much memory will be needed. A
1932 bit of new cunning has reduced the workspace needed for groups with
1933 alternatives. The 1000-alternative test pattern now uses 12 bytes of
1934 workspace instead of running out of the 4096 that are available.
1936 10. Inserted some missing (unsigned int) casts to get rid of compiler warnings.
1938 11. Applied patch from Google to remove an optimization that didn't quite work.
1939 The report of the bug said:
1941 pcrecpp::RE("a*").FullMatch("aaa") matches, while
1942 pcrecpp::RE("a*?").FullMatch("aaa") does not, and
1943 pcrecpp::RE("a*?\\z").FullMatch("aaa") does again.
1945 12. If \p or \P was used in non-UTF-8 mode on a character greater than 127
1946 it matched the wrong number of bytes.
1949 Version 7.1 24-Apr-07
1950 ---------------------
1952 1. Applied Bob Rossi and Daniel G's patches to convert the build system to one
1953 that is more "standard", making use of automake and other Autotools. There
1954 is some re-arrangement of the files and adjustment of comments consequent
1955 on this.
1957 2. Part of the patch fixed a problem with the pcregrep tests. The test of -r
1958 for recursive directory scanning broke on some systems because the files
1959 are not scanned in any specific order and on different systems the order
1960 was different. A call to "sort" has been inserted into RunGrepTest for the
1961 approprate test as a short-term fix. In the longer term there may be an
1962 alternative.
1964 3. I had an email from Eric Raymond about problems translating some of PCRE's
1965 man pages to HTML (despite the fact that I distribute HTML pages, some
1966 people do their own conversions for various reasons). The problems
1967 concerned the use of low-level troff macros .br and .in. I have therefore
1968 removed all such uses from the man pages (some were redundant, some could
1969 be replaced by .nf/.fi pairs). The 132html script that I use to generate
1970 HTML has been updated to handle .nf/.fi and to complain if it encounters
1971 .br or .in.
1973 4. Updated comments in configure.ac that get placed in config.h.in and also
1974 arranged for config.h to be included in the distribution, with the name
1975 config.h.generic, for the benefit of those who have to compile without
1976 Autotools (compare pcre.h, which is now distributed as pcre.h.generic).
1978 5. Updated the support (such as it is) for Virtual Pascal, thanks to Stefan
1979 Weber: (1) pcre_internal.h was missing some function renames; (2) updated
1980 makevp.bat for the current PCRE, using the additional files
1981 makevp_c.txt, makevp_l.txt, and pcregexp.pas.
1983 6. A Windows user reported a minor discrepancy with test 2, which turned out
1984 to be caused by a trailing space on an input line that had got lost in his
1985 copy. The trailing space was an accident, so I've just removed it.
1987 7. Add -Wl,-R... flags in pcre-config.in for *BSD* systems, as I'm told
1988 that is needed.
1990 8. Mark ucp_table (in ucptable.h) and ucp_gentype (in pcre_ucp_searchfuncs.c)
1991 as "const" (a) because they are and (b) because it helps the PHP
1992 maintainers who have recently made a script to detect big data structures
1993 in the php code that should be moved to the .rodata section. I remembered
1994 to update Builducptable as well, so it won't revert if ucptable.h is ever
1995 re-created.
1997 9. Added some extra #ifdef SUPPORT_UTF8 conditionals into pcretest.c,
1998 pcre_printint.src, pcre_compile.c, pcre_study.c, and pcre_tables.c, in
1999 order to be able to cut out the UTF-8 tables in the latter when UTF-8
2000 support is not required. This saves 1.5-2K of code, which is important in
2001 some applications.
2003 Later: more #ifdefs are needed in pcre_ord2utf8.c and pcre_valid_utf8.c
2004 so as not to refer to the tables, even though these functions will never be
2005 called when UTF-8 support is disabled. Otherwise there are problems with a
2006 shared library.
2008 10. Fixed two bugs in the emulated memmove() function in pcre_internal.h:
2010 (a) It was defining its arguments as char * instead of void *.
2012 (b) It was assuming that all moves were upwards in memory; this was true
2013 a long time ago when I wrote it, but is no longer the case.
2015 The emulated memove() is provided for those environments that have neither
2016 memmove() nor bcopy(). I didn't think anyone used it these days, but that
2017 is clearly not the case, as these two bugs were recently reported.
2019 11. The script PrepareRelease is now distributed: it calls 132html, CleanTxt,
2020 and Detrail to create the HTML documentation, the .txt form of the man
2021 pages, and it removes trailing spaces from listed files. It also creates
2022 pcre.h.generic and config.h.generic from pcre.h and config.h. In the latter
2023 case, it wraps all the #defines with #ifndefs. This script should be run
2024 before "make dist".
2026 12. Fixed two fairly obscure bugs concerned with quantified caseless matching
2027 with Unicode property support.
2029 (a) For a maximizing quantifier, if the two different cases of the
2030 character were of different lengths in their UTF-8 codings (there are
2031 some cases like this - I found 11), and the matching function had to
2032 back up over a mixture of the two cases, it incorrectly assumed they
2033 were both the same length.
2035 (b) When PCRE was configured to use the heap rather than the stack for
2036 recursion during matching, it was not correctly preserving the data for
2037 the other case of a UTF-8 character when checking ahead for a match
2038 while processing a minimizing repeat. If the check also involved
2039 matching a wide character, but failed, corruption could cause an
2040 erroneous result when trying to check for a repeat of the original
2041 character.
2043 13. Some tidying changes to the testing mechanism:
2045 (a) The RunTest script now detects the internal link size and whether there
2046 is UTF-8 and UCP support by running ./pcretest -C instead of relying on
2047 values substituted by "configure". (The RunGrepTest script already did
2048 this for UTF-8.) The configure.ac script no longer substitutes the
2049 relevant variables.
2051 (b) The debugging options /B and /D in pcretest show the compiled bytecode
2052 with length and offset values. This means that the output is different
2053 for different internal link sizes. Test 2 is skipped for link sizes
2054 other than 2 because of this, bypassing the problem. Unfortunately,
2055 there was also a test in test 3 (the locale tests) that used /B and
2056 failed for link sizes other than 2. Rather than cut the whole test out,
2057 I have added a new /Z option to pcretest that replaces the length and
2058 offset values with spaces. This is now used to make test 3 independent
2059 of link size. (Test 2 will be tidied up later.)
2061 14. If erroroffset was passed as NULL to pcre_compile, it provoked a
2062 segmentation fault instead of returning the appropriate error message.
2064 15. In multiline mode when the newline sequence was set to "any", the pattern
2065 ^$ would give a match between the \r and \n of a subject such as "A\r\nB".
2066 This doesn't seem right; it now treats the CRLF combination as the line
2067 ending, and so does not match in that case. It's only a pattern such as ^$
2068 that would hit this one: something like ^ABC$ would have failed after \r
2069 and then tried again after \r\n.
2071 16. Changed the comparison command for RunGrepTest from "diff -u" to "diff -ub"
2072 in an attempt to make files that differ only in their line terminators
2073 compare equal. This works on Linux.
2075 17. Under certain error circumstances pcregrep might try to free random memory
2076 as it exited. This is now fixed, thanks to valgrind.
2078 19. In pcretest, if the pattern /(?m)^$/g<any> was matched against the string
2079 "abc\r\n\r\n", it found an unwanted second match after the second \r. This
2080 was because its rules for how to advance for /g after matching an empty
2081 string at the end of a line did not allow for this case. They now check for
2082 it specially.
2084 20. pcretest is supposed to handle patterns and data of any length, by
2085 extending its buffers when necessary. It was getting this wrong when the
2086 buffer for a data line had to be extended.
2088 21. Added PCRE_NEWLINE_ANYCRLF which is like ANY, but matches only CR, LF, or
2089 CRLF as a newline sequence.
2091 22. Code for handling Unicode properties in pcre_dfa_exec() wasn't being cut
2092 out by #ifdef SUPPORT_UCP. This did no harm, as it could never be used, but
2093 I have nevertheless tidied it up.
2095 23. Added some casts to kill warnings from HP-UX ia64 compiler.
2097 24. Added a man page for pcre-config.
2100 Version 7.0 19-Dec-06
2101 ---------------------
2103 1. Fixed a signed/unsigned compiler warning in pcre_compile.c, shown up by
2104 moving to gcc 4.1.1.
2106 2. The -S option for pcretest uses setrlimit(); I had omitted to #include
2107 sys/time.h, which is documented as needed for this function. It doesn't
2108 seem to matter on Linux, but it showed up on some releases of OS X.
2110 3. It seems that there are systems where bytes whose values are greater than
2111 127 match isprint() in the "C" locale. The "C" locale should be the
2112 default when a C program starts up. In most systems, only ASCII printing
2113 characters match isprint(). This difference caused the output from pcretest
2114 to vary, making some of the tests fail. I have changed pcretest so that:
2116 (a) When it is outputting text in the compiled version of a pattern, bytes
2117 other than 32-126 are always shown as hex escapes.
2119 (b) When it is outputting text that is a matched part of a subject string,
2120 it does the same, unless a different locale has been set for the match
2121 (using the /L modifier). In this case, it uses isprint() to decide.
2123 4. Fixed a major bug that caused incorrect computation of the amount of memory
2124 required for a compiled pattern when options that changed within the
2125 pattern affected the logic of the preliminary scan that determines the
2126 length. The relevant options are -x, and -i in UTF-8 mode. The result was
2127 that the computed length was too small. The symptoms of this bug were
2128 either the PCRE error "internal error: code overflow" from pcre_compile(),
2129 or a glibc crash with a message such as "pcretest: free(): invalid next
2130 size (fast)". Examples of patterns that provoked this bug (shown in
2131 pcretest format) are:
2133 /(?-x: )/x
2134 /(?x)(?-x: \s*#\s*)/
2135 /((?i)[\x{c0}])/8
2136 /(?i:[\x{c0}])/8
2138 HOWEVER: Change 17 below makes this fix obsolete as the memory computation
2139 is now done differently.
2141 5. Applied patches from Google to: (a) add a QuoteMeta function to the C++
2142 wrapper classes; (b) implement a new function in the C++ scanner that is
2143 more efficient than the old way of doing things because it avoids levels of
2144 recursion in the regex matching; (c) add a paragraph to the documentation
2145 for the FullMatch() function.
2147 6. The escape sequence \n was being treated as whatever was defined as
2148 "newline". Not only was this contrary to the documentation, which states
2149 that \n is character 10 (hex 0A), but it also went horribly wrong when
2150 "newline" was defined as CRLF. This has been fixed.
2152 7. In pcre_dfa_exec.c the value of an unsigned integer (the variable called c)
2153 was being set to -1 for the "end of line" case (supposedly a value that no
2154 character can have). Though this value is never used (the check for end of
2155 line is "zero bytes in current character"), it caused compiler complaints.
2156 I've changed it to 0xffffffff.
2158 8. In pcre_version.c, the version string was being built by a sequence of
2159 C macros that, in the event of PCRE_PRERELEASE being defined as an empty
2160 string (as it is for production releases) called a macro with an empty
2161 argument. The C standard says the result of this is undefined. The gcc
2162 compiler treats it as an empty string (which was what was wanted) but it is
2163 reported that Visual C gives an error. The source has been hacked around to
2164 avoid this problem.
2166 9. On the advice of a Windows user, included <io.h> and <fcntl.h> in Windows
2167 builds of pcretest, and changed the call to _setmode() to use _O_BINARY
2168 instead of 0x8000. Made all the #ifdefs test both _WIN32 and WIN32 (not all
2169 of them did).
2171 10. Originally, pcretest opened its input and output without "b"; then I was
2172 told that "b" was needed in some environments, so it was added for release
2173 5.0 to both the input and output. (It makes no difference on Unix-like
2174 systems.) Later I was told that it is wrong for the input on Windows. I've
2175 now abstracted the modes into two macros, to make it easier to fiddle with
2176 them, and removed "b" from the input mode under Windows.
2178 11. Added pkgconfig support for the C++ wrapper library, libpcrecpp.
2180 12. Added -help and --help to pcretest as an official way of being reminded
2181 of the options.
2183 13. Removed some redundant semicolons after macro calls in pcrecpparg.h.in
2184 and pcrecpp.cc because they annoy compilers at high warning levels.
2186 14. A bit of tidying/refactoring in pcre_exec.c in the main bumpalong loop.
2188 15. Fixed an occurrence of == in configure.ac that should have been = (shell
2189 scripts are not C programs :-) and which was not noticed because it works
2190 on Linux.
2192 16. pcretest is supposed to handle any length of pattern and data line (as one
2193 line or as a continued sequence of lines) by extending its input buffer if
2194 necessary. This feature was broken for very long pattern lines, leading to
2195 a string of junk being passed to pcre_compile() if the pattern was longer
2196 than about 50K.
2198 17. I have done a major re-factoring of the way pcre_compile() computes the
2199 amount of memory needed for a compiled pattern. Previously, there was code
2200 that made a preliminary scan of the pattern in order to do this. That was
2201 OK when PCRE was new, but as the facilities have expanded, it has become
2202 harder and harder to keep it in step with the real compile phase, and there
2203 have been a number of bugs (see for example, 4 above). I have now found a
2204 cunning way of running the real compile function in a "fake" mode that
2205 enables it to compute how much memory it would need, while actually only
2206 ever using a few hundred bytes of working memory and without too many
2207 tests of the mode. This should make future maintenance and development
2208 easier. A side effect of this work is that the limit of 200 on the nesting
2209 depth of parentheses has been removed (though this was never a serious
2210 limitation, I suspect). However, there is a downside: pcre_compile() now
2211 runs more slowly than before (30% or more, depending on the pattern). I
2212 hope this isn't a big issue. There is no effect on runtime performance.
2214 18. Fixed a minor bug in pcretest: if a pattern line was not terminated by a
2215 newline (only possible for the last line of a file) and it was a
2216 pattern that set a locale (followed by /Lsomething), pcretest crashed.
2218 19. Added additional timing features to pcretest. (1) The -tm option now times
2219 matching only, not compiling. (2) Both -t and -tm can be followed, as a
2220 separate command line item, by a number that specifies the number of
2221 repeats to use when timing. The default is 50000; this gives better
2222 precision, but takes uncomfortably long for very large patterns.
2224 20. Extended pcre_study() to be more clever in cases where a branch of a
2225 subpattern has no definite first character. For example, (a*|b*)[cd] would
2226 previously give no result from pcre_study(). Now it recognizes that the
2227 first character must be a, b, c, or d.
2229 21. There was an incorrect error "recursive call could loop indefinitely" if
2230 a subpattern (or the entire pattern) that was being tested for matching an
2231 empty string contained only one non-empty item after a nested subpattern.
2232 For example, the pattern (?>\x{100}*)\d(?R) provoked this error
2233 incorrectly, because the \d was being skipped in the check.
2235 22. The pcretest program now has a new pattern option /B and a command line
2236 option -b, which is equivalent to adding /B to every pattern. This causes
2237 it to show the compiled bytecode, without the additional information that
2238 -d shows. The effect of -d is now the same as -b with -i (and similarly, /D
2239 is the same as /B/I).
2241 23. A new optimization is now able automatically to treat some sequences such
2242 as a*b as a*+b. More specifically, if something simple (such as a character
2243 or a simple class like \d) has an unlimited quantifier, and is followed by
2244 something that cannot possibly match the quantified thing, the quantifier
2245 is automatically "possessified".
2247 24. A recursive reference to a subpattern whose number was greater than 39
2248 went wrong under certain circumstances in UTF-8 mode. This bug could also
2249 have affected the operation of pcre_study().
2251 25. Realized that a little bit of performance could be had by replacing
2252 (c & 0xc0) == 0xc0 with c >= 0xc0 when processing UTF-8 characters.
2254 26. Timing data from pcretest is now shown to 4 decimal places instead of 3.
2256 27. Possessive quantifiers such as a++ were previously implemented by turning
2257 them into atomic groups such as ($>a+). Now they have their own opcodes,
2258 which improves performance. This includes the automatically created ones
2259 from 23 above.
2261 28. A pattern such as (?=(\w+))\1: which simulates an atomic group using a
2262 lookahead was broken if it was not anchored. PCRE was mistakenly expecting
2263 the first matched character to be a colon. This applied both to named and
2264 numbered groups.
2266 29. The ucpinternal.h header file was missing its idempotency #ifdef.
2268 30. I was sent a "project" file called libpcre.a.dev which I understand makes
2269 building PCRE on Windows easier, so I have included it in the distribution.
2271 31. There is now a check in pcretest against a ridiculously large number being
2272 returned by pcre_exec() or pcre_dfa_exec(). If this happens in a /g or /G
2273 loop, the loop is abandoned.
2275 32. Forward references to subpatterns in conditions such as (?(2)...) where
2276 subpattern 2 is defined later cause pcre_compile() to search forwards in
2277 the pattern for the relevant set of parentheses. This search went wrong
2278 when there were unescaped parentheses in a character class, parentheses
2279 escaped with \Q...\E, or parentheses in a #-comment in /x mode.
2281 33. "Subroutine" calls and backreferences were previously restricted to
2282 referencing subpatterns earlier in the regex. This restriction has now
2283 been removed.
2285 34. Added a number of extra features that are going to be in Perl 5.10. On the
2286 whole, these are just syntactic alternatives for features that PCRE had
2287 previously implemented using the Python syntax or my own invention. The
2288 other formats are all retained for compatibility.
2290 (a) Named groups can now be defined as (?<name>...) or (?'name'...) as well
2291 as (?P<name>...). The new forms, as well as being in Perl 5.10, are
2292 also .NET compatible.
2294 (b) A recursion or subroutine call to a named group can now be defined as
2295 (?&name) as well as (?P>name).
2297 (c) A backreference to a named group can now be defined as \k<name> or
2298 \k'name' as well as (?P=name). The new forms, as well as being in Perl
2299 5.10, are also .NET compatible.
2301 (d) A conditional reference to a named group can now use the syntax
2302 (?(<name>) or (?('name') as well as (?(name).
2304 (e) A "conditional group" of the form (?(DEFINE)...) can be used to define
2305 groups (named and numbered) that are never evaluated inline, but can be
2306 called as "subroutines" from elsewhere. In effect, the DEFINE condition
2307 is always false. There may be only one alternative in such a group.
2309 (f) A test for recursion can be given as (?(R1).. or (?(R&name)... as well
2310 as the simple (?(R). The condition is true only if the most recent
2311 recursion is that of the given number or name. It does not search out
2312 through the entire recursion stack.
2314 (g) The escape \gN or \g{N} has been added, where N is a positive or
2315 negative number, specifying an absolute or relative reference.
2317 35. Tidied to get rid of some further signed/unsigned compiler warnings and
2318 some "unreachable code" warnings.
2320 36. Updated the Unicode property tables to Unicode version 5.0.0. Amongst other
2321 things, this adds five new scripts.
2323 37. Perl ignores orphaned \E escapes completely. PCRE now does the same.
2324 There were also incompatibilities regarding the handling of \Q..\E inside
2325 character classes, for example with patterns like [\Qa\E-\Qz\E] where the
2326 hyphen was adjacent to \Q or \E. I hope I've cleared all this up now.
2328 38. Like Perl, PCRE detects when an indefinitely repeated parenthesized group
2329 matches an empty string, and forcibly breaks the loop. There were bugs in
2330 this code in non-simple cases. For a pattern such as ^(a()*)* matched
2331 against aaaa the result was just "a" rather than "aaaa", for example. Two
2332 separate and independent bugs (that affected different cases) have been
2333 fixed.
2335 39. Refactored the code to abolish the use of different opcodes for small
2336 capturing bracket numbers. This is a tidy that I avoided doing when I
2337 removed the limit on the number of capturing brackets for 3.5 back in 2001.
2338 The new approach is not only tidier, it makes it possible to reduce the
2339 memory needed to fix the previous bug (38).
2341 40. Implemented PCRE_NEWLINE_ANY to recognize any of the Unicode newline
2342 sequences (http://unicode.org/unicode/reports/tr18/) as "newline" when
2343 processing dot, circumflex, or dollar metacharacters, or #-comments in /x
2344 mode.
2346 41. Add \R to match any Unicode newline sequence, as suggested in the Unicode
2347 report.
2349 42. Applied patch, originally from Ari Pollak, modified by Google, to allow
2350 copy construction and assignment in the C++ wrapper.
2352 43. Updated pcregrep to support "--newline=any". In the process, I fixed a
2353 couple of bugs that could have given wrong results in the "--newline=crlf"
2354 case.
2356 44. Added a number of casts and did some reorganization of signed/unsigned int
2357 variables following suggestions from Dair Grant. Also renamed the variable
2358 "this" as "item" because it is a C++ keyword.
2360 45. Arranged for dftables to add
2362 #include "pcre_internal.h"
2364 to pcre_chartables.c because without it, gcc 4.x may remove the array
2365 definition from the final binary if PCRE is built into a static library and
2366 dead code stripping is activated.
2368 46. For an unanchored pattern, if a match attempt fails at the start of a
2369 newline sequence, and the newline setting is CRLF or ANY, and the next two
2370 characters are CRLF, advance by two characters instead of one.
2373 Version 6.7 04-Jul-06
2374 ---------------------
2376 1. In order to handle tests when input lines are enormously long, pcretest has
2377 been re-factored so that it automatically extends its buffers when
2378 necessary. The code is crude, but this _is_ just a test program. The
2379 default size has been increased from 32K to 50K.
2381 2. The code in pcre_study() was using the value of the re argument before
2382 testing it for NULL. (Of course, in any sensible call of the function, it
2383 won't be NULL.)
2385 3. The memmove() emulation function in pcre_internal.h, which is used on
2386 systems that lack both memmove() and bcopy() - that is, hardly ever -
2387 was missing a "static" storage class specifier.
2389 4. When UTF-8 mode was not set, PCRE looped when compiling certain patterns
2390 containing an extended class (one that cannot be represented by a bitmap
2391 because it contains high-valued characters or Unicode property items, e.g.
2392 [\pZ]). Almost always one would set UTF-8 mode when processing such a
2393 pattern, but PCRE should not loop if you do not (it no longer does).
2394 [Detail: two cases were found: (a) a repeated subpattern containing an
2395 extended class; (b) a recursive reference to a subpattern that followed a
2396 previous extended class. It wasn't skipping over the extended class
2397 correctly when UTF-8 mode was not set.]
2399 5. A negated single-character class was not being recognized as fixed-length
2400 in lookbehind assertions such as (?<=[^f]), leading to an incorrect
2401 compile error "lookbehind assertion is not fixed length".
2403 6. The RunPerlTest auxiliary script was showing an unexpected difference
2404 between PCRE and Perl for UTF-8 tests. It turns out that it is hard to
2405 write a Perl script that can interpret lines of an input file either as
2406 byte characters or as UTF-8, which is what "perltest" was being required to
2407 do for the non-UTF-8 and UTF-8 tests, respectively. Essentially what you
2408 can't do is switch easily at run time between having the "use utf8;" pragma
2409 or not. In the end, I fudged it by using the RunPerlTest script to insert
2410 "use utf8;" explicitly for the UTF-8 tests.
2412 7. In multiline (/m) mode, PCRE was matching ^ after a terminating newline at
2413 the end of the subject string, contrary to the documentation and to what
2414 Perl does. This was true of both matching functions. Now it matches only at
2415 the start of the subject and immediately after *internal* newlines.
2417 8. A call of pcre_fullinfo() from pcretest to get the option bits was passing
2418 a pointer to an int instead of a pointer to an unsigned long int. This
2419 caused problems on 64-bit systems.
2421 9. Applied a patch from the folks at Google to pcrecpp.cc, to fix "another
2422 instance of the 'standard' template library not being so standard".
2424 10. There was no check on the number of named subpatterns nor the maximum
2425 length of a subpattern name. The product of these values is used to compute
2426 the size of the memory block for a compiled pattern. By supplying a very
2427 long subpattern name and a large number of named subpatterns, the size
2428 computation could be caused to overflow. This is now prevented by limiting
2429 the length of names to 32 characters, and the number of named subpatterns
2430 to 10,000.
2432 11. Subpatterns that are repeated with specific counts have to be replicated in
2433 the compiled pattern. The size of memory for this was computed from the
2434 length of the subpattern and the repeat count. The latter is limited to
2435 65535, but there was no limit on the former, meaning that integer overflow
2436 could in principle occur. The compiled length of a repeated subpattern is
2437 now limited to 30,000 bytes in order to prevent this.
2439 12. Added the optional facility to have named substrings with the same name.
2441 13. Added the ability to use a named substring as a condition, using the
2442 Python syntax: (?(name)yes|no). This overloads (?(R)... and names that
2443 are numbers (not recommended). Forward references are permitted.
2445 14. Added forward references in named backreferences (if you see what I mean).
2447 15. In UTF-8 mode, with the PCRE_DOTALL option set, a quantified dot in the
2448 pattern could run off the end of the subject. For example, the pattern
2449 "(?s)(.{1,5})"8 did this with the subject "ab".
2451 16. If PCRE_DOTALL or PCRE_MULTILINE were set, pcre_dfa_exec() behaved as if
2452 PCRE_CASELESS was set when matching characters that were quantified with ?
2453 or *.
2455 17. A character class other than a single negated character that had a minimum
2456 but no maximum quantifier - for example [ab]{6,} - was not handled
2457 correctly by pce_dfa_exec(). It would match only one character.
2459 18. A valid (though odd) pattern that looked like a POSIX character
2460 class but used an invalid character after [ (for example [[,abc,]]) caused
2461 pcre_compile() to give the error "Failed: internal error: code overflow" or
2462 in some cases to crash with a glibc free() error. This could even happen if
2463 the pattern terminated after [[ but there just happened to be a sequence of
2464 letters, a binary zero, and a closing ] in the memory that followed.
2466 19. Perl's treatment of octal escapes in the range \400 to \777 has changed
2467 over the years. Originally (before any Unicode support), just the bottom 8
2468 bits were taken. Thus, for example, \500 really meant \100. Nowadays the
2469 output from "man perlunicode" includes this:
2471 The regular expression compiler produces polymorphic opcodes. That
2472 is, the pattern adapts to the data and automatically switches to
2473 the Unicode character scheme when presented with Unicode data--or
2474 instead uses a traditional byte scheme when presented with byte
2475 data.
2477 Sadly, a wide octal escape does not cause a switch, and in a string with
2478 no other multibyte characters, these octal escapes are treated as before.
2479 Thus, in Perl, the pattern /\500/ actually matches \100 but the pattern
2480 /\500|\x{1ff}/ matches \500 or \777 because the whole thing is treated as a
2481 Unicode string.
2483 I have not perpetrated such confusion in PCRE. Up till now, it took just
2484 the bottom 8 bits, as in old Perl. I have now made octal escapes with
2485 values greater than \377 illegal in non-UTF-8 mode. In UTF-8 mode they
2486 translate to the appropriate multibyte character.
2488 29. Applied some refactoring to reduce the number of warnings from Microsoft
2489 and Borland compilers. This has included removing the fudge introduced
2490 seven years ago for the OS/2 compiler (see 2.02/2 below) because it caused
2491 a warning about an unused variable.
2493 21. PCRE has not included VT (character 0x0b) in the set of whitespace
2494 characters since release 4.0, because Perl (from release 5.004) does not.
2495 [Or at least, is documented not to: some releases seem to be in conflict
2496 with the documentation.] However, when a pattern was studied with
2497 pcre_study() and all its branches started with \s, PCRE still included VT
2498 as a possible starting character. Of course, this did no harm; it just
2499 caused an unnecessary match attempt.
2501 22. Removed a now-redundant internal flag bit that recorded the fact that case
2502 dependency changed within the pattern. This was once needed for "required
2503 byte" processing, but is no longer used. This recovers a now-scarce options
2504 bit. Also moved the least significant internal flag bit to the most-
2505 significant bit of the word, which was not previously used (hangover from
2506 the days when it was an int rather than a uint) to free up another bit for
2507 the future.
2509 23. Added support for CRLF line endings as well as CR and LF. As well as the
2510 default being selectable at build time, it can now be changed at runtime
2511 via the PCRE_NEWLINE_xxx flags. There are now options for pcregrep to
2512 specify that it is scanning data with non-default line endings.
2514 24. Changed the definition of CXXLINK to make it agree with the definition of
2515 LINK in the Makefile, by replacing LDFLAGS to CXXFLAGS.
2517 25. Applied Ian Taylor's patches to avoid using another stack frame for tail
2518 recursions. This makes a big different to stack usage for some patterns.
2520 26. If a subpattern containing a named recursion or subroutine reference such
2521 as (?P>B) was quantified, for example (xxx(?P>B)){3}, the calculation of
2522 the space required for the compiled pattern went wrong and gave too small a
2523 value. Depending on the environment, this could lead to "Failed: internal
2524 error: code overflow at offset 49" or "glibc detected double free or
2525 corruption" errors.
2527 27. Applied patches from Google (a) to support the new newline modes and (b) to
2528 advance over multibyte UTF-8 characters in GlobalReplace.
2530 28. Change free() to pcre_free() in pcredemo.c. Apparently this makes a
2531 difference for some implementation of PCRE in some Windows version.
2533 29. Added some extra testing facilities to pcretest:
2535 \q<number> in a data line sets the "match limit" value
2536 \Q<number> in a data line sets the "match recursion limt" value
2537 -S <number> sets the stack size, where <number> is in megabytes
2539 The -S option isn't available for Windows.
2542 Version 6.6 06-Feb-06
2543 ---------------------
2545 1. Change 16(a) for 6.5 broke things, because PCRE_DATA_SCOPE was not defined
2546 in pcreposix.h. I have copied the definition from pcre.h.
2548 2. Change 25 for 6.5 broke compilation in a build directory out-of-tree
2549 because pcre.h is no longer a built file.
2551 3. Added Jeff Friedl's additional debugging patches to pcregrep. These are
2552 not normally included in the compiled code.
2555 Version 6.5 01-Feb-06
2556 ---------------------
2558 1. When using the partial match feature with pcre_dfa_exec(), it was not
2559 anchoring the second and subsequent partial matches at the new starting
2560 point. This could lead to incorrect results. For example, with the pattern
2561 /1234/, partially matching against "123" and then "a4" gave a match.
2563 2. Changes to pcregrep:
2565 (a) All non-match returns from pcre_exec() were being treated as failures
2566 to match the line. Now, unless the error is PCRE_ERROR_NOMATCH, an
2567 error message is output. Some extra information is given for the
2569 probably the only errors that are likely to be caused by users (by
2570 specifying a regex that has nested indefinite repeats, for instance).
2571 If there are more than 20 of these errors, pcregrep is abandoned.
2573 (b) A binary zero was treated as data while matching, but terminated the
2574 output line if it was written out. This has been fixed: binary zeroes
2575 are now no different to any other data bytes.
2577 (c) Whichever of the LC_ALL or LC_CTYPE environment variables is set is
2578 used to set a locale for matching. The --locale=xxxx long option has
2579 been added (no short equivalent) to specify a locale explicitly on the
2580 pcregrep command, overriding the environment variables.
2582 (d) When -B was used with -n, some line numbers in the output were one less
2583 than they should have been.
2585 (e) Added the -o (--only-matching) option.
2587 (f) If -A or -C was used with -c (count only), some lines of context were
2588 accidentally printed for the final match.
2590 (g) Added the -H (--with-filename) option.
2592 (h) The combination of options -rh failed to suppress file names for files
2593 that were found from directory arguments.
2595 (i) Added the -D (--devices) and -d (--directories) options.
2597 (j) Added the -F (--fixed-strings) option.
2599 (k) Allow "-" to be used as a file name for -f as well as for a data file.
2601 (l) Added the --colo(u)r option.
2603 (m) Added Jeffrey Friedl's -S testing option, but within #ifdefs so that it
2604 is not present by default.
2606 3. A nasty bug was discovered in the handling of recursive patterns, that is,
2607 items such as (?R) or (?1), when the recursion could match a number of
2608 alternatives. If it matched one of the alternatives, but subsequently,
2609 outside the recursion, there was a failure, the code tried to back up into
2610 the recursion. However, because of the way PCRE is implemented, this is not
2611 possible, and the result was an incorrect result from the match.
2613 In order to prevent this happening, the specification of recursion has
2614 been changed so that all such subpatterns are automatically treated as
2615 atomic groups. Thus, for example, (?R) is treated as if it were (?>(?R)).
2617 4. I had overlooked the fact that, in some locales, there are characters for
2618 which isalpha() is true but neither isupper() nor islower() are true. In
2619 the fr_FR locale, for instance, the \xAA and \xBA characters (ordmasculine
2620 and ordfeminine) are like this. This affected the treatment of \w and \W
2621 when they appeared in character classes, but not when they appeared outside
2622 a character class. The bit map for "word" characters is now created
2623 separately from the results of isalnum() instead of just taking it from the
2624 upper, lower, and digit maps. (Plus the underscore character, of course.)
2626 5. The above bug also affected the handling of POSIX character classes such as
2627 [[:alpha:]] and [[:alnum:]]. These do not have their own bit maps in PCRE's
2628 permanent tables. Instead, the bit maps for such a class were previously
2629 created as the appropriate unions of the upper, lower, and digit bitmaps.
2630 Now they are created by subtraction from the [[:word:]] class, which has
2631 its own bitmap.
2633 6. The [[:blank:]] character class matches horizontal, but not vertical space.
2634 It is created by subtracting the vertical space characters (\x09, \x0a,
2635 \x0b, \x0c) from the [[:space:]] bitmap. Previously, however, the
2636 subtraction was done in the overall bitmap for a character class, meaning
2637 that a class such as [\x0c[:blank:]] was incorrect because \x0c would not
2638 be recognized. This bug has been fixed.
2640 7. Patches from the folks at Google:
2642 (a) pcrecpp.cc: "to handle a corner case that may or may not happen in
2643 real life, but is still worth protecting against".
2645 (b) pcrecpp.cc: "corrects a bug when negative radixes are used with
2646 regular expressions".
2648 (c) pcre_scanner.cc: avoid use of std::count() because not all systems
2649 have it.
2651 (d) Split off pcrecpparg.h from pcrecpp.h and had the former built by
2652 "configure" and the latter not, in order to fix a problem somebody had
2653 with compiling the Arg class on HP-UX.
2655 (e) Improve the error-handling of the C++ wrapper a little bit.
2657 (f) New tests for checking recursion limiting.
2659 8. The pcre_memmove() function, which is used only if the environment does not
2660 have a standard memmove() function (and is therefore rarely compiled),
2661 contained two bugs: (a) use of int instead of size_t, and (b) it was not
2662 returning a result (though PCRE never actually uses the result).
2664 9. In the POSIX regexec() interface, if nmatch is specified as a ridiculously
2665 large number - greater than INT_MAX/(3*sizeof(int)) - REG_ESPACE is
2666 returned instead of calling malloc() with an overflowing number that would
2667 most likely cause subsequent chaos.
2669 10. The debugging option of pcretest was not showing the NO_AUTO_CAPTURE flag.
2671 11. The POSIX flag REG_NOSUB is now supported. When a pattern that was compiled
2672 with this option is matched, the nmatch and pmatch options of regexec() are
2673 ignored.
2675 12. Added REG_UTF8 to the POSIX interface. This is not defined by POSIX, but is
2676 provided in case anyone wants to the the POSIX interface with UTF-8
2677 strings.
2679 13. Added CXXLDFLAGS to the Makefile parameters to provide settings only on the
2680 C++ linking (needed for some HP-UX environments).
2682 14. Avoid compiler warnings in get_ucpname() when compiled without UCP support
2683 (unused parameter) and in the pcre_printint() function (omitted "default"
2684 switch label when the default is to do nothing).
2686 15. Added some code to make it possible, when PCRE is compiled as a C++
2687 library, to replace subject pointers for pcre_exec() with a smart pointer
2688 class, thus making it possible to process discontinuous strings.
2690 16. The two macros PCRE_EXPORT and PCRE_DATA_SCOPE are confusing, and perform
2691 much the same function. They were added by different people who were trying
2692 to make PCRE easy to compile on non-Unix systems. It has been suggested
2693 that PCRE_EXPORT be abolished now that there is more automatic apparatus
2694 for compiling on Windows systems. I have therefore replaced it with
2695 PCRE_DATA_SCOPE. This is set automatically for Windows; if not set it
2696 defaults to "extern" for C or "extern C" for C++, which works fine on
2697 Unix-like systems. It is now possible to override the value of PCRE_DATA_
2698 SCOPE with something explicit in config.h. In addition:
2700 (a) pcreposix.h still had just "extern" instead of either of these macros;
2701 I have replaced it with PCRE_DATA_SCOPE.
2703 (b) Functions such as _pcre_xclass(), which are internal to the library,
2704 but external in the C sense, all had PCRE_EXPORT in their definitions.
2705 This is apparently wrong for the Windows case, so I have removed it.
2706 (It makes no difference on Unix-like systems.)
2708 17. Added a new limit, MATCH_LIMIT_RECURSION, which limits the depth of nesting
2709 of recursive calls to match(). This is different to MATCH_LIMIT because
2710 that limits the total number of calls to match(), not all of which increase
2711 the depth of recursion. Limiting the recursion depth limits the amount of
2712 stack (or heap if NO_RECURSE is set) that is used. The default can be set
2713 when PCRE is compiled, and changed at run time. A patch from Google adds
2714 this functionality to the C++ interface.
2716 18. Changes to the handling of Unicode character properties:
2718 (a) Updated the table to Unicode 4.1.0.
2720 (b) Recognize characters that are not in the table as "Cn" (undefined).
2722 (c) I revised the way the table is implemented to a much improved format
2723 which includes recognition of ranges. It now supports the ranges that
2724 are defined in UnicodeData.txt, and it also amalgamates other
2725 characters into ranges. This has reduced the number of entries in the
2726 table from around 16,000 to around 3,000, thus reducing its size
2727 considerably. I realized I did not need to use a tree structure after
2728 all - a binary chop search is just as efficient. Having reduced the
2729 number of entries, I extended their size from 6 bytes to 8 bytes to
2730 allow for more data.
2732 (d) Added support for Unicode script names via properties such as \p{Han}.
2734 19. In UTF-8 mode, a backslash followed by a non-Ascii character was not
2735 matching that character.
2737 20. When matching a repeated Unicode property with a minimum greater than zero,
2738 (for example \pL{2,}), PCRE could look past the end of the subject if it
2739 reached it while seeking the minimum number of characters. This could
2740 happen only if some of the characters were more than one byte long, because
2741 there is a check for at least the minimum number of bytes.
2743 21. Refactored the implementation of \p and \P so as to be more general, to
2744 allow for more different types of property in future. This has changed the
2745 compiled form incompatibly. Anybody with saved compiled patterns that use
2746 \p or \P will have to recompile them.
2748 22. Added "Any" and "L&" to the supported property types.
2750 23. Recognize \x{...} as a code point specifier, even when not in UTF-8 mode,
2751 but give a compile time error if the value is greater than 0xff.
2753 24. The man pages for pcrepartial, pcreprecompile, and pcre_compile2 were
2754 accidentally not being installed or uninstalled.
2756 25. The pcre.h file was built from pcre.h.in, but the only changes that were
2757 made were to insert the current release number. This seemed silly, because
2758 it made things harder for people building PCRE on systems that don't run
2759 "configure". I have turned pcre.h into a distributed file, no longer built
2760 by "configure", with the version identification directly included. There is
2761 no longer a pcre.h.in file.
2763 However, this change necessitated a change to the pcre-config script as
2764 well. It is built from pcre-config.in, and one of the substitutions was the
2765 release number. I have updated configure.ac so that ./configure now finds
2766 the release number by grepping pcre.h.
2768 26. Added the ability to run the tests under valgrind.
2771 Version 6.4 05-Sep-05
2772 ---------------------
2774 1. Change 6.0/10/(l) to pcregrep introduced a bug that caused separator lines
2775 "--" to be printed when multiple files were scanned, even when none of the
2776 -A, -B, or -C options were used. This is not compatible with Gnu grep, so I
2777 consider it to be a bug, and have restored the previous behaviour.
2779 2. A couple of code tidies to get rid of compiler warnings.
2781 3. The pcretest program used to cheat by referring to symbols in the library
2782 whose names begin with _pcre_. These are internal symbols that are not
2783 really supposed to be visible externally, and in some environments it is
2784 possible to suppress them. The cheating is now confined to including
2785 certain files from the library's source, which is a bit cleaner.
2787 4. Renamed pcre.in as pcre.h.in to go with pcrecpp.h.in; it also makes the
2788 file's purpose clearer.
2790 5. Reorganized pcre_ucp_findchar().
2793 Version 6.3 15-Aug-05
2794 ---------------------
2796 1. The file libpcre.pc.in did not have general read permission in the tarball.
2798 2. There were some problems when building without C++ support:
2800 (a) If C++ support was not built, "make install" and "make test" still
2801 tried to test it.
2803 (b) There were problems when the value of CXX was explicitly set. Some
2804 changes have been made to try to fix these, and ...
2806 (c) --disable-cpp can now be used to explicitly disable C++ support.
2808 (d) The use of @CPP_OBJ@ directly caused a blank line preceded by a
2809 backslash in a target when C++ was disabled. This confuses some
2810 versions of "make", apparently. Using an intermediate variable solves
2811 this. (Same for CPP_LOBJ.)
2813 3. $(LINK_FOR_BUILD) now includes $(CFLAGS_FOR_BUILD) and $(LINK)
2814 (non-Windows) now includes $(CFLAGS) because these flags are sometimes
2815 necessary on certain architectures.
2817 4. Added a setting of -export-symbols-regex to the link command to remove
2818 those symbols that are exported in the C sense, but actually are local
2819 within the library, and not documented. Their names all begin with
2820 "_pcre_". This is not a perfect job, because (a) we have to except some
2821 symbols that pcretest ("illegally") uses, and (b) the facility isn't always
2822 available (and never for static libraries). I have made a note to try to
2823 find a way round (a) in the future.
2826 Version 6.2 01-Aug-05
2827 ---------------------
2829 1. There was no test for integer overflow of quantifier values. A construction
2830 such as {1111111111111111} would give undefined results. What is worse, if
2831 a minimum quantifier for a parenthesized subpattern overflowed and became
2832 negative, the calculation of the memory size went wrong. This could have
2833 led to memory overwriting.
2835 2. Building PCRE using VPATH was broken. Hopefully it is now fixed.
2837 3. Added "b" to the 2nd argument of fopen() in dftables.c, for non-Unix-like
2838 operating environments where this matters.
2840 4. Applied Giuseppe Maxia's patch to add additional features for controlling
2841 PCRE options from within the C++ wrapper.
2843 5. Named capturing subpatterns were not being correctly counted when a pattern
2844 was compiled. This caused two problems: (a) If there were more than 100
2845 such subpatterns, the calculation of the memory needed for the whole
2846 compiled pattern went wrong, leading to an overflow error. (b) Numerical
2847 back references of the form \12, where the number was greater than 9, were
2848 not recognized as back references, even though there were sufficient
2849 previous subpatterns.
2851 6. Two minor patches to pcrecpp.cc in order to allow it to compile on older
2852 versions of gcc, e.g. 2.95.4.
2855 Version 6.1 21-Jun-05
2856 ---------------------
2858 1. There was one reference to the variable "posix" in pcretest.c that was not
2859 surrounded by "#if !defined NOPOSIX".
2861 2. Make it possible to compile pcretest without DFA support, UTF8 support, or
2862 the cross-check on the old pcre_info() function, for the benefit of the
2863 cut-down version of PCRE that is currently imported into Exim.
2865 3. A (silly) pattern starting with (?i)(?-i) caused an internal space
2866 allocation error. I've done the easy fix, which wastes 2 bytes for sensible
2867 patterns that start (?i) but I don't think that matters. The use of (?i) is
2868 just an example; this all applies to the other options as well.
2870 4. Since libtool seems to echo the compile commands it is issuing, the output
2871 from "make" can be reduced a bit by putting "@" in front of each libtool
2872 compile command.
2874 5. Patch from the folks at Google for configure.in to be a bit more thorough
2875 in checking for a suitable C++ installation before trying to compile the
2876 C++ stuff. This should fix a reported problem when a compiler was present,
2877 but no suitable headers.
2879 6. The man pages all had just "PCRE" as their title. I have changed them to
2880 be the relevant file name. I have also arranged that these names are
2881 retained in the file doc/pcre.txt, which is a concatenation in text format
2882 of all the man pages except the little individual ones for each function.
2884 7. The NON-UNIX-USE file had not been updated for the different set of source
2885 files that come with release 6. I also added a few comments about the C++
2886 wrapper.
2889 Version 6.0 07-Jun-05
2890 ---------------------
2892 1. Some minor internal re-organization to help with my DFA experiments.
2894 2. Some missing #ifdef SUPPORT_UCP conditionals in pcretest and printint that
2895 didn't matter for the library itself when fully configured, but did matter
2896 when compiling without UCP support, or within Exim, where the ucp files are
2897 not imported.
2899 3. Refactoring of the library code to split up the various functions into
2900 different source modules. The addition of the new DFA matching code (see
2901 below) to a single monolithic source would have made it really too
2902 unwieldy, quite apart from causing all the code to be include in a
2903 statically linked application, when only some functions are used. This is
2904 relevant even without the DFA addition now that patterns can be compiled in
2905 one application and matched in another.
2907 The downside of splitting up is that there have to be some external
2908 functions and data tables that are used internally in different modules of
2909 the library but which are not part of the API. These have all had their
2910 names changed to start with "_pcre_" so that they are unlikely to clash
2911 with other external names.
2913 4. Added an alternate matching function, pcre_dfa_exec(), which matches using
2914 a different (DFA) algorithm. Although it is slower than the original
2915 function, it does have some advantages for certain types of matching
2916 problem.
2918 5. Upgrades to pcretest in order to test the features of pcre_dfa_exec(),
2919 including restarting after a partial match.
2921 6. A patch for pcregrep that defines INVALID_FILE_ATTRIBUTES if it is not
2922 defined when compiling for Windows was sent to me. I have put it into the
2923 code, though I have no means of testing or verifying it.
2925 7. Added the pcre_refcount() auxiliary function.
2927 8. Added the PCRE_FIRSTLINE option. This constrains an unanchored pattern to
2928 match before or at the first newline in the subject string. In pcretest,
2929 the /f option on a pattern can be used to set this.
2931 9. A repeated \w when used in UTF-8 mode with characters greater than 256
2932 would behave wrongly. This has been present in PCRE since release 4.0.
2934 10. A number of changes to the pcregrep command:
2936 (a) Refactored how -x works; insert ^(...)$ instead of setting
2937 PCRE_ANCHORED and checking the length, in preparation for adding
2938 something similar for -w.
2940 (b) Added the -w (match as a word) option.
2942 (c) Refactored the way lines are read and buffered so as to have more
2943 than one at a time available.
2945 (d) Implemented a pcregrep test script.
2947 (e) Added the -M (multiline match) option. This allows patterns to match
2948 over several lines of the subject. The buffering ensures that at least
2949 8K, or the rest of the document (whichever is the shorter) is available
2950 for matching (and similarly the previous 8K for lookbehind assertions).
2952 (f) Changed the --help output so that it now says
2954 -w, --word-regex(p)
2956 instead of two lines, one with "regex" and the other with "regexp"
2957 because that confused at least one person since the short forms are the
2958 same. (This required a bit of code, as the output is generated
2959 automatically from a table. It wasn't just a text change.)
2961 (g) -- can be used to terminate pcregrep options if the next thing isn't an
2962 option but starts with a hyphen. Could be a pattern or a path name
2963 starting with a hyphen, for instance.
2965 (h) "-" can be given as a file name to represent stdin.
2967 (i) When file names are being printed, "(standard input)" is used for
2968 the standard input, for compatibility with GNU grep. Previously
2969 "<stdin>" was used.
2971 (j) The option --label=xxx can be used to supply a name to be used for
2972 stdin when file names are being printed. There is no short form.
2974 (k) Re-factored the options decoding logic because we are going to add
2975 two more options that take data. Such options can now be given in four
2976 different ways, e.g. "-fname", "-f name", "--file=name", "--file name".
2978 (l) Added the -A, -B, and -C options for requesting that lines of context
2979 around matches be printed.
2981 (m) Added the -L option to print the names of files that do not contain
2982 any matching lines, that is, the complement of -l.
2984 (n) The return code is 2 if any file cannot be opened, but pcregrep does
2985 continue to scan other files.
2987 (o) The -s option was incorrectly implemented. For compatibility with other
2988 greps, it now suppresses the error message for a non-existent or non-
2989 accessible file (but not the return code). There is a new option called
2990 -q that suppresses the output of matching lines, which was what -s was
2991 previously doing.
2993 (p) Added --include and --exclude options to specify files for inclusion
2994 and exclusion when recursing.
2996 11. The Makefile was not using the Autoconf-supported LDFLAGS macro properly.
2997 Hopefully, it now does.
2999 12. Missing cast in pcre_study().
3001 13. Added an "uninstall" target to the makefile.
3003 14. Replaced "extern" in the function prototypes in Makefile.in with
3004 "PCRE_DATA_SCOPE", which defaults to 'extern' or 'extern "C"' in the Unix
3005 world, but is set differently for Windows.
3007 15. Added a second compiling function called pcre_compile2(). The only
3008 difference is that it has an extra argument, which is a pointer to an
3009 integer error code. When there is a compile-time failure, this is set
3010 non-zero, in addition to the error test pointer being set to point to an
3011 error message. The new argument may be NULL if no error number is required
3012 (but then you may as well call pcre_compile(), which is now just a
3013 wrapper). This facility is provided because some applications need a
3014 numeric error indication, but it has also enabled me to tidy up the way
3015 compile-time errors are handled in the POSIX wrapper.
3017 16. Added VPATH=.libs to the makefile; this should help when building with one
3018 prefix path and installing with another. (Or so I'm told by someone who
3019 knows more about this stuff than I do.)
3021 17. Added a new option, REG_DOTALL, to the POSIX function regcomp(). This
3022 passes PCRE_DOTALL to the pcre_compile() function, making the "." character
3023 match everything, including newlines. This is not POSIX-compatible, but
3024 somebody wanted the feature. From pcretest it can be activated by using
3025 both the P and the s flags.
3027 18. AC_PROG_LIBTOOL appeared twice in Makefile.in. Removed one.
3029 19. libpcre.pc was being incorrectly installed as executable.
3031 20. A couple of places in pcretest check for end-of-line by looking for '\n';
3032 it now also looks for '\r' so that it will work unmodified on Windows.
3034 21. Added Google's contributed C++ wrapper to the distribution.
3036 22. Added some untidy missing memory free() calls in pcretest, to keep
3037 Electric Fence happy when testing.
3041 Version 5.0 13-Sep-04
3042 ---------------------
3044 1. Internal change: literal characters are no longer packed up into items
3045 containing multiple characters in a single byte-string. Each character
3046 is now matched using a separate opcode. However, there may be more than one
3047 byte in the character in UTF-8 mode.
3049 2. The pcre_callout_block structure has two new fields: pattern_position and
3050 next_item_length. These contain the offset in the pattern to the next match
3051 item, and its length, respectively.
3053 3. The PCRE_AUTO_CALLOUT option for pcre_compile() requests the automatic
3054 insertion of callouts before each pattern item. Added the /C option to
3055 pcretest to make use of this.
3057 4. On the advice of a Windows user, the lines
3059 #if defined(_WIN32) || defined(WIN32)
3060 _setmode( _fileno( stdout ), 0x8000 );
3061 #endif /* defined(_WIN32) || defined(WIN32) */
3063 have been added to the source of pcretest. This apparently does useful
3064 magic in relation to line terminators.
3066 5. Changed "r" and "w" in the calls to fopen() in pcretest to "rb" and "wb"
3067 for the benefit of those environments where the "b" makes a difference.
3069 6. The icc compiler has the same options as gcc, but "configure" doesn't seem
3070 to know about it. I have put a hack into configure.in that adds in code
3071 to set GCC=yes if CC=icc. This seems to end up at a point in the
3072 generated configure script that is early enough to affect the setting of
3073 compiler options, which is what is needed, but I have no means of testing
3074 whether it really works. (The user who reported this had patched the
3075 generated configure script, which of course I cannot do.)
3077 LATER: After change 22 below (new libtool files), the configure script
3078 seems to know about icc (and also ecc). Therefore, I have commented out
3079 this hack in configure.in.
3081 7. Added support for pkg-config (2 patches were sent in).
3083 8. Negated POSIX character classes that used a combination of internal tables
3084 were completely broken. These were [[:^alpha:]], [[:^alnum:]], and
3085 [[:^ascii]]. Typically, they would match almost any characters. The other
3086 POSIX classes were not broken in this way.
3088 9. Matching the pattern "\b.*?" against "ab cd", starting at offset 1, failed
3089 to find the match, as PCRE was deluded into thinking that the match had to
3090 start at the start point or following a newline. The same bug applied to
3091 patterns with negative forward assertions or any backward assertions
3092 preceding ".*" at the start, unless the pattern required a fixed first
3093 character. This was a failing pattern: "(?!.bcd).*". The bug is now fixed.
3095 10. In UTF-8 mode, when moving forwards in the subject after a failed match
3096 starting at the last subject character, bytes beyond the end of the subject
3097 string were read.
3099 11. Renamed the variable "class" as "classbits" to make life easier for C++
3100 users. (Previously there was a macro definition, but it apparently wasn't
3101 enough.)
3103 12. Added the new field "tables" to the extra data so that tables can be passed
3104 in at exec time, or the internal tables can be re-selected. This allows
3105 a compiled regex to be saved and re-used at a later time by a different
3106 program that might have everything at different addresses.
3108 13. Modified the pcre-config script so that, when run on Solaris, it shows a
3109 -R library as well as a -L library.
3111 14. The debugging options of pcretest (-d on the command line or D on a
3112 pattern) showed incorrect output for anything following an extended class
3113 that contained multibyte characters and which was followed by a quantifier.
3115 15. Added optional support for general category Unicode character properties
3116 via the \p, \P, and \X escapes. Unicode property support implies UTF-8
3117 support. It adds about 90K to the size of the library. The meanings of the
3118 inbuilt class escapes such as \d and \s have NOT been changed.
3120 16. Updated pcredemo.c to include calls to free() to release the memory for the
3121 compiled pattern.
3123 17. The generated file chartables.c was being created in the source directory
3124 instead of in the building directory. This caused the build to fail if the
3125 source directory was different from the building directory, and was
3126 read-only.
3128 18. Added some sample Win commands from Mark Tetrode into the NON-UNIX-USE
3129 file. No doubt somebody will tell me if they don't make sense... Also added
3130 Dan Mooney's comments about building on OpenVMS.
3132 19. Added support for partial matching via the PCRE_PARTIAL option for
3133 pcre_exec() and the \P data escape in pcretest.
3135 20. Extended pcretest with 3 new pattern features:
3137 (i) A pattern option of the form ">rest-of-line" causes pcretest to
3138 write the compiled pattern to the file whose name is "rest-of-line".
3139 This is a straight binary dump of the data, with the saved pointer to
3140 the character tables forced to be NULL. The study data, if any, is
3141 written too. After writing, pcretest reads a new pattern.
3143 (ii) If, instead of a pattern, "<rest-of-line" is given, pcretest reads a
3144 compiled pattern from the given file. There must not be any
3145 occurrences of "<" in the file name (pretty unlikely); if there are,
3146 pcretest will instead treat the initial "<" as a pattern delimiter.
3147 After reading in the pattern, pcretest goes on to read data lines as
3148 usual.
3150 (iii) The F pattern option causes pcretest to flip the bytes in the 32-bit
3151 and 16-bit fields in a compiled pattern, to simulate a pattern that
3152 was compiled on a host of opposite endianness.
3154 21. The pcre-exec() function can now cope with patterns that were compiled on
3155 hosts of opposite endianness, with this restriction:
3157 As for any compiled expression that is saved and used later, the tables
3158 pointer field cannot be preserved; the extra_data field in the arguments
3159 to pcre_exec() should be used to pass in a tables address if a value
3160 other than the default internal tables were used at compile time.
3162 22. Calling pcre_exec() with a negative value of the "ovecsize" parameter is
3163 now diagnosed as an error. Previously, most of the time, a negative number
3164 would have been treated as zero, but if in addition "ovector" was passed as
3165 NULL, a crash could occur.
3167 23. Updated the files ltmain.sh, config.sub, config.guess, and aclocal.m4 with
3168 new versions from the libtool 1.5 distribution (the last one is a copy of
3169 a file called libtool.m4). This seems to have fixed the need to patch
3170 "configure" to support Darwin 1.3 (which I used to do). However, I still
3171 had to patch ltmain.sh to ensure that ${SED} is set (it isn't on my
3172 workstation).
3174 24. Changed the PCRE licence to be the more standard "BSD" licence.
3177 Version 4.5 01-Dec-03
3178 ---------------------
3180 1. There has been some re-arrangement of the code for the match() function so
3181 that it can be compiled in a version that does not call itself recursively.
3182 Instead, it keeps those local variables that need separate instances for
3183 each "recursion" in a frame on the heap, and gets/frees frames whenever it
3184 needs to "recurse". Keeping track of where control must go is done by means
3185 of setjmp/longjmp. The whole thing is implemented by a set of macros that
3186 hide most of the details from the main code, and operates only if
3187 NO_RECURSE is defined while compiling pcre.c. If PCRE is built using the
3188 "configure" mechanism, "--disable-stack-for-recursion" turns on this way of
3189 operating.
3191 To make it easier for callers to provide specially tailored get/free
3192 functions for this usage, two new functions, pcre_stack_malloc, and
3193 pcre_stack_free, are used. They are always called in strict stacking order,
3194 and the size of block requested is always the same.
3196 The PCRE_CONFIG_STACKRECURSE info parameter can be used to find out whether
3197 PCRE has been compiled to use the stack or the heap for recursion. The
3198 -C option of pcretest uses this to show which version is compiled.
3200 A new data escape \S, is added to pcretest; it causes the amounts of store
3201 obtained and freed by both kinds of malloc/free at match time to be added
3202 to the output.
3204 2. Changed the locale test to use "fr_FR" instead of "fr" because that's
3205 what's available on my current Linux desktop machine.
3207 3. When matching a UTF-8 string, the test for a valid string at the start has
3208 been extended. If start_offset is not zero, PCRE now checks that it points
3209 to a byte that is the start of a UTF-8 character. If not, it returns
3210 PCRE_ERROR_BADUTF8_OFFSET (-11). Note: the whole string is still checked;
3211 this is necessary because there may be backward assertions in the pattern.
3212 When matching the same subject several times, it may save resources to use
3213 PCRE_NO_UTF8_CHECK on all but the first call if the string is long.
3215 4. The code for checking the validity of UTF-8 strings has been tightened so
3216 that it rejects (a) strings containing 0xfe or 0xff bytes and (b) strings
3217 containing "overlong sequences".
3219 5. Fixed a bug (appearing twice) that I could not find any way of exploiting!
3220 I had written "if ((digitab[*p++] && chtab_digit) == 0)" where the "&&"
3221 should have been "&", but it just so happened that all the cases this let
3222 through by mistake were picked up later in the function.
3224 6. I had used a variable called "isblank" - this is a C99 function, causing
3225 some compilers to warn. To avoid this, I renamed it (as "blankclass").
3227 7. Cosmetic: (a) only output another newline at the end of pcretest if it is
3228 prompting; (b) run "./pcretest /dev/null" at the start of the test script
3229 so the version is shown; (c) stop "make test" echoing "./RunTest".
3231 8. Added patches from David Burgess to enable PCRE to run on EBCDIC systems.
3233 9. The prototype for memmove() for systems that don't have it was using
3234 size_t, but the inclusion of the header that defines size_t was later. I've
3235 moved the #includes for the C headers earlier to avoid this.
3237 10. Added some adjustments to the code to make it easier to compiler on certain
3238 special systems:
3240 (a) Some "const" qualifiers were missing.
3241 (b) Added the macro EXPORT before all exported functions; by default this
3242 is defined to be empty.
3243 (c) Changed the dftables auxiliary program (that builds chartables.c) so
3244 that it reads its output file name as an argument instead of writing
3245 to the standard output and assuming this can be redirected.
3247 11. In UTF-8 mode, if a recursive reference (e.g. (?1)) followed a character
3248 class containing characters with values greater than 255, PCRE compilation
3249 went into a loop.
3251 12. A recursive reference to a subpattern that was within another subpattern
3252 that had a minimum quantifier of zero caused PCRE to crash. For example,
3253 (x(y(?2))z)? provoked this bug with a subject that got as far as the
3254 recursion. If the recursively-called subpattern itself had a zero repeat,
3255 that was OK.
3257 13. In pcretest, the buffer for reading a data line was set at 30K, but the
3258 buffer into which it was copied (for escape processing) was still set at
3259 1024, so long lines caused crashes.
3261 14. A pattern such as /[ab]{1,3}+/ failed to compile, giving the error
3262 "internal error: code overflow...". This applied to any character class
3263 that was followed by a possessive quantifier.
3265 15. Modified the Makefile to add libpcre.la as a prerequisite for
3266 libpcreposix.la because I was told this is needed for a parallel build to
3267 work.
3269 16. If a pattern that contained .* following optional items at the start was
3270 studied, the wrong optimizing data was generated, leading to matching
3271 errors. For example, studying /[ab]*.*c/ concluded, erroneously, that any
3272 matching string must start with a or b or c. The correct conclusion for
3273 this pattern is that a match can start with any character.
3276 Version 4.4 13-Aug-03
3277 ---------------------
3279 1. In UTF-8 mode, a character class containing characters with values between
3280 127 and 255 was not handled correctly if the compiled pattern was studied.
3281 In fixing this, I have also improved the studying algorithm for such
3282 classes (slightly).
3284 2. Three internal functions had redundant arguments passed to them. Removal
3285 might give a very teeny performance improvement.
3287 3. Documentation bug: the value of the capture_top field in a callout is *one
3288 more than* the number of the hightest numbered captured substring.
3290 4. The Makefile linked pcretest and pcregrep with -lpcre, which could result
3291 in incorrectly linking with a previously installed version. They now link
3292 explicitly with libpcre.la.
3294 5. configure.in no longer needs to recognize Cygwin specially.
3296 6. A problem in pcre.in for Windows platforms is fixed.
3298 7. If a pattern was successfully studied, and the -d (or /D) flag was given to
3299 pcretest, it used to include the size of the study block as part of its
3300 output. Unfortunately, the structure contains a field that has a different
3301 size on different hardware architectures. This meant that the tests that
3302 showed this size failed. As the block is currently always of a fixed size,
3303 this information isn't actually particularly useful in pcretest output, so
3304 I have just removed it.
3306 8. Three pre-processor statements accidentally did not start in column 1.
3307 Sadly, there are *still* compilers around that complain, even though
3308 standard C has not required this for well over a decade. Sigh.
3310 9. In pcretest, the code for checking callouts passed small integers in the
3311 callout_data field, which is a void * field. However, some picky compilers
3312 complained about the casts involved for this on 64-bit systems. Now
3313 pcretest passes the address of the small integer instead, which should get
3314 rid of the warnings.
3316 10. By default, when in UTF-8 mode, PCRE now checks for valid UTF-8 strings at
3317 both compile and run time, and gives an error if an invalid UTF-8 sequence
3318 is found. There is a option for disabling this check in cases where the
3319 string is known to be correct and/or the maximum performance is wanted.
3321 11. In response to a bug report, I changed one line in Makefile.in from
3323 -Wl,--out-implib,.libs/lib@WIN_PREFIX@pcreposix.dll.a \
3324 to
3325 -Wl,--out-implib,.libs/@WIN_PREFIX@libpcreposix.dll.a \
3327 to look similar to other lines, but I have no way of telling whether this
3328 is the right thing to do, as I do not use Windows. No doubt I'll get told
3329 if it's wrong...
3332 Version 4.3 21-May-03
3333 ---------------------
3335 1. Two instances of @WIN_PREFIX@ omitted from the Windows targets in the
3336 Makefile.
3338 2. Some refactoring to improve the quality of the code:
3340 (i) The utf8_table... variables are now declared "const".
3342 (ii) The code for \cx, which used the "case flipping" table to upper case
3343 lower case letters, now just substracts 32. This is ASCII-specific,
3344 but the whole concept of \cx is ASCII-specific, so it seems
3345 reasonable.
3347 (iii) PCRE was using its character types table to recognize decimal and
3348 hexadecimal digits in the pattern. This is silly, because it handles
3349 only 0-9, a-f, and A-F, but the character types table is locale-
3350 specific, which means strange things might happen. A private
3351 table is now used for this - though it costs 256 bytes, a table is
3352 much faster than multiple explicit tests. Of course, the standard
3353 character types table is still used for matching digits in subject
3354 strings against \d.
3356 (iv) Strictly, the identifier ESC_t is reserved by POSIX (all identifiers
3357 ending in _t are). So I've renamed it as ESC_tee.
3359 3. The first argument for regexec() in the POSIX wrapper should have been
3360 defined as "const".
3362 4. Changed pcretest to use malloc() for its buffers so that they can be
3363 Electric Fenced for debugging.
3365 5. There were several places in the code where, in UTF-8 mode, PCRE would try
3366 to read one or more bytes before the start of the subject string. Often this
3367 had no effect on PCRE's behaviour, but in some circumstances it could
3368 provoke a segmentation fault.
3370 6. A lookbehind at the start of a pattern in UTF-8 mode could also cause PCRE
3371 to try to read one or more bytes before the start of the subject string.
3373 7. A lookbehind in a pattern matched in non-UTF-8 mode on a PCRE compiled with
3374 UTF-8 support could misbehave in various ways if the subject string
3375 contained bytes with the 0x80 bit set and the 0x40 bit unset in a lookbehind
3376 area. (PCRE was not checking for the UTF-8 mode flag, and trying to move
3377 back over UTF-8 characters.)
3380 Version 4.2 14-Apr-03
3381 ---------------------
3383 1. Typo "#if SUPPORT_UTF8" instead of "#ifdef SUPPORT_UTF8" fixed.
3385 2. Changes to the building process, supplied by Ronald Landheer-Cieslak
3386 [ON_WINDOWS]: new variable, "#" on non-Windows platforms
3387 [NOT_ON_WINDOWS]: new variable, "#" on Windows platforms
3388 [WIN_PREFIX]: new variable, "cyg" for Cygwin
3389 * Makefile.in: use autoconf substitution for OBJEXT, EXEEXT, BUILD_OBJEXT
3391 Note: automatic setting of the BUILD variables is not yet working
3392 set CPPFLAGS and BUILD_CPPFLAGS (but don't use yet) - should be used at
3393 compile-time but not at link-time
3394 [LINK]: use for linking executables only
3395 make different versions for Windows and non-Windows
3396 [LINKLIB]: new variable, copy of UNIX-style LINK, used for linking
3397 libraries
3398 [LINK_FOR_BUILD]: new variable
3399 [OBJEXT]: use throughout
3400 [EXEEXT]: use throughout
3401 <winshared>: new target
3402 <wininstall>: new target
3403 <dftables.o>: use native compiler
3404 <dftables>: use native linker
3405 <install>: handle Windows platform correctly
3406 <clean>: ditto
3407 <check>: ditto
3408 copy DLL to top builddir before testing
3410 As part of these changes, -no-undefined was removed again. This was reported
3411 to give trouble on HP-UX 11.0, so getting rid of it seems like a good idea
3412 in any case.
3414 3. Some tidies to get rid of compiler warnings:
3416 . In the match_data structure, match_limit was an unsigned long int, whereas
3417 match_call_count was an int. I've made them both unsigned long ints.
3419 . In pcretest the fact that a const uschar * doesn't automatically cast to
3420 a void * provoked a warning.
3422 . Turning on some more compiler warnings threw up some "shadow" variables
3423 and a few more missing casts.
3425 4. If PCRE was complied with UTF-8 support, but called without the PCRE_UTF8
3426 option, a class that contained a single character with a value between 128
3427 and 255 (e.g. /[\xFF]/) caused PCRE to crash.
3429 5. If PCRE was compiled with UTF-8 support, but called without the PCRE_UTF8
3430 option, a class that contained several characters, but with at least one
3431 whose value was between 128 and 255 caused PCRE to crash.
3434 Version 4.1 12-Mar-03
3435 ---------------------
3437 1. Compiling with gcc -pedantic found a couple of places where casts were
3438 needed, and a string in dftables.c that was longer than standard compilers are
3439 required to support.
3441 2. Compiling with Sun's compiler found a few more places where the code could
3442 be tidied up in order to avoid warnings.
3444 3. The variables for cross-compiling were called HOST_CC and HOST_CFLAGS; the
3445 first of these names is deprecated in the latest Autoconf in favour of the name
3446 CC_FOR_BUILD, because "host" is typically used to mean the system on which the
3447 compiled code will be run. I can't find a reference for HOST_CFLAGS, but by
3448 analogy I have changed it to CFLAGS_FOR_BUILD.
3450 4. Added -no-undefined to the linking command in the Makefile, because this is
3451 apparently helpful for Windows. To make it work, also added "-L. -lpcre" to the
3452 linking step for the pcreposix library.
3454 5. PCRE was failing to diagnose the case of two named groups with the same
3455 name.
3457 6. A problem with one of PCRE's optimizations was discovered. PCRE remembers a
3458 literal character that is needed in the subject for a match, and scans along to
3459 ensure that it is present before embarking on the full matching process. This
3460 saves time in cases of nested unlimited repeats that are never going to match.
3461 Problem: the scan can take a lot of time if the subject is very long (e.g.
3462 megabytes), thus penalizing straightforward matches. It is now done only if the
3463 amount of subject to be scanned is less than 1000 bytes.
3465 7. A lesser problem with the same optimization is that it was recording the
3466 first character of an anchored pattern as "needed", thus provoking a search
3467 right along the subject, even when the first match of the pattern was going to
3468 fail. The "needed" character is now not set for anchored patterns, unless it
3469 follows something in the pattern that is of non-fixed length. Thus, it still
3470 fulfils its original purpose of finding quick non-matches in cases of nested
3471 unlimited repeats, but isn't used for simple anchored patterns such as /^abc/.
3474 Version 4.0 17-Feb-03
3475 ---------------------
3477 1. If a comment in an extended regex that started immediately after a meta-item
3478 extended to the end of string, PCRE compiled incorrect data. This could lead to
3479 all kinds of weird effects. Example: /#/ was bad; /()#/ was bad; /a#/ was not.
3481 2. Moved to autoconf 2.53 and libtool 1.4.2.
3483 3. Perl 5.8 no longer needs "use utf8" for doing UTF-8 things. Consequently,
3484 the special perltest8 script is no longer needed - all the tests can be run
3485 from a single perltest script.
3487 4. From 5.004, Perl has not included the VT character (0x0b) in the set defined
3488 by \s. It has now been removed in PCRE. This means it isn't recognized as
3489 whitespace in /x regexes too, which is the same as Perl. Note that the POSIX
3490 class [:space:] *does* include VT, thereby creating a mess.
3492 5. Added the class [:blank:] (a GNU extension from Perl 5.8) to match only
3493 space and tab.
3495 6. Perl 5.005 was a long time ago. It's time to amalgamate the tests that use
3496 its new features into the main test script, reducing the number of scripts.
3498 7. Perl 5.8 has changed the meaning of patterns like /a(?i)b/. Earlier versions
3499 were backward compatible, and made the (?i) apply to the whole pattern, as if
3500 /i were given. Now it behaves more logically, and applies the option setting
3501 only to what follows. PCRE has been changed to follow suit. However, if it
3502 finds options settings right at the start of the pattern, it extracts them into
3503 the global options, as before. Thus, they show up in the info data.
3505 8. Added support for the \Q...\E escape sequence. Characters in between are
3506 treated as literals. This is slightly different from Perl in that $ and @ are
3507 also handled as literals inside the quotes. In Perl, they will cause variable
3508 interpolation. Note the following examples:
3510 Pattern PCRE matches Perl matches
3512 \Qabc$xyz\E abc$xyz abc followed by the contents of $xyz
3513 \Qabc\$xyz\E abc\$xyz abc\$xyz
3514 \Qabc\E\$\Qxyz\E abc$xyz abc$xyz
3516 For compatibility with Perl, \Q...\E sequences are recognized inside character
3517 classes as well as outside them.
3519 9. Re-organized 3 code statements in pcretest to avoid "overflow in
3520 floating-point constant arithmetic" warnings from a Microsoft compiler. Added a
3521 (size_t) cast to one statement in pcretest and one in pcreposix to avoid
3522 signed/unsigned warnings.
3524 10. SunOS4 doesn't have strtoul(). This was used only for unpicking the -o
3525 option for pcretest, so I've replaced it by a simple function that does just
3526 that job.
3528 11. pcregrep was ending with code 0 instead of 2 for the commands "pcregrep" or
3529 "pcregrep -".
3531 12. Added "possessive quantifiers" ?+, *+, ++, and {,}+ which come from Sun's
3532 Java package. This provides some syntactic sugar for simple cases of what my
3533 documentation calls "once-only subpatterns". A pattern such as x*+ is the same
3534 as (?>x*). In other words, if what is inside (?>...) is just a single repeated
3535 item, you can use this simplified notation. Note that only makes sense with
3536 greedy quantifiers. Consequently, the use of the possessive quantifier forces
3537 greediness, whatever the setting of the PCRE_UNGREEDY option.
3539 13. A change of greediness default within a pattern was not taking effect at
3540 the current level for patterns like /(b+(?U)a+)/. It did apply to parenthesized
3541 subpatterns that followed. Patterns like /b+(?U)a+/ worked because the option
3542 was abstracted outside.
3544 14. PCRE now supports the \G assertion. It is true when the current matching
3545 position is at the start point of the match. This differs from \A when the
3546 starting offset is non-zero. Used with the /g option of pcretest (or similar
3547 code), it works in the same way as it does for Perl's /g option. If all
3548 alternatives of a regex begin with \G, the expression is anchored to the start
3549 match position, and the "anchored" flag is set in the compiled expression.
3551 15. Some bugs concerning the handling of certain option changes within patterns
3552 have been fixed. These applied to options other than (?ims). For example,
3553 "a(?x: b c )d" did not match "XabcdY" but did match "Xa b c dY". It should have
3554 been the other way round. Some of this was related to change 7 above.
3556 16. PCRE now gives errors for /[.x.]/ and /[=x=]/ as unsupported POSIX
3557 features, as Perl does. Previously, PCRE gave the warnings only for /[[.x.]]/
3558 and /[[=x=]]/. PCRE now also gives an error for /[:name:]/ because it supports
3559 POSIX classes only within a class (e.g. /[[:alpha:]]/).
3561 17. Added support for Perl's \C escape. This matches one byte, even in UTF8
3562 mode. Unlike ".", it always matches newline, whatever the setting of
3563 PCRE_DOTALL. However, PCRE does not permit \C to appear in lookbehind
3564 assertions. Perl allows it, but it doesn't (in general) work because it can't
3565 calculate the length of the lookbehind. At least, that's the case for Perl
3566 5.8.0 - I've been told they are going to document that it doesn't work in
3567 future.
3569 18. Added an error diagnosis for escapes that PCRE does not support: these are
3570 \L, \l, \N, \P, \p, \U, \u, and \X.
3572 19. Although correctly diagnosing a missing ']' in a character class, PCRE was
3573 reading past the end of the pattern in cases such as /[abcd/.
3575 20. PCRE was getting more memory than necessary for patterns with classes that
3576 contained both POSIX named classes and other characters, e.g. /[[:space:]abc/.
3578 21. Added some code, conditional on #ifdef VPCOMPAT, to make life easier for
3579 compiling PCRE for use with Virtual Pascal.
3581 22. Small fix to the Makefile to make it work properly if the build is done
3582 outside the source tree.
3584 23. Added a new extension: a condition to go with recursion. If a conditional
3585 subpattern starts with (?(R) the "true" branch is used if recursion has
3586 happened, whereas the "false" branch is used only at the top level.
3588 24. When there was a very long string of literal characters (over 255 bytes
3589 without UTF support, over 250 bytes with UTF support), the computation of how
3590 much memory was required could be incorrect, leading to segfaults or other
3591 strange effects.
3593 25. PCRE was incorrectly assuming anchoring (either to start of subject or to
3594 start of line for a non-DOTALL pattern) when a pattern started with (.*) and
3595 there was a subsequent back reference to those brackets. This meant that, for
3596 example, /(.*)\d+\1/ failed to match "abc123bc". Unfortunately, it isn't
3597 possible to check for precisely this case. All we can do is abandon the
3598 optimization if .* occurs inside capturing brackets when there are any back
3599 references whatsoever. (See below for a better fix that came later.)
3601 26. The handling of the optimization for finding the first character of a
3602 non-anchored pattern, and for finding a character that is required later in the
3603 match were failing in some cases. This didn't break the matching; it just
3604 failed to optimize when it could. The way this is done has been re-implemented.
3606 27. Fixed typo in error message for invalid (?R item (it said "(?p").
3608 28. Added a new feature that provides some of the functionality that Perl
3609 provides with (?{...}). The facility is termed a "callout". The way it is done
3610 in PCRE is for the caller to provide an optional function, by setting
3611 pcre_callout to its entry point. Like pcre_malloc and pcre_free, this is a
3612 global variable. By default it is unset, which disables all calling out. To get
3613 the function called, the regex must include (?C) at appropriate points. This
3614 is, in fact, equivalent to (?C0), and any number <= 255 may be given with (?C).
3615 This provides a means of identifying different callout points. When PCRE
3616 reaches such a point in the regex, if pcre_callout has been set, the external
3617 function is called. It is provided with data in a structure called
3618 pcre_callout_block, which is defined in pcre.h. If the function returns 0,
3619 matching continues; if it returns a non-zero value, the match at the current
3620 point fails. However, backtracking will occur if possible. [This was changed
3621 later and other features added - see item 49 below.]
3623 29. pcretest is upgraded to test the callout functionality. It provides a
3624 callout function that displays information. By default, it shows the start of
3625 the match and the current position in the text. There are some new data escapes
3626 to vary what happens:
3628 \C+ in addition, show current contents of captured substrings
3629 \C- do not supply a callout function
3630 \C!n return 1 when callout number n is reached
3631 \C!n!m return 1 when callout number n is reached for the mth time
3633 30. If pcregrep was called with the -l option and just a single file name, it
3634 output "<stdin>" if a match was found, instead of the file name.
3636 31. Improve the efficiency of the POSIX API to PCRE. If the number of capturing
3637 slots is less than POSIX_MALLOC_THRESHOLD, use a block on the stack to pass to
3638 pcre_exec(). This saves a malloc/free per call. The default value of
3639 POSIX_MALLOC_THRESHOLD is 10; it can be changed by --with-posix-malloc-threshold
3640 when configuring.
3642 32. The default maximum size of a compiled pattern is 64K. There have been a
3643 few cases of people hitting this limit. The code now uses macros to handle the
3644 storing of links as offsets within the compiled pattern. It defaults to 2-byte
3645 links, but this can be changed to 3 or 4 bytes by --with-link-size when
3646 configuring. Tests 2 and 5 work only with 2-byte links because they output
3647 debugging information about compiled patterns.
3649 33. Internal code re-arrangements:
3651 (a) Moved the debugging function for printing out a compiled regex into
3652 its own source file (printint.c) and used #include to pull it into
3653 pcretest.c and, when DEBUG is defined, into pcre.c, instead of having two
3654 separate copies.
3656 (b) Defined the list of op-code names for debugging as a macro in
3657 internal.h so that it is next to the definition of the opcodes.
3659 (c) Defined a table of op-code lengths for simpler skipping along compiled
3660 code. This is again a macro in internal.h so that it is next to the
3661 definition of the opcodes.
3663 34. Added support for recursive calls to individual subpatterns, along the
3664 lines of Robin Houston's patch (but implemented somewhat differently).
3666 35. Further mods to the Makefile to help Win32. Also, added code to pcregrep to
3667 allow it to read and process whole directories in Win32. This code was
3668 contributed by Lionel Fourquaux; it has not been tested by me.
3670 36. Added support for named subpatterns. The Python syntax (?P<name>...) is
3671 used to name a group. Names consist of alphanumerics and underscores, and must
3672 be unique. Back references use the syntax (?P=name) and recursive calls use
3673 (?P>name) which is a PCRE extension to the Python extension. Groups still have
3674 numbers. The function pcre_fullinfo() can be used after compilation to extract
3675 a name/number map. There are three relevant calls:
3677 PCRE_INFO_NAMEENTRYSIZE yields the size of each entry in the map
3678 PCRE_INFO_NAMECOUNT yields the number of entries
3679 PCRE_INFO_NAMETABLE yields a pointer to the map.
3681 The map is a vector of fixed-size entries. The size of each entry depends on
3682 the length of the longest name used. The first two bytes of each entry are the
3683 group number, most significant byte first. There follows the corresponding
3684 name, zero terminated. The names are in alphabetical order.
3686 37. Make the maximum literal string in the compiled code 250 for the non-UTF-8
3687 case instead of 255. Making it the same both with and without UTF-8 support
3688 means that the same test output works with both.
3690 38. There was a case of malloc(0) in the POSIX testing code in pcretest. Avoid
3691 calling malloc() with a zero argument.
3693 39. Change 25 above had to resort to a heavy-handed test for the .* anchoring
3694 optimization. I've improved things by keeping a bitmap of backreferences with
3695 numbers 1-31 so that if .* occurs inside capturing brackets that are not in
3696 fact referenced, the optimization can be applied. It is unlikely that a
3697 relevant occurrence of .* (i.e. one which might indicate anchoring or forcing
3698 the match to follow \n) will appear inside brackets with a number greater than
3699 31, but if it does, any back reference > 31 suppresses the optimization.
3701 40. Added a new compile-time option PCRE_NO_AUTO_CAPTURE. This has the effect
3702 of disabling numbered capturing parentheses. Any opening parenthesis that is
3703 not followed by ? behaves as if it were followed by ?: but named parentheses
3704 can still be used for capturing (and they will acquire numbers in the usual
3705 way).
3707 41. Redesigned the return codes from the match() function into yes/no/error so
3708 that errors can be passed back from deep inside the nested calls. A malloc
3709 failure while inside a recursive subpattern call now causes the
3710 PCRE_ERROR_NOMEMORY return instead of quietly going wrong.
3712 42. It is now possible to set a limit on the number of times the match()
3713 function is called in a call to pcre_exec(). This facility makes it possible to
3714 limit the amount of recursion and backtracking, though not in a directly
3715 obvious way, because the match() function is used in a number of different
3716 circumstances. The count starts from zero for each position in the subject
3717 string (for non-anchored patterns). The default limit is, for compatibility, a
3718 large number, namely 10 000 000. You can change this in two ways:
3720 (a) When configuring PCRE before making, you can use --with-match-limit=n
3721 to set a default value for the compiled library.
3723 (b) For each call to pcre_exec(), you can pass a pcre_extra block in which
3724 a different value is set. See 45 below.
3726 If the limit is exceeded, pcre_exec() returns PCRE_ERROR_MATCHLIMIT.
3728 43. Added a new function pcre_config(int, void *) to enable run-time extraction
3729 of things that can be changed at compile time. The first argument specifies
3730 what is wanted and the second points to where the information is to be placed.
3731 The current list of available information is:
3735 The output is an integer that is set to one if UTF-8 support is available;
3736 otherwise it is set to zero.
3740 The output is an integer that it set to the value of the code that is used for
3741 newline. It is either LF (10) or CR (13).
3745 The output is an integer that contains the number of bytes used for internal
3746 linkage in compiled expressions. The value is 2, 3, or 4. See item 32 above.
3750 The output is an integer that contains the threshold above which the POSIX
3751 interface uses malloc() for output vectors. See item 31 above.
3755 The output is an unsigned integer that contains the default limit of the number
3756 of match() calls in a pcre_exec() execution. See 42 above.
3758 44. pcretest has been upgraded by the addition of the -C option. This causes it
3759 to extract all the available output from the new pcre_config() function, and to
3760 output it. The program then exits immediately.
3762 45. A need has arisen to pass over additional data with calls to pcre_exec() in
3763 order to support additional features. One way would have been to define
3764 pcre_exec2() (for example) with extra arguments, but this would not have been
3765 extensible, and would also have required all calls to the original function to
3766 be mapped to the new one. Instead, I have chosen to extend the mechanism that
3767 is used for passing in "extra" data from pcre_study().
3769 The pcre_extra structure is now exposed and defined in pcre.h. It currently
3770 contains the following fields:
3772 flags a bitmap indicating which of the following fields are set
3773 study_data opaque data from pcre_study()
3774 match_limit a way of specifying a limit on match() calls for a specific
3775 call to pcre_exec()
3776 callout_data data for callouts (see 49 below)
3778 The flag bits are also defined in pcre.h, and are
3784 The pcre_study() function now returns one of these new pcre_extra blocks, with
3785 the actual study data pointed to by the study_data field, and the
3786 PCRE_EXTRA_STUDY_DATA flag set. This can be passed directly to pcre_exec() as
3787 before. That is, this change is entirely upwards-compatible and requires no
3788 change to existing code.
3790 If you want to pass in additional data to pcre_exec(), you can either place it
3791 in a pcre_extra block provided by pcre_study(), or create your own pcre_extra
3792 block.
3794 46. pcretest has been extended to test the PCRE_EXTRA_MATCH_LIMIT feature. If a
3795 data string contains the escape sequence \M, pcretest calls pcre_exec() several
3796 times with different match limits, until it finds the minimum value needed for
3797 pcre_exec() to complete. The value is then output. This can be instructive; for
3798 most simple matches the number is quite small, but for pathological cases it
3799 gets very large very quickly.
3801 47. There's a new option for pcre_fullinfo() called PCRE_INFO_STUDYSIZE. It
3802 returns the size of the data block pointed to by the study_data field in a
3803 pcre_extra block, that is, the value that was passed as the argument to
3804 pcre_malloc() when PCRE was getting memory in which to place the information
3805 created by pcre_study(). The fourth argument should point to a size_t variable.
3806 pcretest has been extended so that this information is shown after a successful
3807 pcre_study() call when information about the compiled regex is being displayed.
3809 48. Cosmetic change to Makefile: there's no need to have / after $(DESTDIR)
3810 because what follows is always an absolute path. (Later: it turns out that this
3811 is more than cosmetic for MinGW, because it doesn't like empty path
3812 components.)
3814 49. Some changes have been made to the callout feature (see 28 above):
3816 (i) A callout function now has three choices for what it returns:
3818 0 => success, carry on matching
3819 > 0 => failure at this point, but backtrack if possible
3820 < 0 => serious error, return this value from pcre_exec()
3822 Negative values should normally be chosen from the set of PCRE_ERROR_xxx
3823 values. In particular, returning PCRE_ERROR_NOMATCH forces a standard
3824 "match failed" error. The error number PCRE_ERROR_CALLOUT is reserved for
3825 use by callout functions. It will never be used by PCRE itself.
3827 (ii) The pcre_extra structure (see 45 above) has a void * field called
3828 callout_data, with corresponding flag bit PCRE_EXTRA_CALLOUT_DATA. The
3829 pcre_callout_block structure has a field of the same name. The contents of
3830 the field passed in the pcre_extra structure are passed to the callout
3831 function in the corresponding field in the callout block. This makes it
3832 easier to use the same callout-containing regex from multiple threads. For
3833 testing, the pcretest program has a new data escape
3835 \C*n pass the number n (may be negative) as callout_data
3837 If the callout function in pcretest receives a non-zero value as
3838 callout_data, it returns that value.
3840 50. Makefile wasn't handling CFLAGS properly when compiling dftables. Also,
3841 there were some redundant $(CFLAGS) in commands that are now specified as
3842 $(LINK), which already includes $(CFLAGS).
3844 51. Extensions to UTF-8 support are listed below. These all apply when (a) PCRE
3845 has been compiled with UTF-8 support *and* pcre_compile() has been compiled
3846 with the PCRE_UTF8 flag. Patterns that are compiled without that flag assume
3847 one-byte characters throughout. Note that case-insensitive matching applies
3848 only to characters whose values are less than 256. PCRE doesn't support the
3849 notion of cases for higher-valued characters.
3851 (i) A character class whose characters are all within 0-255 is handled as
3852 a bit map, and the map is inverted for negative classes. Previously, a
3853 character > 255 always failed to match such a class; however it should
3854 match if the class was a negative one (e.g. [^ab]). This has been fixed.
3856 (ii) A negated character class with a single character < 255 is coded as
3857 "not this character" (OP_NOT). This wasn't working properly when the test
3858 character was multibyte, either singly or repeated.
3860 (iii) Repeats of multibyte characters are now handled correctly in UTF-8
3861 mode, for example: \x{100}{2,3}.
3863 (iv) The character escapes \b, \B, \d, \D, \s, \S, \w, and \W (either
3864 singly or repeated) now correctly test multibyte characters. However,
3865 PCRE doesn't recognize any characters with values greater than 255 as
3866 digits, spaces, or word characters. Such characters always match \D, \S,
3867 and \W, and never match \d, \s, or \w.
3869 (v) Classes may now contain characters and character ranges with values
3870 greater than 255. For example: [ab\x{100}-\x{400}].
3872 (vi) pcregrep now has a --utf-8 option (synonym -u) which makes it call
3873 PCRE in UTF-8 mode.
3875 52. The info request value PCRE_INFO_FIRSTCHAR has been renamed
3876 PCRE_INFO_FIRSTBYTE because it is a byte value. However, the old name is
3877 retained for backwards compatibility. (Note that LASTLITERAL is also a byte
3878 value.)
3880 53. The single man page has become too large. I have therefore split it up into
3881 a number of separate man pages. These also give rise to individual HTML pages;
3882 these are now put in a separate directory, and there is an index.html page that
3883 lists them all. Some hyperlinking between the pages has been installed.
3885 54. Added convenience functions for handling named capturing parentheses.
3887 55. Unknown escapes inside character classes (e.g. [\M]) and escapes that
3888 aren't interpreted therein (e.g. [\C]) are literals in Perl. This is now also
3889 true in PCRE, except when the PCRE_EXTENDED option is set, in which case they
3890 are faulted.
3892 56. Introduced HOST_CC and HOST_CFLAGS which can be set in the environment when
3893 calling configure. These values are used when compiling the dftables.c program
3894 which is run to generate the source of the default character tables. They
3895 default to the values of CC and CFLAGS. If you are cross-compiling PCRE,
3896 you will need to set these values.
3898 57. Updated the building process for Windows DLL, as provided by Fred Cox.
3901 Version 3.9 02-Jan-02
3902 ---------------------
3904 1. A bit of extraneous text had somehow crept into the pcregrep documentation.
3906 2. If --disable-static was given, the building process failed when trying to
3907 build pcretest and pcregrep. (For some reason it was using libtool to compile
3908 them, which is not right, as they aren't part of the library.)
3911 Version 3.8 18-Dec-01
3912 ---------------------
3914 1. The experimental UTF-8 code was completely screwed up. It was packing the
3915 bytes in the wrong order. How dumb can you get?
3918 Version 3.7 29-Oct-01
3919 ---------------------
3921 1. In updating pcretest to check change 1 of version 3.6, I screwed up.
3922 This caused pcretest, when used on the test data, to segfault. Unfortunately,
3923 this didn't happen under Solaris 8, where I normally test things.
3925 2. The Makefile had to be changed to make it work on BSD systems, where 'make'
3926 doesn't seem to recognize that ./xxx and xxx are the same file. (This entry
3927 isn't in ChangeLog distributed with 3.7 because I forgot when I hastily made
3928 this fix an hour or so after the initial 3.7 release.)
3931 Version 3.6 23-Oct-01
3932 ---------------------
3934 1. Crashed with /(sens|respons)e and \1ibility/ and "sense and sensibility" if
3935 offsets passed as NULL with zero offset count.
3937 2. The config.guess and config.sub files had not been updated when I moved to
3938 the latest autoconf.
3941 Version 3.5 15-Aug-01
3942 ---------------------
3944 1. Added some missing #if !defined NOPOSIX conditionals in pcretest.c that
3945 had been forgotten.
3947 2. By using declared but undefined structures, we can avoid using "void"
3948 definitions in pcre.h while keeping the internal definitions of the structures
3949 private.
3951 3. The distribution is now built using autoconf 2.50 and libtool 1.4. From a
3952 user point of view, this means that both static and shared libraries are built
3953 by default, but this can be individually controlled. More of the work of
3954 handling this static/shared cases is now inside libtool instead of PCRE's make
3955 file.
3957 4. The pcretest utility is now installed along with pcregrep because it is
3958 useful for users (to test regexs) and by doing this, it automatically gets
3959 relinked by libtool. The documentation has been turned into a man page, so
3960 there are now .1, .txt, and .html versions in /doc.
3962 5. Upgrades to pcregrep:
3963 (i) Added long-form option names like gnu grep.
3964 (ii) Added --help to list all options with an explanatory phrase.
3965 (iii) Added -r, --recursive to recurse into sub-directories.
3966 (iv) Added -f, --file to read patterns from a file.
3968 6. pcre_exec() was referring to its "code" argument before testing that
3969 argument for NULL (and giving an error if it was NULL).
3971 7. Upgraded Makefile.in to allow for compiling in a different directory from
3972 the source directory.
3974 8. Tiny buglet in pcretest: when pcre_fullinfo() was called to retrieve the
3975 options bits, the pointer it was passed was to an int instead of to an unsigned
3976 long int. This mattered only on 64-bit systems.
3978 9. Fixed typo (3.4/1) in pcre.h again. Sigh. I had changed pcre.h (which is
3979 generated) instead of pcre.in, which it its source. Also made the same change
3980 in several of the .c files.
3982 10. A new release of gcc defines printf() as a macro, which broke pcretest
3983 because it had an ifdef in the middle of a string argument for printf(). Fixed
3984 by using separate calls to printf().
3986 11. Added --enable-newline-is-cr and --enable-newline-is-lf to the configure
3987 script, to force use of CR or LF instead of \n in the source. On non-Unix
3988 systems, the value can be set in config.h.
3990 12. The limit of 200 on non-capturing parentheses is a _nesting_ limit, not an
3991 absolute limit. Changed the text of the error message to make this clear, and
3992 likewise updated the man page.
3994 13. The limit of 99 on the number of capturing subpatterns has been removed.
3995 The new limit is 65535, which I hope will not be a "real" limit.
3998 Version 3.4 22-Aug-00
3999 ---------------------
4001 1. Fixed typo in pcre.h: unsigned const char * changed to const unsigned char *.
4003 2. Diagnose condition (?(0) as an error instead of crashing on matching.
4006 Version 3.3 01-Aug-00
4007 ---------------------
4009 1. If an octal character was given, but the value was greater than \377, it
4010 was not getting masked to the least significant bits, as documented. This could
4011 lead to crashes in some systems.
4013 2. Perl 5.6 (if not earlier versions) accepts classes like [a-\d] and treats
4014 the hyphen as a literal. PCRE used to give an error; it now behaves like Perl.
4016 3. Added the functions pcre_free_substring() and pcre_free_substring_list().
4017 These just pass their arguments on to (pcre_free)(), but they are provided
4018 because some uses of PCRE bind it to non-C systems that can call its functions,
4019 but cannot call free() or pcre_free() directly.
4021 4. Add "make test" as a synonym for "make check". Corrected some comments in
4022 the Makefile.
4024 5. Add $(DESTDIR)/ in front of all the paths in the "install" target in the
4025 Makefile.
4027 6. Changed the name of pgrep to pcregrep, because Solaris has introduced a
4028 command called pgrep for grepping around the active processes.
4030 7. Added the beginnings of support for UTF-8 character strings.
4032 8. Arranged for the Makefile to pass over the settings of CC, CFLAGS, and
4033 RANLIB to ./ltconfig so that they are used by libtool. I think these are all
4034 the relevant ones. (AR is not passed because ./ltconfig does its own figuring
4035 out for the ar command.)
4038 Version 3.2 12-May-00
4039 ---------------------
4041 This is purely a bug fixing release.
4043 1. If the pattern /((Z)+|A)*/ was matched agained ZABCDEFG it matched Z instead
4044 of ZA. This was just one example of several cases that could provoke this bug,
4045 which was introduced by change 9 of version 2.00. The code for breaking
4046 infinite loops after an iteration that matches an empty string was't working
4047 correctly.
4049 2. The pcretest program was not imitating Perl correctly for the pattern /a*/g
4050 when matched against abbab (for example). After matching an empty string, it
4051 wasn't forcing anchoring when setting PCRE_NOTEMPTY for the next attempt; this
4052 caused it to match further down the string than it should.
4054 3. The code contained an inclusion of sys/types.h. It isn't clear why this
4055 was there because it doesn't seem to be needed, and it causes trouble on some
4056 systems, as it is not a Standard C header. It has been removed.
4058 4. Made 4 silly changes to the source to avoid stupid compiler warnings that
4059 were reported on the Macintosh. The changes were from
4061 while ((c = *(++ptr)) != 0 && c != '\n');
4062 to
4063 while ((c = *(++ptr)) != 0 && c != '\n') ;
4065 Totally extraordinary, but if that's what it takes...
4067 5. PCRE is being used in one environment where neither memmove() nor bcopy() is
4068 available. Added HAVE_BCOPY and an autoconf test for it; if neither
4069 HAVE_MEMMOVE nor HAVE_BCOPY is set, use a built-in emulation function which
4070 assumes the way PCRE uses memmove() (always moving upwards).
4072 6. PCRE is being used in one environment where strchr() is not available. There
4073 was only one use in pcre.c, and writing it out to avoid strchr() probably gives
4074 faster code anyway.
4077 Version 3.1 09-Feb-00
4078 ---------------------
4080 The only change in this release is the fixing of some bugs in Makefile.in for
4081 the "install" target:
4083 (1) It was failing to install pcreposix.h.
4085 (2) It was overwriting the pcre.3 man page with the pcreposix.3 man page.
4088 Version 3.0 01-Feb-00
4089 ---------------------
4091 1. Add support for the /+ modifier to perltest (to output $` like it does in
4092 pcretest).
4094 2. Add support for the /g modifier to perltest.
4096 3. Fix pcretest so that it behaves even more like Perl for /g when the pattern
4097 matches null strings.
4099 4. Fix perltest so that it doesn't do unwanted things when fed an empty
4100 pattern. Perl treats empty patterns specially - it reuses the most recent
4101 pattern, which is not what we want. Replace // by /(?#)/ in order to avoid this
4102 effect.
4104 5. The POSIX interface was broken in that it was just handing over the POSIX
4105 captured string vector to pcre_exec(), but (since release 2.00) PCRE has
4106 required a bigger vector, with some working space on the end. This means that
4107 the POSIX wrapper now has to get and free some memory, and copy the results.
4109 6. Added some simple autoconf support, placing the test data and the
4110 documentation in separate directories, re-organizing some of the
4111 information files, and making it build pcre-config (a GNU standard). Also added
4112 libtool support for building PCRE as a shared library, which is now the
4113 default.
4115 7. Got rid of the leading zero in the definition of PCRE_MINOR because 08 and
4116 09 are not valid octal constants. Single digits will be used for minor values
4117 less than 10.
4119 8. Defined REG_EXTENDED and REG_NOSUB as zero in the POSIX header, so that
4120 existing programs that set these in the POSIX interface can use PCRE without
4121 modification.
4123 9. Added a new function, pcre_fullinfo() with an extensible interface. It can
4124 return all that pcre_info() returns, plus additional data. The pcre_info()
4125 function is retained for compatibility, but is considered to be obsolete.
4127 10. Added experimental recursion feature (?R) to handle one common case that
4128 Perl 5.6 will be able to do with (?p{...}).
4130 11. Added support for POSIX character classes like [:alpha:], which Perl is
4131 adopting.
4134 Version 2.08 31-Aug-99
4135 ----------------------
4137 1. When startoffset was not zero and the pattern began with ".*", PCRE was not
4138 trying to match at the startoffset position, but instead was moving forward to
4139 the next newline as if a previous match had failed.
4141 2. pcretest was not making use of PCRE_NOTEMPTY when repeating for /g and /G,
4142 and could get into a loop if a null string was matched other than at the start
4143 of the subject.
4145 3. Added definitions of PCRE_MAJOR and PCRE_MINOR to pcre.h so the version can
4146 be distinguished at compile time, and for completeness also added PCRE_DATE.
4148 5. Added Paul Sokolovsky's minor changes to make it easy to compile a Win32 DLL
4149 in GnuWin32 environments.
4152 Version 2.07 29-Jul-99
4153 ----------------------
4155 1. The documentation is now supplied in plain text form and HTML as well as in
4156 the form of man page sources.
4158 2. C++ compilers don't like assigning (void *) values to other pointer types.
4159 In particular this affects malloc(). Although there is no problem in Standard
4160 C, I've put in casts to keep C++ compilers happy.
4162 3. Typo on pcretest.c; a cast of (unsigned char *) in the POSIX regexec() call
4163 should be (const char *).
4165 4. If NOPOSIX is defined, pcretest.c compiles without POSIX support. This may
4166 be useful for non-Unix systems who don't want to bother with the POSIX stuff.
4167 However, I haven't made this a standard facility. The documentation doesn't
4168 mention it, and the Makefile doesn't support it.
4170 5. The Makefile now contains an "install" target, with editable destinations at
4171 the top of the file. The pcretest program is not installed.
4173 6. pgrep -V now gives the PCRE version number and date.
4175 7. Fixed bug: a zero repetition after a literal string (e.g. /abcde{0}/) was
4176 causing the entire string to be ignored, instead of just the last character.
4178 8. If a pattern like /"([^\\"]+|\\.)*"/ is applied in the normal way to a
4179 non-matching string, it can take a very, very long time, even for strings of
4180 quite modest length, because of the nested recursion. PCRE now does better in
4181 some of these cases. It does this by remembering the last required literal
4182 character in the pattern, and pre-searching the subject to ensure it is present
4183 before running the real match. In other words, it applies a heuristic to detect
4184 some types of certain failure quickly, and in the above example, if presented
4185 with a string that has no trailing " it gives "no match" very quickly.
4187 9. A new runtime option PCRE_NOTEMPTY causes null string matches to be ignored;
4188 other alternatives are tried instead.
4191 Version 2.06 09-Jun-99
4192 ----------------------
4194 1. Change pcretest's output for amount of store used to show just the code
4195 space, because the remainder (the data block) varies in size between 32-bit and
4196 64-bit systems.
4198 2. Added an extra argument to pcre_exec() to supply an offset in the subject to
4199 start matching at. This allows lookbehinds to work when searching for multiple
4200 occurrences in a string.
4202 3. Added additional options to pcretest for testing multiple occurrences:
4204 /+ outputs the rest of the string that follows a match
4205 /g loops for multiple occurrences, using the new startoffset argument
4206 /G loops for multiple occurrences by passing an incremented pointer
4208 4. PCRE wasn't doing the "first character" optimization for patterns starting
4209 with \b or \B, though it was doing it for other lookbehind assertions. That is,
4210 it wasn't noticing that a match for a pattern such as /\bxyz/ has to start with
4211 the letter 'x'. On long subject strings, this gives a significant speed-up.
4214 Version 2.05 21-Apr-99
4215 ----------------------
4217 1. Changed the type of magic_number from int to long int so that it works
4218 properly on 16-bit systems.
4220 2. Fixed a bug which caused patterns starting with .* not to work correctly
4221 when the subject string contained newline characters. PCRE was assuming
4222 anchoring for such patterns in all cases, which is not correct because .* will
4223 not pass a newline unless PCRE_DOTALL is set. It now assumes anchoring only if
4224 DOTALL is set at top level; otherwise it knows that patterns starting with .*
4225 must be retried after every newline in the subject.
4228 Version 2.04 18-Feb-99
4229 ----------------------
4231 1. For parenthesized subpatterns with repeats whose minimum was zero, the
4232 computation of the store needed to hold the pattern was incorrect (too large).
4233 If such patterns were nested a few deep, this could multiply and become a real
4234 problem.
4236 2. Added /M option to pcretest to show the memory requirement of a specific
4237 pattern. Made -m a synonym of -s (which does this globally) for compatibility.
4239 3. Subpatterns of the form (regex){n,m} (i.e. limited maximum) were being
4240 compiled in such a way that the backtracking after subsequent failure was
4241 pessimal. Something like (a){0,3} was compiled as (a)?(a)?(a)? instead of
4242 ((a)((a)(a)?)?)? with disastrous performance if the maximum was of any size.
4245 Version 2.03 02-Feb-99
4246 ----------------------
4248 1. Fixed typo and small mistake in man page.
4250 2. Added 4th condition (GPL supersedes if conflict) and created separate
4251 LICENCE file containing the conditions.
4253 3. Updated pcretest so that patterns such as /abc\/def/ work like they do in
4254 Perl, that is the internal \ allows the delimiter to be included in the
4255 pattern. Locked out the use of \ as a delimiter. If \ immediately follows
4256 the final delimiter, add \ to the end of the pattern (to test the error).
4258 4. Added the convenience functions for extracting substrings after a successful
4259 match. Updated pcretest to make it able to test these functions.
4262 Version 2.02 14-Jan-99
4263 ----------------------
4265 1. Initialized the working variables associated with each extraction so that
4266 their saving and restoring doesn't refer to uninitialized store.
4268 2. Put dummy code into study.c in order to trick the optimizer of the IBM C
4269 compiler for OS/2 into generating correct code. Apparently IBM isn't going to
4270 fix the problem.
4272 3. Pcretest: the timing code wasn't using LOOPREPEAT for timing execution
4273 calls, and wasn't printing the correct value for compiling calls. Increased the
4274 default value of LOOPREPEAT, and the number of significant figures in the
4275 times.
4277 4. Changed "/bin/rm" in the Makefile to "-rm" so it works on Windows NT.
4279 5. Renamed "deftables" as "dftables" to get it down to 8 characters, to avoid
4280 a building problem on Windows NT with a FAT file system.
4283 Version 2.01 21-Oct-98
4284 ----------------------
4286 1. Changed the API for pcre_compile() to allow for the provision of a pointer
4287 to character tables built by pcre_maketables() in the current locale. If NULL
4288 is passed, the default tables are used.
4291 Version 2.00 24-Sep-98
4292 ----------------------
4294 1. Since the (>?) facility is in Perl 5.005, don't require PCRE_EXTRA to enable
4295 it any more.
4297 2. Allow quantification of (?>) groups, and make it work correctly.
4299 3. The first character computation wasn't working for (?>) groups.
4301 4. Correct the implementation of \Z (it is permitted to match on the \n at the
4302 end of the subject) and add 5.005's \z, which really does match only at the
4303 very end of the subject.
4305 5. Remove the \X "cut" facility; Perl doesn't have it, and (?> is neater.
4307 6. Remove the ability to specify CASELESS, MULTILINE, DOTALL, and
4308 DOLLAR_END_ONLY at runtime, to make it possible to implement the Perl 5.005
4309 localized options. All options to pcre_study() were also removed.
4311 7. Add other new features from 5.005:
4313 $(?<= positive lookbehind
4314 $(?<! negative lookbehind
4315 (?imsx-imsx) added the unsetting capability
4316 such a setting is global if at outer level; local otherwise
4317 (?imsx-imsx:) non-capturing groups with option setting
4318 (?(cond)re|re) conditional pattern matching
4320 A backreference to itself in a repeated group matches the previous
4321 captured string.
4323 8. General tidying up of studying (both automatic and via "study")
4324 consequential on the addition of new assertions.
4326 9. As in 5.005, unlimited repeated groups that could match an empty substring
4327 are no longer faulted at compile time. Instead, the loop is forcibly broken at
4328 runtime if any iteration does actually match an empty substring.
4330 10. Include the RunTest script in the distribution.
4332 11. Added tests from the Perl 5.005_02 distribution. This showed up a few
4333 discrepancies, some of which were old and were also with respect to 5.004. They
4334 have now been fixed.
4337 Version 1.09 28-Apr-98
4338 ----------------------
4340 1. A negated single character class followed by a quantifier with a minimum
4341 value of one (e.g. [^x]{1,6} ) was not compiled correctly. This could lead to
4342 program crashes, or just wrong answers. This did not apply to negated classes
4343 containing more than one character, or to minima other than one.
4346 Version 1.08 27-Mar-98
4347 ----------------------
4349 1. Add PCRE_UNGREEDY to invert the greediness of quantifiers.
4351 2. Add (?U) and (?X) to set PCRE_UNGREEDY and PCRE_EXTRA respectively. The
4352 latter must appear before anything that relies on it in the pattern.
4355 Version 1.07 16-Feb-98
4356 ----------------------
4358 1. A pattern such as /((a)*)*/ was not being diagnosed as in error (unlimited
4359 repeat of a potentially empty string).
4362 Version 1.06 23-Jan-98
4363 ----------------------
4365 1. Added Markus Oberhumer's little patches for C++.
4367 2. Literal strings longer than 255 characters were broken.
4370 Version 1.05 23-Dec-97
4371 ----------------------
4373 1. Negated character classes containing more than one character were failing if
4374 PCRE_CASELESS was set at run time.
4377 Version 1.04 19-Dec-97
4378 ----------------------
4380 1. Corrected the man page, where some "const" qualifiers had been omitted.
4382 2. Made debugging output print "{0,xxx}" instead of just "{,xxx}" to agree with
4383 input syntax.
4385 3. Fixed memory leak which occurred when a regex with back references was
4386 matched with an offsets vector that wasn't big enough. The temporary memory
4387 that is used in this case wasn't being freed if the match failed.
4389 4. Tidied pcretest to ensure it frees memory that it gets.
4391 5. Temporary memory was being obtained in the case where the passed offsets
4392 vector was exactly big enough.
4394 6. Corrected definition of offsetof() from change 5 below.
4396 7. I had screwed up change 6 below and broken the rules for the use of
4397 setjmp(). Now fixed.
4400 Version 1.03 18-Dec-97
4401 ----------------------
4403 1. A erroneous regex with a missing opening parenthesis was correctly
4404 diagnosed, but PCRE attempted to access brastack[-1], which could cause crashes
4405 on some systems.
4407 2. Replaced offsetof(real_pcre, code) by offsetof(real_pcre, code[0]) because
4408 it was reported that one broken compiler failed on the former because "code" is
4409 also an independent variable.
4411 3. The erroneous regex a[]b caused an array overrun reference.
4413 4. A regex ending with a one-character negative class (e.g. /[^k]$/) did not
4414 fail on data ending with that character. (It was going on too far, and checking
4415 the next character, typically a binary zero.) This was specific to the
4416 optimized code for single-character negative classes.
4418 5. Added a contributed patch from the TIN world which does the following:
4420 + Add an undef for memmove, in case the the system defines a macro for it.
4422 + Add a definition of offsetof(), in case there isn't one. (I don't know
4423 the reason behind this - offsetof() is part of the ANSI standard - but
4424 it does no harm).
4426 + Reduce the ifdef's in pcre.c using macro DPRINTF, thereby eliminating
4427 most of the places where whitespace preceded '#'. I have given up and
4428 allowed the remaining 2 cases to be at the margin.
4430 + Rename some variables in pcre to eliminate shadowing. This seems very
4431 pedantic, but does no harm, of course.
4433 6. Moved the call to setjmp() into its own function, to get rid of warnings
4434 from gcc -Wall, and avoided calling it at all unless PCRE_EXTRA is used.
4436 7. Constructs such as \d{8,} were compiling into the equivalent of
4437 \d{8}\d{0,65527} instead of \d{8}\d* which didn't make much difference to the
4438 outcome, but in this particular case used more store than had been allocated,
4439 which caused the bug to be discovered because it threw up an internal error.
4441 8. The debugging code in both pcre and pcretest for outputting the compiled
4442 form of a regex was going wrong in the case of back references followed by
4443 curly-bracketed repeats.
4446 Version 1.02 12-Dec-97
4447 ----------------------
4449 1. Typos in pcre.3 and comments in the source fixed.
4451 2. Applied a contributed patch to get rid of places where it used to remove
4452 'const' from variables, and fixed some signed/unsigned and uninitialized
4453 variable warnings.
4455 3. Added the "runtest" target to Makefile.
4457 4. Set default compiler flag to -O2 rather than just -O.
4460 Version 1.01 19-Nov-97
4461 ----------------------
4463 1. PCRE was failing to diagnose unlimited repeat of empty string for patterns
4464 like /([ab]*)*/, that is, for classes with more than one character in them.
4466 2. Likewise, it wasn't diagnosing patterns with "once-only" subpatterns, such
4467 as /((?>a*))*/ (a PCRE_EXTRA facility).
4470 Version 1.00 18-Nov-97
4471 ----------------------
4473 1. Added compile-time macros to support systems such as SunOS4 which don't have
4474 memmove() or strerror() but have other things that can be used instead.
4476 2. Arranged that "make clean" removes the executables.
4479 Version 0.99 27-Oct-97
4480 ----------------------
4482 1. Fixed bug in code for optimizing classes with only one character. It was
4483 initializing a 32-byte map regardless, which could cause it to run off the end
4484 of the memory it had got.
4486 2. Added, conditional on PCRE_EXTRA, the proposed (?>REGEX) construction.
4489 Version 0.98 22-Oct-97
4490 ----------------------
4492 1. Fixed bug in code for handling temporary memory usage when there are more
4493 back references than supplied space in the ovector. This could cause segfaults.
4496 Version 0.97 21-Oct-97
4497 ----------------------
4499 1. Added the \X "cut" facility, conditional on PCRE_EXTRA.
4501 2. Optimized negated single characters not to use a bit map.
4503 3. Brought error texts together as macro definitions; clarified some of them;
4504 fixed one that was wrong - it said "range out of order" when it meant "invalid
4505 escape sequence".
4507 4. Changed some char * arguments to const char *.
4509 5. Added PCRE_NOTBOL and PCRE_NOTEOL (from POSIX).
4511 6. Added the POSIX-style API wrapper in pcreposix.a and testing facilities in
4512 pcretest.
4515 Version 0.96 16-Oct-97
4516 ----------------------
4518 1. Added a simple "pgrep" utility to the distribution.
4520 2. Fixed an incompatibility with Perl: "{" is now treated as a normal character
4521 unless it appears in one of the precise forms "{ddd}", "{ddd,}", or "{ddd,ddd}"
4522 where "ddd" means "one or more decimal digits".
4524 3. Fixed serious bug. If a pattern had a back reference, but the call to
4525 pcre_exec() didn't supply a large enough ovector to record the related
4526 identifying subpattern, the match always failed. PCRE now remembers the number
4527 of the largest back reference, and gets some temporary memory in which to save
4528 the offsets during matching if necessary, in order to ensure that
4529 backreferences always work.
4531 4. Increased the compatibility with Perl in a number of ways:
4533 (a) . no longer matches \n by default; an option PCRE_DOTALL is provided
4534 to request this handling. The option can be set at compile or exec time.
4536 (b) $ matches before a terminating newline by default; an option
4537 PCRE_DOLLAR_ENDONLY is provided to override this (but not in multiline
4538 mode). The option can be set at compile or exec time.
4540 (c) The handling of \ followed by a digit other than 0 is now supposed to be
4541 the same as Perl's. If the decimal number it represents is less than 10
4542 or there aren't that many previous left capturing parentheses, an octal
4543 escape is read. Inside a character class, it's always an octal escape,
4544 even if it is a single digit.
4546 (d) An escaped but undefined alphabetic character is taken as a literal,
4547 unless PCRE_EXTRA is set. Currently this just reserves the remaining
4548 escapes.
4550 (e) {0} is now permitted. (The previous item is removed from the compiled
4551 pattern).
4553 5. Changed all the names of code files so that the basic parts are no longer
4554 than 10 characters, and abolished the teeny "globals.c" file.
4556 6. Changed the handling of character classes; they are now done with a 32-byte
4557 bit map always.
4559 7. Added the -d and /D options to pcretest to make it possible to look at the
4560 internals of compilation without having to recompile pcre.
4563 Version 0.95 23-Sep-97
4564 ----------------------
4566 1. Fixed bug in pre-pass concerning escaped "normal" characters such as \x5c or
4567 \x20 at the start of a run of normal characters. These were being treated as
4568 real characters, instead of the source characters being re-checked.
4571 Version 0.94 18-Sep-97
4572 ----------------------
4574 1. The functions are now thread-safe, with the caveat that the global variables
4575 containing pointers to malloc() and free() or alternative functions are the
4576 same for all threads.
4578 2. Get pcre_study() to generate a bitmap of initial characters for non-
4579 anchored patterns when this is possible, and use it if passed to pcre_exec().
4582 Version 0.93 15-Sep-97
4583 ----------------------
4585 1. /(b)|(:+)/ was computing an incorrect first character.
4587 2. Add pcre_study() to the API and the passing of pcre_extra to pcre_exec(),
4588 but not actually doing anything yet.
4590 3. Treat "-" characters in classes that cannot be part of ranges as literals,
4591 as Perl does (e.g. [-az] or [az-]).
4593 4. Set the anchored flag if a branch starts with .* or .*? because that tests
4594 all possible positions.
4596 5. Split up into different modules to avoid including unneeded functions in a
4597 compiled binary. However, compile and exec are still in one module. The "study"
4598 function is split off.
4600 6. The character tables are now in a separate module whose source is generated
4601 by an auxiliary program - but can then be edited by hand if required. There are
4602 now no calls to isalnum(), isspace(), isdigit(), isxdigit(), tolower() or
4603 toupper() in the code.
4605 7. Turn the malloc/free funtions variables into pcre_malloc and pcre_free and
4606 make them global. Abolish the function for setting them, as the caller can now
4607 set them directly.
4610 Version 0.92 11-Sep-97
4611 ----------------------
4613 1. A repeat with a fixed maximum and a minimum of 1 for an ordinary character
4614 (e.g. /a{1,3}/) was broken (I mis-optimized it).
4616 2. Caseless matching was not working in character classes if the characters in
4617 the pattern were in upper case.
4619 3. Make ranges like [W-c] work in the same way as Perl for caseless matching.
4621 4. Make PCRE_ANCHORED public and accept as a compile option.
4623 5. Add an options word to pcre_exec() and accept PCRE_ANCHORED and
4624 PCRE_CASELESS at run time. Add escapes \A and \I to pcretest to cause it to
4625 pass them.
4627 6. Give an error if bad option bits passed at compile or run time.
4629 7. Add PCRE_MULTILINE at compile and exec time, and (?m) as well. Add \M to
4630 pcretest to cause it to pass that flag.
4632 8. Add pcre_info(), to get the number of identifying subpatterns, the stored
4633 options, and the first character, if set.
4635 9. Recognize C+ or C{n,m} where n >= 1 as providing a fixed starting character.
4638 Version 0.91 10-Sep-97
4639 ----------------------
4641 1. PCRE was failing to diagnose unlimited repeats of subpatterns that could
4642 match the empty string as in /(a*)*/. It was looping and ultimately crashing.
4644 2. PCRE was looping on encountering an indefinitely repeated back reference to
4645 a subpattern that had matched an empty string, e.g. /(a|)\1*/. It now does what
4646 Perl does - treats the match as successful.
4648 ****


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

  ViewVC Help
Powered by ViewVC 1.1.5